Sorghum Biology

Klaus Ammann, [email protected], Version April 20, 2010 and July 2011 With a contribution on a gene flow experiment on Sorghum in Africa from Mary Mgonja, Nairobi

Fig. 1 Oklahoma Farm Bureau, Galleries Grain Sorghum http://www.okfarmbureau.org/press_pass/galleries/grainSorghum/Sorghum2.jpg

2

Fig. 2 Schools in West Africa: Students threshing and winnowing sorghum earlier harvested from the school farm. http://wassumbee.blogspot.com/2006/01/schools-in-west-africa-1.html

Fig. 3 Lysine Biosynthesis in Sorghum bicolor: GenomeNet Database Service, from KEGG: Kyoto Encyclopedia of Genes and Genomes http://www.genome.jp/dbget-bin/get_pathway?org_name=esbi&mapno=00300 3

Sorghum Biology ...... 1

1. Preface ...... 20

2a. Executive Summary ...... 22

2a.1. General Remarks ...... 22 2a.2. Taxonomy ...... 22 2a.3. Evolutionary dynamics and Landraces of Sorghum bicolor ...... 23 2a.4. Sorghum Breeding Activities ...... 23 2a.5. Gene flow in Sorghum and related species ...... 24 2a.6. Mitigation of Gene Flow in Sorghum and related species...... 24 2b. Extended summary Report Sorghum Biology ...... 27

2b.1. General Remarks ...... 27 2b.2. Taxonomy of Sorghum, the wider picture ...... 27 2b.3. Sorghum species ...... 28 2b.4. Sorghum halepense, Johnsongrass ...... 28 2b.5. Sorghum propinquum ...... 29 2b.6. Sorghum bicolor (L.) Moench ...... 29 2b.7. Numerical taxonomy of Sorghum ...... 32 2b.8. Molecular taxonomy of Sorghum...... 32 2b.9. Distribution of Sorghum ...... 32 2b.10. Centers of crop origin ...... 32 2b.11. Earliest evidence of Sorghum cultivation in Africa 8000 years ago ...... 32 2b.12. Centers of biodiversity generally more robust against alien invasions ...... 33 2b.13. Preservation of landraces through participative breeding programs ...... 33 2b.14. Development of Sorghum breeding ...... 34 2b.15. Evolutionary dynamics of cultivated Sorghums ...... 35 2b.16. Gene flow from Sorghum cultivars to wild and feral species ...... 35 2b.17. Gene flow in Sorghum from crop to crop ...... 36 2b.18. Gene flow from weedy to cultivated Sorghums ...... 37 2b.19. Assessment of gene flow of cultivated Sorghums in Africa ...... 37 2b.20. The agricultural reality ...... 39 2b.21. A summary of gene flow in Sorghum cultivars and its wild relatives ...... 39 2b.22. Consequences and mitigation of gene flow in African Sorghum ...... 41 2b.23. Coexistence rules to be followed ...... 41 2b.24. How to avoid gene flow in cultivated Sorghums ...... 42 3. Taxonomy of Sorghum ...... 46

3.1. Wider taxonomic range of the genus Sorghum ...... 46

3.1.1. The position of the Andropogoneae and Sorghum within the system of the Poaceae ...... 46 3.1.2. The genetic comparison between Maize, Sugarcane and Sorghum ...... 47

3.2. The genus Sorghum Moench ...... 58

3.2.1. Genetics within the genus Sorghum ...... 58 3.2.1. Section Sorghum within the genus ...... 64 3.2.2. Summary taxonomy and systematics of Sorghum ...... 68 3.2.3. Bibliographic references taxonomy and systematics of Sorghum ...... 69

3.3. Sorghum halepense (L.) Pers...... 70

3.3.1. Taxonomic description ...... 70 3.3.2. Evolution of Sorghum halepense ...... 72 3.3.3. Distribution of Sorghum halepense ...... 76 4

3.3.4. Summary Sorghum halepense ...... 76 3.3.5. Bibliographic References Sorghum halepense ...... 77

3.4. Sorghum propinquum (Kunth) Hitchc. Ling. Sci. J. 7: 249, 1929...... 77

3.4.1. Taxonomic description ...... 77 3.4.2. Origin and nature of Sorghum propinquum ...... 78 3.4.3. Hybrids of Sorghum propinquum ...... 78 3.4.4. Genetic maps of Sorghum bicolor and Sorghum propinquum ...... 79 3.4.5. Summary ...... 82 3.4.6. Bibliographic references Sorghum propinquum ...... 82

3.5. Sorghum bicolor (L.) Moench, Meth. Pl. 207, 1794...... 82

3.5.1. Taxonomy of Sorghum bicolor in general ...... 82 3.5.2. Morphological description of Sorghum bicolor (L.) Moench ...... 83 3.5.3. Original taxonomic work on wild and cultivated Sorghum bicolor ...... 84 3.5.4. Sorghum bicolor subsp. bicolor ...... 87 3.5.4.1. Race bicolor ...... 93 3.5.4.2. Race kafir ...... 93 3.5.4.3. Race caudatum...... 95 3.5.4.4. Race durra ...... 96 3.5.4.5. Race guinea ...... 99 3.5.4.6. Intermediate races ...... 101 3.5.5. S. bicolor subsp. drummondii (Nees ex. Steud.) de Wet & Harlan or ex Davidse, Sudangrass ...... 102 3.5.6. S. bicolor subsp. verticilliflorum (Steudel) Piper (Cui et al., 1995; Doggett, 1988) ...... 104 3.5.6.1. Race verticilliflorum ...... 106 3.5.6.2. Race virgatum ...... 106 3.5.6.3. Race aethiopicum ...... 107 3.5.7. Recent research on Sorghum bicolor taxonomy ...... 107 3.5.8. Summary Sorghum bicolor ...... 116 3.5.9. Bibliography Sorghum bicolor ...... 118

3.6. Numerical taxonomy within the genus Sorghum ...... 118

3.6.1. Numerical taxonomy with defined lines of type specimens ...... 118 3.6.2. Numerical taxonomy with land races in Eastern Africa ...... 120 3.6.3. Summary ...... 123 3.6.4. Bibliography on numerical taxonomy of Sorghum ...... 123

3.7. Molecular taxonomic analysis ...... 123

3.7.1. RFLP diversity in cultivated Sorghum in relation to racial differentiation ...... 123 3.7.2. Molecular analysis on the basis of RFLP-assay within the genus Sorghum ...... 124 3.7.3. Comparative analysis on the genetic relatedness of Sorghum bicolor accessions from Southern Africa by RAPDs, AFLPs and SSRs ...... 126 3.7.4. 16QTLs affecting rhizomatousness and tillering of Sorghum ...... 127 3.7.5. Comparative physical mapping links conservation of microsynteny to chromosome structure and recombination in grasses ...... 129 3.7.6. Summary molecular taxonomy of Sorghum ...... 132 3.7.7. Bibliographic references on molecular taxonomy of Sorghum ...... 132

3.8. Summary of the Sorghum bicolor taxonomy ...... 133

3.8.1. The basics, an oversimplified summary ...... 133 3.8.2. Some helpful summary illustrations for the races of S. bicolor ...... 136 3.8.3. Simplified Key for the subspecific taxa of S. bicolor ...... 137 3.8.4. Illustrations of the main traits of Sorghum bicolor ...... 139 3.8.5. Recent taxonomic treatment of Sorghum bicolor (L.) Moench ...... 143 4. Biogeography of Sorghum ...... 144 5

4.1. Distribution of the genus Sorghum ...... 144

4.1.1. Word wide distribution of annual cultivated Sorghum ...... 144 4.1.2. Sorghum, geographical distribution of the staple crop in Africa ...... 147 4.1.3. Distribution of Sorghum bicolor taxa and Sorghum halepense ...... 148 4.1.3.1. Wild Sorghums in Europe and Africa ...... 148 4.1.3.1.1. Sorghum halepense ...... 148 4.1.3.1.2. Distribution of the weedy taxa of Sorghum in Africa ...... 150 4.1.3.1.2.1. Sorghum bicolor var. arundinaceum (in many publications now under S. verticilliflorum) ...... 152 4.1.3.1.2.2. Sorghum bicolor var. aethiopicum ...... 152 4.1.3.1.2.3. Sorghum bicolor var. verticilliflorum ...... 152 4.1.3.1.2.4. Sorghum bicolor var. virgatum ...... 152 4.1.3.2. Distribution of cultivated races of S. bicolor in Africa ...... 154 4.1.3.2.1. S. bicolor race guinea in Africa ...... 154 4.1.3.2.2. S. bicolor race bicolor, kafir and durra in Africa ...... 155 4.1.3.2.3. Race bicolor...... 156 4.1.3.2.4. Race kafir ...... 156 4.1.4. Summary distribution of Sorghum ...... 156 4.1.5. Bibliography distribution of Sorghum ...... 156

4.2. Centers of biodiversity and centers of crop origin, the generalities ...... 156

4.2.1. Metric map of global biodiversity ...... 157 4.2.2. Hot spots of conservation and centers of biodiversity ...... 157 4.2.3. Centers of crop biodiversity ...... 158 4.2.4. Centers of crop origin ...... 160 4.2.5. Summary ...... 161

4.2. Centers of origin of the genus Sorghum according to present day agriculture and landrace distribution ...... 162

4.2.1. One or several centers of origin for Sorghum ? ...... 162 4.2.2. Example from Somalia: Distribution and loss of landraces ...... 163 4.2.3. Domestication of Sorghum started in North-East and Central Africa, the classic model ...... 164 4.2.4. Ancestral farmers chose for the earliest domestication efforts crops which grow in monodominant stands...... 167 4.2.5. Good arguments that important domestication processes in Sorghum have happened in Southeastern Asia ...... 167

4.3. Possible origin of Sorghum in the ancient green Sahara ...... 171

4.3.1. Archaeobotany of Sorghum ...... 172 4.3.2. History of domestication of Sorghum ...... 174 4.3.3. Climate history in relation to the history of African agriculture ...... 179 4.3.4. Recent biogeographical and migration concepts of Sorghum races ...... 182 4.3.6. Summary: ...... 183 4.3.6. Bibliography Sorghum centers of origin ...... 184

4.4. Centers of biodiversity robust against introgression, landraces are a dynamic system...... 184

4.4.1. Centers of biodiversity are less susceptible to biological invasions ...... 184 4.4.2. The case of maize landraces in Mexico ...... 188 4.4.3. Conclusion: Landraces can only be preserved with active and participative breeding programs ...... 189 4.4.4. Summary robustness of biodiversity centers and conservation of landraces ...... 190 5. Reproduction of Sorghum ...... 191

5.1. Inflorescence general ...... 191 5.2. Panicle structure ...... 191 5.3. The pedicelled spikelets ...... 192 5.4. Anthesis ...... 193 5.5. Pollination ...... 195 5.6. Mating systems in Sorghum in general ...... 196 5.6.1. Cleistogamy ...... 196 6

5.6.2. Apomixis ...... 196 5.6.4. Grain development ...... 197 5.6.5. Bibliography reproduction of Sorghum ...... 198 6. Sorghum breeding in relation to its biology ...... 199

6.1. Introduction, Summary ...... 199

6.1.1. The scope of this chapter ...... 199 6.1.3. Basic remarks about Sorghum breeding and previous breeding efforts in Sorghum...... 199 6.1.4. Gradual transition from traditional breeding towards genetic engineering ...... 202 6.1.5. Main purposes of breeding programs in Sorghum ...... 203 6.1.6. Hybridization barriers in Sorghum ...... 204 6.1.7. Sorghum as a model for forage grass improvement ...... 205 6.1.8. Breeding program of the University of Hohenheim ...... 205 6.2.1. Summary of modern breeding situation in Sorghum ...... 205 6.2.2. The present day breeding programs of ICRISAT ...... 209 6.2.3. More efficient testing methods for seed collections proposed...... 210 6.2.4. Breeding strategy of ICRISAT ...... 211 6.2.5. Regionalization of breeding strategy of ICRISAT ...... 212

6.3. Development of modern Sorghum breeding ...... 215

6.3.1. The first Sorghum transformations ...... 215 6.3.2. Difficulties of the transformation of monototylenonous plants ...... 215 6.3.3. Progress in Sorghum transformation methods ...... 216 6.3.4. Situation in the year 2004 in transformed Sorghum cultivars ...... 219 6.3.5. Present day situation in transgenic Sorghums ...... 220 6.3.6. Linkage maps of Sorghum ...... 221

6.4. Present and future breeding efforts ...... 224

6.4.1. Biofortification of Sorghum crops ...... 224 6.4.2. Sorghum breeding for biofuel production ...... 236 The extrusion process is an effective pretreatment method of Sorghum for fermentation. Chemical changes (such as thermal degradation, depolymerization of starch, dietary fiber and proteins, and recombination of depolymerized fragments), and physicochemical changes (such as destruction of native starch and protein structures) could occur during the extrusion process (Camire, 1998)...... 237 6.4.3. Future Breeding Opportunities in general ...... 237 6.5. Bibliography Sorghum breeding ...... 239 6.6. Summary Sorghum breeding ...... 239 7. Evolution and gene flow of Sorghum ...... 242

7.1. Evolutionary dynamics and history of Sorghum domestication ...... 243

7.1.1. Crop – Wild complex, ferality, the basics, related to Sorghum...... 243 7.1.2. Wild and weedy Sorghums in Africa ...... 246 7.1.3. Call for more hybridizing experiments and morphological analysis ...... 247 7.1.4. Evolutionary dynamics on the genomic level ...... 249 7.1.5. Genetic relationships of landraces of Sorghum ...... 251 7.1.6. Summary ...... 254 7.1.7. Bibliography Sorghum landraces ...... 254

7.2. Gene flow and hybridization of Sorghum, general remarks ...... 254

7.2.1. A field experiment from the USA shows the difference between (high) potential outcrossing rate and actual (low) hybridization rate ...... 255 7.2.2. Fitness and many other factors influencing the outcome and persistence of outcrossing events, experience from the USA ...... 257 7

7.3. Gene flow from Sorghum cultivars to wild relatives in Africa ...... 260

7.3.1. General remarks ...... 260 7.3.2. Gene flow scenarios are different from region to region ...... 261 7.3.3. Summary gene flow within Sorghum cultivars ...... 262 7.3.4. Bibliography on outcrossing Sorghum cultivars to wild relatives ...... 262

7.4. Gene flow from weedy to cultivated Sorghums ...... 263

7.4.1. No direct experimental proof of gene flow from weedy to cultivated Sorghums ...... 263 7.4.2. Indirect proof of low gene flow from wild to cultivated Sorghums ...... 263 7.4.3. Summary of gene flow from wild to cultivated Sorghums ...... 265 7.4.4. Bibliography not available on direct evidence of weed to crop gene flow in Sorghum ...... 265

7.5. Gene flow in Sorghum from crop to crop ...... 265

7.5.1. The agricultural reality: Gene flow from Sorghum crop to crop is low, individuality of land races have been remained intact over centuries...... 265 7.5.2. Short term gene flow experiment from Sorghum crop to crop under artificial conditions ...... 271 7.5.3. Summary: ...... 274 7.5.5. Bibliographic References on Outcrossing in Sorghum ...... 275

7.6. Consequences of gene flow and regulation ...... 276

7.6.1. General remarks: Gene flow and hybridization in agriculture,...... 276 7.6.1.1. Conventional crops ...... 276 7.6.1.2. Gene flow of transgenic crops ...... 276 8. Transgenic Sorghum ...... 277

8.1. Biofortified Sorghum ...... 278

8.2. Regulation of transgenic crops ...... 278

8.3. Gene flow and its consequences in the Americas ...... 279

8.3.1. Outbreeding depression of rare species of origin of crops ...... 280 8.3.2. Loss of genetic integrity of rare species of crop origin ...... 281

8.4. Gene flow and its consequences in Africa ...... 282

8.4.1. The field trial of (Schmidt & Bothma, 2006), critical comments in the light of regulatory questions for the SuperSorghum project ...... 283 8.4.2. Gene flow in the genus Sorghum has to be assessed case by case ...... 285 8.4.3. Summary ...... 287

8.5. Mitigation of gene flow ...... 288

8.5.1. Coexistence, general remarks ...... 288 8.5.2. The choice of inbreeding landraces of Sorghum...... 288 8.5.3. Safety distances ...... 289 8.5.4. Tandem constructs for a sustainable mitigation of gene flow ...... 290 8.5.5. Male sterility and apomixis in Sorghum, the prevention of gene flow ...... 290 8.5.6. Gene switching as a prevention method to avoid gene flow ...... 292 8.5.7. Summary ...... 293 9. Geneflow assessment with morphometric methods ...... 294 8

9.1. General remarks ...... 294

9.2. The methodology proposed ...... 295

9.2.1. General situation ...... 295 9.2.2. Early developments of the methodology ...... 295 9.2.2.3. Dutch-Swiss method ...... 296 9.2.2.4. Conditions for successful pollination...... 296 9.2.2.5. Gene flow indices ...... 296 9.2.2.6. Recent improvements of the Dutch-Swiss Method, now called Dutch-Swiss-Irish Method ...... 296 9.2.3. Study elements of Dutch-Swiss-Irish method ...... 297 9.2.3.1. Strand CSV (Crop seed-to-volunteer gene flow) ...... 298 9.2.3.2. Strand CSF (Crop seed-to-feral gene flow) ...... 298

9.3. The organization of the gene flow study part morphometrics ...... 298

9.3.1. Assessing the biosystematics of Sorghum traits and wild relatives, which are relevant for the project ...... 299 9.3.3. Fixing the framework of the geneflow study, start with study with preparative ...... 301 10. Cited references ...... 303

11. Bibliographies on publications on the biology of the genus Sorghum ...... 350

9

Fig. 1 Oklahoma Farm Bureau, Galleries Grain Sorghum http://www.okfarmbureau.org/press_pass/galleries/grainSorghum/Sorghum2.jpg ...... 1 Fig. 2 Schools in West Africa: Students threshing and winnowing sorghum earlier harvested from the school farm. http://wassumbee.blogspot.com/2006/01/schools-in-west-africa-1.html ...... 2 Fig. 3 Lysine Biosynthesis in Sorghum bicolor: GenomeNet Database Service, from KEGG: Kyoto Encyclopedia of Genes and Genomes http://www.genome.jp/dbget-bin/get_pathway?org_name=esbi&mapno=00300 ...... 2 Fig. 4 Suggested relationships among the major groups of grasses. A = Arundinelleae, I = Isachneae, 0 = Oryzeae, R = Aristideae. C, metabolism is indicated by stippling; it is divided into the MS and PS types of Brown (1977). From (Clayton, 1981) ...... 46 Fig. 5 Diagram showing the evolutionary interrelationships of the principal subfamilies and tribes of the Poaceae (Gramineae). The irregular outlines represent approximately the relative size and diversity of the groups named in them, and the distance of a group from the star in the center is a rough indication of its degree of evolutionary specialization. The numbers represent single genera or small clusters of genera which are not easily placed in any of the major groupings, as follows: 1, Streptochaeta; 2, Pariana; 3, Pharus; 4, Centotheceae; 5, Arundinelleae; 6, Uniola, Brylkinia; 7, Distichlis, Aleuropus, Vaseyochloa, Ectosperma; 8, Orcuttia; 9, Neostapfia; 10, Aristicla; 11, Melica, Glyceria, Schizachne; 12, Nardus; 13, Monerma; 14, Scribneria. From (Stebbins, 1956) ...... 47 Fig. 6 A-C Phylogenetic hypothesis generated by analysis of rbcLatpB sequences. A Fifty percent majority rule of 79 equally parsimonious trees generated from analysis of 664 nucleotides and 55 insertion/deletion events scored as unordered binary characters (1,O); numbers on branches refer to number of times (in percentage) in the 79 trees in which the bifurcation was supported. B Semistrict consensus of 79 trees, as in A. C Example of 1 of the 79 equally parsimonious trees, represented as a phylogram in which branch lengths (shown above lines) are proportional to genetic distances calculated in PAUP. In these trees, the following terminal taxa represent more than one accession: Z. mays (2 genotypes sequenced); Saccharum robustum = S. barberi = S. edule = S. officinarum NG 5 1-13 1; S. officinarum Black Cheribon = S. sinense From (Al-Janabi et al., 1994) .. 48 Fig. 7 Alignment of the genomes of six major grass crop species with 19 rice linkage segments, whose order reflects the circularized ancestral grass genome. The data have been redrawn as a series of rice linkage segments (defined by radiating lines) formed into chromosomes (broken color-coded and numbered lines). The thin dashed lines correspond to the duplicated segments. Inversions of sets of sequences within a linkage segment (such as the inversion of segments 3a and 3b in maize chromosome 5) are not shown. Linkage segments forming parts (pt) of Triticeae chromosome 5 are shown as a series of segments connected by colored lines. The alignment is based on the genetic map of the D genome of wheat. The red line indicates the duplicated segments shown as blocks 11 b and 12b. Chromosomes formed by the insertion of one segment into another are shown by black lines with arrows indicating the direction and point of insertion. The points of chromosome breakage involved with insertion events are indicated by black bisected circles. The 'haploid' chromosome number of each species is shown in the column marked 'x'. The haploid DNA content of each species, shown in the column of 1C values, is per 109 bases. From (Moore et al., 1995) ...... 49 Fig. 8 Comparisons of cereal genome evolution based on rice linkage segments. (a) Rice chromosomes dissected into linkage blocks [3]. (b) Wheat, (c) maize, (d) foxtail millet, (e) sugar cane and (f) Sorghum chromosomes represented as 'rice blocks' on the basis of homology and/or conservation of gene order. Connecting lines indicate duplicated segments within the maize chromosomes. (g) An ancestral 'single chromosome' reconstructed on the basis of these linkage blocks [3]. The designation of the different Sorghum linkage groups has varied between laboratories and publications. The data in several publications [8-11] have been assembled in the form presented by Grivet [12,13]. For the purpose of this figure, the Sorghum linkage groups have then been assigned (S1-S10) on the basis of their rice segments and the order of rice segments in 10

the ancestral 'chromosome'. Open boxed linkage groups indicate that the order of segments within each linkage group needs further confirmation. From (Moore et al., 1995) ...... 50 Fig. 9 Saccharum consensus linkage map and corresponding Sorghum linkage groups (LGs). Loci connected by a line are detected by the same probe in both genomes. The underlined markers were tandemly duplicated loci on a sugarcane linkage group. The italic markers, on the basis of their relative positions on different sugarcane linkage groups, might have been duplicated loci or might have been different alleles of the same loci on different homologs. This type of markers was referred as repeated loci to distinguish them from those tandemly duplicated loci on a single linkage group. Tandemly duplicated markers were connected by a line to the corresponding Sorghum markers, but repeated markers were not connected. Markers on the right side of HGs 2 and 3 were approximately at the same location with the markers on the consensus map they aligned to. Markers mapped on a different Sorghum LG were indicated by the Sorghum LG in parentheses following the markers. From (Ming et al., 2002a) ...... 51 Fig. 10 Comparative mapping of sugar yield-related QTLs. Solid lines connect homologous loci on different sugarcane and Sorghum linkage groups. Individual Sorghum linkage groups (LGs) are represented by LGs A to J. Sugarcane linkage groups (Lgs, to be distinguished from Sorghum LGs) from four parental varieties are indicated by the last letter of the marker name: G (Green German); M (Muntok Java); I (IND 81-146); P (PIN 84-1). Approximate map positions of double-dose (#) markers are inferred by the method of (Da Silva et al., 1995). The letters in parenthesis following the marker name represent the Sorghum linkage groups where the marker was mapped, if different from the corresponding location shown. Only regions that contain, or are homologous to, QTLs are shown. Bars and whiskers indicate 1 and 2 LOD-likelihood intervals. Sugar-content QTLs (Ming et al. 2001) are shown to the left of the Sorghum linkage groups, from (Ming et al., 2002b) ...... 52 Fig. 11 Diagram of classification of the Andropogoneae, as proposed by (Hackel, 1889), redrawn by (Spangler et al., 1999). Subtribes are denoted by dashed lines with subtribal names in bold boxes. Arrows show putative direction of evolution and size of arrow indicates certainty of relatedness...... 54 Fig. 12 One of 5661 equally most parsimonious trees based on ndhF sequences. Branch lengths are indicated above each branch and decay values listed below nodes. Decay values of zero indicate that the node collapses in the strict consensus. Letters indicate nodes affected by the inclusion of sex expression data relative to the strict consensus tree. See Results for explanation. Fig. 3b. The strict consensus of all equally most parsimonious trees. Bootstrap values are above resolved branches. The arrow marks the clade comprising tribe Andropogoneae. From (Spangler et al., 1999) ...... 56 Fig. 13 The current classification scheme for the taxa sampled on one of the most parsimonious ndhF trees. Branch shading is an optimization of haploid chromosome numbers. Awnless taxa are also indicated for the Andropogoneae. Stiposorghum = S, Parasorghum = P. From (Spangler et al., 1999) ...... 57 Fig. 14 Plot of the multiple correspondence analysis of 41 Sorghum accessions showing the distribution of species clusters. bicolor laxiflorum (Heterosorghum) and S. nitidum (Parasorghum) appeared as outliers, and species in section Eusorghum appeared to fall into four groups: (i) sweet Sorghums and shattercane; (ii) S. nigricans, S. hewisonii, S. miliaceum, and S. aethiopicum, which represent S. verticilliflorum, S, sudanese, S. halepense, and 21 inbreds, kaoliangs, and land races (they had the same number of restriction fragments and cannot be separated from S. aethiopicum), (iii) S. propinquum, a diploid rhizomatus species; and (iv) S. niloticum, S. controversum, S. arundinaceum, C-401 and 5-27, and S. almum. From (Guo et al., 1996) ...... 60 Fig. 15 The strict consensus tree of the three most parsimonious trees identified by branch and bound analysis of the twenty five Sorghum species. Bootstrap values are shown above the branches, and node numbers shown below the branches. (Dillon et al., 2001)...... 62 Fig. 16 The strict consensus tree of six equally parsimonious solutions of 202 steps (CI = 0_792). Bootstrap support (%) for various nodes from 10 000 replications is indicated above the corresponding node. Only bootstrap values of >50 are shown. Numbers in parentheses above each node represent unambiguous nucleotide substitutions. The tree shows bootstrap support, 56 % and 58 % respectively, for lineages consisting of (a) S. 11

bicolor through S. laxiflorum and (b) S. brachypodum through S. nitidum. The tree was rooted using Zea mays. The 2CDNAcontent, 2n chromosome number (in parenthesis), and x = 5 genome size are denoted next to each species. The DNA content for corn (4C = 10_31 pg) is for inbred line Va35 (Laurie & Bennett, 1985). From (Price et al., 2005a) ...... 63 Fig. 17 The strict consensus trees generated using PAUP branch and bound maximum parsimony analysis. Letters A–D designate clades discussed in the text. 1a. ITS1 data: strict consensus of 60 equally parsimonious trees of 127 steps and consistency index (CI) of 0.764. 1b. ndhF data: strict consensus of 100 equally most parsimonious trees of 92 steps and consistency index (CI) of 0.826. Numbers above branches are percentages of 10,000 bootstrap replicates in which each clade was recovered. Trees were rooted using Zea mays. Letters in parenthesis indicate taxonomic sections within Sorghum where P=Parasorghum, S=Stiposorghum, C`=Chaetosorghum, H=Heterosorghum and E=Eu-Sorghum. (Dillon et al., 2004) ...... 67 Fig. 18 The Sorghum taxa according to (De Wet, 1978) The genus Sorghum is divided into five major subgenera in the Poaceae family. Species Sorghum is the co genitor of the crop-wild-weed complex in the genus Sorghum. (Ejeta & Grenier, 2005) ...... 68 Fig. 19 Sorghum halepense. A: base of plant with rhizomes. B: inflorescence, C: ligule D: terminal end of branch showing two sessile and three pedicellate spikelets, E-H: sessile spikelet, E: whole spikelet at anthesis, lateral ventral and dorsal view, F: lemma, G: palea, H: stamens and pistils, I,J: pedicellate spikelet, I: whole spikelet, ventral view, J: lemma, K: lodicules, L: sessile spikelet in fruit, grain enclosed by glumes, M: grain or caryopsis, scale A-B: ½, C-M: x 8. From (Warwick & Black, 1983) ...... 71 Fig. 20 16QTLs affecting rhizomatousness and tillering of Sorghum. The 10 LGs of Sorghum are denoted A-J (12). DNA markers indicated bylines crossing a LG were used in QTL mapping; those indicated by arrows were mapped on a subset of 56 progeny (12), and locations were inferred relative to flanking markers. Chromosomal locations of selected markers in maize (M) (17, 18), rice (R) (18, 19), and wheat (W) (20) are indicated. Maximum-likelihood locations (arrowhead) and 1-lod (box) and 2-lod (whiskers) likelihood intervals for each QTL are to the left of appropriate linkage groups. The pattern within the 1-lod likelihood interval indicates trait. From (Paterson et al., 1995b) ...... 74 Fig. 21 Phenotypic characterization of the seed (top panel), Panicles (middle Panels) and roots (bottom) of Sorghum (left), Johnsongrass (right), and the F1 hybrid (middle). From (Dweikat, 2005) ...... 75 Fig. 22 Alignment of two Sorghum genetic maps and locations of putative QTLs. Ten chromosomal genetic maps for BTx623/IS3620C (top line) and BTx623/S. propinquum (bottom line) populations are shown as horizontal thin grey lines connected by common markers. The middle two horizontal lines represent bridge maps derived from common markers and centimorgan positions from high-density genetic maps (see Materials and methods). QTLs are shown as thick horizontal lines with BTx623/IS3620C QTLs above the chromosome (labeled with txs) and BTx623/S. propinquum and shown below the chromosome (labeled as uga). Alignments for the entire genome are available from Gramene. (Feltus et al., 2006) ...... 80 Fig. 23 continued, see caption in Fig. 1 (Feltus et al., 2006) ...... 81 Fig. 24 The table of (De Wet & Huckabay, 1967) actually summarizes the concept of the ‘Snowdenian species’. Genetic analysis of recent years show reasonable compatibility with the present day concepts of race denomination, except that geographical differentiation of the same race may have been underestimated (Tao et al., 1993). See the dendrogram related to this table in chapter on numerical taxonomy, and from group C of this dendrogram the figure constructed by (Doggett, 1988) on the clusters of the cultivated ‘Snowdenian species’ D-K...... 85 Fig. 25 Scatter diagram indicating the range of morphological variability of the wild varieties of S. bicolor, and some segregating populations of hybrids among them. The vertical axis represents peduncle-sheath length in cm, and the horizontal axis indicates primary axis length of the inflorescence in cm. Each dot represents the average value of 10 plants studied belonging to a natural collection. The morphological values of the hybrids are based on measurements of individual plants. From (De Wet et al., 1970)...... 86 12

Fig. 26 Pictorial scatter diagram relating African wild and cultivated Sorghums. Panicle length against pedicelled spikelet length demonstrate clearly hybrid introgressions between wild and cultivated Sorghum. Explanation of the dot graphs given above in the figure. From (Doggett & Majisu, 1968) given in (Doggett, 1988)...... 87 Fig. 27 S. bicolor (Linn.) Moench var. bicolor (Pers.) Snowden (Snowden, 1936) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 7: Grain left proximal (in)side with hilum, right: distal (out)side with embryo mark. 8: Longitudinal section of grain. 9: Transverse section of grain. From (Doggett, 1988) ...... 90 Fig. 28 Sorghum bicolor, pedicellate and sessile spikelets, from Agrostology text from the internet: http://gemini.oscs.montana.edu/~mlavin/b434/lab7.htm Montana State University ...... 91 Fig. 29 Sorghum bicolor, Department of Horticulture and Crop Science, The Ohio State University, upper and right grain: distal side with embryo mark, 2 grains to the right below: proximal side with hilum, from http://www.oardc.ohio-state.edu/seedid/single.asp?strID=312 ...... 92 Fig. 30 Race Kafir Sorghum caffrorum Beauv. var. albofuscum (Koern.) Snowden (Snowden, 1936) in (Doggett, 1988). 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 6: Mature fruiting spikelet with two pedicelled spikelets. 7: Grain distal (out)side with embryo mark. 8: Longitudinal section of grain. 9:Transverse section of grain. From (Doggett, 1988)...... 94 Fig. 31 Sorghum bicolor race kafir from Lost Crops of Africa Vol. 1, Grains, 1996 darwin.nap.edu/books/0309049903/html/158.html National Academic Press ...... 94 Fig. 32 Race caudatum. Sorghum caudatum Stapf var. caudatum (Hack.) Snowden (Snowden, 1936) in (Doggett, 1988) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 4: Mature fruiting spikelet. 7: Grain, left distal (out)side with embryo mark, right: proximal (in) side with hilum. From (Doggett, 1988) ...... 96 Fig. 33 Race durra. Sorghum durra (Forsk.) Stapf var. aegyptiacum (Koern.) Snowden (Snowden, 1936), in (Doggett, 1988) 1: Mature fruiting spikelet with two pedicelled spikelets. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 7: Grain, left: distal (out)side with embryo mark, right: proximal (in)side with hilum. 8: Longitudinal section of grain. 9: Transverse section of grain. From (Doggett, 1988) ...... 97 Fig. 34 The 5 races of Sorghum in Africa. For each race four views are illustrated: (a) The spikelet showing grain and glumes gl. (b) View of the side of the grain on which the embryo e is visible. (c) View of the side of the grain opposite the embryo. (d) Longitudinal section through the grain showing the endosperm en and embryo e. (1) Race durra is characterized by a crease cr in one or both glumes gl. The grain is broad and blunt at the top and has a relatively small base. (2) Race caudatum is characterized by asymmetrical grains. The embryo side of the grain bulges. The side of the grain opposite is flattened or may even be concave. (3) Race bicolor is characterized by grains which do not extend beyond the glumes. (4) Race kafir has broadly ellipsoid grains. (5) Race guinea has grains which are discoid, i.e. round when viewing the embryo side, but flattened in longitudinal section. Glumes of race guinea are long and gape at maturity. Grains twist with respect to the glumes. From (Stemler et al., 1975a) ...... 98 Fig. 35 The evolutionary relationship of bicolor, durra, and durra-bicolors in Ethiopia. 3a Race bicolor is characterized by grains which do not extend beyond the glumes. 4. Race durra is characterized by a crease cr in one or both glumes gl. The grain is broad and blunt at the top and has a relatively small base. 5. Race durra-bicolor is characterized by a crease (wrinkle) in one or both glumes and ellipsoid to obovoid grains. From (Stemler et al., 1975b) ...... 99 Fig. 36 Race guinea of S. bicolor subsp. bicolor. 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 7: Grain: left proximal (in)side with hilum. 8: Longitudinal section of grain. 9: Transverse section of grain. From: (Doggett, 1988) . 101 Fig. 37 Sorghum verticilliflorum. NRCS Natural Resources United States Department of Agriculture, http://www.pr.nrcs.usda.gov/technical/Plants/pla46.html ...... 105 13

Fig. 38 Neighbor-joining phenogram depicting genetic relationships among S. bicolor accessions. Wild accessions and the various cultivated races of Sorghum are color-coded. Numbers along branches denote bootstrap support (shown only for values greater than 50) Group I: landraces, Group II: race bicolor, Group III: other racial types, Group IV: guinea and caudata accessions. From (Casa et al., 2005)...... 109 Fig. 39 Relationship between field altitude in m and Sorghum landrace diversity on polynomial regression analysis, all statistical parameters are significant. From (Teshome et al., 1999b) ...... 110 Fig. 40 Genetic diversity of cultivated Sorghums revealed by morphological markers. The tree is constructed from the Sokal and Mitchener index. From (Deu et al., 2003) ...... 113 Fig. 41 Neighbor-joining analysis based on RFLP data among 205 cultivated accessions using the Nei and Li similarity index. The numbers on the branches indicate bootstrap values (expressed in percentages) and are shown for all clusters with > 60%) bootstrap support. Cl, cluster. From (Deu et al., 2006) ...... 115 Fig. 42 Correspondence analysis on morphological traits. C1 and C2 shows again the two substructures of the caudatum race. Comments above. From (Deu et al., 2003) ...... 115 Fig. 43 Dendrogram showing relationships between the 52 Sorghum species recognized by Snowden. See Table 1 in chapter on S. bicolor for code numbers. Fig 1 from (De Wet & Huckabay, 1967) ...... 119 Fig. 44 Cultivated ‘Snowdenian species’ of group C in the above dendrogram arranged in clusters by a further iteration on the data in the Table (see chapter on the taxonomy of Sorghum bicolor) of (De Wet & Huckabay, 1967) ...... 120 Fig. 45 Sorghum landrace ordination by canonical discriminant analysis Sorghum landrace ordination by canonical discriminant analysis, using farmers-naming of Sorghum accessions as group criterion. The landraces named by the farmers are supported by the 14 morphological characters and form distinct groups on the ordination plot as well as in the analysis. Variation explained by axes 1,2 and 3 were 58.18% and 12.05% respectively, from (Teshome et al., 1997) ...... 121 Fig. 46 Scatter diagram of the first three canonical variables (CAN) based on mean values of observations in Sorghum landraces, from (Abdi et al., 2002)...... 122 Fig. 47 PAUP computer analysis – derived dendrogram A PAUP computer program generated dendrogram of all 53 accessions of S. bicolor based on RFLPs detected with 62 Sorghum genomic clones. Tree length = 1633. CI=0.206. HI = 0.794. First column - accession code or name; second column - original classification; third column-Geographical origin; fourth column - RFLP cluster (Cui et al., 1995) ...... 125 Fig. 48 AFPL, B: RAPD, C: SSR, D: Δμ-SSR. UPGMA-Clusters of 46 Sorghum accessions derived from southern Africa. LR: Landrace, BV: breeding variety, RB: Botswana, LS: Lesoto, MW: Malawi, ZA: South Africa, IS: ICRISAT, NP: Northern Province, South Africa. From (Uptmoor et al., 2003) ...... 127 Fig. 49 MDS plot indicating the genetic relationships between 100 Guinea-race bicolor accessions. Accessions are labeled according to their eco-geographical origin. Fig. 2 from (Folkertsma et al., 2005) ...... 128 Fig. 50 MDS plot indicating the genetic relationships between 92 Guinea-race bicolor accessions. Accessions are labeled according to their Snowden Guinea-race classification (Folkertsma et al., 2005) ...... 129 Fig. 51 Comparative physical maps. A segment of the rice chromosome 4 pseudomolecule is compared with BAC contigs from SP, SB, and Zea mays based on hybridization anchors (Table 1). The scales (MBP) of the rice pseudomolecule and sorghum contigs are equal, whereas maize contigs are shown at a 1:5 scale. Sorghum and maize contig lengths were estimated by multiplying the average FPC band size (4,740 bp, the observed average for rice) by the length of the contig in FPC consensus band (CB) units. Red lines represent cases for which loci were inferred in one genome (where no dot is shown) due to missing data. Maize contigs were from a recent release (www.genome.arizona.edu_fpc_maize, release 10_25_04), incorporating hybridization anchors fromTable1. Current Sorghum contig assemblies are available online (www.plantgenomega.edu_projects.htm) , from (Bowers et al., 2005) ...... 130 Fig. 52 Coverage of the rice genome by syntenic sorghum BAC contigs. Only hybridization markers that hit two or more BACs in the same contig were considered for microsynteny comparisons. Sorghum contigs were aligned 14

to the rice sequence based on the criteria that two or more low copy probe loci (detecting five contigs or fewer) showed best matches between 5 and 500 kb apart in rice (anchoring 456 SP and 303 SB contigs). Sorghum chromosome (32) and linkage group designations (16) are as cited. Cytologically identified heterochromatic regions were approximated from refs. 19 and 25 based on relative distance from the centromere, assigning approximate base-pair locations without accounting for sequence gaps. From (Bowers et al., 2005) ...... 131 Fig. 53 The five major races of grain Sorghum as described above, from: (Bantilan et al., 2004) ...... 134 Fig. 54 Classification of S. bicolor, Table 1 from (De Wet & Huckabay, 1967) ...... 136 Fig. 55 Five basic spikelet types of cultivated Sorghum; wild type and shattercane not shown, from (Harlan & De Wet, 1972): ...... 136 Fig. 56 Head types of cultivated Sorghum. Type 1 is reserved for wild races and is considerably more diffuse than type 2. From (Harlan & De Wet, 1972) Sorghum and guinea: types 2-4 kafir and durra: types 5-7 caudatum: wide range of types broomcorn and half-broomcorn types: 8-9 from (Harlan & De Wet, 1972): ...... 136 Fig. 57 Diagram of a spikelet pair showing floret positions scored for sex expression. 1: sessile spikelet , proximal floret; 2: sessile spikelet, distal floret; 3: pedicellate spikelet, proximal floret; 4: pedicellate spikelet, distal floret...... 138 Fig. 58 The inflorescence and spikelets of Sorghum bicolor var. bicolor. 1: Part of panicle: a: internode of rhachis, b: node with branches, c: branch with several racemes. 2: Raceme: a: node, b: internode, c: sessile spikelet, d: pedicel, e: pedicelled spikelet, f: terminal pedicelled spikelets, g: awn. 3: Upper glume: a: keel, b: incurved margin. 4: Lower glumes: a: keel, b: keel-wing, c: minute tooth terminating keel. 5: Lower gemma: a: nerves. 6: Upper lemma: a: nerves, b: awn. 7: Palea. 8: Lodicules. 9: Flower: a: ovary, b: stigmas, c: anthers. 10: Grain: a: hilum. 11: Grain: a: embryo mark, b: lateral lines (1x 2/4, 4x 2-9, 5x 10-11 (Doggett, 1988) ...... 140 Fig. 59 S. bicolor (Linn.) Moench var. bicolor (Pers.) Snowden (Snowden, 1936) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 4: Mature fruiting spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 6: Mature fruiting spikelet with two pedicelled spikelets. 7: Grain. 8: Longitudinal section of grain. 9: Transverse section of grain. (Doggett, 1988) ...... 141 Fig. 60 Sorghum conspicuum var. conspicuum (Snowden) race guinea (Snowden, 1936) ...... 141 Fig. 61 Sorghum durra (Forsk.) Stapf var. aegyptiacum (Koern.) Snowden (Snowden, 1936) ...... 142 Fig. 62 Sorghum caudatum Stapf var. caudatum (Hack.) Snowden (Snowden, 1936) ...... 142 Fig. 63 Sorghum caffrorum Beauv. var. albofuscum (Koern.) Snowden (Snowden, 1936) ...... 143 Fig. 64 Global map on annual Sorghum cultivated around the world. Source: University of Hohenheim, http://www.uni-hohenheim.de/~ipspwww/350b/index.html ...... 144 Fig. 65 Asian and Australian distribution of cultivated annual Sorghum. From the website of http://www.tropicalforages.info/key/Forages/Media/Html/Sorghum_(annual).htm ...... 145 Fig. 66 Central and South American Distribution of cultivated annual Sorghum, source: http://www.tropicalforages.info/key/Forages/Media/Html/Sorghum_(annual).htm ...... 145 Fig. 67 African distribution of annual cultivated Sorghum. Source: http://www.tropicalforages.info/key/Forages/Media/Html/Sorghum_(annual).htm ...... 146 Fig. 68 Distribution of cultivated Sorghum races, from (Deu et al., 2003) ...... 146 Fig. 69 Sorghum the leading crop of Africa as a whole and one of the four outstanding cereal grains in the world (along with maize, rice, and wheat), was first brought under cultivation by the Negroes of the western Sudan, probably in the fifth millennium before the Christian era. Of its numerous subspecies, Guinea corn (var. guineense) is especially prominent in the region of origin and in the central Sudan; Durra (var. durra), developed in Ethiopia more than a thousand years later, assumes the leading position in East Africa; and Kafir corn (var. caffrorum), presumably evolved fairly early in the Christian era, holds first place from southern Tanganyika to Natal and Bechuanaland. Still other varieties have been developed subsequently in countries to which the plant has spread, notably Malaya, China, and the United States. In Africa, the only introduced crop 15

to which Sorghum appears to have lost substantial ground is maize, particularly in the south. Original caption from (Murdock, 1960), the caveats see the comments in this study on the race Durra cited from (Stemler et al., 1975a) ...... 147 Fig. 70 Distribution of S. bicolor, 13: S. halepense 14: S. bicolor var. verticilliflorum (arundinaceum), 15: S. bicolor var. aethiopicum. From (De Wet & Huckabay, 1967) ...... 148 Fig. 71 Distribution of the series spontanea. (Sorghum volelianum should read Sorghum vogelianum) From (De Wet et al., 1970) ...... 150 Fig. 72 Distribution of the wild varieties of Sorghum bicolor, original caption from (De Wet et al., 1970) caption see above. See the full list of the taxa in the previous figure. The wild varieties in this map comprise Sorghum race aethiopicum, arundinaceum, verticilliflorum and virgatum...... 151 Fig. 73 Distribution of S. bicolor 16: subsp. bicolor var. verticilliflorum, 17: subsp. bicolor race guinea, 18: subsp. bicolor var. bicolor race kafir. From (De Wet & Huckabay, 1967) ...... 153 Fig. 74 Distribution of cultivated Sorghums in Africa, from (De Wet et al., 1972) ...... 154 Fig. 75 Distribution of S. bicolor. 19: subsp. bicolor var Sorghum race durra; 20: subsp. bicolor var Sorghum race bicolor. 21: Distribution of rice (horizontal lines), root crops (vertical lines), and Sorghum (solid circles). From (De Wet & Huckabay, 1967) ...... 155 Fig. 76 A metric map of biodiversity, based on counting per area: It shows surprising results, which are not in all aspects matching other world map views such as centers of biodiversity or centers of crop diversity. Source: http://www.nhm.ac.uk/research-curation/projects/worldmap/ Wilhelm Bartlott, Bonn, Germany (Kier et al., 2005; Lughadha et al., 2005) ...... 157 Fig. 77 Centers of 25 hotspots of biodiversity and conservation (Myers, 2000) ...... 158 Fig. 78 The original eight centers of crop diversity according to Vavilov, N.I. (Hawkes, 1983; Hawkes, 1990, 1991, 1999; Hawkes & Harris, 1990; Vavilov, 1987; Vavilov, 1940; Williams, 1990) ...... 159 Fig. 79 P.M. Zhukovsky's alterations (solid lines) and additions (broken lines) to Vavilovs original concept of crop diversity, from (Zeven & Zhukovsky, 1975) ...... 160 Fig. 80 Centers and noncenters of agricultural origins (A1, Near East center; A2, African noncenter; B1, North Chinese center; B2, Southeast Asian and South Pacific noncenter; C1, Mesoamerican center; C2, South American noncenter, from (Harlan, 1971) ...... 161 Fig. 81 Probable areas of domestication of selected African crops: 1, Brachiaria deflexa; 2, Digitaria exilis and Digitaria iburua; 3, Oryza glaberrima; 4, Dioscorea rotundata; 5, Musa ensete and Guizotia abyssinica; 6, Eragrostis tef; Voandzeia and Kerstingiella; 8, S. bicolor; 9, Pennisetum americanum; 10, Eleusine coracana, from (Harlan, 1971) ...... 162 Fig. 82 Graph of the first two components of a principal component analysis based on isozyme allele frequency data from individual accessions of subsp. arundinaceum and subsp. drummondii. from Countries are abbreviated as follows: AF Afghanistan, AL Algeria, AN Angola, BE Benin, BG Bangladesh, BO Botswana, B U Burma, CD Chad, CH China, CM Cameroon, CO Congo, CR Central African Republic, EG Egypt, ET Ethiopia, GH Ghana, GM Gambia, IC Ivory Coast, ID Indonesia, IN India, IQ Iraq, IR Iran, IS Israel, JA Japan, KE Kenya, KO Korea, LB Lebanon, LS Lesotho, MA Mall, ML Malagasky Republic, MW Malawi, MZ Mozambique, NG Niger, NI Nigeria, NP Nepal, PK Pakistan, SA South Africa, SE Senegal, SI Sierra Leone, SL Sri Lanka, SO Somalia, SU Sudan, SW Swaziland, TA Taiwan, TH Thailand, TU Turkey, TZ Tanzania, UG Uganda; UR USSR, UV Upper Volta, YE Yemen, ZA Zaire, ZI Zimbabwe, ZM Zambia. From (Aldrich et al., 1992) ...... Error! Bookmark not defined. Fig. 83 Graph of the first two components of a principal component analysis based on wild and cultivated isozyme allele frequency data pooled by country of origin. Abbreviations for countries of origin are listed in the caption of the Figure above, from (Aldrich et al., 1992) ...... Error! Bookmark not defined. Fig. 84 Average linkage cluster analysis based on isozyme allele frequency data for wild and cultivated accessions pooled by country of origin using modified Rogers' distance (Wright, 1978), from (Aldrich et al., 1992) ... Error! Bookmark not defined. 16

Fig. 85 Pattern of Domestication and Spread of the Genus Sorghum. (Ejeta & Grenier, 2005) ... Error! Bookmark not defined. Fig. 86 Comparison of advantages and disadvantages of in-situ versus ex-situ conservation, from (Hawkes, 1991) ...... 188 Fig. 87 Head types of cultivated Sorghum. Type 1 is reserved for wild races and is considerably more diffuse than type 2. From (Harlan & De Wet, 1972) Sorghum and guinea: types 2-4 kafir and durra: types 5-7 caudatum: wide range of types broomcorn and half-broomcorn types: 8-9 ...... 192 Fig. 88 Diagram of a spikelet pair showing floret positions scored for sex expression. 1: sessile spikelet , proximal floret; 2: sessile spikelet, distal floret; 3: pedicellate spikelet, proximal floret; 4: pedicellate spikelet, distal floret...... 193 Fig. 89 Sorghum bicolor (L.) Moench, from website Missouri Plants http://www.missouriplants.com/Grasses/Sorghum_bicolor_page.html...... 195 Fig. 90 Sorghum bicolor (L.) Moench from website Missouri Plants http://www.missouriplants.com/Grasses/Sorghum_bicolor_page.html...... 197 Fig. 91 Fig. 7.3 from Jensen 1970 demonstrates without extensive comments how tedious traditional crop breeding really is...... 200 Fig. 92 The Sorghum breeding program of ESIP in Ethiopia, as described in detail by (Doggett, 1988). The program comprises in any one year five kinds of activities, each one with a different objective: 1. backcross crossing block, 2. Dented seed recurrent selection block, 3. ms3 crossing block, 4. Hybrid crossing block, 5. Pedigree crossing block. The diagram summarizes the interrelationships and flow of ESIPs programs. * DSBN = dented seed breeding method, ** Keremt = main rainy season, June – September, in which the evaluations are made. From (Doggett, 1988) p. 223...... 201 Fig. 93 Global trends in sorghum production, 1979-94. From (ICRISAT & FAO, 1996) ...... 207 Fig. 94 Development of Yield per 1000 ha in Western and Central Africa, Southern Africa and Eastern Africa. From (Bantilan et al., 2004) ...... 207 Fig. 95 Dendrogram for cumulative classification of 38 sites based on grain yield per hectare of Sorghum varieties planted during 1987/1988–1992/1993 and 1999/2000 using weighted environment-standardized squared Euclidean distance as dissimilarity measure and incremental sum of squares as clustering strategy. Site codes in Table 1. Italics indicate sites added to site groups based on nearest centroid criterion...... 210 Fig. 96 ICRISAT’s Sorghum breeding strategy from 1972 onwards (Bantilan et al., 2004) ...... 212 Fig. 97 Sorghum research domains in Africa and India (Bantilan et al., 2004) ...... 213 Fig. 98 Average yield and yield gain in Sorghum in different countries (Bantilan et al., 2004) ...... 214 Fig. 99 Characteristics of Sorghum research domains, from (Bantilan et al., 2004) ...... 220 Fig. 100 Gene constructs used for rice (O. sativa) transformation. tRNAlys (CUA), A. thaliana tRNAlys gene with anticodon sequence altered from CTT to CTA; UBI, maize ubiquitin promoter; GUS, b-glucuronidase gene; LUCam, firefly luciferase gene with TAG codon at position 206; BAR, Phosphinothricin acetyltransferase(PAT) gene; GST, Glutathione S-transferase gene; NOS, terminator of the nopaline synthase gene; HA, Haemagglutinin epitope coding sequence. From (Wu et al., 2003) ...... 227 Fig. 101 The three steps determining the effective supply of nutrients, the four current strategies (a–d) and a fifth new strategy impacting on these steps. (Source: C.E. West, unpublished). From (Slingerland et al., 2006) ... 228 Fig. 102 Analytical framework for the research program: food chain approach and context. From (Slingerland et al., 2006) ...... 229 Fig. 103 Effect of Zn and P fertilizer on the relation between sorghum grain yield and (A) Zn in the grain, (B) phytic acid in the grain, and (C) the phytic acid / Zn molar ratio. Results from the Somyaga field experiment 2003. (Source: K. Traore, unpublished results) ...... 231 17

Fig. 104 Zinc, iron and phytic acid contents of the grain of six contrasting sorghum varieties selected from the CIRAD sorghum collection at their field station at Samanke, Mali, in the 2002–2003 season. (Source: M.A. Slingerland, unpublished results)...... 232 Fig. 105 Fig 2" Light micrographs of thin section of protein-stained developing sorghum grain (30 days after half- bloom/pollination): Modified HPD/hl mutant endosperm (A); normal vitreous endosperm (B). Note the lack of protein (dark-stained lines) surrounding starch granules in the section of the modified mutant endosperm with high starch packing (white arrows) contrary to the complete protein matrix surrounding starch granules in the normal vitreous endosperm (black arrows)...... 234 Fig. 106 Microstructures of Sorghum: (A) non-extruded Sorghum and (B) supercritical-fluid-extrusion cooked Sorghum after (Zhan et al., 2006) ...... 237 Fig. 107 Bantilan, M., Deb, U., Gowda, C., Reddy, B., Obilana, A., & Evenson, R., eds. (2004) Sorghum Genetic Enhancement: Research Process, Dissemination and Impacts, pp 312, ICRISAT,Patancheru, Andhra Pradesh, India, plate p. 224, frontispiece to chapter ‘Impact of Improved Sorghum Cultivars on Genetic Diversity and Yield Stability’...... 242 Fig. 108 Plants under human influence: The situation is very complex, if the time axis is fully followed as has been shown by (Gladis, 1966; Hammer et al., 2003), this is demonstrated in a single graph with all details: Feral crops are (or can be) of multiple origin...... 243 Fig. 109 The original gene pool concept, established by Harlan and de Wet and modified by (Hammer et al., 2003). GP 1 The biological species, including wild, weedy and cultivated races. GP 2 All species that can be crossed with GP 1, with some fertility in individuals of the F1 generation; gene transfer is possible but may be difficult. GP 3 Hybrids with GP 1 do not occur in nature; they are anomalous, lethal, or completely sterile; gene transfer is not possible without applying radical techniques. GP 4 Any synthetic strains with nucleic acid, i.e., DNA or RNA, frequencies that do not occur in nature...... 244 Fig. 110 (Ejeta & Grenier, 2005): The adaptation to the gene pool of Sorghum based on classification of a 'biological species' as per (Harlan & De Wet, 1971, 1972) . S. bicolor and propinquum: the biological species ...... 245 Fig. 111 Example of an organismoid or a hypothetically designed crop with a genome composed of different gene pools and synthetic genes, for the explanation of this complicated matter, see (Gladis & Hammer, 2000). As complex as this situation is depicted, it might well come close to reality also with specific cases of Sorghum...... 246 Fig. 112 Spikelets and flowers of Johnson and Sudan grass. 1: terminal group of Sudan grass having one sessile and two pedicelled spikelets. 2: same for Johnson grass. 3: pair of spikelets of Johnson grass. 4: flower of sessile spikelet of Johnson grass with abaxial lodicules (lod). 5: staminate flower of pedicelled spikelet showing aborted pistil (p). (Long, 1930) ...... 248 Fig. 113 Schematic diagram of the simplest evolutionary model for events that took place in this orthologous region of the rice, Sorghum, and maize genomes after their divergence from a common ancestor. Arrows indicate genes, whereas gray bars in the two maize segments show truncated gene fragments. The genetic composition of the common ancestor of the three species is proposed, with 11 genes in this region, and is structurally very similar to the rice region. (Ilic et al., 2003) ...... 249 Fig. 114 Charred plant remains from Site E-75-6. a-c, Grains of Sorghum sp showing the ventral (a) and dorsal (b, c) surfaces, with consistent radiocarbon dates 8000 BP, from Fig. 2, (Wendorf et al., 1992) . Error! Bookmark not defined. Fig. 115 Ancestral stock of Sorghum after (Garber, 1950), taken from (Doggett, 1988). Compare the revised view of (Spangler et al., 1999; Spangler, 2003) in chapter on Taxonomy of Sorghum in general. ... Error! Bookmark not defined. Fig. 116 Suggested relationship among the genera of Sorghastrae, after (Celarier, 1958), from (Doggett, 1988)...... Error! Bookmark not defined. 18

Fig. 117 Graph of the first two components from a principal component analysis based on RFLP allele frequency data from 56 wild and cultivated Sorghum accessions. Countries of origin are abbreviated as follows: CH China, EG Egypt, ET Ethiopia, IN India, IC Ivory Coast, ICE Kenya, NI Nigeria, SA South Africa, SE Senegal, SU Sudan, UG Uganda. Fig. 1 from (Aldrich & Doebley, 1992) ...... 252 Fig. 118 Average linkage cluster analysis based on RFLP allele frequency data from 56 wild and cultivated Sorghum accessions using modified Rogers' distance (Wright 1978) ...... 253 Fig. 119 Pattern of Domestication and Spread of the Genus Sorghum. From (Ejeta & Grenier, 2005) see chapter on Centers of Origin of the Genus Sorghum for more comments...... Error! Bookmark not defined. Fig. 120 shows the scheme of the field experiment performed by (Arriola & Ellstrand, 1996), the diagram is not to scale...... 256 Fig. 121 Left: Percentage of hybrids produced among total progeny sampled. 4: mean rate at each distance class at Moreno Valley Field Station. 5: Mean rate at each distance class at South Coast Field Station...... 256 Fig. 122 Plot of the frequency of occurrence of each cultivar-specific allele (sorted and ranked) in each Johnsongrass population vs. the proportion of individuals in each population carrying each allele. The plot demonstrates that Johnsongrass populations differ both in the proportion of cultivar-specific alleles present in the population and in allele frequency at individual loci. (Morrell et al., 2005) ...... 259 Fig. 123 Distributions of the estimate of outcrossing rate among progeny families sampled from two fields of Sorghum. (Djè et al., 2004) ...... 266 Fig. 124 Scatter plots of 54 individuals of Sorghum belonging to six fields on principal components 1 (PC 1) and 2 (PC 2) based of morphological variation. For each field, ninety-five percent confidence ellipses are given. From (Djè et al., 1998) ...... 267 Fig. 125 Geographical distribution of enzymatic groups. From (Zongo et al., 2005) fig.5 ...... 268 Fig. 126 Principal component analysis on the total Sudanese landrace collection and for nine quantitative morpho- agronomic characters. Graphical representation of phenotypic diversity in the dimensional space as defined from the two first extracted factors (Varimax procedure). Separated scatter plots for each region and for the unknown source of landraces. (Grenier et al., 2004) ...... 270 Fig. 127 Plan of the Sorghum gene flow trial layout. Numbers were assigned clockwise to radiating arms, starting with the north-eastern arm. The central block with Redlan B-line (male fertile) was the pollen source, the arms with Redlan A-line (male sterile) were the pollen receptors. Each small square represents a block with Sorghum plants. The first 10 blocks closest to the central field were located 13 m, the blocks farther away were located 28 m from each other. Veld grass were located between the arms not containing wild Sorghum grasses. (Schmidt & Bothma, 2006)...... 272 Fig. 128 Amount of seeds produced per plant and model of seed production gradient, calculated with a power fit model from three data sets: (i) all data, (ii) 95 percentile, and (iii) 99 percentile. Note the double logarithmic scale. The results were calculated based on field experiments with pollen sterile receiving fields and do not reflect agricultural reality in Africa. From (Schmidt & Bothma, 2006)...... 273 Fig. 129 Relative rate of pollen flow at male sterile plants. Values represent the amount of seeds produced 2pr, where r is equal to the distance from the pollen-donating field. These results were obtained under artificial conditions experimenting with pollen sterile receiving fields. They do not reflect agricultural reality in Africa. From (Schmidt & Bothma, 2006)...... 273 Fig. 130 Distance-dependent hybridization rate of pollinated Sorghum plants. The dark solid line represents the average rate and the grey lines indicate the error interval. It has to be stressed here that the results are obtained with experiments of pollen sterile receiving fields, and do not reflect agricultural reality of Sorghum outcrossing from crop to crop in Africa. From (Schmidt & Bothma, 2006)...... 274 Fig. 131 Distance-dependent hybridization rate of pollinated Sorghum plants. The dark solid line represents the average rate and the grey lines indicate the error interval. The small peak at 100m can be interpreted as a 19

small patch of Sorghum where the cytoplasmatic sterility failed among the receptor plants. (Schmidt & Bothma, 2006) ...... 284 Fig. 132 The measures required to operate below a GM presence threshold lower than 0.9% will be the same as those needed for the 0.9% threshold, except that longer crop separation distances would have to be applied for oilseed rape and maize. Based on Table 6, the aim would be to limit cross-pollination to no more than 0.1%, and on the face of it the NIAB report provides recommended distances for this purpose (Annex C). However, there is a general point here about the difficulty of establishing measures for a 0.1% threshold. The NIAB data is robust for cross-pollination thresholds down to 0.3%, but at the 0.1% level any data has to be treated with caution. Margins of error and issues of statistical sensitivity and uncertainty become much more pronounced at such a very low level. It is possible that appreciably longer distances would be required than those in the NIAB report to be confident of routinely meeting a 0.1% cross-pollination threshold. Defra will consider this further as new scientific information becomes available. Annex C Taken with the comment in footnote No. 126, p. 42 from (DEFRA, 2006) ...... 289 Fig. 133 Sorghum verticilliflorum from the National Herbarium of Nairobi, a specimen of a wild taxa included in the species Sorghum bicolor. Herbarium specimens offer lots of morphological characters ready to be included in statistical morphometrics, also they can be precious sources for genetic analysis. Furthermore, authors of new descriptions and new definitions of taxa, based on their taxonomic studies are obliged to deposit in a Herbarium open to public access a type specimen...... 294 Fig. 134 Study elements of the Morphometric method, from (Flannery et al., 2005) ...... 297 Fig. 135 Scattergram Aegilops ovata x Triticum aestivum, based upon unpublished data of Jacot, J. and P. Rufener Al Mazyad. From (Jacot et al., 2004) ...... 299 Fig. 136 Dendrogram showing the clustering patterns of the 34 Sorghum (Sorghum bicolor (L.) Moench) landraces in North Shewa and South Welo, Ethiopia (key to legend: for the local name of Sorghum landraces, refer to Table below. From (Abdi et al., 2002) ...... 300 Fig. 137 Table with morphological details, from (Abdi et al., 2002) ...... 301

20

1. Preface This text, now grown over some years has its origin in a short report on general Sorghum biology aspects on biodiversity during a brief participation in the Super-Sorghum Project of Africa Harvest in 2005-2006. The study is in its present scope is restricted to aspects of the biology of this fascinating genus, it is dealing with agronomic aspects only in context with its biology in a broad sense and it is not dealing with the Super-Sorghum Project, but will give some basic information on Sorghum in Africa and breeding, specifically also on biofortification activities worldwide. This work has been encouraged through a contribution of Africa Harvest (for the short 60pp biodiversity report) and with support of the Delft University of Technology, in the framework of a guest professorship 2006-2007.

It is written in a review style, as a result of an extensive literature review with some efficient electronic instruments such as the Web of Science (Knowledge) and the Regensburg Consortium journal retrieval system, the citations are kept in databases built with Endnote Version X2, the document has been written in Word from Office professional for Vista Business.

Some 15000 publications have been screened, over 6000 citations have been transferred to the Sorghum database, the majority on a basis of reading the summary, some 400 papers are inserted in the database with the full original text as a link.

Screening and compiling and the selecting the citations are done by the author alone and he takes full responsibility for the correctness and interpretation of the texts and possible unintentional omissions. It should be mentioned explicitly that this is a compilation text and often authors are fully cited from their abstracts, in order to avoid bias and to give credit to the authors opinion of the original papers. There are only a few cases where the author did not agree with the conclusions of publications in the field of gene flow, which is then explicitly mentioned. The author refrained from making taxonomic decisions in the sense to unify all names under a new system. This is especially blatant with the treatment of two Sorghum races S. arundinaceum and S. verticilliflorum, where various authors follow different taxonomic concepts: Whereas some follow the opinion that both races intergrade so heavily, that distinction is impossible (De Wet et al., 1970), others like (Aldrich & Doebley, 1992; Deu et al., 1995) insist on grounds of genomic analysis that in most populations the two races can be separated properly.

One important source needs to be mentioned explicitly: it’s the monograph on Sorghum written by H. Doggett and published in its second edition in 1988, it cannot be replaced by this report in its broad scope. Nevertheless, this report also demonstrates that science has made a lot of progress in the years after 1988, but still - we all stand on the shoulders of this great author. The other major source is the monograph of (Smith & Frederickson, 2000), dealing primarily with agronomic aspects.

21

The manuscript has been developed over 3 years until September 5, 2007, with a major revision from September 2009 and January 2010. Neuchâtel, Monruz 20, Switzerland, September 2009, [email protected] Delft University of Technology, Faculty of Applied Sciences, Department of Biotechnology, Kluyver Center, NL-2628 BC Delft, Netherlands and Sabanci University Istanbul, Turkey April 1, 2009 22

2a. Executive Summary

2a.1. General Remarks To designate Sorghum as a "lost" crop, on the face of it, seems like a gross mistake. After all, the plant is Africa's contribution to the world's top crops. It belongs to the elite handful of plants that collectively provide more than 85 percent of all human energy. Globally, it produces approximately 70 million metric tons of grain from about 50 million hectares of land. Today, it is the dietary staple of more than 500 million people in more than 30 countries, let alone in Africa it is an important crop for 300 million people. Only rice, wheat, maize, and potatoes surpass it in its importance for feeding the human race. More than 11’000 references are listed in the Web of Science, and alone in the last three years 2007- 2009 there are nearly a 1000 references registered. The reference figures for wheat: 67’000, rice 48’000, maize 33’000 are strikingly higher of course. For all that, Sorghum now receives merely a fraction of the attention it warrants and produces merely a fraction of what it could. Not only is it inadequately supported for the world's fifth major crop, it is under-estimated considering its vast and untapped potential. Viewed in this light it is indeed a "lost crop" – up to now. But this situation may not continue much longer. A growing number of researchers in governmental institutions and companies already see that a new and enlightened Sorghum era is just around the corner. Accorded research support at a level comparable to that devoted worldwide to wheat or rice or maize, Sorghum could contribute a great deal more to food supplies than it does at present. And it would contribute most to those regions and peoples in greatest need. Indeed, if the twentieth century has been the century of wheat, rice, and maize, the twenty-first century could become the century of Sorghum. The Bill and Melinda Gates Foundation, in close collaboration with Africa Harvest will make it possible to foster biofortification of this wonderful crop, a crop, which has already proven its big potential for subsistence and commercial farming.

The book presented helps to synthesize the literature and achievements in the research on the biology of Sorghum in the wake of the 21st century. May it be of help for future research and development of Sorghum.

2a.2. Taxonomy Sorghum is genetically closely related to other well known crop grasses such as maize, sugar cane and rice, it is obvious from this that transformation among those genera will greatly enhance the agronomic characteristics of those crops. Sorghum bicolor comprise a still not well defined number of cultivars including many landraces and interbreeding wild relatives. Three species of Sorghum are recognized, two are rhizomatous taxa: S. halepense and S. propinquum, and the large and complex S. bicolor to include all annual wild, weedy and cultivated taxa. The latest interpretation about true Sorghums: They originate from Africa and Asia with 2n=20 and 40 chromosomes, and the known progenitor of cultivated S. bicolor is a Sorghum verticilliflorum (arundinaceum) of unknown origin and makeup, most probably of polyploid nature.

23

2a.3. Evolutionary dynamics and Landraces of Sorghum bicolor Sorghum bicolor has been cultivated since at least 8000 years, the origins are not yet definitely clarified, but the region between Sudan and Ethiopia seems to harbour many ancestral landraces, although it has to be admitted that early migration between South Africa, Equatorial Africa and the dryer regions of Subsaharian Africa have blurred a clear picture. Strict farming traditions and feeding habits, often related to language and tribes, and an important percentage of inbreeding traits, combined with strictly followed agricultural traditions have maintained clearcut delimitations of landraces over centuries, despite of a proven gene flow from cultivars to wild relatives in Africa. There are considerable differences in reproduction biology, biogeography and environmental adaptation for many of the landraces. The Africa Harvest project should concentrate on inbreeding Sorghum races in order to reduce considerably gene flow. According to some authors, the guinea-race is a predominantly inbreeding, diploid cereal crop, but more research is needed.

2a.4. Sorghum Breeding Activities Encouraged by the results from natural hybridization and selection, attempts at deliberate crossing were made, wherein it was found that crosses between divergent cultivars exhibited high levels of heterosis. The discovery of the cytoplasmic genetic male sterility based on the milo-kafir system was a milestone in Sorghum breeding and research. It is widely used in the commercial exploitation of heterosis. Today, more than 30% of the Sorghum area is under hybrids (in the USA!), which have yields about twice that of any local cultivar. Species outside the Eu-Sorghum section are sources of important genes for Sorghum improvement, including those for insect and disease resistance, but these have not been used because of the failure of these species to cross with Sorghum. Most breeding programs consistently select for tolerance to abiotic stresses (such as drought and low temperatures) and biotic stresses (such as Sorghum midge, grain mold, anthracnose, charcoal rot and Striga and bird resistance). Finally, the integration of molecular genetic technology supports Sorghum improvement by providing a genetic basis for many important traits and through marker-assisted selection. Biofortification has recently become part of many breeding research and development programmes: Preliminary results suggest that a feasible chain solution consists of breeding for high Fe and moderate phytic acid contents and using soil organic amendments and P fertilization to increase yields but that this needs to be followed by improved food processing to remove phytic acid. Further research on timing of application of phosphate, Fe fertilizer and soil organic amendments is needed to improve phytic acid-Fe molar ratios in the grain. Research on the exact distribution of Fe, phosphate, phytic acid and tannins within the Sorghum grain is needed to enable the development of more effective combinations of food processing methods aiming for more favorable phytic acid-Fe molar ratios in Sorghum-based food. In another joint project between the university of Ouagadougou and the agricultural research center of Wageningen Sorghum will be better adapted to hard environments and enhancement of starch levels (amylose and amylopectin) and starch depolymerizing enzymes are breeding targets. Biofortification of Sorghum is achieved in the same partnership by selection of suitable traits, taking into account a holistic implementation approach. Current research activities focus on identifying varieties meeting specific agricultural and food requirements from the great biodiversity of Sorghums to insure food security. 24

An important effort is now made with the biofortification for a higher lysine content of Sorghum cultivars within the framework of a Africa Harvest research and development project financed mainly by the Bill and Melinda Gates Project (www.SuperSorghum.org), again with the collaboration of ICRISAT. There is still a lot of work to be done, in all domains of research and development, the maps and statistics on Sorghum yields show this clearly. Yield gain has been achieved in many developed countries on a large scale, in Africa only in Egypt.

2a.5. Gene flow in Sorghum and related species Gene flow from Sorghum cultivars to wild and feral species is well documented in the United States. In Africa and elsewhere experimental studies are still needed, but gene flow is well documented indirectly with numerous genetic screenings. Basically, one should distinguish between the high potential of crossability and the low hybridization rate in the reality of agricultural practice. Situation in the USA: Sorghum halepense is a proven conduit between Sorghum cultivars and the wild and feral populations, gene flow has been studied in field experiments. Although the rates are very low compared to maize as an example, there seems to be persistent establishment possible on a long term, and also seed exchange might play a role. In Africa, the situation is very different: up to now there are only indirect hints that there is gene flow from wild Sorghums to cultivated ones, and there are only a few reports of spontaneous hybrids between African wild and cultivated Sorghums, which could be due to lack of precise scientific knowledge. Outcrossing rates have been detected, specifically bound to local situations and specific landraces from minimal amounts to 15-nearly 30%. Still, feral complexes of wild and cultivated Sorghums seem to be a reality to such a degree that often harvest is hampered and measures should be taken in order to reduce gene flow in African Sorghum agriculture. On the other hand, in many countries in-depth genetic analysis has revealed a often striking constance of landraces over centuries, even in cases, where there was no geographical separation. This is normally interpreted by the constancy of agricultural practices and traditional use of the various Sorghum races. There are even examples known in Central Niger and in a village in Cameroon where researchers were able to find all the 5 African Sorghum races ecxept kafir in sympatry without much gene flow and hybridization.

2a.6. Mitigation of Gene Flow in Sorghum and related species There are a variety of methods available to mitigate the problem of gene flo. To assess the environmental risk of possible future deployment of transgenic Sorghum, more data on the outward flow of genes through Sorghum pollen to non-target relatives is needed.

Moreover, available information indicates existence of variation in outcrossing rates among varieties and locations. Knowledge on actual outcrossing rates in selected and specific cases will be essential to come up with recommendations for variety maintenance and on-farm seed production.

This study on Sorghum Biology proposes to determine the extent of seed and pollen mediated gene flow in selected experiments using both molecular and morphological methods. Present day morphogenetic 25 analysis of landraces, often intricately growing together, shows, that the individuality of the landraces is maintained despite selection and seed exchange activity of the farmers over a long time.

Coexistence rules to be followed No zero tolerance: Zero values with intermixing seeds or yield cannot be achieved in most cases (and is not necessary, since the introgressed genes stem from cultivars which have been approved for environmental and food and feed safety anyway). So it is a misleading expression to talk about “contamination”, we should use terms like intermixing, and threshold values instead. Field size influences results The impact of gene flow and the result of intermixing tracable genes in the yield is depending on the size of the fields involved: Big fields of many hectares usually do not pose a problem, since intermixing will be homogenized to minute traces well below the threshold values fixed by law. Safety distances: Coexistence depends on safety distances, which have to be determined in a case by case decision. A multitude of characters of crop reproduction biology will have to be evaluated in order to set up the individual rules of lawful segregation Biology and production characteristics of seeds: Seed biology, production (and selection practices and exchange) also has to be involved in coexistence ruling Time scale: There is, depending on the crop biology, a need to look at several years of crop production because of seed exchange by farmers (which traditionally is low), volunteer plants, tillering and dormant seeds, all these factors depend on the specific characters of a given crop, soil and region. Traditional knowledge In the majority of cases, these rules do not have to be invented and developed from scratch, they can be derived from the practices of traditional and modern farming.

Mitigation needs a case to case approach Mitigation of gene flow and the insurance of coexistence will depend on a variety of methods, which have to be adapted from case to case.

How to avoid gene flow in cultivated Sorghums In order to avoid gene flow, there are several possibilities open for future development and regulation: Regulation through establishing safety distances Working with safety distances is often the easiest strategy of choice, this has been shown with many different crops, the safety distances depending on the field size according to experimental data, see below the table of DEFRA, based on a multitude of field experiments cited in this extensive report. Safety distances have to be assessed in the case of Africa, they will have to be assessed by field experiments with marker genes and with morphometric methods, details see chapters on proposed field experiments of Africa Harvest SuperSorghum project at the end of the study. Transgenic mitigation (TM), Tandem constructs, where a desired primary gene is tandemly coupled with mitigating genes that are positive or neutral to the crop but deleterious to hybrids and their progeny. 26 propose tandem constructs, which would avoid or at least reduce considerably the spread of transgenes, the concept has been proven to produce stable gene constructs and sustainable mitigation: Transgenic mitigation (TM), where a desired primary gene is tandemly coupled with mitigating genes that are positive or neutral to the crop but deleterious to hybrids and their progeny. The linked unfitness through TM would be continuously manifested in future generations, keeping the transgene at a low frequency. Male sterility and apomixis in Sorghum, the prevention of gene flow Several studies demonstrate the effects of cytoplasmatic pollen sterility with Sorghum, a valid method to mitigate gene flow. Gene flow through pollen can be severely restricted in Sorghum by adaptation of this technique aiming at A3 cytoplasmic male sterility. The low levels of seed set on selfed A3 F2 individuals (0.04%) as compared with A1 F2 individuals (74%) is a clear indication of greatly reduced amounts of viable pollen production from volunteer progeny of seed that escapes harvest in the cropping year. In field scale research and crop production, it is unlikely that all risks associated with new technologies can be completely eliminated. Apomixis has been studied for years in other crops, and it could be promising to develop such outcrossing prevention also for Sorghum. Application of gene switching technologies, a method to avoid gene flow Gene switching might develop in a real opportunity, first studies for other purposes show, that mechanisms could be developed for a crop protection technology. A Chinese study claims that chemical regulation of transgene expression presents a powerful tool for basic research in plant biology and biotechnological applications. Various chemical-inducible systems based on de-repression, activation and inactivation of the target gene have been described. The utility of inducible promoters has been successfully demonstrated by the development of a marker-free transformation system and large scale gene profiling. In addition, field applications appear to be promising through the use of registered agrochemicals (e.g. RH5992) as inducers. most of the systems reported thus far are unsuitable for field applications because of the chemical nature of the inducers. Further work should focus on systems suitable for applications with transgenic crop plants, with particular emphasis on agricultural chemicals (e.g. insecticides and safeners; the latter chemicals are used in agriculture to render crops tolerate to herbicides) that have already been registered for field usage. An additional interest would be to develop multiple-inducible systems to independently regulate several target genes.

27

2b. Extended summary Report Sorghum Biology

2b.1. General Remarks To designate Sorghum as a "lost" crop, on the face of it, seems like a gross mistake. After all, the plant is Africa's contribution to the world's top crops. Indeed, according to the numerous authors contributing in the last decades it belongs to the elite handful of plants that collectively provide more than 85 percent of all human energy. Globally, it produces approximately 70 million metric tons of grain from about 50 million hectares of land. Today, it is the dietary staple crop of more than 500 million people in more than 30 countries, let alone in Africa it is an important crop for 300 million people. Only rice, wheat, maize, and potatoes surpass it in feeding the human race. For all that, however, Sorghum now receives merely a fraction of the attention it warrants and produces merely a fraction of what it could. Not only is it inadequately supported for the world's fifth major crop, it is under-supported considering its vast and untapped potential. Viewed in this light it is indeed a "lost crop". But this situation may not continue much longer. A growing number of researchers in governmental institutions and companies already see that a new and enlightened Sorghum era is just around the corner. Accorded research support at a level comparable to that devoted worldwide to wheat or rice or maize, Sorghum could contribute a great deal more to food supplies than it does at present. And it would contribute most to those regions and peoples in greatest need. Indeed, if the twentieth century has been the century of wheat, rice, and maize, the twenty-first could become the century of Sorghum. The Bill and Melinda Gates Foundation, in close collaboration with Africa Harvest will make it possible to foster biofortification of this wonderful crop, a crop, which has already proven its big potential for subsistence and commercial farming. Research on Sorghum has received a big boost, let alone the Web of Science reveals some 700 publications for the year 2010.

The book presented should help to synthesize the literature and achievements in the research on the biology of Sorghum in the wake of the 21st century. May it be of help for future research and development of Sorghum.

2b.2. Taxonomy of Sorghum, the wider picture Sorghum Moench is a heterogeneous genus. In the taxonomic literature, usually the following sections of the genus are recognized (Section = defined group of species within one genus): Stiposorghum, Parasorghum, Sorghum, Heterosorghum and Chaetosorghum. Stiposorghums are characterized by distinct rings of hairs at each culm node, with the awns over 65mm long. those of Parasorghums are substantially shorter. Maize and even more so Sugarcane are both genetically closely related to Sorghum, opening breeding prospects both ways with modern molecular tools. 28

It is important to know why interspecific hybridization does not occur within the genus Sorghum: The growth of alien pollen tubes is inhibited in Sorghum pistils. For three species where limited fertilization and embryo formation occurred, the endosperm in immature seed aborted and no viable seed were produced. Conclusions from DNA sequence studies about the phylogeny split Sorghum into two lineages, one comprising the 2n = 10 species with large genomes and their polyploid relatives, and the other with the 2n = 20, 40 species with relatively small genomes. An apparent phylogenetic reduction in genome size has occurred in the 2n = 10 lineage. Genome size evolution in the genus Sorghum apparently did not involve a one way ticket to genomic obesity as has been proposed for the grasses.

2b.3. Sorghum species Three species of Sorghum are recognized, two are rhizomatous taxa: S. halepense and S. propinquum, and the large and complex S. bicolor to include all annual wild, weedy and cultivated taxa. The latest interpretation about Eu-Sorghum: They originate from Africa and Asia with 2n=20 and 40 chromosomes, and the known progenitor of cultivated S. bicolor is a Sorghum verticilliflorum (arundinaceum) of unknown origin and makeup, most probably of polyploid nature.

2b.4. Sorghum halepense, Johnsongrass Sorghum (x) halepense (Johnsongrass) is derived from a doubling of the chromosomes in a natural cross between S. verticilliflorum (arundinaceum) and S. propinquum. Still, and rightly so, Sorghum halepense is recognized as a species due to its worldwide establishment, originating from the temperate regions of the Mediterranean region eastwards. Sorghum x drummondii originated from a natural cross between S. bicolor and S. arundinaceum and Sorghum x almum is the natural cross between S. bicolor and S. halepense. The close genetic relationship and inter-crossability between the Eu-Sorghum species is well documented. Although S. bicolor (2n = 2x = 20) and S. halepense (2n = 4x = 40 and higher) differ in ploidy, numerous artificial crossing studies have demonstrated that S. bicolor can serve as the pollen parent of triploid and tetraploid hybrids. Naturalized populations of S. halepense occur in many regions where Sorghum is cultivated. In the United States, S. halepense is a very common roadside weed. It also invades cultivated fields of many warm season crops and can reduce agricultural productivity by as much as 45%. The two species frequently grow in close physical proximity have overlapping flowering periods. Hybridization with cultivated (non-transgenic) S. bicolor has been proposed as a potential cause of increased aggressiveness in the weed. Based on morphological evidence, it has been suggested that ‘Johnsongrass’ in North America may actually be an S. halepense × S. bicolor hybrid similar to S. almum (Columbus grass) (2n = 4x = 40) of South America. A relatively high frequency of occurrence of cultivar-specific alleles in S. halepense populations far from Sorghum fields and outside the region of the United States where S. bicolor is cultivated has been found. Sorghum halepense has only recently invaded the more northerly portions of its introduced range in the United States and Canada and it is possible that alleles have been maintained in populations that experienced previous introgressive hybridization. 29

The existence of intermediary forms in most Sorghum growing areas of the warm temperate zones in the Americas offers an empirical evidence for weedy forms arising from continued introgression (exo- ferality) among different Sorghum types.

2b.5. Sorghum propinquum S. propinquum is a species related to S. bicolor in the broad sense. It belongs to the six Eu-Sorghum species and it is reputedly a perennial rhizomatous relative of S. bicolor. Molecular data have shown that there is good transferability from Sorghum propinquum for interesting genetic sites to the cultigens, and breeders will use these new insights. The natural distribution of Sorghum bicolor, however, is strictly African, and since Sorghum propinquum is spatially isolated from the annual S. bicolor, there are no spontaneous African hybrids between S. propinquum and African wild S. bicolor reported.

2b.6. Sorghum bicolor (L.) Moench The species S. bicolor (L.) Moench includes all annual taxa of the section Sorghum as recognized by early authors. It comprises an extremely variable complex of cultivated taxa, including a widely distributed and ecologically variable wild African complex, and also stabilized weedy derivatives (a complex of feral taxa) originating from introgression between domesticated grain Sorghums and their closest wild relatives.

There are three complexes recognized as subspecies: Sorghum bicolor subsp. bicolor see 3.6.1. Sorghum bicolor subsp. drummondii see 3.6.2. Sorghum bicolor subsp. verticilliflorum see 3.6.3.

2b.6.1 Sorghum bicolor subsp. bicolor: the cultivated grain Sorghums The variation of cultivated African Sorghum is discontinuous and justifies the above designation of five major races with some varieties included, the definite number of races and varieties will have to be decided for in a new comprehensive monograph of the whole genus. Cultivated grain Sorghums were earlier divided into the Guineensia, Nervosa, Bicoloria, Caffra and Durra complexes. A modern monograph, based on sturdy molecular analysis is still lacking, but it will be a major undertaking involving many laboratories and scientists. The variability in cultivated Sorghums can be described in terms of four roughly geographic races (guinea, durra, kafir and caudatum) and one widely distributed morphologically primitive race bicolor.

3b.6.1. Race bicolor Race bicolor is morphologically seen the most primitive. It has small, ellipsoidal grains which are tightly enclosed by the glumes. This race occurs widely but is not grown much for grain. Most bicolors are now grown for their sweet stalks or for production of beer or dye. The primitive morphology of bicolor Sorghums and wide geographic distribution suggest that a bicolor type was the progenitor of the more highly derived races. 30

2b.6.1.1. Race guinea Race guinea is distinguished by long glumes that gape at maturity and discoid grains that twist up to go degrees at maturity. Guineas are adapted to areas with a long rainy season, areas too wet for other Sorghums and most other cereals and do not yield as much grain as some other Sorghums, but many varieties are highly prized because they yield a superior nonbitter white flour. According to recent studies, the guinea-race is a predominantly inbreeding, diploid cereal crop. For more details about inbreeding guinea races see chapter on gene flow avoidance. It originated from West Africa and appears to have spread throughout Africa and South Asia, where it is now the dominant Sorghum race, via ancient trade routes.

2b.6.1.1.1. Race guinea variety margaritiferum Margaritiferum varieties of the race guinea are characterized through very small, hard grains (3-5mm) eaten like rice, grown in parts of West Africa, receiving up to 120 inches of rain during the growing season.

2b.6.1.1.2. Race guinea variety conspicuum Conspicuum varieties are characterized through very large grains (5-8mm), grown in the fog belt of Mozambique, Malawi, and Swaziland.

2b.6.1.1.3. Race guinea variety roxburghii Roxburghii Sorghums are distinguished from gambicum varieties by glumes larger than the grain, and have intermediate grains in size (between margaritiferum and conspicuum varieties) and are grown in East and South Africa.

2b.6.1.1.4. Race guinea variety gambicum Gambicum Sorghums are distinguished from variety roxburghii by glumes subequalling the grains, they have also intermediate grains in size and grown in tropical West Africa.

2b.6.1.2. Race durra Sorghums of race durra have glumes which are usually creased horizontally. The grains are very broad and blunt at the top and taper to a relatively small, wedge-shaped base. Some mature in as little as three months and thus can be grown in areas with a very short rainy season. Durras are grown in India and south-west Asia as well as the northern fringe of the African savanna. The pattern of distribution strongly suggests that durras were brought into Africa under Islamic influences.

2b.6.1.3. Race kafir Sorghums of race kafir have grains that are broadly ellipsoid. The part of the grain which extends beyond the clasping glumes is symmetrical and round. Kafir Sorghums are found in parts of Africa south of the equator, mainly in areas with one well-defined rainy season such as Natal, Transvaal, and Orange Free State.

2b.6.1.4. Race caudatum The caudatums are distinguished by its very asymmetrical grain. The embryo side of the grain bulges; the opposite side is flat or may even be concave; and the tip of the grain often comes to a point. This combination of characters gives race caudatum a very distinctive appearance. Caudatum Sorghums are grown in Hyparrhenia and Andropogon dominated regions of north-central Africa receiving 15 to 50 31 inches of rain per year. They are the staff of life in much of Sudan and Chad, and parts of Cameroon and Uganda. Caudatums have not, however, been widely accepted as human food by peoples outside traditionally caudatum-growing areas possibly because most caudatum varieties contain polyphenolic compounds often called 'tannins' that make caudatum flour bitter and dark in color. This biochemical characteristic makes caudatum inferior in a culinary sense to most guinea, kafir, and durra Sorghums. Though caudatum is not very popular as food for humans outside the areas listed, caudatum varieties such as 'feterita' and 'hegari' or hybrid derivatives of these varieties are important in sub-humid to semi- arid regions of the world where they are grown for use as stock feed. It is typical for caudatum Sorghums that its more difficult to define them clearly on morphological grounds. Caudatum, more than any other indigenous African crop, can be counted on to produce a crop despite high water, drought, parasites such as witch weed, and every other kind of hazard, a considerable overall contribution to African food safety.

2b.6.2. Sorghum bicolor subsp. drummondii, (Nees ex. Steud.) de Wet & Harlan or ex Davidse, Sudangrass Its an annual weed thought to be a natural hybrid between subsp. bicolor and verticilliflorum, grouped mainly with a subset of the cultivated subsp. bicolor (namely, race bicolor). Subspecies drummondii occurs as a weed in Africa wherever cultivated grain Sorghums and their closest wild relatives are sympatric. All races of subspecies Sorghum and all wild kinds of S. bicolor hybridize to produce weedy derivatives. Morphologically stabilized derivatives often accompany grain Sorghums as weeds beyond the natural range of S. bicolor.

2b.6.3. Sorghum bicolor subsp. verticilliflorum, Wild Sorghum bicolors This subspecies includes also Sorghum arundinaceum and is defined as the wild subspecies of Sorghum bicolor. The following ‘races’ intergrade heavily, but are still often distinguished and named and also found with typical morphology:

2b.6.3.1. Race verticilliflorum can be distinguished from other wild Sorghums by its large and open inflorescences with spreading, but not pendulous branches. This race grades into race arundinaceum which is distributed across tropical Africa as a forest grass. In the broad leaf savannah race arundinaceum is often difficult to distinguish from race verticilliflorum except by inflorescences in which the branches become pendulous at maturity.

3.6.3.2. Race virgatum resembles race verticilliflorum except that the inflorescence branches are more erect, and the leaf blades are narrowly linear. Race virgatum occurs along stream banks and irrigation ditches in arid northeastern Africa. It is mentioned in various studies on genetic diversity and agronomic characterization.

2b.6.3.3. Race aethiopicum is a desert grass, and is easily recognized by its large ovate-lanceolate, densely tomentose sessile spikelets. It occurs across the Sahel from Mauritania to the Sudan. It is mentioned as a possible source for Striga resistance.

32

The close relationships identified between S. bicolor and several wild species caused gene flow over centuries, but they have also exciting prospects in utilizing this diverse gene pool for improving Sorghum production.

2b.7. Numerical taxonomy of Sorghum If characters known for their moderate variability, excluding leaf and inflorescence sizes and plant height, are used exclusively from certified type specimens, numerical taxonomy shows reasonable clustering and fits well with the view of earlier taxonomists.

2b.8. Molecular taxonomy of Sorghum There are some 750 papers published on molecular analysis of Sorghum. Progress in molecular knowledge is considerable, and growing rapidly in the last few years, helping to understand taxonomic relationships, but also providing basic knowledge on useful details for future breeding efforts.

2b.9. Distribution of Sorghum The genus Sorghum is strictly Old-World in its original distribution, extending almost continuously from Southern Africa to subtropical Australia. The cultivated-weed complex of Sorghum reached Australia and the New World only after the colonization of the Europeans. The cultivated Sorghums are even more variable than the feral complexes through continuous selection and extensive exchange of germplasm, including hybrid plants. Still it is possible to distinguish geographically between four distinct cultivated complexes. Centers of Biodiversity and Crop Biodiversity are not the same, although sometimes they overlap.

2b.10. Centers of crop origin These centers have changed in concept and are still in debate, since in many cases reliable archaeobotanical data still lack. The best solution is to divide in centers and non-centers. There is no real proof of any genuine place or country of origin of Sorghum. But clearly, the region of Sudan and partly Ethiopia harbor a series of ancestral landraces, show a high biodiversity and certainly can be taken as the centers of Sorghum landrace biodiversity of the present time. Present day understanding is based on a center of origin as a wide zone in the broad-leaved savanna belt that stretches from about lake Chad to eastern central Sudan. Vast amounts of wild Sorghum are found along the Sudan-Ethiopia border, but there is no indication that the area was ever farmed before government settlement projects were established. Variations in Sorghum do not suggest that its homeland is Ethiopia; by far the bulk of Ethiopian Sorghums are durras, which are the most specialized traits with a narrow genome and certainly derived from cultivated Sorghum.

2b.11. Earliest evidence of Sorghum cultivation in Africa 8000 years ago Early domestication in Africa and India is often cited, but the data on which these conclusions were until recently at best circumstantial. 33

Although not generally known, in the early nineties, a discovery pushed back the origin of Sorghum domestication for many thousand years, this opens new perspectives for the history of domestication: Archeobotanical results allow to trace back Sorghum grains to much earlier times than assumed before: Excavations at an early Holocene archaeological site in southernmost Egypt, 100 km west of Abu Simbel, have yielded hundreds of carbonized seeds of Sorghum and millets, with consistent radiocarbon dates of 8,000 years before present (BP), thus providing the earliest evidence for the use of these plants. They are morphologically wild, but the lipid fraction of the Sorghum grains shows a closer relationship to domestication than to wild varieties. Whatever their domestication status is, the use of these plants 8,000 years BP suggests that the African plant-food complex developed independently of the Levantine wheat and barley complex. This ancient history, going back maybe even to the old Saharan population which existed before the dramatic desertification of huge regions, would also explain the high degree of biodiversity of landraces which does not follow the rules of their present day distribution properly.

2b.12. Centers of biodiversity generally more robust against alien invasions Centers of crop biodiversity usually are rich in plant species. Although it is clear, that centers of biodiversity in general suffer from species loss due to many different factors, there are enough data in population ecology to show that plant communities with a rich set of species are less susceptible to alien species invasions, except in cases where those species have a considerable advantage in competitiveness. The rule is that in heavily disturbed areas one finds the most dramatic invasion impacts of alien species.

2b.13. Preservation of landraces through participative breeding programs Landraces cannot be preserved with a conservative system, since they are dynamic populations, having been subject of spontaneous breeding and seed interchange for centuries. In consequence, it will be important to establish participative breeding programs in order to preserve landraces, since landraces have only a chance of survival, if there is still a market for the products. An important consequence of environmental and human selection is that each race of Sorghum has become ecologically and culturally specialized. A race of Sorghum resulting from generations of selection in one climate will not do well in another. And a race of Sorghum selected by one group of people is usually considered undesirable by another group of people. Predilections of Africans of the present in favor of their own Sorghums and against Sorghums of other peoples are so strong as to suggest that traditional African farmers do not and have not casually borrowed and adopted unfamiliar Sorghums raised by unrelated peoples. The usual reaction to an unfamiliar kind of Sorghum is that it is unfit for human consumption. And in fact very little exchange of Sorghum varieties of different races has been observed among unrelated groups of people. But this should not be mistaken for an appeal of an inconsiderate traditional approach, since history of agriculture has shown major shifts of cultivation tradition, especially in the face of threatening malnutrition situations. Limitations in ecological amplitude of the staple crop must have imposed limitations on the direction and extent of migration of peoples who depended on Sorghum for their food. Looking at this more positively, we might say that the ability of a group of Sorghum-growing peoples to occupy an area or to expand into 34 new areas has probably been determined in part by the productivity, genetic potential, and ecological amplitude of the Sorghum cultivars they created and depended on for food. This all can be taken as a serious caveat when new Sorghum with definitely better agronomic characteristics should be introduced in Africa. If the African farmers should benefit from new Sorghum breeding, the new traits will have to be carefully adapted to the local needs of the population. Nevertheless, (Deu et al., 2006) state, that the genetic distinctness of rnargaritiferums from other guinea Sorghums from western Africa is remarkable, all the more both are interfertile and cultivated in sympatry in the same season by the same farmers. This can also be taken as a hint that we do not yet understand Sorghum cultivation in Africa in all important population genetics aspects.

2b.14. Development of Sorghum breeding Sorghum breeding general remarks Depending on the region of production, the type of Sorghum and the purpose for its production varies widely. Whether they are breeding varieties or hybrids, the primary focus of Sorghum breeders throughout the world are yield, adaptation and quality. In addition to breeding for these factors, reducing losses due to stress is equally important. Most breeding programs consistently select for tolerance to abiotic stresses (such as drought and low temperatures) and biotic stresses (such as Sorghum midge, grain mold, anthracnose, charcoal rot and Striga and bird resistance). Finally, the integration of molecular genetic technology is enhancing Sorghum improvement by providing a genetic basis for many important traits and through marker-assisted selection. Species outside the Eu-Sorghum section are sources of important genes for Sorghum improvement, including those for insect and disease resistance, but these have not been used because of the failure of these species to cross with Sorghum. An understanding of the biological nature of the incompatibility system(s) that prevent hybridization and/or seed development is necessary for the successful hybridization and introgression between Sorghum and divergent Sorghum species. Studies of S. bicolor (L.) Moench are beginning to link physiological behavior to specific genes and hormone-based regulatory systems in ways that suggest specific strategies for improvement. Findings from several other grasses like maize and sugarcane are adding to the pool of information being derived from Sorghum. This information relates to flowering and floral development, maturity and senescence, temperature effects via the biological clock, shade avoidance behavior, apical dominance, shoot elongation, and root development including constitutive aerenchyma formation. These studies, along with others, offer a number of options for conventional plant breeding and genetic transformations to improve grass-based crops and satisfy part of the projected human food needs of coming decades. The global economic importance of Sorghum makes it a prime candidate for genetic transformation.

2b.14.1. The various breeding programs Encouraged by the results from natural hybridization and selection, attempts at deliberate crossing were made, wherein it was found that crosses between divergent cultivars exhibited high levels of heterosis. The discovery of the cytoplasmic genetic male sterility based on the milo-kafir system was a milestone in Sorghum breeding and research. It is widely used in the commercial exploitation of heterosis. Today, 35 more than 30% of the Sorghum area is under hybrids (in the USA!), which have yields about twice that of any local cultivar.

2b.14.2. Biofortification is now belonging to several breeding programs Preliminary results suggest that a feasible chain solution consists of breeding for high Fe and moderate phytic acid contents and using soil organic amendments and P fertilization to increase yields but that this needs to be followed by improved food processing to remove phytic acid. Further research on timing of application of phosphate, Fe fertilizer and soil organic amendments is needed to improve phytic acid-Fe molar ratios in the grain. Research on the exact distribution of Fe, phosphate, phytic acid and tannins within the Sorghum grain is needed to enable the development of more effective combinations of food processing methods aiming for more favorable phytic acid-Fe molar ratios in Sorghum-based food. In another joint project between the university of Ouagadougou and the agricultural research center of Wageningen Sorghum will be better adapted to hard environments and enhancement of starch levels (amylose and amylopectin) and starch depolymerizing enzymes are breeding targets. Biofortification of Sorghum is achieved in the same partnership by selection of suitable traits, taking into account a holistic approach. Current research activities focus on identifying varieties meeting specific agricultural and food requirements from the great biodiversity of Sorghums to insure food security. An important effort is now made with the biofortification for a higher lysine content of Sorghum cultivars within the framework of a Africa Harvest research and development project financed mainly by the Bill and Melinda Gates Project (www.SuperSorghum.org), again with the collaboration of ICRISAT. The future of Sorghum breeding There is still a lot of work to be done, in all domains of research and development, the maps and statistics on Sorghum yields show this clearly. Yield gain has been achieved in many developed countries on a large scale, in Africa only in Egypt. In all other African countries there is still a lot of work to be done.

2b.15. Evolutionary dynamics of cultivated Sorghums Molecular data support previous concepts of Sorghum evolution; namely, that multiple origins, diverse environments and human involvement have contributed to the existence of different types of wild and cultivated Sorghum. Outcrossing has led to gene introgression and gene flow among the natural populations. Then polymorphic subpopulations develop, and disruptive selection started. Intermediate types may have existed for a period, but differentiation continued until a number of distinct, separate and adaptive populations are subsequently formed. In summary, the population structure of modern Sorghums seems to fit well into Wright's "shifting balance" theory of adaptation.

2b.16. Gene flow from Sorghum cultivars to wild and feral species Gene flow from Sorghum cultivars to wild and feral species is well documented in the United States. In Africa and elsewhere experimental studies are still needed, but gene flow is well documented indirectly with numerous genetic screenings. Basically, one should distinguish between the high potential of crossability and the low hybridization rate in the reality of agricultural practice. 36

Situation in the USA: Sorghum halepense is a proven conduit between Sorghum cultivars and the wild and feral populations, gene flow has been studied in field experiments. Although the rates are very low compared to maize as an example, there seems to be persistent establishment possible on a long term, and also seed exchange might play a role. In Africa, up to now there are only indirect hints that there is gene flow from wild Sorghums to cultivated ones, and there are only a few reports of spontaneous hybrids between African wild and cultivated Sorghums, which could be due to lack of precise scientific knowledge. Still, feral complexes of wild and cultivated Sorghums seem to be a reality to such a degree that harvest is hampered and measures should be taken in order to reduce gene flow in African Sorghum agriculture.

2b.17. Gene flow in Sorghum from crop to crop Gene flow between different varieties of the same crop is almost as old as agriculture itself. Native plant genes as well as genes from transgenics can be dispersed either in seed or pollen, but in the case of the partially autogamous crop Sorghum the outbreeding effects remain low, the result is that landraces of Sorghum in Africa were maintaining their individual genomes over centuries. There are even examples known in Central Niger and in a village in Cameroon where Monique Deu and her team were able to find all the 5 African Sorghum races exept kafir in sympatry without much gene flow and hybridization (personal communication). Pollen mediated gene flow and the separation distances employed to minimize it typically generate more public interest than that for seed dispersal. For the most part, gene flow takes place within a few meters of the plant, but the up to now published data sets cannot be taken for reality: The cases documented for gene flow with Sorghum cultivars and Johnsongrass from the US cannot be compared to African situations. The existing studies from the US, although very detailed, are of a more theoretical interest, since the receiving fields has been planted with pollen sterile traits, in order to get a maximum rate of outcrossing. In reality it is possible to choose traits with a very high degree of autogamy (self-pollination), a similar situation as in wheat, which has no real outcrossing problems from crop to crop. From Africa, a few studies demonstrate highly variable outcrossing rates from minimal to 16 until nearly 30%, verified with genomic analysis methods. Distance from a pollen source and cross-pollination frequency with neighboring crops has been identified for major crops and employed in the production of certified seeds. In the case of Sorghum, certified seed production fields are isolated by at least 100 meters from other Sorghum fields, sufficient to maintain genetic purity of at least 99 per cent over many years of production. However, restricting gene flow between GM and non-GM varieties needs to be still better understood, and several methods already developed should be put in practice. These will efficiently help to maintain co-existence as it was practiced for many decades in modern agriculture without transgenic crops. There are a variety of methods available to mitigate the problem of gene flow, this will be dealt with in a separate chapter on the consequences and mitigation of gene flow and hybridization, and the SuperSorghum project will be able to take care of it. To assess the environmental risk of possible future deployment of transgenic Sorghum, more data on the outward flow of genes through Sorghum pollen to non-target relatives is needed. Moreover, available information indicates existence of considerable variation in outcrossing rates among varieties and locations. Knowledge on actual outcrossing rates in selected and specific cases will be essential to come up with recommendations for variety maintenance and on-farm seed production. 37

2b.18. Gene flow from weedy to cultivated Sorghums Only a few studies have been undertaken on assessing weed to crop gene flow in Sorghum, all refer to weed-to-crop gene flow only in an indirect manner. There are no field studies which aim at the direct assessment of weed to crop gene flow. Therefore, we still have insufficient information on the rate and the extent of genetic exchange in situ. It is probable, based on the potential of cross pollination and the overlap in natural habitats, that there is a continual transfer of an array of fitness and other genes from weedy types to cultivated Sorghum. Allele exchange has been shown in numerous cases in Africa. However, this is the result of century old neighborhood of wild and cultivated Sorghum.

2b.19. Assessment of gene flow of cultivated Sorghums in Africa

2b.19.1. Distribution of Landraces Race bicolor, most widely distributed and least different in morphology from wild Sorghum, is presumed to be the primitive cultivated type from which the more highly differentiated races were selected. It is not known whether cultivated bicolor Sorghums arose in one or several places in Africa. The four remaining other races of Africa, which are centered in different regions and different climates, differ from each other in morphology and some aspects of physiology. These races appear to be the products of two kinds of selection:

2b.19.2. Segregation and persistence of Sorghum landraces Natural (environmental) selection in various regions has resulted in different physiological characteristics such as photoperiodic responses, maturity cycles, moisture requirements, and some spikelet and inflorescence characters. Human selection is undoubtedly responsible for environmentally neutral aspects of the size, shape, and color of the grain and some culinary qualities. Thus, natural and human selection have simultaneously acted to produce cultivars which are physiologically suited to perform well in the environment where selection has occurred and culturally suited to the needs, expectations, and aesthetics of the people who have raised the Sorghum for countless generations.

An important consequence of environmental and human selection is that each Sorghum race has become ecologically and culturally specialized. A race resulting from generations of selection in one climate will not do well in another. And a race selected by one group of people is usually considered undesirable by another group of people.

Present day predilections of Africans in favor of their own Sorghums and against Sorghums of other people are so strong as to suggest that traditional African farmers do not and have not casually borrowed and adopted unfamiliar Sorghums raised by unrelated peoples. The usual reaction to an unfamiliar kind of Sorghum is that it is unfit for human consumption. In fact very little exchange of Sorghum varieties of different races has been observed among unrelated groups of people. One would guess that ecological and cultural specialization of races of Sorghum has had profound historical implications for Sorghum-growing peoples. It seems clear that the fate of 38

Sorghum-growing people has been inextricably linked to the Sorghum cultivars created by their forefathers.

2b.19.3. Case studies from Somalia and Ethiopia Sudan has the largest land area in Africa, yet the population base in low, only a third of the population in Ethiopia. Per-capita holdings of arable land are higher in the Sudan than all other African countries. In northern Sudan, where human settlement has been historically very light, the genetic identity of wild Sorghums may have been further protected by their isolation from human disturbance. Thorough genetic studies indicate the strong differentiation among the Sorghum materials. In the central clay plains of Somalia where Sorghum farming is practiced under irrigation and in rotation with other crops, wild Sorghums have also survived as weed in cotton and wheat fields and along the irrigation ditches. Studies in those regions give insight to Sorghum landrace management regarding biodiversity and distribution of S. bicolor: The problems of vanishing landraces and their invaluable characteristics are addressed, having adapted to harsh conditions over centuries. Today this traditional inter-dependence is at risk with catastrophic consequences for local poor resources farmers and maintenance of genetic diversity within Sorghum. Traditional landraces are being relegated to marginal and risk prone areas as they are replaced by improved varieties. This can lead to the loss of local knowledge of traditional landraces and to an erosion of genetic diversity. This should be taken into account by all breeding programs, including those on biofortification within humanitarian projects. Although the empirical evidence described from Somalia and Ethiopia is still in many places circumstantial and only based on genomic analysis, gene flow ultimately needs to be assessed experimentally, which needs to be done in the framework of the differential effects of gene flow in contrast to ecological and demographical factors and to farming practices. Sudan and Ethiopia are the presumed birth places of the crop and have witnessed the evolution of wild and primitive forms of Sorghum, although empirical data from Ethiopia show that Sorghum durra is the predominant race with a rather narrow genome. Another example on why the risk assessment approaches must be carried out in a case by case manner is demonstrated from studies in Somalia, where the situation is different from Ethiopia: In both the rainfed and irrigated Sudanese agriculture, genetic exchange between Sorghum and its wild relatives has resulted in formation of two widely recognized forms of crop-wild hybrids: Aggressive forms of weedy Sorghum bicolor have evolved that are readily identified and recognized by most farmers as feral weeds, and known under a local name of “adar”. This is a form of shattercane that is widely distributed and almost accepted as unavoidable. In spite of continual weeding and selective roguing this weedy S. bicolor has not been easy to eradicate in Sudan. The second form of intermediate is equally feral, but appears to be more similar to cultivated Sorghum and produces grains that only slowly shatter. Continued introgression of cultivated Sorghum genes into wild forms has resulted in this hybrid form called “kerketita”. Farmers selectively harvest these types and encourage their continued existence, for they rely on them as feed and food depending on the harvest prospect: In bad years, these fast growing intermediates provide the only harvest possible, particularly for fodder. So in conclusion gene flow between wild and cultivated Sorghums in Africa cannot be a really serious problem: over centuries obviously wild Sorghums and cultivated species maintained their identity well, 39 although there are regional differences, as discussed in the chapter on the origin of Sorghum. They are manifested some genetic studies. Some geographic portions of the wild gene pool are genetically more similar to the cultivars than others. Collections of wild Sorghum from Uganda, Sudan, and the Ivory Coast exhibit the highest genetic similarity with the cultivars. Although this latter collection from the Ivory Coast is quite similar to the cultivars, the other 4 wild collections from northwest Africa (Ivory Coast and Nigeria) do not share this relationship and are isolated in the principal component plot at the end opposite the cultivars. Similarly, wild Sorghum from southern Africa is quite distinct from the cultivars. In contrast, the majority of wild collections from northeast Africa (Egypt, Ethiopia, and Sudan) and central Africa (Kenya and Uganda) are fairly similar to the cultivars.

Thus, the majority of wild Sorghum from both southern and northwest Africa share less resemblance with the cultivars of the same region than does Sorghum of central and northeast Africa. Therefore, most of the cultivated Sorghum was probably domesticated from wild progenitors of the northeast and central African regions.

In chapter on molecular taxonomy the results of genomic analysis have been summed up, it is now fact that the races caudatum and kafir are not uniform and need to be studied further on.

2b.20. The agricultural reality A picture close to agricultural reality of outcrossing rates of landraces has been developed in Morocco, building on population genetic data analysis in the field. The outcrossing rate of Sorghum landraces was assessed by sampling in situ from two fields under traditional cultivation in north-western Morocco using genotypic data from five micro satellite loci. Assuming a mixed mating model, they outcrossing parameters were estimated by two methods based on progeny analyses, which demonstrated, that Sorghum landraces in this region are predominantly autogamous but still show measurable outcrossing rates. An other study presents molecular taxonomic data which demonstrate gene flow among landraces: Outcrossing has led to gene introgression and gene flow among the natural populations. Then polymorphic subpopulations develop, and disruptive selection starts. Intermediate types may exist for a period, but differentiation continues until a number of distinct, separate and adaptive populations are formed. In summary, the population structure of modern Sorghums seems to fit well into Wright's "shifting balance" theory of adaptation, which assumes that genetic drift and selection operating on subpopulations leads to a number of genotypes occupying different adaptive peaks, even though gene flow can occur between the subpopulations. A study from Burkina Faso gives real time and geography based data on the genetic relationship among Sahelian Sorghum bicolor landraces. It showed that in this Sahelian region there is a high level of diversity between Sorghum landraces and a low rate of outcrossing.

2b.21. A summary of gene flow in Sorghum cultivars and its wild relatives Gene flow between different varieties of the same crop is almost as old as agriculture itself. Native plant genes as well as genes from transgenics can be dispersed either in seed or pollen, but in the case of the 40 partially autogamous crop Sorghum the outbreeding effects remain low, the result is that landraces of Sorghum in Africa were maintaining their individual genomes over centuries. Several papers deal in all detail with the gene flow from crop to crop, and gene flow has been documented on a morphological and genetic basis, its so minimal, that it should not pose any problem with the introduction of new traits in Africa, provided some mitigation measures are taken, see chapter on consequences and prevention of gene flow.

It has to be stated very clearly, that gene flow with most Sorghum species and cultivars did always happen in the last centuries, but there are many data available to show that landraces have been remained stable, despite of free breeding and selection activity by the farmers. To erect absolute gene flow barriers would mean the genetic stabilization of the landraces and in this way ultimately lead to the end of their existence. This has been shown in the case of the Mexican landraces of maize.

Pollen mediated gene flow and the separation distances employed to minimize it typically generate more public interest than that for seed dispersal. For the most part, gene flow takes place within a few meters of the plant, but the up to now published data sets cannot be taken for reality: The case of gene flow with Sorghum cultivars and Johnsongrass cannot be compared to African situations, it is restricted to the Americas. The now often cited results of Schmidt et al. do not apply to agricultural reality: They are of a more theoretical interest, since the receiving field has been planted with pollen sterile traits, in order to get a maximum rate of outcrossing. In reality it is possible to choose traits with a very high degree of autogamy (self-pollination), a similar situation as in wheat, which has no real outcrossing problems from crop to crop. Distance from a pollen source and cross-pollination frequency with neighboring crops has been identified for major crops and employed in the production of certified seeds. In the case of Sorghum, certified seed production fields are isolated by at least 100 meters from other Sorghum fields, sufficient to maintain genetic purity of at least 99 per cent over many years of production. However, restricting gene flow between GM and non-GM varieties needs to be still better understood, and several methods already developed should be put in practice. These will efficiently help to maintain co-existence as it was practiced for many decades in modern agriculture without transgenic crops.

However, intermediate forms such as the Shattercanes appear also in the agro-ecosystems of Africa where Sorghum cultivation is practiced without sufficient isolation from wild forms and in sympatry with weedy relatives. It has to be stressed that this is not a new situation, and it will be interesting to see and learn how African farmers have coped up with such phenomena over centuries. But we have no reason to romanticize the situation: Up to now, this was often accompanied by considerable yield losses. At the end of the day it is not the introgression itself, it’s the consequences of gene flow which counts. There is some experience from several studies existing about the importance of taking in farmers knowledge on various levels of screening landrace diversity and studying farmers management methods and their influence on agricultural ecology: African landraces are according to many authors threatened through the shift in agriculture and not through gene flow. The modern world is placing a range of pressures on wild areas and on traditional agricultural communities, and external interests (often dominated by economic or political issues) 41 strongly impinge. The major external forces relevant to this discussion advocate the introduction of high- yield varieties, accompanied by mechanization and major chemical inputs, as the means to increase total production and economic return. These forces change the nature of the decision-making process dramatically; the farmer is encouraged to grow high-yield varieties in monoculture using inputs of fertilizer and pesticides.

Consequences of recurrent gene flow between cultivated Sorghum and its wild relatives should therefore not be generalized in the case of Africa: Sorghum biodiversity, mating characteristics and diversity of farmers management methods of the many landraces are just too diverse: Many of the African Sorghum varieties are strong inbreeders (especially som of the true Guineas growing in wet tropical Africa). (For more details see chapter on gene flow avoidance). And for sure we should not make the mistake of copying the concerns of the American Sorghum breeders with their problems with Sorghum halepense. Sorghum gene flow should always be considered on a case by case basis and in individual countries where the crop is grown.

2b.22. Consequences and mitigation of gene flow in African Sorghum There are a variety of methods available to mitigate the problem of gene flow, this is dealt with in a separate chapter on the consequences and mitigation of gene flow and hybridization, and the SuperSorghum project will be able to take care of it.

To assess the environmental risk of possible future deployment of transgenic Sorghum, more data on the outward flow of genes through Sorghum pollen to non-target relatives is needed.

Moreover, available information indicates existence of variation in outcrossing rates among varieties and locations. Knowledge on actual outcrossing rates in selected and specific cases will be essential to come up with recommendations for variety maintenance and on-farm seed production.

This study on Sorghum Biology proposes to determine the extent of seed and pollen mediated gene flow in selected experiments using both molecular and morphological methods. Present day morphogenetic analysis of landraces, often intricately growing together, shows, that the individuality of the landraces is maintained despite selection and seed exchange activity of the farmers over a long time.

2b.23. Coexistence rules to be followed

2b.23.1. No zero tolerance: Zero values with intermixing seeds or yield cannot be achieved in most cases (and is not necessary, since the introgressed genes stem from cultivars which have been approved for environmental and food and feed safety anyway). So it is a misleading expression to talk about “contamination”, we should use terms like intermixing, and threshold values instead. 42

2b.23.2. Field size influences results: The impact of gene flow and the result of intermixing tracable genes in the yield is depending on the size of the fields involved: Big fields of many hectares usually do not pose a problem, since intermixing will be homogenized to minute traces well below the threshold values fixed by law.

2b.23.3. Safety distances: Coexistence depends on safety distances, which have to be determined in a case by case decision. A multitude of characters of crop reproduction biology will have to be evaluated in order to set up the individual rules of lawful segregation

2b.23.4. Biology and production characteristics of seeds: Seed biology, production (and selection practices and exchange) also has to be involved in coexistence ruling

2b.23.5. Time scale: There is, depending on the crop biology, a need to look at several years of crop production because of seed exchange by farmers, volunteer plants, tillering and dormant seeds, all these factors depend on the specific characters of a given crop, soil and regional factors.

2b.23.6. Traditional knowledge In the majority of cases, these rules do not have to be invented and developed from scratch, they can be derived from the practices of traditional and modern farming.

3.23.7. Mitigation needs a case to case approach Mitigation of gene flow and the insurance of coexistence will depend on a variety of methods, which have to be adapted from case to case.

2b.24. How to avoid gene flow in cultivated Sorghums In order to avoid gene flow, there are several possibilities open for future development and regulation:

2b.24.1. Regulation through establishing safety distances Working with safety distances is often the easiest strategy of choice, this has been shown with many different crops, the safety distances depending on the field size according to experimental data, see below the table of DEFRA, based on a multitude of field experiments cited in this extensive report. Safety distances have to be assessed in the case of Africa, they will have to be assessed by field experiments with marker genes and with morphometric methods, details see chapters on proposed field experiments of Africa Harvest SuperSorghum project at the end of the study.

2b.24.1.1. Transgenic mitigation (TM), Tandem constructs, where a desired primary gene is tandemly coupled with mitigating genes that are positive or neutral to the crop but deleterious to hybrids and their progeny. propose tandem constructs, which would avoid or at least reduce considerably the spread of transgenes, the concept has been proven to produce stable gene constructs and sustainable mitigation: Transgenic mitigation (TM), where a desired primary gene is tandemly coupled with mitigating genes that are 43 positive or neutral to the crop but deleterious to hybrids and their progeny. This was tested experimentally by the team of Gressel from Rehovot in Israel as a mechanism to mitigate transgene introgression. Dwarfism, which typically increases crop yield while decreasing the ability to compete, was used as a mitigator. A construct of a dominant ahasR (acetohydroxy acid synthase) gene conferring herbicide resistance in tandem with the semidominant mitigator dwarfing Δ gai (gibberellic acid- insensitive) gene was transformed into tobacco (Nicotiana tabacchum). The highest reproductive TM fitness relative to the wild type was 17%. The results demonstrate the suppression of crop–weed hybrids when competing with wild type weeds, or such crops as volunteer weeds, in seasons when the selector (herbicide) is not used. The linked unfitness would be continuously manifested in future generations, keeping the transgene at a low frequency.

2b.24.1.2. Male sterility and apomixis in Sorghum, the prevention of gene flow Several studies demonstrate the effects of cytoplasmatic pollen sterility with Sorghum, a valid method to mitigate gene flow. Gene flow through pollen can be severely restricted in Sorghum by adaptation of this technique aiming at A3 cytoplasmic male sterility. The low levels of seed set on selfed A3 F2 individuals (0.04%) as compared with A1 F2 individuals (74%) is a clear indication of greatly reduced amounts of viable pollen production from volunteer progeny of seed that escapes harvest in the cropping year. In field scale research and crop production, it is unlikely that all risks associated with new technologies can be completely eliminated. In the event that transgenic Sorghum is developed to the stage that field testing becomes appropriate, utilization of this technique should be considered in combination with cultural controls including spatial isolation, crop rotation or fallowing in subsequent cropping seasons, and herbicides active against Sorghum hybrids to reduce risk of gene flow to weedy relatives. Related to yield, the experiments from the ETH in Zurich Switzerland indicate that such associations can bring about grain production as high or even higher than those produced by pure male-fertile maize crops, especially when the male-sterile component is pollinated non-isogenically. The grain yield benefits from cytoplasmic male sterility and xenia as well as the fact that seed of male-sterile varieties can be produced cheaply and reliably in large quantities. This would facilitate the implementation of the proposed system in agricultural practice. Exploiting this technique has proven successful for the production of high oil maize and high grain quality maize and may be used to restrict gene flow in genetically modified Sorghum as well.

2b.24.1.3. Application of gene switching technologies, a method to avoid gene flow Gene switching might develop in a real opportunity, first studies for other purposes show, that mechanisms could be developed for a crop protection technology. A Chinese study claims that chemical regulation of transgene expression presents a powerful tool for basic research in plant biology and biotechnological applications. Various chemical-inducible systems based on de-repression, activation and inactivation of the target gene have been described. The utility of inducible promoters has been successfully demonstrated by the development of a marker-free transformation system and large scale gene profiling. In addition, field applications appear to be promising through the use of registered agrochemicals (e.g. RH5992) as inducers. most of the systems reported thus far are unsuitable for field applications because of the chemical nature of the inducers. Further work should focus on systems suitable for applications with transgenic crop plants, with particular emphasis on agricultural chemicals (e.g. insecticides and safeners; the latter chemicals are used in agriculture to render crops tolerate to 44 herbicides) that have already been registered for field usage. An additional interest would be to develop multiple-inducible systems to independently regulate several target genes.

2b.24.2. Proposed gene flow studies The study on Sorghum Biology proposes to determine the extent of seed and pollen mediated gene flow in selected experiments using both molecular and morphological methods. Present day morphogenetic analysis of landraces, often intricately growing together, shows, that the individuality of the landraces is maintained despite selection and seed exchange activity of the farmers over a long time.

The situation in the USA and Africa is different regarding gene flow: In the Americas Sorghum halepense and Sorghum almum pose problems with some gene flow and their weedy character, causing also problems in cultures like maize and soya. In Africa a steady but slow gene flow is indirectly documented through genomic analysis. Wild Sorghums are widespread and cause problems for the local farmers, weeding is often not enough and in some years yield reduction is considerable. It will be crucial to learn from the local farmers about their specific problems with wild Sorghums.

A case by case view needs to be applied, it will also include local agricultural management tradition. S. bicolor offers an excellent example of the sympatric association and interaction of a crop-wild-weed complex of a species in an agro-ecosystem. The nature of genetic interaction among forms of the taxa and the consequences of these exchanges depend not only on the power of the genes involved but also on several other associated factors. Many of the African landraces are autogamous. Prevalence of wild relatives naturally varies from region to region based on extent of inherent genetic diversity, existence of selective pressures, and the farming systems in the region.

2b.24.3. Assessment projects for gene flow summary Two methods are proposed for the assessment of gene flow for the AHBFI-Project:

Marker gene based assessment with field data Marker genes will be selected and tested in field experiments in order to quantify gene flow in real time and real field situations.

Morphometric assessment of hybridization dynamics with field data In gradients of gene flow in the field morphometric data allow to trace down gene flow which produced hybrid progeny, a data set relevant to agricultural reality.

For the chapters of the full document, see the interactive contents index and the interactive list of figures at the beginning of the document

A chapter with several Sorghum bibliographies is appended

45

46

3. Taxonomy of Sorghum

3.1. Wider taxonomic range of the genus Sorghum

3.1.1. The position of the Andropogoneae and Sorghum within the system of the Poaceae The grasses (Poaceae, formerly Gramineae) are a monophyletic family of monocotyledonous flowering plants with 10,000 species, 9000 of them registered in (Clayton et al., 2002 ff), including the worlds most important cereal crops such as rice, maize, wheat, barley, and Sorghum.

Along with maize (Zea mays), sugarcane (Saccharum subsp.), and all the millets (Pennisetum, Eleusine, Eragrostis, Setaria, etc.), it falls into the tribe Andropogoneae with ca. 116 genera according to (Clayton & Renvoize, 1986, repr. 1999; Tzvelev, 1989). (tribe = defined group of genera).

(Clayton, 1981) propose relationships among the major groups of grasses

Fig. 4 Suggested relationships among the major groups of grasses. A = Arundinelleae, I = Isachneae, 0 = Oryzeae, R = Aristideae. C, metabolism is indicated by stippling; it is divided into the MS and PS types of Brown (1977). From (Clayton, 1981)

The classic evolutionary scheme of (Stebbins, 1956) still maintains the incomplete separation of the Paniceae and the Andropogoneae. Note the twig of the Maydiae.

47

Fig. 5 Diagram showing the evolutionary interrelationships of the principal subfamilies and tribes of the Poaceae (Gramineae). The irregular outlines represent approximately the relative size and diversity of the groups named in them, and the distance of a group from the star in the center is a rough indication of its degree of evolutionary specialization. The numbers represent single genera or small clusters of genera which are not easily placed in any of the major groupings, as follows: 1, Streptochaeta; 2, Pariana; 3, Pharus; 4, Centotheceae; 5, Arundinelleae; 6, Uniola, Brylkinia; 7, Distichlis, Aleuropus, Vaseyochloa, Ectosperma; 8, Orcuttia; 9, Neostapfia; 10, Aristicla; 11, Melica, Glyceria, Schizachne; 12, Nardus; 13, Monerma; 14, Scribneria. From (Stebbins, 1956)

3.1.2. The genetic comparison between Maize, Sugarcane and Sorghum A comparison between maize, rice and Sorghum reveals some relationship, based on an genomic analysis, see further comments in the chapter on evolutionary dynamics. Here an attempt to bring 48 together the relationship of closely related genera in the Andropogoneae by (Al-Janabi et al., 1994), The phylogenetic hypothesis shown indicates that Sorghum, maize, rice, wheat, and barley occupied outgroup positions, as expected, and that the ingroup, composed of Saccharum complex accessions, was monophyletic and displayed extremely low sequence variability. Furthermore, cultivated sugarcane displayed no detectable chloroplast diversity, also in agreement with (Sobral et al., 1994), suggesting a world-wide cytoplasmic monoculture for sugarcane. This is in contrast to the situation in the closely related annual S. bicolor (Duvall & Doebley, 1990), for which cytoplasmic diversity has been demonstrated and shown to be at least as great as the diversity revealed in this study among Saccharum complex members.

Fig. 6 A-C Phylogenetic hypothesis generated by analysis of rbcLatpB sequences. A Fifty percent majority rule of 79 equally parsimonious trees generated from analysis of 664 nucleotides and 55 insertion/deletion events scored as unordered binary characters (1,O); numbers on branches refer to number of times (in percentage) in the 79 trees in which the bifurcation was supported. B Semistrict consensus of 79 trees, as in A. C Example of 1 of the 79 equally parsimonious trees, represented as a phylogram in which branch lengths (shown above lines) are proportional to genetic distances calculated in PAUP. In these trees, the following terminal taxa represent more than one accession: Z. mays (2 genotypes sequenced); Saccharum robustum = S. barberi = S. edule = S. officinarum NG 5 1-13 1; S. officinarum Black Cheribon = S. sinense From (Al-Janabi et al., 1994)

It may be surprising that taxa within the Saccharum complex, which contain plants with diverse chromosome numbers and proposed geographic origin actually display such little variation within their chloroplast genome. This may be caused by the relatively recent evolution of the complex, although other explanations are possible. It has been shown that morphological variation within a species and chloroplast DNA sequence variation do not necessarily correlate (Soltis & Soltis, 1993, 1995; Song et al., 1995), although some authors have considered many of these genera to be synonymous (Clayton & Renvoize, 1986). On purely speculative grounds, it is tempting to imagine that some differences among these plants, despite large variations in ploidy, may 49 be due to a few genes with large effects, such as has been observed for some morphological variation between cultivated maize and teosinte (Doebley & Stec, 1991).

A tremendous amount of genetic work has been done on Zea mays and Sorghum bicolor. Recent data suggest high genomic colinearity among all grasses (Moore et al., 1995), and a phylogenetic context can serve to expedite studies of morphological pattern throughout the tribe based on genetic discoveries in the model systems. The high similarity of genomes in grasses coupled with the apparent lack of phylogenetic divergence in the Andropogoneae provide an exciting context for the application of genetic findings in model systems to evolutionary explanations of variability.

Fig. 7 Alignment of the genomes of six major grass crop species with 19 rice linkage segments, whose order reflects the circularized ancestral grass genome. The data have been redrawn as a series of rice linkage segments (defined by radiating lines) formed into chromosomes (broken color-coded and numbered lines). The thin dashed lines correspond to the duplicated segments. Inversions of sets of sequences within a linkage segment (such as the inversion of segments 3a and 3b in maize chromosome 5) are not shown. Linkage segments forming parts (pt) of Triticeae chromosome 5 are shown as a series of segments connected by colored lines. The alignment is based on the genetic map of the D genome of wheat. The red line indicates the duplicated segments shown as blocks 11 b and 12b. Chromosomes formed by the insertion of one segment into another are shown by black lines with arrows indicating the direction and point of insertion. The points of chromosome breakage involved with insertion events are indicated by black bisected circles. The 'haploid' chromosome number of each species is shown in the column marked 'x'. The haploid DNA content of each species, shown in the column of 1C values, is per 109 bases. From (Moore et al., 1995) 50

Fig. 8 Comparisons of cereal genome evolution based on rice linkage segments. (a) Rice chromosomes dissected into linkage blocks [3]. (b) Wheat, (c) maize, (d) foxtail millet, (e) sugar cane and (f) Sorghum chromosomes represented as 'rice blocks' on the basis of homology and/or conservation of gene order. Connecting lines indicate duplicated segments within the maize chromosomes. (g) An ancestral 'single chromosome' reconstructed on the basis of these linkage blocks [3]. The designation of the different Sorghum linkage groups has varied between laboratories and publications. The data in several publications [8- 11] have been assembled in the form presented by Grivet [12,13]. For the purpose of this figure, the Sorghum linkage groups have then been assigned (S1-S10) on the basis of their rice segments and the order of rice segments in the ancestral 'chromosome'. Open boxed linkage groups indicate that the order of segments within each linkage group needs further confirmation. From (Moore et al., 1995)

(Ming et al., 1998) came up with a detailed genetic linkage map: The complex polyploid genomes of three Saccharum species have been aligned with the compact diploid genome of Sorghum (2n = 2x = 20).

Based on previous work, (Ming et al., 2002a) constructed a consensus map of homologous DNA linkage groups from two genotypes in each of two Saccharum species by aligning them with the compact diploid genome of S. bicolor (L.) Moench.

The overwhelming correspondence among the HG markers of four Saccharum linkage maps to a particular Sorghum linkage group prompted us to assemble a Saccharum map. The authors used a minimum of two common markers to connect corresponding HGs to form a single unified HG. Thirty-six of the 41 pair-wise comparisons shared more than half the markers on individual HGs ranging from three to 25 (Table 4). Among the 13 consensus sugarcane HGs, 11 HGs were shared by both S. officinarum and S. spontaneum, and only one short HG each was specific to either S. officinarum or S. spontaneum. Only 30 (10%) of the 286 loci on the Saccharum consensus map failed to match corresponding Sorghum LGS. 51

Sorghum is a close relative of sugarcane and it has been suggested that these two species may have diverged as little as 5 million years ago (Al-Janabi et al., 1994). See the fig below, taken from (Ming et al., 2002a).

Fig. 9 Saccharum consensus linkage map and corresponding Sorghum linkage groups (LGs). Loci connected by a line are detected by the same probe in both genomes. The underlined markers were tandemly duplicated loci on a sugarcane linkage group. The italic markers, on the basis of their relative positions on different sugarcane linkage groups, might have been duplicated loci or might have been different alleles of the same loci on different homologs. This type of markers was referred 52 as repeated loci to distinguish them from those tandemly duplicated loci on a single linkage group. Tandemly duplicated markers were connected by a line to the corresponding Sorghum markers, but repeated markers were not connected. Markers on the right side of HGs 2 and 3 were approximately at the same location with the markers on the consensus map they aligned to. Markers mapped on a different Sorghum LG were indicated by the Sorghum LG in parentheses following the markers. From (Ming et al., 2002a)

Fig. 10 Comparative mapping of sugar yield-related QTLs. Solid lines connect homologous loci on different sugarcane and Sorghum linkage groups. Individual Sorghum linkage groups (LGs) are represented by LGs A to J. Sugarcane linkage groups (Lgs, to be distinguished from Sorghum LGs) from four parental varieties are indicated by the last letter of the marker name: 53

G (Green German); M (Muntok Java); I (IND 81-146); P (PIN 84-1). Approximate map positions of double-dose (#) markers are inferred by the method of (Da Silva et al., 1995). The letters in parenthesis following the marker name represent the Sorghum linkage groups where the marker was mapped, if different from the corresponding location shown. Only regions that contain, or are homologous to, QTLs are shown. Bars and whiskers indicate 1 and 2 LOD-likelihood intervals. Sugar-content QTLs (Ming et al. 2001) are shown to the left of the Sorghum linkage groups, from (Ming et al., 2002b) (Ming et al., 2002b): Active breeding work in Saccharum using plant materials derived from mapping populations are in progress, and markers linked to the QTLs could be used directly to incorporate positive alleles and eliminate negative alleles of the sugar yield components. To balance the high sugar content with reasonable stalk-strength, fiber-content and ash-content QTLs would also be useful in selection. The use of DNA markers in selection would allow the identification of potentially superior materials and the elimination of undesirable ones in the early stages of a breeding program.

(Mathews et al., 2002): Andropogoneae are a monophyletic tribe of 85 genera that includes Zea and Sorghum. All members exhibit C-4 photosynthesis and have inflorescences of paired spikelets. Previous studies of the chloroplast gene ndhF and the nuclear gene GBSSI identified numerous mutations that distinguish genera of the tribe but do not indicate relationships among them; the deep branches of the trees are quite short. The authors add newly collected data from phytochrome B to the data from the other two genes. The same pattern holds, with very short branches along the backbone of the tree indicating that the tribe resulted from rapid radiation. The phylogeny shows a single origin of a disarticulating rachis, which is a synapomorphy for the tribe. Strong support is found for a core Andropogoneae that includes Andropogon, Bothriochloa, Capillipedium, Cymbopogon, Dichanthium, Heteropogon, Hyparrhenia, and Schizachyrium and support for its relationship with an expanded Saccharinae that includes Microstegium. The combined data reject the monophyly of subtribes Andropogoninae and Anthistiriinae and provide evidence that subtribes Sorghinae, Saccharinae, and Rottboelliinae are paraor polyphyletic, as is the traditional Maydeae. A relationship with Zea and Tripsacum is indicated for Elionurus, while Chionachne and Phacelurus are shown to diverge early in the history of the tribe. Arundinella hirta and Arundinella nepalensis can be included in an expanded Andropogoneae. Thus, the earlier findings of (Spangler et al., 1999) are confirmed.

(Spangler et al., 1999) redraw the scheme of (Hackel, 1889), it provides insight in the complex system of the tribe Andropogoneae.

54

Fig. 11 Diagram of classification of the Andropogoneae, as proposed by (Hackel, 1889), redrawn by (Spangler et al., 1999). Subtribes are denoted by dashed lines with subtribal names in bold boxes. Arrows show putative direction of evolution and size of arrow indicates certainty of relatedness. 55

(Spangler et al., 1999) used DNA sequences of the chloroplast gene ndhF to estimate the phylogeny of the grass tribe Andropogoneae. Previous hypotheses of relationships in the tribe were based on cytological and morphological characters such as the presence/absence of awns, monoecy vs. andromonoecy, or inflorescence characteristics. Classifications were subsequently proposed based on those ideas of relationships, and these are examined in the context of the molecular data. Some of the molecular data are summarized in the following two figures by (Spangler et al., 1999):

Note that the monophyly of the subtribus Sorghinae is not supported by this analysis, but this might have been influenced by the exclusion of Sorghastrum and other genera. For more interpretation read the extensive comments in (Spangler et al., 1999). We concentrate in this report on the genus Sorghum. The basics of the systematics of the relatives of Sorghum is made clear in this graph: There are good relationships on the molecular and the morphological level within the genera of the Andropogoneae, which opens great breeding potential, as will be shown in the chapters on breeding and evolutionary dynamics of Sorghum.

It is clear, that Saccharum is closer to S. bicolor than to Zea mays, see (Nair et al., 2006) in chapter on Sorghum breeding.

56

Fig. 12 One of 5661 equally most parsimonious trees based on ndhF sequences. Branch lengths are indicated above each branch and decay values listed below nodes. Decay values of zero indicate that the node collapses in the strict consensus. Letters indicate nodes affected by the inclusion of sex expression data relative to the strict consensus tree. See Results for explanation. Fig. 3b. The strict consensus of all equally most parsimonious trees. Bootstrap values are above resolved branches. The arrow marks the clade comprising tribe Andropogoneae. From (Spangler et al., 1999) 57

Fig. 13 The current classification scheme for the taxa sampled on one of the most parsimonious ndhF trees. Branch shading is an optimization of haploid chromosome numbers. Awnless taxa are also indicated for the Andropogoneae. Stiposorghum = S, Parasorghum = P. From (Spangler et al., 1999)

58

3.2. The genus Sorghum Moench Sorghum Moench is a heterogeneous genus. In the taxonomic literature, usually the following sections of the genus are recognized according to (Ejeta & Grenier, 2005; Garber, 1950) (Section = defined group of species within one genus). Stiposorghum, Parasorghum, Sorghum, Heterosorghum and Chaetosorghum. Stiposorghum are characterized by distinct rings of hairs at each culm node, with the awns of Stiposorghum over 65mm long and those of ParaSorghum substantially shorter. The culm nodes are glabrous or hairy in Sorghum, Chaetosorghum and Heterosorghum, but never do the hairs form a nodal ring. The pedicellate spikelets are reduced to subequal glumes in Heterosorghum and unequal glumes in Chaetosorghum. Section Sorghum (incl. S. bicolor) is characterized by better developed pedicellate spikelets that are often male fertile but almost always female sterile. Sessile spikelets in the genus Sorghum are bisexual. This introduction into the Taxonomy and Evolution of Sorghum followes closely some previous publications: (De Wet, 1978) and (Doggett, 1988) The wider taxonomic range besides Sorghum has been examined genetically by (Dillon et al., 2001; Dillon et al., 2004). They used ribosomal ITS to determine phylogenetic relationships within Sorghum. The result can be summarized in the figure below. The authors conclude: The morphological, cytological and genetic information presented for all Sorghum species clearly indicates that classification problems exist within this genus although there is strong bootstrap support for maintaining them on the genus level.

Further details of genetic analysis needs to be done to determine the relationship between S. angustum and S. bicolor. The genetic similarities and similar morphologies of the Chaetosorghum and Heterosorghum species indicate that a separate sectional status is not justified, and they should be combined to form one new section. Problems are also seen in the separation of Parasorghum and Stiposorgum as sections: Callus structure and articulation joint are not very useful characters. As there are still molecular data supporting their separation, morphology needs to be re-assessed.

3.2.1. Genetics within the genus Sorghum Basically, there seems to be widespread incompatibility between cultivated Sorghums. (Hodnett et al., 2005) used in a compatibility study the following cultivated Sorghums from different sub- genera: Chaeto-Sorghum: S. macrospermum Garber Para-Sorghum: S. leiocladum (Hack.) C.E. Hubb. S. matarankense Garber and Snyder S. nitidum (Vahl) Pers. S. purpureo-sericeum (A.Rich.) Aschers. and Schweinf. Stipo-Sorghum: S. amplum Lazarides S. angustum S.T.Blake S. brachypodum Lazarides S. ecarinatum Lazarides S. interjectum Lazarides S. intrans F.Muell. ex Benth. S. plumosum (R.Br.) P.Beauv. 59

S. timorense (Kunth) Buse Eu-Sorghum: S. bicolor (L.) Moench

(Hodnett et al., 2005) determined the reason(s) for reproductive isolation between Sorghum species. The study utilized 14 alien Sorghum species and established that pollen-pistil incompatibilities are the primary reasons that hybrids with Sorghum are not obtained. The alien pollen tubes showed major inhibition of growth in Sorghum pistils and seldom grew beyond the stigma. Pollen tubes of only three species grew into the ovary of Sorghum. Fertilization and subsequent embryo development were not common. Seeds with developing embryos aborted before maturation, apparently because of breakdown of the endosperm. In conclusion, the primary reason why interspecific hybridization does not occur in Sorghum is that growth of alien pollen tubes is inhibited in Sorghum pistils. For three species where limited fertilization and embryo formation occurred, the endosperm in immature seed aborted and no viable seed were produced. Thus pollen - pistil interactions and post-fertilization events are the reasons why hybrids have not been produced.

An understanding of the biological nature of the incompatibility system(s) that prevent hybridization and/or seed development is necessary for understanding potential gene flow and also for the successful hybridization and introgression between Sorghum and divergent Sorghum species for breeding purposes.

(Guo et al., 1996) summarize the relationships within the genus Sorghum, giving an instructive PCA diagram:

60

Fig. 14 Plot of the multiple correspondence analysis of 41 Sorghum accessions showing the distribution of species clusters. bicolor laxiflorum (Heterosorghum) and S. nitidum (Parasorghum) appeared as outliers, and species in section Eusorghum appeared to fall into four groups: (i) sweet Sorghums and shattercane; (ii) S. nigricans, S. hewisonii, S. miliaceum, and S. aethiopicum, which represent S. verticilliflorum, S, sudanese, S. halepense, and 21 inbreds, kaoliangs, and land races (they had the same number of restriction fragments and cannot be separated from S. aethiopicum), (iii) S. propinquum, a diploid rhizomatus species; and (iv) S. niloticum, S. controversum, S. arundinaceum, C-401 and 5-27, and S. almum. From (Guo et al., 1996) From the phylogenetic tree (Fig. above), the accessions can be divided into three major groups, one consisting of S. laxiflorum, the second of S nitidum, and the third including all Eusorghum species. The Eusorghum species can be further divided into four groups: one group including S. sudanense (S. bicolor), S. halepense, S. aethiopicum, S. verticilliflorum, S. hewisonii, S. nigricans, the Kaoliangs, and the 21 cultivated Sorghums from the U.S.A. and China; the second group containing S. miliaceum, a tetraploid; the third comprising C-401 and 5-27, S. niloticum, S. arundinaceum, S. almum, and S. controversum; and the fourth with only one species, S. propinquum. From the phylogenetic tree, S. laxiflorum and S. nitidum formed separate lines of evolution and were distantly related to the accessions in section Eusorghum. The phylogenetic analysis was congruent with that of multiple correspondence in showing clusters of closely related accessions and it was able to distinguish accessions at the species, subspecies, and even race level.

The close relationships identified between S. bicolor and several wild species caused gene flow over centuries, but they have also exciting prospects in utilizing this diverse gene pool for improving Sorghum production.

61

One example, why the segregation of Heterosorghum is not justified, is given by (Wu, 1993): Sorghum laxiflorum (2n = 40), a single tetraploid species accommodated in the subgenus Heterosorghum, is a clearly marked species with advanced characteristics shown by its absent lemma pedicellate spikelet and a well-developed awn, whereas the cultivated Eu-Sorghum species S. bicolor (2n = 20) possesses primitive characteristics, such as its persistent lemma pedicellate spikelet and a poorly developed awn. The former is consequently considered to be more advanced than the latter in the line of evolutionary development. Comparative studies of the chromosome complements between S. laxiflorum and S. bicolor show karyotypes that are essentially alike, except that the former is a tetraploid and the latter is a diploid. The meiotic behavior of S. laxiflorum is regular, pairing so that 20 bivalents are formed, resulting in good fertility. Based on these morphological and cytological observations, (Wu, 1993) speculates that the Heterosorghum species is probably an allopolyploid that might have originated as a result of a cross between two unknown 20-chromosome Eu-Sorghum species, followed by chromosome doubling.

This is also confirmed by (Dillon et al., 2004) and already stated by (Dillon et al., 2001): The genetic similarities and similar morphologies of the Chaetosorghum and the Heterosorghum species indicate that a separate sectional status for each is unnecessary, and they should be combined to form one new section. Definite problems exist in sectional boundaries between Parasorghum and Stiposorghum sections, with the structure of the callus and articulation joint unable to accurately separate these species. These boundaries need to be redefined , as there are no molecular data to support combining these species to form one enlarged section. The analysis of all Sorghum species has confirmed the close relationships between Cleistachne sorghoides and many of the Sorghum species, indicating that Cleistachne sorghoides should probably be included into the genus Sorghum. The grouping of Saccharum within the major S. bicolor clade shows the closer relationship of Saccharum to S. bicolor than to Zea mays. The diverse nature of Erianthus and Saccharum species have been shown, with Erianthus appearing more closely related to Zea mays than to the Saccharum or Sorghum species.

62

Fig. 15 The strict consensus tree of the three most parsimonious trees identified by branch and bound analysis of the twenty five Sorghum species. Bootstrap values are shown above the branches, and node numbers shown below the branches. (Dillon et al., 2001) 63

In another attempt to clarify the situation of the possible evolutionary relationship of the of the genus Sorghum and its relatives, (Price et al., 2005a) have constructed strict consensus trees based on combined nuclear ITS and chloroplast ndhF DNA sequences, determined by flow cytometry, see the figure below:

Fig. 16 The strict consensus tree of six equally parsimonious solutions of 202 steps (CI = 0_792). Bootstrap support (%) for various nodes from 10 000 replications is indicated above the corresponding node. Only bootstrap values of >50 are shown. Numbers in parentheses above each node represent unambiguous nucleotide substitutions. The tree shows bootstrap support, 56 % and 58 % respectively, for lineages consisting of (a) S. bicolor through S. laxiflorum and (b) S. brachypodum through S. nitidum. The tree was rooted using Zea mays. The 2CDNAcontent, 2n chromosome number (in parenthesis), and x = 5 genome size are denoted next to each species. The DNA content for corn (4C = 10_31 pg) is for inbred line Va35 (Laurie & Bennett, 1985). From (Price et al., 2005a) 64

Possibly, phylogenies based on sequence analysis suggest that the Sorghum section designations may not correspond to evolutionary relationships. Recent sequence and systematic data have led (Spangler et al., 1999; Spangler, 2003) to split Sorghum into three genera, Sorghum, Sarga and Vacoparis. However, the limitations of the available sequence- based phylogenies suggest that this reclassification is premature (Price et al., 2005a). The probably most comprehensive summary of the taxonomy of the genus Sorghum is given by (Price et al., 2005a), based on the analysis of DNA genomic analysis, see Fig. above.

Background and Aims: The roles of variation in DNA content in plant evolution and adaptation remain a major biological enigma. Chromosome number and 2C DNA content were determined for 21 of the 25 species of the genus Sorghum and analysed from a phylogenetic perspective.

Methods: DNA content was determined by flow cytometry. A Sorghum phylogeny was constructed based on combined nuclear ITS and chloroplast ndhF DNA sequences.

Key Results: Chromosome counts (2n = 10, 20, 30, 40) were, with few exceptions, concordant with published numbers. New chromosome numbers were obtained for S. amplum (2n = 30) and S. leiocladum (2n = 10). 2C DNA content varies 8.1-fold (1.27– 10.30 pg) among the 21 Sorghum species. 2C DNA content varies 3.6-fold from 1.27 pg to 4.60 pg among the 2n = 10 species and 5.8-fold (1.52–8.79 pg) among the 2n = 20 species. The x = 5 genome size varies over an 8.8-fold range from 0.26 pg to 2.30 pg. The mean 2C DNA content of perennial species (6.20 pg) is significantly greater than the mean (2.92 pg) of the annuals. Among the 21 species studied, the mean x = 5 genome size of annuals (1.15 pg) and of perennials (1.29 pg) is not significantly different. Statistical analysis of Australian species showed: (a) mean 2C DNA content of annual (2_89 pg) and perennial (7.73 pg) species is significantly different; (b) mean x = 5 genome size of perennials (1_66 pg) is significantly greater than that of the annuals (1.09 pg); (c) the mean maximum latitude at which perennial species grow (-25.4 degrees) is significantly greater than the mean maximum latitude (-17.6) at which annual species grow. The sequence-based phylogenetic trees of Spangler et al. (1999) and that are presented in Fig. 3 leave unresolved the base chromosome number of the genus Sorghum. Spangler et al. (1999) showed chromosome numbers in an ndhF tree of 35 Andropogoneae species that included 12 species of Sorghum. There were two inaccuracies regarding chromosome numbers of Sorghum species in their report. First, they used a chromosome number of n = 10 for S. leiocladum. Here a chromosome number of n = 5 is reported for this assumed without documentation that its chromosome number was n = 10. When this species is removed and a correct chromosome number for S. leiocladum is used, the Spangler et al. (1999) tree has Sorghum split into two branches diverging from a node with equivocal chromosome numbers. One branch contains Sorghum with chromosome numbers of 2n = 20, 40. The other branch contains the 2n = 10 chromosome species.

Conclusions: The DNA sequence phylogeny splits Sorghum into two lineages, one comprising the 2n = 10 species with large genomes and their polyploid relatives, and the other with the 2n = 20, 40 species with relatively small genomes. An apparent phylogenetic reduction in genome size has occurred in the 2n = 10 lineage. Genome size evolution in the genus Sorghum apparently did not involve a ‘one way ticket to genomic obesity’ as has been proposed for the grasses.

3.2.1. Section Sorghum within the genus Section Sorghum (in older literature called Eu-Sorghums, that means the true Sorghums) includes cultivated grain Sorghum, a complex of closely related annual taxa from Africa, and a complex of perennial taxa from southern Europe and Asia. This section has received considerable attention from systematists (Snowden, 1936, 1955). As is true for most domesticated taxa, the genus is over-classified, 65 there are too many micro-taxa related to traditional names. The perennial complex is usually recognized to include four species, spontaneous taxa are divided among 17 species, and (Snowden, 1936) recognized 31 domesticated species. Cultivated and spontaneous annual taxa are all interfertile, perennial polyploid taxa cross readily to produce fertile hybrids, and diploid perennial taxa are interfertile with all diploid annual taxa. (De Wet, 1978) summarizes experimental data over two decades of biosystematic studies in Sorghum. The Eu-Sorghums originate from Africa and Asia with 2n=20 and 40 chromosomes, and the known progenitor of cultivated S. bicolor is S. arundinaceum , (De Wet & Harlan, 1971; Doggett, 1976; Duvall & Doebley, 1990).

New research supports the concept of a reduced number of species within the Section Sorghum (which is the only section summarized here). Three species are recognized, two are rhizomatous taxa: S. halepense and S. propinquum, and the large and complex species S. bicolor to include all annual wild, weedy and cultivated taxa. (Dillon et al., 2004) gives the latest interpretation about Eu-Sorghum: The species S. bicolor includes all annual taxa of the section Sorghum as recognized by (Snowden, 1936, 1955). It comprises an extremely variable complex of cultivated taxa, including a widely distributed and ecologically variable wild African complex, and stabilized weedy derivatives originating from introgression between domesticated grain Sorghums and their closest wild relatives. The lack of genetic barriers between the below enumerated taxa indicates that they basically belong to a single species (De Wet & Huckabay, 1967) A summarizing table is given below from (De Wet & Huckabay, 1967) is given in the chapter on S. bicolor.

The plethora of traits, micro-species and putative hybrids can be reduced for practical purposes and with a strict view on agriculture to 6 races:

Cultivated grain Sorghums belong to Sorghum bicolor and were divided into the Guineensia, Nervosa, Bicoloria, Caffra and Durra complexes by (Snowden, 1936). Genetic diversity of those groups have been studied by (Menz et al., 2004), but the results were mixed, clusters obtained only fitted to the pedigree information after some adaptation of the marker systems used. For more comments see under Sorghum bicolor.

A modern monograph, based on sturdy molecular analysis is still lacking, but it will be a major undertaking involving many laboratories and scientists, since it needs to take account of hundreds of accessions with dozens of doubtful and local names – as said before, this is sometimes the destiny of taxonomists revising overclassified groups.

66

S. propinquum is reputedly a perennial rhizomatous form of S. bicolor (Chittenden et al., 1994; Doggett, 1976; Sun et al., 1994).

Sorghum (x) halepense (Johnsongrass) is derived from a doubling of the chromosomes in a natural cross between S. arundinaceum and S. propinquum. Still, and rightly so, Sorghum halepense is recognized as a species due to its worldwide establishment, originating from the temperate regions of the Mediterranean region eastwards.

Sorghum x drummondii originated from a natural cross between S. bicolor and S. arundinaceum. (De Wet & Harlan, 1971; Doggett, 1976; Duvall & Doebley, 1990)

Sorghum x almum is the natural cross between S. bicolor and S. halepense (Doggett, 1976) having been studied in populations from South America.

The close genetic relationships and inter-crossability between the Eu-Sorghum species is well known (Chittenden et al., 1994; Magoon & Shambuli.Kg, 1961; Paterson et al., 1995b; Stenhouse et al., 1997; Wu, 1979).

(Dillon et al., 2004): ITS1 analysis showing a single resolved lineage in Sorghum. Aligned ITS1 sequences were 252 bp in length, and contained 47 parsimony informative characters that on maximum parsimony analysis generated 60 trees of length 127 and consistency index (CI) of 0.764. The strict consensus of the 60 trees is shown in Fig. 1a, with the bootstrap support for each clade shown above the branches. A single lineage (A) was resolved with moderate bootstrap support (70%) that contained the Eu- Sorghum species (clade B, 100% bootstrap) and the Australian natives S. laxiflorum and S. macrospermum (clade C, 100% bootstrap). NdhF consensus tree showing four lineages. The sequence length of the aligned chloroplast ndhF data set was 2014 bp with only 26 parsimony informative characters. One hundred trees of length 92 and CI of 0.628 were generated by maximum parsimony analysis, with the strict consensus tree generated from these 100 trees showing four lineages (Fig. 1b). Clade B retains the Eu-Sorghum species as in Fig. 1a. Sorghum laxiflorum and S. macrospermum form clade C and includes the Australian native species S. nitidum (bootstrap =79%). The African grass Cleistachne sorghoides and S. versicolor (of African origin) form lineage D with 81% bootstrap support. All remaining Sorghum species form lineage E that is very strongly supported by bootstrap data (98%). Internal relationships within lineage E are either weakly supported by bootstrap data (55–59%) or remain unresolved (Fig. 1b below).

67

Fig. 17 The strict consensus trees generated using PAUP branch and bound maximum parsimony analysis. Letters A–D designate clades discussed in the text. 1a. ITS1 data: strict consensus of 60 equally parsimonious trees of 127 steps and consistency index (CI) of 0.764. 1b. ndhF data: strict consensus of 100 equally most parsimonious trees of 92 steps and consistency index (CI) of 0.826. Numbers above branches are percentages of 10,000 bootstrap replicates in which each clade was recovered. Trees were rooted using Zea mays. Letters in parenthesis indicate taxonomic sections within Sorghum where P=Parasorghum, S=Stiposorghum, C`=Chaetosorghum, H=Heterosorghum and E=Eu-Sorghum. (Dillon et al., 2004) 68

More details on the evolution of the genus Sorghum are given in the chapter on evolutionary dynamics of the genus Sorghum.

3.2.2. Summary taxonomy and systematics of Sorghum Sorghum Moench is a heterogeneous genus. In the taxonomic literature, usually the following sections of the genus are recognized according to (Ejeta & Grenier, 2005; Garber, 1950) (Section = defined group of species within one genus) Stiposorghum, Parasorghum, Sorghum, Heterosorghum and Chaetosorghum. Stiposorghums are characterized by distinct rings of hairs at each culm node, with the awns of Stiposorghum over 65mm long and those of Parasorghum substantially shorter. The relationships can be summarized as follows in a figure below: (Ejeta & Grenier, 2005), with the caveats given by (Dillon et al., 2001; Dillon et al., 2004) above.

Fig. 18 The Sorghum taxa according to (De Wet, 1978) The genus Sorghum is divided into five major subgenera in the Poaceae family. Species Sorghum is the co genitor of the crop-wild-weed complex in the genus Sorghum. (Ejeta & Grenier, 2005) It is important to cite (Hodnett et al., 2005): The primary reason why interspecific hybridization does not occur in Sorghum is that growth of alien pollen tubes is inhibited in Sorghum pistils. For three species where limited 69 fertilization and embryo formation occurred, the endosperm in immature seed aborted and no viable seed were produced.

The most important summary comes from (Price et al., 2005a): Conclusions The DNA sequence phylogeny splits Sorghum into two lineages, one comprising the 2n = 10 species with large genomes and their polyploid relatives, and the other with the 2n = 20, 40 species with relatively small genomes. An apparent phylogenetic reduction in genome size has occurred in the 2n = 10 lineage. Genome size evolution in the genus Sorghum apparently did not involve a ‘one way ticket to genomic obesity’ as has been proposed for the grasses.

Three species of Sorghum are recognized, two are rhizomatous taxa: S. halepense and S. propinquum, and the large and complex S. bicolor to include all annual wild, weedy and cultivated taxa. (Dillon et al., 2004) gives the latest interpretation about Eu-Sorghum: The Eu-Sorghums originate from Africa and Asia with 2n=20 and 40 chromosomes, and the known progenitor of cultivated S. bicolor is S. arundinaceum , (De Wet & Harlan, 1971; Doggett, 1976; Duvall & Doebley, 1990).

The species S. bicolor includes all annual taxa of the section Sorghum as recognized by (Snowden, 1936, 1955). It comprises an extremely variable complex of cultivated taxa, including a widely distributed and ecologically variable wild African complex, and stabilized weedy derivatives originating from introgression between domesticated grain Sorghums and their closest wild relatives. Cultivated grain Sorghums were divided into the Guineensia, Nervosa, Bicoloria, Caffra and Durra complexes by (Snowden, 1936). Genetic diversity of those groups have been studied by (Menz et al., 2004), but the results were mixed, clusters obtained only fitted to the pedigree information after some adaptation of the marker systems used. A modern monograph, based on sturdy molecular analysis is still lacking, but it will be a major undertaking involving many laboratories and scientists.

S. propinquum is reputedly a perennial rhizomatous form of S. bicolor (Chittenden et al., 1994; Doggett, 1976; Sun et al., 1994).

Sorghum (x) halepense (Johnsongrass) is derived from a doubling of the chromosomes in a natural cross between S. arundinaceum and S. propinquum. Still, and rightly so, Sorghum halepense is recognized as a species due to its worldwide establishment, originating from the temperate regions of the Mediterranean region eastwards.

Sorghum x drummondii originated from a natural cross between S. bicolor and S. arundinaceum. (De Wet & Harlan, 1971; Doggett, 1976; Duvall & Doebley, 1990)

Sorghum x almum is the natural cross between S. bicolor and S. halepense (Doggett, 1976).

The close genetic relationships and inter-crossability between the Eu-Sorghum species are well known (Chittenden et al., 1994; Magoon & Shambuli.Kg, 1961; Paterson et al., 1995b; Stenhouse et al., 1997; Wu, 1979).

3.2.3. Bibliographic references taxonomy and systematics of Sorghum http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Taxonomy-Systematics-20061123.pdf

70

3.3. Sorghum halepense (L.) Pers.

3.3.1. Taxonomic description Sorghum halepense (L.) Pers., Synops. 1, 101: 1805, commonly called Johnsongrass, is based on Holcus halapensis L and other taxa in many floras, the synonymy is discussed in detail in (Snowden, 1955).

Plants perennial with well developed creeping rhizomes. Culms slender to robust, erect, mostly simple, 0.5-3.5m tall and up to 2cm wide near the base. Leaf blades linear up to 90cm long and 4cm wide, more or less smooth. Sheaths essentially glabrous. Inflorescence a large and more or less open panicle 10-60cm long and 5-25cm wide, with the lower branches up to 8cm long in small inflorescences and up to 25cm long in large ones. Panicle branches slender, often somewhat pendulous, usually naked for 2- 5cm from the base, divided upwards. Racemes fragile, 1-5 noded, up to 2.5cm long, pedicels densely ciliate. Sessile spikelets 4.0-6.5mm long, more or less elliptic- lanceolate, almost glabrous to densely hairy, awned or awnless. Glumes coriaceous, lower 7-12nerved, with the wings of the keels ending in minute teeth, forming with the pointed apex a 3-toothed tip; upper glume not 3-toothed. Lemmas ciliate, lower 3-5mm long, upper 3-4mm long, minutely bilobed with usually a 10-16mm long awn. Stamens three. Grains oblong-ovate, 2-3mm long. Pedicellate spikelets male or neuter, 4-7mm long.

Johnsongrass can have ploidy levels ranging from triploid to octoploid, although it is generally accepted to be tetraploid (Hadley, 1958b; Hoangtang & Liang, 1988).

Johnsongrass is one of the worst weeds in the US and other places (Chao et al., 2005; Toler et al., 1996), because of its aggressive tillering down to 1m in the soil, to a lower degree because its seed production (Scopel et al., 1988) and due to its seed shattering caused by a single gene (Paterson et al., 1995b). The seeds can persist up to 10 years in the soil (Zeller, 2000). (Vila et al., 2004) warn that measuring competition effect of weedy species alone without evaluating the crop situation could lead to wrong conclusions.

71

Fig. 19 Sorghum halepense. A: base of plant with rhizomes. B: inflorescence, C: ligule D: terminal end of branch showing two sessile and three pedicellate spikelets, E-H: sessile spikelet, E: whole spikelet at anthesis, lateral ventral and dorsal view, F: 72 lemma, G: palea, H: stamens and pistils, I,J: pedicellate spikelet, I: whole spikelet, ventral view, J: lemma, K: lodicules, L: sessile spikelet in fruit, grain enclosed by glumes, M: grain or caryopsis, scale A-B: ½, C-M: x 8. From (Warwick & Black, 1983)

3.3.2. Evolution of Sorghum halepense (Hadley, 1953; Hadley & Mahan, 1956) established the working hypothesis that Sorghum halepense arose as a hybrid between two species with 2n = 20, whose chromosome complements were similar but not identical. (Paterson et al., 1995b) demonstrates that Sorghum halepense is a vigorous hybrid between S. bicolor and Sorghum propinquum with its tillering habit.

It is not clear, whether the tillering characteristics should be derived from three QTLs (Paterson et al., 1995b) or whether it is caused by a single dominant gene (Yim & Bayer, 1997). (Paterson et al., 1995b) found 16QTLs affecting rhizomatousness and tillering of Sorghum, two important factors of weediness with Johnsongrass Sorghum halepense. Variation in regrowth (ratooning) after overwintering was associated with QTLs accounting for additional rhizomatous growth and with QTLs influencing tillering. Vegetative buds that become rhizomes are similar to those that become tillers. See fig. below from (Paterson et al., 1995a).

The alternative would be a single dominant gene, as (Yim & Bayer, 1997) claim: The segregating F2 population from the hybrid between grain Sorghum (S. bicolor), chromosome number doubled, and Johnsongrass (S. halepense) was examined for expression of rhizomes. A rhizome to non-rhizome segregation ratio of 3:1 was observed suggesting a single dominant gene regulation in the rhizome phenotype of S. bicolor and S. halepense. Approximately 72% of the F2 population overwintered and regrew from the rhizomes the following spring. When comparing these data with data in the literature, the rhizome regulating gene in the Sorghum genus may be an incomplete dominant gene requiring a few genes to be expressed additively that result in different degrees of rhizome development in the Sorghum genus.

73

74

Fig. 20 16QTLs affecting rhizomatousness and tillering of Sorghum. The 10 LGs of Sorghum are denoted A-J (12). DNA markers indicated bylines crossing a LG were used in QTL mapping; those indicated by arrows were mapped on a subset of 56 progeny (12), and locations were inferred relative to flanking markers. Chromosomal locations of selected markers in maize (M) (17, 18), rice (R) (18, 19), and wheat (W) (20) are indicated. Maximum-likelihood locations (arrowhead) and 1-lod (box) and 2-lod (whiskers) likelihood intervals for each QTL are to the left of appropriate linkage groups. The pattern within the 1-lod likelihood interval indicates trait. From (Paterson et al., 1995b)

(Dweikat, 2005) write, that previous attempts to hybridize S. bicolor with S. halepense have proven unsuccessful in generating agronomically viable hybrids, particularly using S. bicolor as female in the cross. Maintaining cultivated Sorghum as female, and cytoplasm donor, is important to preventing nuclear cytoplasm incompatibilities in downstream breeding efforts (Kaul, 1988). Early research on hybrids between diploid Sorghum and Johnsongrass (Hadley, 1953, 1958b; Hadley & Mahan, 1956; Hadley & Openshaw, 1980) using S. halepense as female, produced two types of hybrids: 30-chromosome plants that although male-sterile could be backcrossed to the diploid parent, and fertile 40-chromosome plants derived from unreduced female gametes in the diploid parent. The 30-chromosome plants were more strongly rhizomatous. (Hadley & Mahan, 1956) identified seven 20-chromosome backcross plants that were rhizomatous, but most were chlorophyll mutants. The ability to hybridize a cultivated diploid species to a weedy tetraploid, and the production of a fully fertile diploid progeny from that cross, is significant. This was achieved in part, by utilizing the nuclear male sterility gene ms3 and by pollination of about 20,000 florets of cultivated Sorghum. Considered one of the 10 most troublesome weeds, Johnsongrass possesses several extremely desirable plant traits including tolerance to pests, cold and drought. Results obtained in these studies support the hypothesis that S. halepense is a segmental allopolyploid. The fact that it is a perennial, may explain its survival in its early existence as perhaps a highly sterile diploid hybrid between two closely related species or as a relatively unstable polyploid derivative of such a cross. From a plant breeding standpoint, it appears that genes can be transferred from tetraploid S. bicolor to Johnsongrass and that recessive phenotypes can be isolated, which would be improbable if Johnsongrass were a strong allopolyploid.

In earlier studies, interspecific crosses of Johnsongrass cultivated Sorghum and the reciprocal produced mostly sterile triploid F1 plants (Casady & Anderson, 1952; Piper & Kulakow, 1994; Sangduen & Hanna, 1984) or both triploid and fertile tetraploid F1 (Hadley, 1953, 1958b; Hadley & Mahan, 1956) due to unreduced gametes produced by the diploid Sorghum. In the study of (Dweikat, 2005) the resulting F1 was a diploid with a chromosome number of 20. Underlying this phenomenon may be the low frequency production of monohaploid gametes by the Johnsongrass line selected for these crosses. Similar observations were reported in crosses between diploid · allohexaploid millet (Hanna, 1990) .

75

Fig. 21 Phenotypic characterization of the seed (top panel), Panicles (middle Panels) and roots (bottom) of Sorghum (left), Johnsongrass (right), and the F1 hybrid (middle). From (Dweikat, 2005)

Also (Morrell et al., 2005) claim that crop-to-weed introgression has impacted allelic composition of Johnsongrass populations with and without recent exposure to cultivated Sorghum. They report molecular analysis of these traits in a cross between cultivated and wild species of Sorghum that are the probable progenitors of the major weed 76

"Johnsongrass." By restriction fragment length polymorphism mapping, variation in the number of rhizomes producing above- ground shoots was associated with three quantitative trait loci (QTLs). Variation in regrowth (ratooning) after overwintering was associated with QTLs accounting for additional rhizomatous growth and with QTLs influencing tillering. (Hoangtang & Liang, 1988) suggest that cultivated Sorghum (S. bicolor) is a tetraploid species with the genomic formula AAB1B1, and Johnsongrass (Sorghum halepense) is a segmental auto-allo-octoploid, AAAA B1B1B2B2. The model is further substantiated by chromosome pairing in amphiploid plants whose proposed genomic formula is AAAAAA B1B1B1B1 B2B2.

3.3.3. Distribution of Sorghum halepense S. halepense is native of southern Eurasia east to India (Monaghan, 1979) and has been introduced as a highly estimated forage grass in the United States at the beginning of the 19th century. Slender plants with relatively small inflorescences and narrow leaf blades occupy the western range of the species. These specimens are commonly classified as S. halepense while more robust plants with large inflorescences and broader leaf-blades, and which occupy the eastern half of the species range are usually recognized as S. miliaceum. The species has also been introduced as a weed to all warm temperate regions of the world.

In the Americas it has introgressed with grain Sorghum to produce the widely distributed Johnsongrass (Celarier, 1958). In Argentina, derivatives of such introgression were described as S. almum Parodi, Columbusgrass (Parodi, 1943). Sorghum almum has also been used for a while in Australia, but it is now replaced there with a triple hybrid S. bicolor x halepense x verticilliflorum (Zeller, 2000). (Chen et al., 1993) found out that the cytoplasmatic sterile plants contain a slightly smaller 3.7 kb HindIII chloroplast DNA fragment, whereas the fertile plants from S. versicolor, S. almum, S. halepense, and Sorghastrum nutans (Yellow Indiangrass) also have the 3.8 kb fragment, and CMS lines of those taxa studied containing Al, A2 and A3 cytoplasms have the 3.7 kb fragment. (Morden et al., 1990) provide a differentiated discussion on the origin of Sorghum halepense, call for further studies, but conclude that S. bicolor (in the broad sense) is one of the parents of Sorghum halepense.

In fact, Johnsongrass in all its derivatives forms with cultivated Sorghums a complex with feral populations, where hybridization is going forth and back and the resulting populations are often difficult to determine to their origin. This is extensively discussed in the chapter on evolutionary dynamics of the genus Sorghum below.

For this chapter it is enough to give the interpretation of (Ejeta & Grenier, 2005; Warwick & Stewart Jr., 2005): the existence of intermediary forms in most Sorghum growing areas of the warm temperate zones in the Americas offers an empirical evidence for noxious weedy forms arising from continued introgression (exo-ferality) among different Sorghum types.

3.3.4. Summary Sorghum halepense Although S. bicolor (2n = 2x = 20) and S. halepense (2n = 4x = 40 and higher) differ in ploidy, numerous artificial crossing studies have demonstrated that S. bicolor can serve as the pollen parent of triploid and tetraploid hybrids (reviewed in (Warwick & Black, 1983). Naturalized populations of S. halepense occur in many regions where Sorghum is cultivated. In the United States, S. halepense is a very common roadside weed. It also invades cultivated fields of many warm season crops and can reduce agricultural productivity by as much as 45% (McWhorter & Anderson, 1981; Millhollon, 2000). Sorghum halepense is a particular problem in and around Sorghum fields because no known herbicide can selectively eliminate S. halepense without damaging cultivated S. bicolor (Paterson et al., 1995b). The two species frequently grow in close physical proximity and have overlapping flowering periods. Hybridization with cultivated (non-transgenic) S. bicolor has been proposed as a potential cause of increased aggressiveness in the 77 weed (e.g. (De Wet & Harlan, 1975); (Holm et al., 1977). Based on morphological evidence, it has been suggested that ‘Johnsongrass’ in North America may actually be an S. halepense × S. bicolor hybrid (Celarier, 1958; De Wet, 1978), similar to S. almum (Columbus grass) (2n = 4x = 40) of South America, which based on morphological and cytological evidence appears to be an S. halepense × S. bicolor hybrid (Parodi, 1943) (Hadley, 1953). While experimental field studies demonstrate the potential for S. halepense × S. bicolor hybrid formation (Arriola & Ellstrand, 1996) and persistence (Arriola & Ellstrand, 1997), existence of naturally occurring hybrids has not been confirmed beyond the observation of morphological intermediacy (Jarvis & Hodgkin, 1999) but see (Hoang-Tang & Liang, 1988).

The extent of hybrid formation and persistence as been addressed by (Morrell et al., 2005), they confirmed gene introgression from S. bicolor into S. halepense, but also not anticipated was the finding that the majority of S. halepense plants in these populations carry cultivar-specific alleles, and thus appear to be S. halepense × S. bicolor hybrids. Also not anticipated was the frequency of occurrence of cultivar-specific alleles in S. halepense populations far from Sorghum fields and outside the region of the United States where S. bicolor is cultivated. Sorghum halepense has only recently invaded the more northerly portions of its introduced range in the United States and Canada (Warwick, 1990; Warwick & Black, 1983) and it is possible that alleles have been maintained in populations that experienced previous introgressive hybridization. After (Ejeta & Grenier, 2005; Warwick & Stewart Jr., 2005) the existence of intermediary forms in most Sorghum growing areas of the warm temperate zones in the Americas offers an empirical evidence for weedy forms arising from continued introgression (exo-ferality) among different Sorghum types.

3.3.5. Bibliographic References Sorghum halepense http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-halepense-20061123.pdf

3.4. Sorghum propinquum (Kunth) Hitchc. Ling. Sci. J. 7: 249, 1929.

3.4.1. Taxonomic description Based on Andropogon propinquus Kunth.

Plants robust, tufted, perennial with stout rhizomes. Culms up to 5m tall, erect. Leaf blades linear, up to 1m long and 3-5cm wide. Leaf sheaths glabrous. Inflorescences large, open, up to 60cm long, with slender branches, the lower ones up to 25cm long, usually bare of spikelets for as much as 5cm at the base. Racemes fragile, 1-5nided, with slender and ciliated pedicels. Sessile spikelets elliptic-lanceolate, 3.5-5.0mm ling, strigose at the base and above the middle, usually awnless. Glumes crustaceous below and membranaceous near the tip. lower keeled, often ending in minute teeth, upper glume similarly keeled. Lemmas thinly ciliate above the middle, lower 3.5-4.5mm long, upper one about 3mm long, usually awnless. Anthers three. Grains obovate 1.5-2.0mm long. Pedicellate spikelets 4.0-5.5mm long, male, rarely neuter.

Distribution:

Sorghum propinquum occurs in Ceylon and southern India, and from Burma eastward to the islands of south- eastern Asia. Specimens from the Southeast Asia islands usually have smaller (4.0-4.5mm) sessile spikelets than collections from the South Asian mainland (4.5-5.0mm). In the Philippines this species crosses with introduced grain Sorghum, and derivatives of such crosses are obnoxious weeds in parts of Luzon and Mindanao islands. 78

3.4.2. Origin and nature of Sorghum propinquum According to (Dillon et al., 2004) Sorghum propinquum belongs to the six Eu-Sorghum species and it is reputedly a perennial rhizomatous form of S. bicolor (Chittenden et al., 1994; Doggett, 1976; Sun et al., 1994)

This view has already been supported by (Magoon & Shambuli.Kg, 1962): They studied the pachytene morphology of S. propinquum on the basis of chromosome length, centromere position and the distribution of deeply staining regions on the chromosome. The present species which belongs to the subsection Halepensia differs from the species of the subsection Arundinacea in that the nucleolus is organized by the shortest chromosome of the complement while in the latter group the nucleolar chromosome is the longest. The significance of this difference in evaluating the probable role of S. propinquum (2n = 20, but rhizomatous) in the origin of 40-chromosome species has been noted by the authors.

(Chittenden et al., 1994) established the first ''complete'' genetic linkage map of Sorghum section Sorghum comprising ten linkage groups putatively corresponding to the ten gametic chromosomes of S. bicolor and S. propinquum. Although prior cytological evidence suggests that the genomes of these species are largely homosequential, a high level of molecular divergence is evidenced by the abundant RFLP and RAPD polymorphisms, the marked deviations from Mendelian segregation in many regions of the genome, and several species-specific DNA probes. The remarkable level of DNA polymorphism between these species will facilitate development of a high-density genetic map. Limited investigation of a small number of RFLPs showed several alleles common to S. bicolor and S. halepense , Johnsongrass, but few alleles common to S. propinquum and S. halepense, raising questions about the origin of S. halepense.

3.4.3. Hybrids of Sorghum propinquum Hybrids between S. propinquum and S. bicolor are fully fertile and these two taxa evidently belong genetically to one biological species. (Paterson et al., 1995b) demonstrates that Sorghum halepense is a vigorous hybrid between S. bicolor and Sorghum propinquum with its tillering habit.

The natural distribution of Sorghum bicolor, however, is strictly African, and since Sorghum propinquum is spatially isolated from the annual S. bicolor, there are no spontaneous African hybrids between S. propinquum and African wild S. bicolor reported.

S. propinquum as a perennial taxon is recognized as a species by (Dillon et al., 2004). Distribution and morphology indicate affinities with S. halepense rather than S. bicolor although it is clear that there is colinearity between S. propinquum and S. bicolor and hybrids are fully fertile. Hybrids between S. halepense (2n=40) and S. propinquum (2n=20) are characterized by a substantial number of trivalent chromosome configurations during meiotic prophase of microsporogenesis. Cytological behavior is completely regular in hybrids between S. propinquum and S. bicolor. Again fig. 1 in (Ejeta & Grenier, 2005) applies to this situation. (Feltus et al., 2006).and (Ejeta & Grenier, 2005) also agree with the interpretation that Sorghum halepense is actually a hybrid between S. bicolor and Sorghum propinquum. From the view of QTL analysis (Feltus et al., 2006) see a considerable amount of colinearity among the interspecific and intraspecific QTL maps related to hybrids between S. propinquum and S. bicolor and from infraspecific hybrids within S. bicolor. This makes S. propinquum a valuable source for important crop specific beneficial genes.

In a detailed molecular analysis of genes influencing dispersal and persistence of Johnsongrass, (Paterson et al., 1995b) conclude that genes for weediness are probably derived from Sorghum propinquum. Genes responsible for rhizome development and tillering in Sorghum may at least partly account for these traits in other grasses, in which up-regulation of rhizomatousness might improve agricultural productivity. Sorghum provides a facile model for detailed investigation of genes controlling rhizomatousness, a trait important to productivity, quality, and protection of agro-ecosystems. See the fig. and more remarks in the chapter on Sorghum halepense.

79

3.4.4. Genetic maps of Sorghum bicolor and Sorghum propinquum (Lin et al., 1999) constructed a large insert Sorghum propinquum BAC library (BAC = bacterial artificial chromosome) to analyze the physical organization of the Sorghum genome and to facilitate positional cloning of Sorghum genes and QTLs associated with the early stages of grain crop domestication. This library was established from 12 different ligations using high-molecular-weight DNA generated from either one cycle or two cycles of size selection. The S. propinquum BAC library can be used for positional cloning of genes associated with domestication. The genes might be identified by transforming candidate BAC clones via the BIBAC vector (BIBAC = Binary BAC), or by transforming candidate cDNAs identified using candidate BAC clones, into cultivars for gene expression.

(Feltus et al., 2006) have aligned genetic maps derived from two Sorghum populations that share one common parent (S. bicolor L. Moench accession BTx623) but differ in morphological and evolutionarily distant alternate parents (S. propinquum or S. bicolor accession IS3620C). A total of 106 well-distributed DNA markers provide for map alignment, revealing only six nominal differences in marker order that are readily explained by sampling variation or mapping of paralogous loci. The authors also report a total of 61 new QTLs detected from 17 traits in these crosses. Among eight corresponding traits (some new, some previously published) that could be directly compared between the two maps, QTLs for two (tiller height and tiller number) were found to correspond in a non-random manner (P<0.05). For several other traits, correspondence of subsets of QTLs narrowly missed statistical significance. In particular, several QTLs for leaf senescence were near loci previously mapped for ‘stay-green’ that have been implicated by others in drought tolerance. These data provide strong validation for the value of molecular tools developed in the interspecific cross for utilization in cultivated Sorghum, and begin to separate QTLs that distinguish among Sorghum species from those that are informative within the cultigen (S. bicolor). Collectively, these data provide strong support for the transferability of molecular tools between interspecific and intraspecific crosses, and begin the process of separating QTLs that contribute to morphological and physiological divergence among Sorghum species from those that contribute to diversity within the cultigen (S. bicolor). Such a ‘categorization’ of QTLs is of value for setting priorities in ongoing studies of botanical diversity, and crop improvement, respectively.

80

Fig. 22 Alignment of two Sorghum genetic maps and locations of putative QTLs. Ten chromosomal genetic maps for BTx623/IS3620C (top line) and BTx623/S. propinquum (bottom line) populations are shown as horizontal thin grey lines connected by common markers. The middle two horizontal lines represent bridge maps derived from common markers and centimorgan positions from high-density genetic maps (see Materials and methods). QTLs are shown as thick horizontal lines with BTx623/IS3620C QTLs above the chromosome (labeled with txs) and BTx623/S. propinquum and shown below the chromosome (labeled as uga). Alignments for the entire genome are available from Gramene. (Feltus et al., 2006) 81

Fig. 23 continued, see caption in Fig. 1 (Feltus et al., 2006)

82

3.4.5. Summary Sorghum propinquum is a species related to S. bicolor in the broad sense. According to (Dillon et al., 2004) Sorghum propinquum belongs to the six Eu-Sorghum species and it is reputedly a perennial rhizomatous relative of S. bicolor This view is also supported by the pachytene morphology of S. propinquum on the basis of chromosome length, centromere position and the distribution of deeply staining regions on the chromosome. Sorghum propinquum which belongs to the subsection Halepensia differs from the species of the subsection Arundinacea in that the nucleolus is organized by the shortest chromosome of the complement while in the latter group the nucleolar chromosome is the longest. Hybrids between S. bicolor and Sorghum propinquum are fully fertile. Molecular data have shown that there is good transferability from Sorghum propinquum for interesting genetic sites to the cultigens, and breeders will use these new insights.

3.4.6. Bibliographic references Sorghum propinquum http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-propinquum-20060806.pdf

3.5. Sorghum bicolor (L.) Moench, Meth. Pl. 207, 1794.

3.5.1. Taxonomy of Sorghum bicolor in general Sorghum bicolor (L.) Moench, Methodus 207. 1794 ≡ Holcus sorghum L., Sp. Pl. 2: 1047. 1753 ≡ H. bicolor L., Mant. Pl. 2: 301. 1771 ≡ Andropogon sorghum (L.) Brot., Fl. Lusit. 1: 88. 1804 – Lectotype (designated by Poilecot [for H. bicolor] in Boissiera 56: 509. 1999 and Davidse [for H. sorghum] in Taxon 49: 251. 2000): Herb. Clifford: 468, Holcus 1 (BM) (summarized according to (Wiersema & Dahlberg, 2007) The name is based on Holcus Sorghum L. Linnaeus described in 1753 a Sorghum collection under the name Holcus: it delineated several species of Holcus, some of which have been moved to the tribe Avenae, where the generic name Holcus now belongs. Adanson used the name Sorghum as an alternative for Holcus. In 1794 Moench distinguished the genus Sorghum from the genus Holcus (Celarier, 1958). (Davidse & Turland, 2001) proposed to reject the name Holcus saccharatum and give extensive justification, providing complete references for this decision.

Some basic remarks about the taxonomy Sorghum bicolor and cultivated plants in general: The synonymy of this species, as the taxonomic knowledge was developed in the time of Snowden, is extensively and fully discussed in (Snowden, 1936), see the table fig. 24 below. In his comprehensive taxonomic review, Snowden distinguished 7 weedy, 13 wild and 28 cultivated species, most of them attached with numerous varieties and forms in order to cope up with the enormous variability of the bicolor group. A refinement of Snowden’s classification was developed by Jakuševskij (1969), and is still used in some parts of the world (Fritsch, & al. 2001), especially those of the former Soviet Union. Difficulties in applying the complicated Snowden classification and the lack of genetic barriers between these taxa however led (De Wet, 1978; De Wet & Huckabay, 1967) to radically treat all of Sorghum 83 subgenus Sorghum, including the perennial members in a “lumper” approach within a single species, S. bicolor (L.) Moench with 5 basic races and 10 hybrid races, a view which also has been adopted (certainly for practical reasons) by major treatments such as (Doggett, 1988).

These views need to be revised from this point. A nomenclatural revision has recently been published by (Wiersema & Dahlberg, 2007) taking into account relevant literature, and in this text we will keep to those decisions, but also we will keep the colloquial names for synonyms for easier communication. Nevertheless, in a later revision the numerous studies published in recent years (1989 – 2010) using modern methods of molecular analysis and numerical taxonomy (see the chapters on numerical and molecular taxonomy) needs to be incorporated. This will be a major undertaking, but it still needs to be professionally linked to the classic methods of taxonomy checking out the type specimens, reinforcing this part with molecular analytical methods which are now available. (Ammann, 1986; Cornelissen et al., 2003; Hetterscheid & Brandenburg, 1995; Hetterscheid et al., 1999; Hetterscheid et al., 1996; Zeven, 1991). As Hetterscheid and Brandenburg point out, this does not mean that the revision of such a complex group of cultivars and wild relatives as S. bicolor should strictly follow the old and often cumbersome paths of classic nomenclature for wild plants – on the contrary, there are many reasons to follow a new scheme with a new unit called the ‘culton’ proposed by the two authors in order to facilitate a new monograph of such groups – since the sole purpose of any taxonomic activity is to produce a system which is falsifiable and thus reproducible in order to establish a reliable base for future work of practical use such as managing core seed collections and organize reasonable and practicable breeding programs. This includes the use of modern molecular methods for the analysis of dried herbarium specimens – specimens, which are often invaluable proof material of classic authors, including type- iso- and lectotype material. This has been shown by (Drabkova et al., 2002; Savolainen et al., 1995). The synthesis of modern breeding and traditional knowledge (Ammann, 2006, 2008; Ammann, 2009; Ammann & van Montagu, 2009) may ultimately require also the above described culton systematics, uniting farmers and plant systematists knowledge.

3.5.2. Morphological description of Sorghum bicolor (L.) Moench Plants annual, often tillering. Culms erect, slender to robust, 0.5m to over 5m tall, branched or unbranched at maturity. Leaf blades linear to linear-lanceolate, glabrous or hairy, up to 1m long and 10cm wide; sheaths glabrous to hirsute or pubescent. Inflorescence an open or contracted panicle, 5- 50cm long and 3-30cm wide, with the branches stiffly ascending or spreading and sometimes pendulous and the lower panicle branches often about half as long as the panicle, usually bare for up to 10cm from the insertion with the primary axis. Racemes fragile, tough or tardily disarticulating at maturity, one-to- several-noded. Sessile spikelets elliptic-lanceolate to obovate, up to 6mm long, glabrous to densely hirsute or pubescent. Glumes coriaceous to membranaceous, shorter than, as long as or longer than the spikelet, usually winged on the keels, sometimes three-toothed. Lemmas of sessile spikelets hyaline, moderately ciliate, lower up to 6mm long, the upper one usually somewhat shorter, often awned. Anthers three. Sessile spikelets bisexual, pedicellate spikelets male or neuter.

84

S. bicolor (L.) Moench is the fifth most important cereal grown worldwide in terms of both production and area planted (FAO, 2004). Like millet and fonio, Sorghum is genetically suited to hot and dry agro ecologies where it is difficult to grow most food grain. S. bicolor, therefore, is a pillar of food security in the semi-arid zones of Western and Central Africa.

3.5.3. Original taxonomic work on wild and cultivated Sorghum bicolor The classic papers of (De Wet et al., 1972; De Wet, 1978; De Wet & Harlan, 1971, 1975; De Wet et al., 1970; De Wet & Huckabay, 1967) contain a lot of useful information about wild Sorghums. Although written still under the impression of the Snowden taxonomy of small units, they show the full complexity of the taxonomy. (Snowden, 1936, 1955) finally subdivided this section into subsection arundinacea with 31 cultivated species (series sativa) and 17 wild or weedy species (series spontanea), and subsection halepensia which includes 4 weedy species. These 52 taxa were combined into a single species by (De Wet & Huckabay, 1967). They studied distribution ranges, adaptive polymorphism, and morphological variability of the 17 taxa belonging to the spontanea in an effort to trace the origins of cultivated Sorghums.

The species S. bicolor (L.) Moench includes all annual taxa of the section Sorghum as recognized by (Snowden, 1936, 1955). It comprises an extremely variable complex of cultivated taxa, including a widely distributed and ecologically variable wild African complex, and stabilized weedy (feral) derivatives originating from introgression between domesticated grain Sorghums and their closest wild relatives. The lack of genetic barriers between the enumerated taxa (see table below) indicates that they basically belong to a single species in the classic lumper sense (De Wet & Huckabay, 1967). For their study they used only type specimens and original collections that fitted Snowden’s type descriptions in detail. For more comments see chapter on numerical taxonomy.

There is room for interpretation about this selection: There is some logic in the conclusions that the numerical analysis follows closely the concepts of classic taxonomy. On the other hand, more comprehensive field sampling could well reveal a somehow different taxonomic picture - closer to reality. But for the present time, we have still reason to follow the advice of the most experienced classic systematists and the taxonomic concept they established. Studies done with modern numerical analysis or with highly refined molecular analysis should follow a careful selection strategy for the material analysed, the best methods cannot come up for errors committed in sampling.

85

Fig. 24 The table of (De Wet & Huckabay, 1967) actually summarizes the concept of the ‘Snowdenian species’. Genetic analysis of recent years show reasonable compatibility with the present day concepts of race denomination, except that geographical differentiation of the same race may have been underestimated (Tao et al., 1993). See the dendrogram related to this table in chapter on numerical taxonomy, and from group C of this dendrogram the figure constructed by (Doggett, 1988) on the clusters of the cultivated ‘Snowdenian species’ D-K.

86

As (De Wet et al., 1970) pointed out already, the complex of the wild S. bicolor taxa has no clear delimitation in morphology, which is nicely illustrated by the two figures below:

Fig. 25 Scatter diagram indicating the range of morphological variability of the wild varieties of S. bicolor, and some segregating populations of hybrids among them. The vertical axis represents peduncle-sheath length in cm, and the horizontal axis indicates primary axis length of the inflorescence in cm. Each dot represents the average value of 10 plants studied belonging to a natural collection. The morphological values of the hybrids are based on measurements of individual plants. From (De Wet et al., 1970).

87

Fig. 26 Pictorial scatter diagram relating African wild and cultivated Sorghums. Panicle length against pedicelled spikelet length demonstrate clearly hybrid introgressions between wild and cultivated Sorghum. Explanation of the dot graphs given above in the figure. From (Doggett & Majisu, 1968) given in (Doggett, 1988).

There are three complexes recognized as subspecies:  Sorghum bicolor subsp. bicolor (see 3.5.3.)  Sorghum bicolor subsp. drummondii (see 3.5.4.) and  Sorghum bicolor subsp. verticilliflorum (see 3.5.5.)

3.5.4. Sorghum bicolor subsp. bicolor Sorghum bicolour subsp. Bicolour is based on S. bicolor (L.) Moench 1794. Meth. Pl. 207. 1794. Here all the diversity of cultivated Sorghums is included.

88

The 28 species of grain Sorghum recognized by Snowden are listed in (De Wet, 1978):

The species of grain Sorghum recognized by Snowden are listed in alphabetical order for convenience. Sorghum ankolib Stapf in Prain, Fl. Trop. Afr. 9: 135. 1917. S. basutorum Snowden, Kew Bull. 1935: 232. 1935. S. bicolor (L.) Moench, Meth. Pl. 207. 1794. S. caffrorum Beauv., Agrost. 131, 178. 1812. S. caudatum Stapf in Prain, Fl. Trop. Afr. 9: 131. 1917. S. cernuum Host. Gram, Austr. 4, 2, t. 3. 1809. S. conspicuum Snowden, Kew Bull. 1935: 226. 1935. S. coriaceum Snowden, Kew Bull. 1935: 240. 1935. S. dochna (Forrsk.) Snowden, Kew Bull. 1935: 234. 1935. S. dulcicaule Snowden, Kew Bull. 1935: 247. 1935. S. durra Stapf in Prain, Fl. Trop. Afr. 9: 129. 1917. S. elegans (Koern.) Snowden, Kew Bull. 1935: 238. 1935. S. exsertum Snowden, Kew Bull. 1935: 230, 1935. S. gambicum Snowden, Kew Bull. 1935: 229. 1935. S. guineense Stapf in Prain, Fl. Trop. Afr. 9: 123. 1917. S. margaritiferum Stapf in Prain, Fl. Trop. Afr. 9: 125. 1917. S. melaleucum Stapf in Prain, Fl. Trop. Afr. 9: 134. 1917. S. mellitum Snowden, Kew Bull. 1935: 225. 1935. S. membranaceum Chiov., Monogr. Rapp. Col. Rome 19: 23-34. 1912. S. miliiforme (Hack.) Snowden, Kew Bull. 1935: 237. 1935. S. nervosum Bess. ex Schult. in Roem., et Schult., Syst. Veg. 2, Mant. 669. 1827. S. nigricans (Ruiz et Pavon) Snowden, Kew Bull. 1935: 244. 1935. S. notabile Snowden, Kew Bull. 1935: 239. 1935. S. rigidum Snowden, Kew Bull. 1935: 248. 1935. S. roxburghii Stapf in Prain, Fl. Trop. Aft. 9: 126. 1917. S. simulans Snowden, Kew Bull. 1935: 237. 1935. S. splendidum (Hack.) Snowden, Kew Bull. 1935: 233. 1935. S. subglabrescens Schweinf. et Aschers. in Schweinf., Beitr. Fl. Aethiop. 302, 306. 1867. The complex synonymy of the various cultivated taxa is listed in (Snowden, 1936).

Morphological description of S. bicolor subsp. bicolor:

Plants annual, with stout culms up to 5m tall, often branched, frequently tillering. Leaf blades up to 90cm long and 12cm wide, usually glabrous, not hairy, sheaths glabrous to sparsely pilose, lower branches of inflorescence slender and pendulous or stiff, the lower ones often almost as long as the panicle. Sessile spikelets extremely variable in size and shape, usually 3-9mm long and 2-5mm wide at maturity, glabrous to pilose. Racemes tough, usually 1-5noded. Glumes equal to subequal, coriaceous to membranaceous, keeled, often terminating into three minute teeth. Lemmas finely ciliate, lower awnless, the upper often awned. Pedicellate spikelets linear-lanceolate, male or neuter, persistent or deciduous. (Glumes, lemmas, and paleas are the bracts which partially or fully enclose the grains of grasses).

Cultivated grain Sorghums were divided into Guineensia, Nervosa, Bicoloria, Caffra and Durra complexes by (Snowden, 1936). Genetic diversity of those groups have been studied by (Menz et al., 2004), but the results were mixed, clusters obtained only fitted to the pedigree information after some adaptation of the marker systems used. 89

A helpful general description for beginners has been given by ICRISAT (ICRISAT, 1993), it can be downloaded at www.ipgri.cgiar.org/.../pubfile+.asp?ID_PUB=251 Another general manual, (Li et al., 1998) with emphasis on agronomic topics can be downloaded under http://ecoport.org/ep?SearchType=earticleView&earticleId=172&page=-2

It is worthwhile to read the general introduction of Sorghum bicolor races by (De Wet et al., 1972) published in his account on the race guinea: “Sorghum is a versatile crop. It is prepared in many different ways as human food, fed to livestock, and the tall stems produced by some kinds are used as fuel or building material. The numerous uses of Sorghum are often mutually exclusive. Different kinds of Sorghum are grown for different uses, and adjacent villages often grow different kind ofSorghums for similar uses. It is therefore not surprising that Sorghum is extremely variable morphologically. Ethnological isolation seems to have given rise to morphologically distinct cultivated complexes (De Wet & Huckabay, 1967). One of the widest distributed is race guinea.” Cultivated Sorghums and their closest wild relatives have been variously classified by (Murty et al., 1967; Snowden, 1936, 1955, 1961) and by (De Wet & Harlan, 1971). Snowden recognized 52 species of cultivated and related spontaneous Sorghums. These Snowdenian species were combined into S. bicolor (L.) Moench by (De Wet & Huckabay, 1967). Cultivated Sorghums are included in subspecies bicolor, and closely allied spontaneous taxa in subsp. arundinaceum (now verticilliflorum) (De Wet & Harlan, 1971). Subspecies bicolor is recognized to include races bicolor, caudatum, durra, guinea and kafir, and ten intermediate races that each combines characteristics of two basic races in all possible combinations (Harlan & De Wet, 1972). Subspecies arundinaceum is divided into races arundinaceum, aethiopicum, drummondii, verticilliflorum and virgatum (De Wet et al., 1970). The reason why the study of (Murty et al., 1967) did not show fitting results, might be hidden in the fact that the authors relied on a world collection of Sorghums with seeds from plants which have been grown for decades in agricultural field collections.

90

Fig. 27 S. bicolor (Linn.) Moench var. bicolor (Pers.) Snowden (Snowden, 1936) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 7: Grain left proximal (in)side with hilum, right: distal (out)side with embryo mark. 8: Longitudinal section of grain. 9: Transverse section of grain. From (Doggett, 1988)

91

Fig. 28 Sorghum bicolor, pedicellate and sessile spikelets, from Agrostology text from the internet: http://gemini.oscs.montana.edu/~mlavin/b434/lab7.htm Montana State University 92

Fig. 29 Sorghum bicolor, Department of Horticulture and Crop Science, The Ohio State University, upper and right grain: distal side with embryo mark, 2 grains to the right below: proximal side with hilum, from http://www.oardc.ohio- state.edu/seedid/single.asp?strID=312

A recent overview on Sorghum races has been delivered by (Deu et al., 2006): They give an analysis of the structure of genetic diversity in cultivated Sorghums. A core collection of 210 landraces representative of race, latitude of origin, response to day length, and production system was analysed with 74 RFLP probes dispersed throughout the genome. Multivariate analyses showed the specificity of the variety guinea margaritiferum, as well as the geographical and racial pattern of genetic diversity. Neighbor-joining analysis revealed a clear differentiation between northern and southern equatorial African accessions. The presence of Asian accessions in these 2 major geographical poles for Sorghum evolution indicated two introductions of Sorghum into Asia. Morphological race also influenced the pattern of Sorghum genetic diversity. A single predominant race was identified in 8 of 10 clusters of accessions, i.e., 1 kafir, 1 durra, 4 guinea, and 2 caudatum clusters. Guinea Sorghums, with the exception of accessions in the margaritiferum variety, clustered in 3 geographical groups, i.e., western African, southern African, and Asian guinea clusters; the latter two appeared more closely related. Caudatum were mainly distributed in 2 clusters, the African Great Lakes caudatum cluster and those African caudatum originating from other African regions. This last differentiation appears related to contrasting photoperiod responses. These results are helpful in the 93 optimization of sampling accessions for introgression in breeding programs, see the fig at the end of this chapter and more comments also the chapter on Sorghum breeding.

3.5.4.1. Race bicolor This race probably represents relics of ancient cultivated kinds. It shows the most ancestral spikelet morphology, the race has usually very small grains that are fully enclosed by the long, clasping glumes which usually enclose the grains at maturity. The inflorescence is usually open, the racemes baring grains are sitting at long panicle branches. They resemble spontaneous weedy Sorghums, but they lack the ability of natural seed dispersal, they do not shatter. They are low yielding as cereals, but they grow nearly always besides Sorghum cultivars. They are rarely an important crop, except in Ethiopia West of the Rift Valley. Most bicolors are now grown for their sweet stalks or for production of beer or dye.

(deOliveira et al., 1996) tested three different molecular marker technologies to determine the relatedness of 84 different lines of Sorghum. Both racial characterization and geographical origin were found to be correlated with relatedness. In some cases, the region of origin was the more significant factor, where samples of different races from the same locality were more closely related than were samples of the same race from different localities. Wild Sorghums were shown to have few novel alleles, suggesting that they would be poor sources of germplasm diversity. The results also indicated that Chinese Sorghums are a narrow and distinctive group that is most closely related to race bicolor.

3.5.4.2. Race kafir This race is based on subseries Caffra of Snowden with the exclusion of S. caudatum and S. nigricans. It is characterized by more or less compact inflorescences that are often cylindrical in outline. Sessile spikelets are typically elliptic with the glumes, at maturity, tightly clasping the usually much longer broadly ellipsoid grain. The part of the grain which extends beyond the clasping glumes is symmetrical and round. Kafir Sorghums are found in parts of Africa south of the equator, mainly in areas with one well-defined rainy season such as Natal, Transvaal, and Orange Free State (Snowden, 1936).

The name is derived from 'kafir' the Arabic for unbeliever, referring to the Bantu who grow this race. Kafir Sorghums are important staples across the eastern and southern savanna from Tanzania to South Africa. The race includes S. caffrorum and S. coriaceum. According to (Stemler et al., 1975a) there is little reason to suppose that race kafir has originated in Ethiopia, since it does not exist there today. (Deu et al., 1994; Deu et al., 2006) showed with a comparison of morphological characters and marker genes included in a multivariate statistical analysis, that among the six races in the S. bicolor complex the race kafir is well defined. (Hicks et al., 2001) had a closer look at the kafirin proteins, using lines and their hybrids of S. bicolor, which were grown under dry land conditions at two locations in Kansas USA using a randomized complete block design. The effects of genotype, location, and males were significant for all kafirins. Wide variations in composition and general combining ability (GCA) for kafirin content were noted among parent lines and hybrids.

(Menkir et al., 1997) found that accessions within races bicolor and guinea had greater genetic diversity than accessions from race kafir. 94

Fig. 30 Race Kafir Sorghum caffrorum Beauv. var. albofuscum (Koern.) Snowden (Snowden, 1936) in (Doggett, 1988). 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 6: Mature fruiting spikelet with two pedicelled spikelets. 7: Grain distal (out)side with embryo mark. 8: Longitudinal section of grain. 9:Transverse section of grain. From (Doggett, 1988).

Fig. 31 Sorghum bicolor race kafir from Lost Crops of Africa Vol. 1, Grains, 1996 darwin.nap.edu/books/0309049903/html/158.html National Academic Press 95

3.5.4.3. Race caudatum. This race is based on S. caudatum including S. nigricans of subseries Caffra. Caudatum Sorghums have characteristic asymmetric turtle-backed grains that are flat or even concave on one side and distinctly curved on the opposite side (bulging on the side where the embryo is visible). The grains are usually exposed between the shorter glumes at maturity. The inflorescences range in shape from compact to open. Caudatum Sorghums are the staple crop and widely grown in the northeast African savanna, in Sudan and Chad, and in parts of Nigeria, Cameroon, Uganda, Ethiopia and Kenya. Most of the people who grow Caudatum Sorghums also raise cattle. This race is closely associated with speakers of Chari-Nile languages in Africa (Stemler et al., 1975a, b). It is an old race of grain Sorghum: Carbonized grains from Daima, dated to the ninth century (Connah, 1971), belong typically to the race caudatum. Several research groups mention explicitly the race caudatum as to be genetically distinguishable (Cisse & Ejeta, 2003; Dahlberg et al., 2002; Deu et al., 1994; Deu et al., 2003; Deu et al., 2006; Rami et al., 1998; Xu et al., 1994).

A biogeographic and historical study by (Stemler et al., 1975a) concluded that Caudatum Sorghums probably originated in the North-East African savannah, possibly in the southern half of Sudan or Chad. They were probably developed there by people who spoke languages in the Chari-Nile family of languages, as do most Caudatum growers today. There is little reason to suppose that people and events on the Ethiopian highlands were of any importance in the development of Caudatum Sorghums. Caudatums have not, however, been widely accepted as human food by peoples outside traditionally caudatum-growing areas, possibly because most caudatum varieties contain polyphenolic compounds often called 'tannins' that make caudatum flour bitter and dark in color. This biochemical characteristic makes caudatum inferior in a culinary sense to most guinea, kafir, and durra Sorghums. Round about five centuries ago, a large group of Luo-speakers moved south to Uganda. There they are said to have replaced earlier rulers of the Bantu agriculturalists. The deposed Bacwezi rulers, possibly related to modern Bahima, left their 'royal kraals' at Ntusi according to Ankole tradition (Posnansky, 1969) and were replaced by Luo conquerors who established the Bito dynasty of Bunyoro. Other Luo clans established themselves in the Nyanza province of Kenya.

Sorghums from Tanzania which combine the features of kafir and caudatum may constitute botanical evidence of this or some other migration of caudatum-growing peoples. No significant numbers of caudatums reach far west of Lake Chad – this may seem strange considering the excellent agronomic properties of caudatum, but, looking further, (De Wet et al., 1970) can understand why people in West Africa with a tradition of growing African rice, yams, and guinea Sorghums are reluctant to adopt caudatum, a crop that yields dark, bitter flour. Also qualities such as drought tolerance which make caudatums so valuable to dwellers of the dry savanna are of little 96 advantage in areas with plentiful rains. Some caudatums are grown in north-eastern Nigeria, west of the westernmost outposts of Chari-Nile languages, but these are grown to feed animals and are known locally by names which suggest they were distributed in Nigeria by Kanuri and Hausa who have contact with Chari-Nile speakers to the east. (LeConte, 1965; Viguier, 1945).

Fig. 32 Race caudatum. Sorghum caudatum Stapf var. caudatum (Hack.) Snowden (Snowden, 1936) in (Doggett, 1988) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 4: Mature fruiting spikelet. 7: Grain, left distal (out)side with embryo mark, right: proximal (in) side with hilum. From (Doggett, 1988)

3.5.4.4. Race durra This race includes S. durra and S. cernuum of subseries Durra. The name is derived from the Arabic name for Sorghum, and the distribution of the race durra is closely associated with Islamic people in Africa. Inflorescences are usually compact. Sessile spikelets are characteristically flattened and ovate in outline, and the lower glume is creased (wrinkled) near the middle or with the tip having a distinctly different texture than the lower two-thirds. Durra Sorghums are widely grown along the fringes of the southern Sahara, across arid West Africa, the Near East and parts of India. A sheath of Sorghum discovered in a building at Qasr Ibrim and dated to Meroitic time (Macdonald et al., 1995) is composed of durra-like Sorghums. Durra Sorghum has also revealed to be the best performant race in resistance tests against one of the important Sorghum pests, the greenbug (Schizaphis graminum (Rondani) (Andrews et al., 1993). See also (Radchenko, 2000) with extensive genetic analysis for greenbug resistance.

97

Fig. 33 Race durra. Sorghum durra (Forsk.) Stapf var. aegyptiacum (Koern.) Snowden (Snowden, 1936), in (Doggett, 1988) 1: Mature fruiting spikelet with two pedicelled spikelets. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 7: Grain, left: distal (out)side with embryo mark, right: proximal (in)side with hilum. 8: Longitudinal section of grain. 9: Transverse section of grain. From (Doggett, 1988) Below some very instructive summarizing figures of the most widespread Sorghum races by (Stemler et al., 1975a)

98

Fig. 34 The 5 races of Sorghum in Africa. For each race four views are illustrated: (a) The spikelet showing grain and glumes gl. (b) View of the side of the grain on which the embryo e is visible. (c) View of the side of the grain opposite the embryo. (d) Longitudinal section through the grain showing the endosperm en and embryo e. (1) Race durra is characterized by a crease cr in one or both glumes gl. The grain is broad and blunt at the top and has a relatively small base. (2) Race caudatum is characterized by asymmetrical grains. The embryo side of the grain bulges. The side of the grain opposite is flattened or may even be concave. (3) Race bicolor is characterized by grains which do not extend beyond the glumes. (4) Race kafir has broadly ellipsoid grains. (5) Race guinea has grains which are discoid, i.e. round when viewing the embryo side, but flattened in longitudinal section. Glumes of race guinea are long and gape at maturity. Grains twist with respect to the glumes. From (Stemler et al., 1975a)

(Stemler et al., 1975a): The terms 'durra' and 'millet' are often used as general terms for Sorghum. Thus, a reference to 'Sorghum of the durra variety' in G. P. Murdock, Africa, Its Peoples and their Culture 99

History (New York, 1959), p. 334, see also (Murdock, 1960) is not durra in the sense that it is normally used. The term 'millet' may mean S. bicolor (great or giant millet), Pennisetum americanum (Linn.) K. Schum. (pearl or bulrush millet), Eleusine coracana (Linn.) Gaertn. (finger millet), Digitaria exilis (Kippist) Stapf. (fonio, acha, or hungry rice), Digitaria iburua Stapf. (black fonio), or Brachiaria deflexa (Schumach.) C. E. Hubbard ex Robyns (guinea millet).

Fig. 35 The evolutionary relationship of bicolor, durra, and durra-bicolors in Ethiopia. 3a Race bicolor is characterized by grains which do not extend beyond the glumes. 4. Race durra is characterized by a crease cr in one or both glumes gl. The grain is broad and blunt at the top and has a relatively small base. 5. Race durra-bicolor is characterized by a crease (wrinkle) in one or both glumes and ellipsoid to obovoid grains. From (Stemler et al., 1975b)

3.5.4.5. Race guinea This race is morphologically the most ancestral race, based on the Snowdenian subseries Guineensia. It is characterized by long and gaping glumes that reveal discoid grains that are obliquely twisted up to 90 degrees at maturity. Inflorescences are usually large and open, with the branches often pendulous at maturity. This character appears to have been selected by cultivators growing Sorghum in wet environmental situations with long rainy seasons. Open inflorescences dry quickly and readily after rain and are thus less vulnerable to damage by fungi that grow on Sorghum inflorescences and grain.

100

The race is primarily a Sorghum of West Africa, where the flinty grains are cooked and eaten as rice. Guineas do not yield as much grain as some other Sorghums but many large grained varieties are highly prized because they yield a superior nonbitter white flour. See also (De Wet et al., 1972) with more details about guinea race. Race guinea is also grown along the East African rift from Malawi south to Swaziland. Some cultivars are adapted to high rainfall areas, and in the fog belt of Mozambique, Malawi and Swaziland. Others are used in decrue agriculture (Harlan & Pasquereau.J, 1969). Guinea Sorghums are grown extensively in northern Nigeria where hybridization with drought-tolerant kinds of race durra is evident, see fig above: 5a. Race guinea includes three well defined cultivated groups (De Wet et al., 1972) . The one complex includes S. margaritiferum with its varieties, and is characterized by small (3.0-5.0 mm long) grains that are shorter than the glumes. This complex is grown in the high rainfall areas of West Africa and the fog- belt of eastern South Africa receiving up to 120 inches of rainfall during the wet season.

According to (Folkertsma et al., 2005), Figures based on work done by Djè (SSR markers revealing a low rate of heterozygosity nicely comply with the observations among the 100 Guinea-race Sorghum accessions reported by Fokkertsma and point to the fact that Guinea-race Sorghum is predominantly inbreeding, resulting in low levels of observed heterozygosity, but that the gene pool as a whole maintains a high level of allelic variation. Race guinea is thus a predominantly inbreeding, diploid cereal crop. It originated from West Africa and appears to have spread throughout Africa and South Asia, where it is now the dominant Sorghum race, via ancient trade routes. But according to Monique Deu (personal communication) race guinea is still predominantly outbreeding. Clearly more research needs to be done.

To elucidate the genetic diversity and differentiation among guinea-race, (Folkertsma et al., 2005) selected 100 accessions from the ICRISAT Sorghum guinea-race Core Collection and genotyped these using 21 simple sequence repeat (SSR) markers. The authors also confirmed earlier reports on the spread of guinea-race bicolor across Africa and South Asia: most of the variation was found among the accessions from semi-arid and Sahelian Africa and the least among accessions from South Asia. In addition, accessions from South Asia most closely resembled those from southern and eastern Africa, supporting earlier suggestions that Sorghum germplasm might have reached South Asia via ancient trade routes along the Arabian Sea coasts of eastern Africa, Arabia and South Asia. Figures see in chapter on molecular taxonomic analysis.

The Snowdenian S. guineense, S. conspicuum and S. gambicum have larger (5.0-9.0 mm long) grains that are about equal in length to the gaping glumes. This complex is widely grown in the broad leaf savanna of West Africa, and the savanna of Malawi. In Malawi the inflorescences are often picked before they mature, and the sweet grains are eaten raw after drying. Sorghum roxburghii is intermediate in spikelet morphology between races guinea and kafir. This complex extends across the range of Sorghum cultivation in Africa and South Asia. There are four publications, on molecular analysis, dealing with Sorghum roxburghii, essentials are commented elsewhere in this chapter already: (Deu et al., 1995; Folkertsma et al., 2005; Mulcahy et al., 1992; Wu, 1979).

101

Fig. 36 Race guinea of S. bicolor subsp. bicolor. 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 7: Grain: left proximal (in)side with hilum. 8: Longitudinal section of grain. 9: Transverse section of grain. From: (Doggett, 1988)

3.5.4.6. Intermediate races It was already (De Wet et al., 1972; De Wet & Harlan, 1971; De Wet et al., 1970) stating that Sorghum roxburghii includes true guineas as well as kafir-guinea intermediates. Margaritiferum, conspicuum, guineense, gambicum, and most roxburghii varieties of Snowden are true guinea Sorghums. Besides S. roxburghii, several other Snowdenian species do not typically belong to any one of the five basic races S. bicolor: S. ankolib, S. rigidum and S. subglabrescens are intermediate in spikelet morphology between races bicolor and durra. Derivatives of hybridization between races guinea and caudatum are represented by S. dulcicaule and some varieties of S. elegans and S. notabile. Sorghum melaleucum and S. mellitum are guinea Sorghum in spikelet morphology, and the intermediate kafir Sorghum complex includes S. basutorum, S. miliiforme, S. simulans and some cultivars of S. elegans and S. nervosum. Some cultivars of S. nigricans as recognized by Snowden are derivatives of kafir-caudatum hybridization. Hybrid races occur primarily where two or more basic races overlap. However, the hybrid morphology of some intermediate races has become fixed and several of these kinds are grown beyond their probable centers of origin. (Folkertsma et al., 2005) found with genetic analysis, that stratification of the accessions according to their Snowden classification indicated clear genetic variation between margaritiferum, conspicuum and roxburghii accessions, whereas the gambicum and guineense accessions were genetically similar. For additional remarks see under race guinea. See the figure 35 above about the morphological intergrading characters in wild S. bicolor races and its hybrids from (Stemler et al., 1975a).

102

According to (Doggett, 1988), (Harlan & De Wet, 1972) put some better order into a previously chaotic system by listing 5 races and 10 intermediate races: basic races race (1) bicolor (B) race (2) guinea (G) race (3) caudatum (C) race (4) kafir (K) race (5) durra (D) intermediate races (6) guinea-bicolor (GB) (7) caudatum-bicolor (CB) (8) kafir-bicolor (KB) (9) durra-bicolor (DB) see fig. 29 (10) guinea-caudatum (GC) (11) guinea-kafir (GK) (12) guinea-durra (GD) (13) kafir-caudatum (KC) (14) durra-caudatum (DC) (15) kafir-durra (KD)

3.5.5. S. bicolor subsp. drummondii (Nees ex. Steud.) de Wet & Harlan or ex Davidse, Sudangrass Based on S. drummondii (Steud.) Millsp. et Chase, Publ. Field Columb. Mus. Bot. 3: 21. 1903.

The species of weedy Sorghums recognized by Snowden are listed in alphabetical order. Sorghum atterrimum Stapf in Prain, Fl. Trop. Afr. 9: 121. 1917. S. drummondii (Steud.) Millsp. et Chase, Publ. Field Columb. Mus. Bot. 3: 21. 1903. S. elliotii Stapf in Prain, Fl. Trop. Aft. 9: 118. 1917. S. hewisonii (Piper) Longley, J. Agric. Res. 44: 318, 319. 1932. S. niloticum (Stapf ex Piper) Snowden, J. Linn. Soc. Bot. 55: 258. 1955. S. nitens (Busse et Pilger) Snowden, Kew Bull. 1935: 224. 1935. S. sudanense (Piper) Stapf in Prain, Fl. Trop. Afr. 9: 113. 1917.

Synonymy of these taxa is listed by (Snowden, 1936, 1955). Plants annual with relatively stout culms up to 4 m tall. Leaf blades lanceolate, up to 50 cm long and 6 cm wide. Panicle variable, usually rather contracted, sometimes similar to that of the cultivated grain Sorghum this subspecies accompanies as a weed, with culms up to 30 cm long and 15 cm wide, branches often somewhat pendulous. Racemes more or less crowded, mostly 3-5-noded, tardily disarticulating at maturity. Sessile spikelets lanceolate-elliptic, 5-6 mm long. Lemmas hyaline, ciliate, lower around 6 mm, upper around 5 mm in length, shortly lobed, and the lower often awned. Grains variable, usually 103 somewhat resembling the shape of the grain of the cultivated race from which the weed was derived. Pedicellate spikelets male or neuter.

S. bicolor subsp. drummondii, an annual weed thought to be a natural hybrid between subsp. bicolor and verticilliflorum (De Wet, 1978), grouped mainly with a subset of the cultivated subsp. bicolor (namely, race bicolor). Race bicolor resembles spontaneous weedy Sorghums but lacks the ability to disperse seeds naturally (i.e., seeds do not shatter) (De Wet, 1978). Because its long, clasping glumes, elongated seed, and open panicles are considered to be primitive characters, Sorghum bicolor is thought to be the race most closely related to wild Sorghums (Harlan & De Wet, 1972). The remaining cultivated types were loosely partitioned into three groups: one consisting primarily of Sorghum bicolor and durra accessions, another containing the kafir accessions and a mixture of several other racial types, and a third group comprising mainly the guinea and caudatum accessions. The authors emphasize that branch support for groupings of cultivated accessions was low, indicating a history of gene flow among the various races and/or recent common ancestry. Subspecies drummondii occurs as a weed in Africa wherever cultivated grain Sorghums and their closest wild relatives are sympatric. All races of subspecies Sorghum and all wild kinds of S. bicolor hybridize to produce weedy derivatives. Morphologically stabilized derivatives often accompany grain Sorghums as weeds beyond the natural range of S. bicolor. These occur widely in India and the highlands of Ethiopia. grain Sorghums also introgress with the widely distributed Eurasian tetraploid S. halepense in South Asia, and in other parts of the world where this rhizomatous perennial became established, to form both annual diploid and perennial tetraploid weedy derivatives. The widely distributed chicken corn of the south-eastern United States probably originated from an abandoned cultivated grain Sorghum. It is characterized by fragile racemes, but disarticulation takes place by breaking of the rachis, rather than through callus formation as is typical for spontaneous Sorghum. Weeds with similar dispersal mechanisms were collected among Sorghum plants in Africa outside the range of wild S. bicolor, and in the Punjab. (Rich et al., 2004) recognize the potential of Striga resistance in many of those wild Sorghums: Witch weeds (Striga spp.) are noxious parasitic weeds that cause considerable crop damage in the semiarid tropics. Genetic control of Striga is effective, although sources of resistance are limited in most crops. Useful resistance sources have been obtained in S. bicolor (L.) Moench, an important host crop that has coevolved with the parasite. Fifty-five wild accessions within the primary gene pool of Sorghum and 20 Sorghum cultivars were screened for resistance to Striga asiatica L. Kuntze in the laboratory. Wild Sorghums assayed included S. almum Parodi, S. bicolor subsp. drummondii (Steud.) De Wet, race drummondii and race hewisonni, S. bicolor subsp. verticilliflorum (Steud.) Piper with races aethiopicum, arundinaceum, verticilliflorum, and virgatum; S. halepense (L.) Pers.; S. miliaceum; S. rhizomatores; S. sorghastrum; and S. usambarense. Wild Sorghum accessions varied in their effects on S. asiatica at the pre-attachment level of association. Potential striga-resistance mechanisms of low germination stimulant production, germination inhibition, and low haustorial initiation activity were observed in this collection of Sorghums. Some of these potential striga-resistance mechanisms, reported here for the first time, appear to be unique to wild Sorghums. The results described in this study offer the possibility of introgressing valuable resistance genes from wild to cultivated Sorghum. There are 55 papers published on Striga related to Sorghum, screen main bibliography of Sorghum (not the cited literature of this report). 104

(Casler et al., 2003) quantified in a study the increase in forage quality and decrease in forage yield and to provide an economic assessment of this dichotomy. Piper and Greenleaf (normal leaves) were compared with their brown-midrib counterparts and to four highly selected brown-midrib (FG) lines at two locations for 2 yr. Brown-midrib lines averaged 9.0% lower in lignin and 7.2% higher in vitro fiber digestibility than normal lines. The reduction in first-harvest forage yield was highly variable across germplasms and locations.

3.5.6. S. bicolor subsp. verticilliflorum (Steudel) Piper (Cui et al., 1995; Doggett, 1988) Based on S. arundinaceum (Desv.) Stapf in Prain, Fl. Trop. Afr. 9: 119. 1917. This is the wild subspecies of S. bicolor: In older literature this taxon runs under S. bicolor subsp. arundinaceum (Desv.) de Wet et Harlan and intergrades with S. bicolor subsp. verticilliflorum (Steud.) Stapf (Harlan & Stemler, 1976). According to (Cui et al., 1995; Doggett, 1988) this subspecies should now be called S. bicolor subsp. vertilliflorum (L.) Moench and includes S. bicolor subsp. arundinaceum (Desv.) de Wet et Harlan.

The species of wild Sorghums usually recognized are listed alphabetically. Sorghum arundinaceum (Desv.) Stapf in Prain, Fl. Trop. Aft. 9: 114. 1917. S. aethiopicum (Hack.) Rupr. ex Stapf in Prain, Fl. Trop. Afr. 9: 119. 1917. S. brevicarinatum Snowden, J. Linn. Soc. Bot. 55: 242. 1955. S. castaneum Hubb. et Snowden, Kew Bull. 1936: 316. 1936. S. lanceolatum Stapf in Prain, Fl. Trop. Aft. 9: 112. 1917. S. macrochaeta Snowden, Cult. Races Sorgh. 237. 1936. S. panicoides Stapf in Prain, Fl. Trop. Afr. 9: 120. 1917. S. pugionifolium Snowden, J. Linn. Soc. Bot. 55: 240. 1955. S. somaliense Snowden, J. Linn. Soc. Bot. 55: 238. 1955. S. usumbarense Snowden, J. Linn. Soc. Bot. 55: 238. 1955. S. verticilliflorum (Steud.) Stapf in Prain, Fl. Trop. Afr. 9: 116. 1917. S. virgatum (Hack.) Stapf in Prain, Fl. Trop. Afr. 9: 111. 1917. S. vogelianum (Piper) Stapf in Prain, Fl. Trop. Afr. 9: 116. 1971. Synonymy is discussed in detail by Snowden (1955).

Description of Sorghum bicolor subsp. verticilliflorum Plants are tufted annuals or weak biannuals, with slender to stout culms that may reach 4 m in height. Leaf blades linear-lanceolate to 75 cm long and 7 cm wide. Panicles usually large, somewhat contracted to loose, up to 60 cm long and 25 cm wide, with the branches obliquely ascending or spreading, sometimes pendulous. Racemes 1-5-noded, fragile. Sessile spikelets lanceolate to elliptic, 5-8 mm long, usually ciliate. Glumes coriaceous, more or less ciliolate at least in the upper third. Lemmas ciliate, lower up to 7 mm, upper around 5 mm long, 2-lobed, the upper awned from between the lobes, rarely awnless. Grains lanceolate to obovate-lanceolate. Pedicelled spikelets male or neuter, often longer than the sessile spikelet. 105

Fig. 37 Sorghum verticilliflorum. NRCS Natural Resources United States Department of Agriculture, http://www.pr.nrcs.usda.gov/technical/Plants/pla46.html

A summary of the old views reveals the uncertainty in the former literature (Doggett & Majisu, 1968): (Snowden, 1936, 1955, 1961) suggested, that the main wild types of Sorghum were S. arundinaceum, S. verticilliflorum, S. aethiopicum and S. sudanense, he considers them to be the progenitors of cultivated Sorghums.

(Price et al., 2005a), based on his own recent and extensive genetic analysis, confirms that the origin of the cultivated Sorghums must be derived from polyploid ancestors. The sequence phylogeny they present, based upon combined ITS/ndhF sequences (Fig. 3), splits Sorghum into two lineages, one with 2n = 10 species with relatively large chromosomes and their related polyploids (2n = 20, 30 and 40) and a second containing 2n = 20 and 2n = 40 species with smaller chromosomes. The ancestral chromosome number of the genus Sorghum (x = 5 vs. x = 10) remains unresolved.

Subspecies arundinaceum (now verticilliflorum) was divided among three varieties (S. arundinaceum, vertilliflorum, aethiopicum by (De Wet & Huckabay, 1967) and a fourth variety S. virgatum was added by 106

(De Wet et al., 1970). These varieties, however, intergrade morphologically and ecologically so completely into one another that they do not deserve formal taxonomic status. They are more or less well defined ecotypes. The widest distributed and morphologically most variable is race verticilliflorum (as defined in the old sense): It extends across the African savannah and was introduced to tropical Australia, parts of India and the New World. A modern account on the genetic variability of S. bicolor subsp. verticilliflorum including S. bicolor subsp. arundinaceum is given by (Cui et al., 1995). It documents close relationship with S. bicolor. About the genetic variation in wild Sorghum (S. bicolor subsp verticilliflorum (L.) Moench) germplasm from Ethiopia, assessed by random amplified polymorphic DNA (RAPD) by (Ayana et al., 2000a) see in chapter 1.1.5 on the molecular taxonomy of Sorghum. The studies (Cluster analysis of genetic distance estimates) further confirmed a low level of differentiation of wild Sorghum populations both on population and regional bases. This statement should not be misunderstood: Regional differences of Ethiopian landraces are important and maintained with the traditional farmers practices over centuries as confirmed by (Grenier et al., 2004; Zongo et al., 2005) in Sudan and Burkina Faso. In order to give full account of the often mentioned names of wild Sorghum races, the below compilation is still given.

3.5.6.1. Race verticilliflorum Race verticilliflorum can be distinguished from other wild Sorghums by its large and open inflorescences with spreading, but not pendulous branches. This race grades into race arundinaceum which is distributed across tropical Africa as a forest grass. In the broad leaf savanna race arundinaceum is often difficult to distinguish from race verticilliflorum except by inflorescences in which the branches become pendulous at maturity. (De Wet et al., 1970) treat Sorghum arundinaceum (Desv.) De Wet et Huckabay separately from S. verticilliflorum, it includes according to them Sorghum arundinaceum (Desv.) Stapf and Sorghum vogelianum (Piper) Stapf and can be distinguished from other taxa by its large open inflorescence with numerous long and flexuous branches. The whole complexity is revealed by the following description of the same authors: S. arundinaceum is closely allied to the extremely variable S. verticilliflorum (Steudel) De Wet and Huckabay, which includes S. lanceolatum Stapf, S. brevicarinatum Snowden, S. castaneum C.E. Hubb. and Snowden, S. usambarense Snowden and possibly also the incompletely known S. somaliense Snowden, S. panicoides Stapf and S. pugionifolium Snowden. Several of these are probably of hybrid origin, see the fig. 22 (fig. 2 in (De Wet et al., 1970).

Crop mimicry has been described by (Barrett, 1983) in the case of rice and its grassy weed Echinochloa, it would be interesting to check the case of the wild Sorghums.

3.5.6.2. Race virgatum Race virgatum resembles race verticilliflorum except that the inflorescence branches are more erect, and the leaf blades are narrowly linear. Race virgatum occurs along stream banks and irrigation ditches in arid north-eastern Africa. It is mentioned in various studies on genetic diversity and agronomic characterization by (Aldrich et al., 1992; Grabber et al., 1992; Hultquist et al., 1996, 1997; Ma et al., 2000; Madakadze et al., 1998; Rich et al., 2004; Vogel et al., 1999). Distribution map in the chapter on the biogeography of wild Sorghum species fig. 62 from (De Wet et al., 1970). 107

3.5.6.3. Race aethiopicum Race aethiopicum is a desert grass, and is easily recognized by its large ovate-lanceolate, densely tomentose sessile spikelets. It occurs across the Sahel from Mauritania to the Sudan. It is mentioned by (Rich et al., 2004) as a possible source for Striga resistance. (Sun et al., 1994) included the taxon in his genetic analysis using Internal Transcribed Spacers of Nuclear Ribosomal DNA. (see chapter on molecular taxonomy of the genus Sorghum.

From another genomic analysis (Ayana et al., 2000a) on Sorghum verticilliflorum in Ethiopia it became again clear that the variability is rather low within the populations of wild species compared to the cultivated Sorghums. The results agree well with the reduced size of the Ethiopian wild Sorghum populations and perhaps it reflects the real situation, as the degree of polymorphism is a rough measure of genetic variation. In other words, results show that the wild Sorghum population of Ethiopia is much more subject to genetic drift than the cultivated Sorghum germplasm of the same country. Therefore, (Ayana et al., 2000a) strongly recommend urgent rescue collections and in situ conservation. A more efficient way forward to establish such collections including a rational decision making basis for setting the priorities is proposed by (Mgonja et al., 2006), described in more details in the chapter on Sorghum breeding.

3.5.7. Recent research on Sorghum bicolor taxonomy (Casa et al., 2005) gave insight in diversity and selection of cultivated and wild Sorghums from the S. bicolor group by using simple sequence repeats. The aim was to quantify and characterize diversity in a panel of cultivated and wild Sorghums (S. bicolor), establish genetic relationships, and, simultaneously, identify selection signals that might be associated with Sorghum domestication. Evaluation of SSR polymorphisms indicated that landraces retained 86% of the diversity observed in the wild Sorghums. The landraces and wilds were moderately differentiated (F st=0.13), but there was little evidence of population differentiation among racial groups of cultivated Sorghums. Neighbor-joining analysis showed that wild Sorghums generally formed a distinct group, and about half the landraces tended to cluster by race. Overall, bootstrap support was low, indicating a history of gene flow among the various cultivated types or recent common ancestry. Interestingly, seven of these loci mapped in or near genomic regions associated with domestication-related QTLs (i.e., shattering, seed weight, and rhizomatousness). The authors anticipate that such population genetics-based statistical approaches will be useful for re- evaluating extant SSR data for mining interesting genomic regions from germplasm collections.

For Sorghum races Sorghum, caudatum, durra, and guinea, there were no significant differences in diversity. The kafir accessions, however, were significantly less diverse than the other races as measured both by allelic richness and gene diversity. Other studies based on SSRs (Djè et al., 1999) and, more recently, on DNA sequences (Hamblin et al., 2004; Hamblin et al., 2005) have shown similar results. This lower diversity could reflect the relative genetic isolation of race kafir in southern Africa. Because it is adapted to a more temperate environment, race kafir tends to be more photoperiod insensitive than the other Sorghum races (Grenier et al., 2001a; Grenier et al., 2001b). Both geographic isolation and temperate adaptation, therefore, may have contributed to limited opportunities for matings between kafir and other racial types. 108

Neighbor-joining analysis indicated that wild Sorghums belonging to subsp. verticilliflorum generally formed a coherent group that failed to cluster with the landraces. Moreover, wild accessions from the same geographic region tended to be genetically more similar to each other than to those from more distant locations, indicating the presence of population structure. While the area of functional genomics is still in its infancy, the authors anticipate that the use of population diversity based approaches will allow the mining of germplasm collections and extant SSR diversity data for identifying interesting genomic regions. Certainly molecular data from species with little population structure and intermediate levels of LD would be well suited for re-analysis. The authors stress that while these approaches may be advantageous for identifying genomic regions that differ from the average observed in the genome, some of these departures may also result from non-equilibrium population history. Functional studies (e.g., mutant screening, genetic complementation, expression analysis, biochemical localization and characterization, etc) are still required to establish causation. See phenogramm below.

109

Fig. 38 Neighbor-joining phenogram depicting genetic relationships among S. bicolor accessions. Wild accessions and the various cultivated races of Sorghum are color-coded. Numbers along branches denote bootstrap support (shown only for values greater than 50) Group I: landraces, Group II: race bicolor, Group III: other racial types, Group IV: guinea and caudata accessions. From (Casa et al., 2005).

(Peng et al., 1999) constructed a restriction fragment length polymorphism (RFLP) linkage map of S. bicolor (L.) Moench in a population of 137 F6-8 recombinant inbred lines using Sorghum, maize, oat, barley and rice DNA clones. The distribution of these loci does not provide support for the hypothesis that S. bicolor (L.) Moench is of tetraploid origin. Comparison of the map with RFLP maps of maize, rice, and oat produced evidence for Sorghum-maize LG rearrangements and homoeologies not reported previously, including evidence that: (1) a segment of maize 5L and a segment of 5S may be homoeologous to Sorghum.

110

(Teshome et al., 1997; Teshome et al., 1999a; Teshome et al., 1999b) report in several publications about the intricate system between landrace biodiversity and farmers management. There seems also to be a measurable interdependence between the altitude of the fields and landrace biodiversity, the optimal height at around 1600m for Ethiopia.

Fig. 39 Relationship between field altitude in m and Sorghum landrace diversity on polynomial regression analysis, all statistical parameters are significant. From (Teshome et al., 1999b)

It is justified to place the dendrogram of a global analysis of (Deu et al., 2006) at the end of this review, it summarizes the molecular view of the complex picture of races of S. bicolor: The discussion of results by (Deu et al., 2006) gives the most up-to-date view available, it is given here in full length:

The apparent first order differentiation between northern and southern equatorial African accessions indicated by the NJ analysis suggests 2 major geographic poles for Sorghum evolution and differentiation. (NJ: A Neighbor-Joining analysis was performed on the dissimilarity matrix to determine the aggregation of the accessions into clusters). This observation is compatible with the postulated origin of this species in the northeast quadrant of Africa (Doggett, 1988; Harlan & Stemler, 1976; Wendorf et al., 1992). The absence of rare alleles in southern equatorial accessions (kafir, guinea, and caudatum found in cluster VII, VIII, and X, respectively) fits also with the expectation that southern equatorial African Sorghums evolved later from other African Sorghums (Doggett, 1988). The suggested bipolar evolution of Sorghums also agrees with indications of ethnic divisions between northern (Nilotic and Sudanian) and southern equatorial Africa (Bantu) (De Wet & Huckabay, 1967; Doggett, 1988; Gourou, 1970); that could have contributed to isolation of gene pools 111 and divergent evolution. The presence of Asian accessions in both the north equatorial group (cluster TV, Fig. 8) and south equatorial group (cluster IX) suggests that Sorghums from both pools could have been introduced to Asia as suggested by (Harlan & Stemler, 1976). The NJ analysis indicated that morphological race also has substantial influence on the pattern of genetic diversity, with 8 of 10 clusters based on a single predominant race (Fig.). Previous studies have also shown associations between racial characterization and the pattern of genetic diversity as measured by RFLP (Cui et al., 1995; Deu et al., 1994) and RAPD (Tao et al., 1993) and by combining RFLP, RAPD, and ISSR molecular markers (deOliveira et al., 1996). However, the racial differentiation observed in most of these studies was less clear than in this study. The stronger association of racial classification with pattern of diversity observed in this study could be due to many factors, including stratification of world germplasm and a greater number of accessions sampled for effectively sampling diversity, inclusion only of landrace accessions, and rigorous verification of racial classification. Genetic diversity within race The high diversity of the Sorghum race, shown by its estimator of gene diversity, its mean number of alleles (%- ble Za), and its presence in multiple clusters (Fig. 8), corresponds with expectations for this race, which is considered the most ancient with such wide geographic distribution and diversity 01' uses (Storage, broom-corn, and sweet sterns) (Doggett, 1988). The guinea race also exhibited high diversity, with 3 main and very distinct groups (western African (non- margaritiferum), margaritiferum, and southern African) (Fig. 8). These groups correspond to those identified previously with isozymes, RFLP, and SSR markers (Cui et al., 1995; Deu et al., 1994; Folkertsma et al., 2005; Ollitrault et al., 1989). The genetic distinctness of guinea margaritiferum Sorghums from other guinea forms was previously discussed by (deOliveira et al., 1996; Deu et al., 1994; Deu et al., 1995) and (Folkertsma et al., 2005). The study showed that margaritiferums differed from other guineas not only by possessing rare and specific alleles, but also by having distinct genotypes at other loci, as indicated by the second PAC performed without the alleles specific to rnargaritiferums. The genetic distinctness of rnargaritiferums from other guinea Sorghums from western Africa is remarkable, since both are interfertile and cultivated in sympatry in the same season by the same farmers. The only margaritiferum from southern Africa (IS 19455) included in this study did not cluster with the western African margaritiferums, but rather clustered with other southern African guinea accessions (Fig.), even though it was found to present a mitotype closely related to western African guinea margaritiferums mitotypes (Deu et al. 1995).

These results suggest that this southern African margaritiferum had a common ancestor with western African margaritiferums and that human selection and geographic isolation resulted in marked changes of its nuclear genetic background.

The study has also revealed an additional small cluster of guinea accessions originating from Asia, not shown in previous studies. The close relationship between these Asian accessions and southern African guinea accessions suggests recent introduction of Asian forms from southern Africa. However, the presence of 3 alleles (frequency <10%)n both the Asian guinea and Asian durra accessions (data not shown) suggests weak gene flow between these races or limited introgression from wild local Sorghums. The differentiation of caudatum accessions into 3 groups, although not previously reported as such, does correspond with previous studies. The 2 main groups, African Great Lakes (cluster X) and remaining 112

African countries (cluster V), were clearly differentiated both in this study and in that of (Deu et al., 2003) based on observations of 21 morphological traits. These 2 groups also showed considerable difference for photoperiod sensitivity when sown in Mali at 1-month intervals, with the Great Lakes accessions showing much higher sensitivity than that of the others (data not shown). The third and smallest group of caudatums and caudatum-Sorghum intermediates (cluster IV) was composed of Chinese accessions. This group showed very restricted genetic diversity, as was previously reported by (deOliveira et al., 1996). This group was not highly distinct from other cultivated Sorghums in contrast to the finding of (deOliveira et al., 1996) although the small number of accessions involved in our study (n = 12) means that results could be influenced by sampling. The observation of durra Sorghums comprising 2 separate groups has not been previously reported. The main group (cluster III) consists of accessions of the most widely cultivated durra Sorghums. These Sorghums are known to have superior adaptation to droughty rain-fed conditions, and are considered to have originated in northeast Africa from where they migrated throughout Africa and on to Asia (Doggett 1988). The other durras, belonging to tkc mixed group composed of durra and caudatum accessions (cluster VI), consist of particular transplanted Sorghums from Chad and Cameroon, which are cultivated in the post-rainy season by transplantation in receding moisture systems. The cultivation of these 2 groups in different seasons would likely limit genetic exchange between them and other Sorghums that did not compete with them in dry conditions. A low genetic diversity was observed in the kafir race as previously reported in studies using isozymes, RFLP, RAPD, or SSR markers (Cui et al., 1995; Deu et al., 1994; Dje et al., 2000; Menkir et al., 1997; Ollitrault et al., 1989). All but one of the kafir accessions constituted a specific cluster. This race exhibited the lowest number of alleles and the unique rare allele was encountered in the accession not included in the kafir cluster. These results are in agreement with the recent origin and restricted geographic distribution of this race (Doggett, 1988).

(Deu et al., 2003) have analysed 230 accessions, comprising a lager proportion of guinea and caudatum but fewer durra, kafir and intermediates than the sample of (Chantereau et al., 1989). Hierarchical clustering gives a more global picture of the organization of genetic diversity and provides indications on the relationships between the groups in terms of proximity and distance. The guinea (G) and bicolor (B) form a highly variable group, the kafir (K) constitute a relatively homogeneous group, the caudatum are separated into two subgroups (C1 AND C2). The subgroup C1 is characterized by varieties presenting a short cycle and few internodes (less than 10), while subgroup C2 covers Sorghums of medium cycle and a larger number of internodes.

113

Fig. 40 Genetic diversity of cultivated Sorghums revealed by morphological markers. The tree is constructed from the Sokal and Mitchener index. From (Deu et al., 2003)

(Deu et al., 2006) report an analysis of the structure of genetic diversity in cultivated Sorghums. A core collection of 210 landraces representative of race, latitude of origin, response to day length, and production system was analysed with 74 RFLP probes dispersed throughout the genome. Multivariate analyses showed the specificity of the subrace guinea margaritiferum, as well as the geographical and racial pattern of genetic diversity. Neighbor joining analysis revealed a clear differentiation between northern and southern equatorial African accessions. The presence of Asian accessions in these 2 major geographical poles for Sorghum evolution indicated two introductions of Sorghum into Asia. Morphological race also influenced the pattern of Sorghum genetic diversity. A single predominant race was identified in 8 of 10 clusters of accessions, i.e., 1 kafir, 1 durra, 4 guinea, and 2 caudatum clusters. Guinea Sorghums, with the exception of accessions in the margaritiferum subrace, clustered in 3 geographical groups, i.e., western African, southern African, and Asian guinea clusters; the latter two appeared more closely related. Caudatum were mainly distributed in 2 clusters, the African Great Lakes caudatum cluster and those African caudatum originating from other African regions. 114

This last differentiation appears related to contrasting photoperiod responses. These results aid in the optimization of sampling accessions for introgression in breeding programs.

115

Fig. 41 Neighbor-joining analysis based on RFLP data among 205 cultivated accessions using the Nei and Li similarity index. The numbers on the branches indicate bootstrap values (expressed in percentages) and are shown for all clusters with > 60%) bootstrap support. Cl, cluster. From (Deu et al., 2006)

Fig. 42 Correspondence analysis on morphological traits. C1 and C2 shows again the two substructures of the caudatum race. Comments above. From (Deu et al., 2003)

The real reason why those landraces have sustained over millennia’s is given by (Stemler et al., 1975a):

An important consequence of environmental and human selection is that each race of Sorghum has become ecologically and culturally specialized. A race of Sorghum resulting from generations of selection in one climate will not do well in another. And a race of Sorghum selected by one group of people is usually considered undesirable by another group of people. Predilections of Africans of the present in favor of their own Sorghums and against Sorghums of other peoples are so strong as to suggest that traditional African farmers do not and have not casually borrowed and adopted unfamiliar Sorghums raised by unrelated peoples. The usual reaction to an unfamiliar kind of Sorghum is that it is unfit for human consumption. And in fact very little exchange of Sorghum varieties of different races has been observed among unrelated groups of people. 116

The authors would guess that ecological and cultural specialization of races of Sorghum has had profound historical implications for Sorghum-growing people. Their impression is that the fate of Sorghum-growing peoples has been inextricably linked to the Sorghum cultivars created by their forefathers. Limitations in ecological amplitude of the staple crop must have imposed limitations on the direction and extent of migration of people who depend on Sorghum for their food. Looking at this more positively, we might say that the ability of a group of Sorghum-growing peoples to occupy an area or to expand into new areas has probably been determined in part by the productivity, genetic potential, and ecological amplitude of the Sorghum cultivars they created and depended on for food. They believe the relationship between Chari-Nilespeaking peoples, caudatum Sorghum and cattle is an example of a partnership which has made food production possible in some of the most difficult settings for agriculture in Africa.

This can be taken as a serious caveat when new Sorghum with definitely better agronomic characteristics should be introduced in Africa. If the African farmers should benefit from new Sorghum breeding, the new traits will have to be carefully adapted to the local needs of the population. Nevertheless, (Deu et al., 2006) state, that the genetic distinctness of rnargaritiferums from other guinea Sorghums from western Africa is remarkable (and somehow astonishing), since both are interfertile and cultivated in sympatry in the same season by the same farmers. This can also be taken as a hint that we do not yet understand Sorghum cultivation in Africa in all important aspects.

3.5.8. Summary Sorghum bicolor The variability in cultivated Sorghums can be described in terms of four roughly geographic races and one widely distributed morphologically primitive race bicolor. Race bicolor is the most primitive morphologically. It has small, ellipsoidal grains which are tightly enclosed by the glumes. This race occurs widely but is not grown much for grain. Most bicolors are now grown for their sweet stalks or for production of beer or dye. The primitive morphology of bicolor Sorghums and wide geographic distribution suggest that a bicolor type was the progenitor of the more highly derived races.

Race guinea is distinguished by long glumes that gape at maturity and discoid grains that twist up to go degrees at maturity. Guineas are adapted to areas with a long rainy season, areas too wet for other Sorghums and most other cereals.

Margaritiferum varieties of race guinea are grown in parts of West Africa, receiving up to 120 inches of rain during the growing season.

Conspicuum varieties are grown in the fog belt of Mozambique, Malawi, and Swaziland. Some guineas have small, hard seeds that are boiled and eaten like rice. Large-grained varieties are used to make Guineas do not yield as much grain as some other Sorghums but many varieties are highly prized because they yield a superior nonbitter white flour.

Sorghums of race durra have glumes which are usually creased horizontally. The grains are very broad and blunt at the top and taper to a relatively small, wedge-shaped base. Some mature in as little as three months and thus can be grown in areas with a very short rainy season. Durras are grown in India and 117 south-west Asia as well as the northern fringe of the African savanna. The pattern of distribution strongly suggests that durras were brought into Africa under Islamic influences.

Sorghums of race kafir have grains that are broadly ellipsoid. The part of the grain which extends beyond the clasping glumes is symmetrical and round. Kafir Sorghums are found in parts of Africa south of the equator, mainly in areas with one well-defined rainy season such as Natal, Transvaal, and Orange Free State.

Race caudatum is distinguished by its very asymmetrical grain. The embryo side of the grain bulges; the opposite side is flat or may even be concave; and the tip of the grain often comes to a point. This combination of characters gives race caudatum a very distinctive appearance. Caudatum Sorghums are grown in Hyparrhenia and Andropogon dominated regions of north-central Africa receiving 15 to 50 inches of rain per year. They are the staff of life in much of Sudan and Chad, and parts of Cameroon and Uganda. Caudatums have not, however, been widely accepted as human food by peoples outside traditionally caudatum-growing areas, possibly because most caudatum varieties contain polyphenolic compounds often called 'tannins' that make caudatum flour bitter and dark in color. This biochemical characteristic makes caudatum inferior in a culinary sense to most guinea, kafir, and durra Sorghums. Though caudatum is not very popular as food for humans outside the areas listed, caudatum varieties such as 'feterita' and 'hegari' or hybrid derivatives of these varieties are important in sub-humid to semi-arid regions of the world where they are grown for use as stock feed. A characteristic of caudatum Sorghums that is more difficult to define and more important to the people who grow it is termed ‘rusticité’ by French speakers in Chad. Caudatum, more than any other indigenous African crop, can be counted on to produce a crop despite high water, drought, parasites such as witch weed, and every other kind of hazard. The many physiological characteristics which contribute to the ‘rusticité’ of caudatum Sorghums make it possible for cultivators of caudatum Sorghums to feed their families in spite of the harsh and often unpredictable environmental conditions prevailing in north- eastern African savanna regions.

In summary: the variation of cultivated African Sorghum is discontinuous and justifies designation of five major races. One race (bicolor) most widely distributed and least different in morphology from wild Sorghum, is presumed to be the primitive cultivated type from which the more highly differentiated races were selected. It is not known whether cultivated bicolor Sorghums arose in one or in many places in Africa. The four remaining races of Africa, which are centered in different regions and different climates, differ from each other in morphology and some aspects of physiology. These races appear to be the products of two kinds of selection. Natural selection in various regions has resulted in different physiological characteristics such as photoperiodic responses, maturity cycles, moisture requirements, and some spikelet and inflorescence characters. Human selection is undoubtedly responsible for environmentally neutral aspects of the size, shape, and color of the grain and some culinary qualities. 118

Thus, natural and human selection have simultaneously acted to produce cultivars which are physiologically suited to perform well in the environment where selection has occurred and culturally suited to the needs, expectations, and aesthetics of the people who have raised the Sorghum for countless generations. An important consequence of environmental and human selection is that each Sorghum race has become ecologically and culturally specialized. A race resulting from generations of selection in one climate will not do well in another. And a race selected by one group of people is usually considered undesirable by another group of people.

Present day predilections of Africans in favor of their own Sorghums and against Sorghums of other people are so strong as to suggest that traditional African farmers do not and have not casually borrowed and adopted unfamiliar Sorghums raised by unrelated peoples. The usual reaction to an unfamiliar kind of Sorghum is that it is unfit for human consumption.

And in fact very little exchange of Sorghum varieties of different races has been observed among unrelated groups of people. One would guess that ecological and cultural specialization of races of Sorghum has had profound historical implications for Sorghum-growing peoples. It seems clear that the fate of Sorghum-growing peoples has been inextricably linked to the Sorghum cultivars created by their forefathers.

Limitations in ecological amplitude of the staple crop must have imposed limitations on the direction and extent of migration of peoples who depended on Sorghum for their food. Looking at this more positively, one might say that the ability of a group of Sorghum-growing peoples to occupy an area or to expand into new areas has probably been determined in part by the productivity, genetic potential, and ecological amplitude of the Sorghum cultivars they created and depended on for food. We believe the relationship between Chari-Nilespeaking peoples, caudatum Sorghum and cattle is an example of a partnership which has made food production possible in some of the most difficult settings for agriculture in Africa. In chapter 1.5.4. the results of modern genomic analysis have been summed up, it is now fact that the races caudatum and kafir are not uniform and need to be studied further on.

3.5.9. Bibliography Sorghum bicolor http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Sorghum-20060608.pdf

3.6. Numerical taxonomy within the genus Sorghum

3.6.1. Numerical taxonomy with defined lines of type specimens (De Wet & Huckabay, 1967) conducted a study in numerical taxonomy, using well defined lines from type specimens and restricted to material from one generation, thus consequently avoiding unwelcome blurring of the data through uncontrolled 119 hybridization. They used a statistical material from 52 taxa and 38 characters. Methodologically they followed the school of Numerical Taxonomy. (Sokal & Mitchener, 1958; Sokal & Sneath, 1963).

Fig. 43 Dendrogram showing relationships between the 52 Sorghum species recognized by Snowden. See Table 1 in chapter on S. bicolor for code numbers. Fig 1 from (De Wet & Huckabay, 1967) In a wise use of characters known for their moderate variability, excluding leaf and inflorescence sizes and plant height, the dendrogram shows reasonable clustering and fits well with the view of taxonomists such as Snowden. In an extensive discussion the authors can shed light to the normally unfathomable complexity of cultivated Sorghum morphology and distribution. The maps produced from the statistical analysis fit well with the views of taxonomists who did not use numerical taxonomy.

In an attempt to arrange the cultivated ‘Snowdenian species’ in clusters with similarity coefficients, (Doggett, 1988) constructed the following figure:

In a further iteration, he obtained clusters of taxa, and the similarity coefficients between them. The fig is not drawn to scale. Results: Guinea race is much dispersed. Numbers 39 and 43 form one cluster, but No. 43 could well have been carried to southern Africa on the old Ghana to Cape Town and Durban slave routes. No. 28 and 29 form one cluster, so it may be misleading to classify exsertum as race bicolor 120 simply on seeds and panicle. No. 30 and 17 are surprisingly distinct, probably old: indeed, the subdivision of the guineense suggests that they are an older subseries or race (De Wet et al., 1972). No. 60, roxburghii, has close links with No. 46, but this could be a type selected in India for fodder out of a chance cross in roxburghii population, with the bicolor grain type favored for seed purposes. Groups E and F justify the separation of caudatum and kafir. Together with roxburghii No. 60 and durra No. 25 they make up the Eastern African complex, emphasizing by the presence of the bicolor types No. 52, 26 and 42, with No. 4 close at hand, the importance of the wild Sorghum influence. Durra is strongly linked to the East African clusters E and F, although some of the material used was of Indian and Asian origin. Beyond durra No. 25 lie types less influenced by wild African Sorghums. No. 44 and 67 have a large component outside Africa; No. 9, 16, 47 and 64 also lie outside Africa. No. 9 contains specialized types grown mainly outside Africa – the broomcorns and many of the sorgos. Group C of the diagram of (De Wet & Huckabay, 1967) therefore gives a credible picture of clusterings within the cultivated Sorghums. Considering the material on which it was based, the results of the numerotaxonomic study are encouraging.

Fig. 44 Cultivated ‘Snowdenian species’ of group C in the above dendrogram arranged in clusters by a further iteration on the data in the Table (see chapter on the taxonomy of Sorghum bicolor) of (De Wet & Huckabay, 1967)

For the biogeographical analysis of the findings based on this study see the chapter on biogeography of the genus Sorghum.

3.6.2. Numerical taxonomy with land races in Eastern Africa (Teshome et al., 1997) used in a study fourteen phenotypic characters for the purpose of obtaining taxonomic evidence on the resemblances of 177 accessions of Sorghum from North Shewa and South Welo regions of Ethiopia. Canonical Discriminant Analysis (CDA) and Modeclus cluster analysis (SAS 121

Instiute, 1992) were conducted to see if the 177 accessions could form clusters based on their morphological characters, and to test the consistency of farmers' naming of the five most common Sorghum landraces represented by 44 accessions. Multivariate analyses grouped the 177 accessions into three clusters linked by a few phenotypic intermediate landraces. The conclusion is that farmers know pretty well how to distinguish landraces with morphological characters:

Fig. 45 Sorghum landrace ordination by canonical discriminant analysis Sorghum landrace ordination by canonical discriminant analysis, using farmers-naming of Sorghum accessions as group criterion. The landraces named by the farmers are supported by the 14 morphological characters and form distinct groups on the ordination plot as well as in the analysis. Variation explained by axes 1,2 and 3 were 58.18% and 12.05% respectively, from (Teshome et al., 1997)

A botanical key was established for easy classification of the Sorghum crop plants grown in the study area in the following way: The number of accessions of the five most common landraces named by the farmers formed dissimilar groups, suggesting that farmers' naming of these Sorghum landraces are consistent. Midrib colour, grain colour, grain size, glume colour, glume hairiness, and grain shape were the leading morphological characters used by the farmers in naming these Sorghum landraces. Based on the combined results of the 'Modeclus' three cluster solution and using grain form as a membership criteria , a botanical key was established for easy classification of the Sorghum plants in north Shewa and south Welo regions of Ethiopia (Teshome et al., 1997)

1) juicy stem : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :Cluster III 11) non-juicy stem : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 2) 2) dimple grain : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :Cluster I 22) plump grain : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : Cluster II 122

Conclusion: It will be easy to assess the locally limited set of landraces by interviewing the farmers and by morphometrics: Each study region will receive their own and simple key for the morphological determination of the present landraces.

(Abdi et al., 2002) have studied Sorghum landraces from the five ecosites, they appeared unequally distributed along the Y-axis (Fun-1) of the analysis (Fig. below). The first three canonical variables accounted for 91 % of the total morphological variation, of which 49 % of the proportion of the variance was explained by Fun-1 (Fig. below) Sorghum landraces from Layignaw ataye were positively correlated to Fun-1. This axis is correlated with grain form and midrib color. Fun-2 maximized the differentiation to which Merewa adere was highly correlated. Awn, grain plumpness and peduncle shape were detected to show high correlation. Fun-3 appears to be more positively correlated with Bati and Merewa adere and endosperm texture showed greatest contribution. (Damania et al., 1996; Pecetti & Damania, 1996); have also used the same method. The dissimilarity matrix of all the 34 Sorghum landraces was examined in an attempt to determine the degree of distance between any two landraces (data not presented). The overall mean diversity index (H_) (0.77_0.04) obtained in the present study was high and similar to that of (Ayana et al., 2001) (overall H_ for Welo=0.75_0.07). This study supports Vavilov’s observation and hypothesis about the high diversity of Sorghum in Ethiopia and which was underlined

Fig. 46 Scatter diagram of the first three canonical variables (CAN) based on mean values of observations in Sorghum landraces, from (Abdi et al., 2002).

123 by (Doggett, 1991; Doggett & Parasada, 1995). One reason for this high diversity of Sorghum landraces in the study area is the fact that these landraces are found in the relatively complex and heterogeneous ecology and the non-uniform pattern of climatic conditions and agricultural management.

There are more publications on numerical taxonomy analysis of Sorghum taxa: (Kabore et al., 2001; Manzelli et al., 2005; Menz et al., 2004; Mickelbart et al., 2003; Pazoutova et al., 2000; White et al., 1995)

3.6.3. Summary In a wise use of characters known for their moderate variability, excluding leaf and inflorescence sizes and plant height, the dendrogram shows reasonable clustering and fits well with the view of taxonomists such as Snowden. In an extensive discussion the authors can shed light to the normally unfathomable complexity of cultivated Sorghum morphology and distribution. The maps produced from the statistical analysis fit well with the views of taxonomists who did not use numerical taxonomy.

3.6.4. Bibliography on numerical taxonomy of Sorghum http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Numerical-Taxonomy-20060806.pdf

3.7. Molecular taxonomic analysis It cannot be the scope of this report to give a full review of molecular taxonomic research on Sorghum, which has been pushed hard in the last two decades. A few typical (maybe not even the best) examples must suffice. Early attempts are summarized including some genomic maps on the CIRAD website with some references which are included in the general bibliographic reference list on Sorghum. http://wwww.cirad.fr/presentation/programmes/biotrop/resultats/biositecirad/maping/sorghugm.htm Some important activities about the sequencing of the Sorghum genome are also summarized in (Kresovich et al., 2005). (Kim et al., 2005a) came up with a comprehensive molecular cytogenetic analysis of the Sorghum genome architecture and gave details of the distribution of euchromatin, heterochromatin, genes and recombination in comparison to rice.

3.7.1. RFLP diversity in cultivated Sorghum in relation to racial differentiation (Deu et al., 1994) analyzed varieties of cultivated Sorghum. Careful assessment of the comparative diversity for molecular markers and for potentially-useful morpho-agronomic traits is paramount to the analysis of a genome through the mapping of favorable geneS. bicolor (S. bicolor subsp. bicolor) varieties are traditionally classified into five races on the basis of morphological traits, especially panicle and grain traits. Isozyme diversity has provided a new insight into genetic diversity, and showed a marked geographic structure. The authors performed RFLP analysis on 94 varieties, chosen to represent the main cross combinations (race x geographic origin), using 35 maize probes that detect polymorphism with at least one of the two restriction enzymes HindIII and XbaI. A 124 total of 50 polymorphic probe-enzyme combinations yielded 158 polymorphic bands. The Sorghum race appeared highly variable and included many rare markers. Among the other races multivariate analysis of the data differentiated six clusters corresponding, by decreasing magnitude of divergence, to: the margaritiferum types (a sub-race of race guinea); the guinea forms from western Africa; race caudatum; race durra; race kafir; and the guinea forms from southern Africa. The apparent geographic differentiation was related to the contrasting distribution of these races and to a higher similarity between races localized in southern Africa. The data agree with the current hypotheses on Sorghum domestication but reveal associations between neutral markers and traits probably highly subjected to human selection. Whether such associations will be observed with other useful traits, and to what extent they are maintained by genetic linkage, is worth exploring.

The same research group presented new confirming data recently: (Deu et al., 2006). A core collection of 210 landraces representative of race, latitude of origin, response to day length, and production system was analysed with 74 RFLP probes dispersed throughout the genome. Multivariate analyses showed the specificity of the subrace guinea margaritiferum, as well as the geographical and racial pattern of genetic diversity. Neighbor-joining analysis revealed a clear differentiation between northern and southern equatorial African accessions. The presence of Asian accessions in these 2 major geographical poles for Sorghum evolution indicated two introductions of Sorghum into Asia. Morphological race also influenced the pattern of Sorghum genetic diversity. A single predominant race was identified in 8 of 10 clusters of accessions, i.e., 1 kafir, 1 durra, 4 guinea, and 2 caudatum clusters. Guinea Sorghums, with the exception of accessions in the margaritiferum subrace, clustered in 3 geographical groups, i.e., western African, southern African, and Asian guinea clusters; the latter two appeared more closely related. Caudatum were mainly distributed in 2 clusters, the African Great Lakes caudatum cluster and those African caudatum originating from other African regions. This last differentiation appears related to contrasting photoperiod responses. These results aid in the optimization of sampling accessions for introgression in breeding programs. See more comments and the main figure at the end of the chapter on recent research on Sorghum bicolor taxonomy.

3.7.2. Molecular analysis on the basis of RFLP-assay within the genus Sorghum In an attempt on the molecular basis (RFLP-assay), (Cui et al., 1995) summarized the genetic diversity in form of a dendrogram as follows:

125

Fig. 47 PAUP computer analysis – derived dendrogram A PAUP computer program generated dendrogram of all 53 accessions of S. bicolor based on RFLPs detected with 62 Sorghum genomic clones. Tree length = 1633. CI=0.206. HI = 0.794. First column - accession code or name; second column - original classification; third column-Geographical origin; fourth column - RFLP cluster (Cui et al., 1995) A PAUP computer program generated dendrogram of all 53 accessions of S. bicolor based on RFLPs detected with 62 Sorghum genomic clones. Tree length = 1633. CI=0.206. HI = 0.794. First column - accession code or name; second column - original classification; third column-Geographical origin; fourth column - RFLP cluster Included were accessions from five morphological races of the cultivated subspecies Sorghum, and four races of the wild subspecies verticilliflorum. From two to twelve alleles were detected with each probe. There was greater nuclear diversity in the wild subspecies (255 alleles in ten accessions) than in the domestic accessions (236 alleles in 37 accessions). Overall, 204 of the 340 alleles (60%) that were detected occurred in both subspecies. Phylogenetic analysis using parsimony separated the subspecies into separate clusters, with one group of intermediate accessions. Though exceptions were common, especially for the race bicolor, accessions classified as the same morphological race tended to group together on the basis of RFLP similarities. The genetic structure of and gene flow between the cultivated-race subspecies and the wild subspecies were tested using data from eight loci that displayed only two alleles per locus (the population size is too small to test multiple-allele loci).

126

By checking hybridization blots of allprobe-enzyme combinations, it was found that 0-10 accessions showed hybrid patterns (two or more fragments) at each locus. The rates of gene flow (Nero) as calculated from FST values for nuclear loci were two to four times higher than for cytoplasmic loci. A possible explanation for this observation is that pollen from neighboring wild accessions more often contributed to variation (gene flow) in the collected accessions than did migrants originating from seeds dispersed from a population with a distinct cytotype. RFLP data support previous concepts of Sorghum evolution; namely, that multiple origins, diverse environments and human involvement have contributed to the existence of different types of wild and cultivated Sorghum. Outcrossing has led to gene introgression and gene flow among the natural populations. Then polymorphic subpopulations develop, and disruptive selection starts. Intermediate types may exist for a period, but differentiation continues until a number of distinct, separate and adaptive populations are formed. In summary, the population structure of modern Sorghums seems to fit well into Wright's "shifting balance" theory of adaptation, which assumes that genetic drift and selection operating on subpopulations leads to a number of genotypes occupying different adaptive peaks, even though gene flow can occur between the subpopulations (Wright, 1990). Wright's theory has been widely accepted to explain plant evolution and speciation (Hartl & Clarc, 1989), including applications to the evolution of Sorghum (Doggett & Majisu, 1968)

3.7.3. Comparative analysis on the genetic relatedness of Sorghum bicolor accessions from Southern Africa by RAPDs, AFLPs and SSRs In order to get an overview on the genetic relatedness of Sorghum (S. bicolor) landraces and cultivars grown in low-input conditions of small-scale farming systems, (Uptmoor et al., 2003) evaluated 46 Sorghum accessions derived from Southern Africa on the basis of amplified fragment length polymorphism (AFLPs), random amplified polymorphic DNAs (RAPDs) and simple sequence repeats (SSRs). By this approach all Sorghum accessions were uniquely fingerprinted by all marker systems. By UPGMA-clustering two main clusters were built on all marker systems comprising landraces on the one hand and newly developed varieties on the other hand. Further sub-groupings were not unequivocal.

127

Fig. 48 AFPL, B: RAPD, C: SSR, D: Δμ-SSR. UPGMA-Clusters of 46 Sorghum accessions derived from southern Africa. LR: Landrace, BV: breeding variety, RB: Botswana, LS: Lesoto, MW: Malawi, ZA: South Africa, IS: ICRISAT, NP: Northern Province, South Africa. From (Uptmoor et al., 2003)

3.7.4. 16QTLs affecting rhizomatousness and tillering of Sorghum

128

Rhizomatousness and tillering, two important factors of weediness of Sorghum, have been studied by (Paterson et al., 1995b) with Johnsongrass Sorghum halepense.

Variation in regrowth (ratooning) after overwintering was associated with QTLs accounting for additional rhizomatous growth and with QTLs influencing tillering. Vegetative buds that become rhizomes are similar to those that become tillers. The still open alternative would be that rhizomatousness and tillering is caused by dominant genes, as claimed by (Yim & Bayer, 1997), comments and figure from (Paterson et al., 1995b) on the QTL map in chapter on Sorghum halepense.

Another result of molecular analysis of Sorghum landraces demonstrate a high degree of individuality of local units, achieved with a statistical method established by (Kruskal, 1964a, b; Shepard & Kruskal, 1964): The resulting similarity matrix was subjected to non-metric multidimensional scaling (MDS).

Despite local agricultural management with selection and seed exchange the landraces can maintain their genetic identity, as shown by (Folkertsma et al., 2005)

Fig. 49 MDS plot indicating the genetic relationships between 100 Guinea-race bicolor accessions. Accessions are labeled according to their eco-geographical origin. Fig. 2 from (Folkertsma et al., 2005)

129

Fig. 50 MDS plot indicating the genetic relationships between 92 Guinea-race bicolor accessions. Accessions are labeled according to their Snowden Guinea-race classification (Folkertsma et al., 2005)

The MDS plots (Fig. 49 + Fig. 50), which were constructed using the Dice similarity index, indicate that the 100 accessions can roughly be divided into two main groups—a cluster of 11 margaritiferum accessions from humid West Africa and semi-arid and Sahelian Africa in the lower-right part of the plot, and a group with the remaining 82 accessions in the left part of the plot. More interpretation given by (Folkertsma et al., 2005) in chapter 1.1.3 S. bicolor under race guinea.

3.7.5. Comparative physical mapping links conservation of microsynteny to chromosome structure and recombination in grasses Nearly finished sequences for model organisms provide a foundation from which to explore genomic diversity among other taxonomic groups. (Bowers et al., 2005) explore genome-wide microsynteny patterns between the rice sequence and two sorghum physical maps that integrate genetic markers, bacterial artificial chromosome (BAC) fingerprints, and BAC hybridization data. The Sorghum maps largely tile a genomic component containing 41% of BACs but 80% of single-copy genes that shows conserved microsynteny with rice and partially tile a nonsyntenic component containing 46% of BACs but only 13% of single-copy genes. The remaining BACs are centromeric (4%) or unassigned (8%). The two genomic components correspond to cytologically discernible ‘‘euchromatin’’ and ‘‘heterochromatin.’’ Gene and repetitive DNA distributions support this classification. Greater microcolinearity in recombinogenic (euchromatic) than nonrecombinogenic (heterochromatic) regions is consistent with the hypothesis that genomic rearrangements are usually deleterious, thus more likely to persist in nonrecombinogenic regions by virtue of Muller’s ratchet. Interchromosomal centromeric rearrangements may have fostered diploidization of a polyploid cereal progenitor. Model plant sequences better guide studies of related genomes in recombinogenic than nonrecombinogenic regions. Bridging of 35 physical gaps in the rice sequence by Sorghum BAC contigs illustrates reciprocal benefits of comparative approaches that extend at least across the cereals and perhaps beyond.

130

Fig. 51 Comparative physical maps. A segment of the rice chromosome 4 pseudomolecule is compared with BAC contigs from SP, SB, and Zea mays based on hybridization anchors (Table 1). The scales (MBP) of the rice pseudomolecule and sorghum contigs are equal, whereas maize contigs are shown at a 1:5 scale. Sorghum and maize contig lengths were estimated by multiplying the average FPC band size (4,740 bp, the observed average for rice) by the length of the contig in FPC consensus band (CB) units. Red lines represent cases for which loci were inferred in one genome (where no dot is shown) due to missing data. Maize contigs were from a recent release (www.genome.arizona.edu_fpc_maize, release 10_25_04), incorporating hybridization anchors fromTable1. Current Sorghum contig assemblies are available online (www.plantgenomega.edu_projects.htm) , from (Bowers et al., 2005) 131

Fig. 52 Coverage of the rice genome by syntenic sorghum BAC contigs. Only hybridization markers that hit two or more BACs in the same contig were considered for microsynteny comparisons. Sorghum contigs were aligned to the rice sequence based on the criteria that two or more low copy probe loci (detecting five contigs or fewer) showed best matches between 5 and 500 kb apart in rice (anchoring 456 SP and 303 SB contigs). Sorghum chromosome (32) and linkage group designations (16) are as cited. Cytologically identified heterochromatic regions were approximated from refs. 19 and 25 based on relative distance from the centromere, assigning approximate base-pair locations without accounting for sequence gaps. From (Bowers et al., 2005)

Identification of a well conserved component of the sorghum and rice genomes will guide contig and sequence assembly in recombinogenic regions of other grasses. Further enrichment of the sorghum physical map with overgos selected using rice as a guide will reduce the population of nonsyntenic BACs_contigs, some of which may just have too few loci to anchor. Targeted overgo enrichment in sorghum segments corresponding to parts of rice chromosomes 3 and 4 (Fig. 52 above) increased coverage of the corresponding regions of the rice sequence from the average 39% to 92% and 97%, respectively (combining SB and SP). In turn, sorghum contigs bridge gaps in the much larger maize genome (Fig. 51). These benefits promise to be reciprocal; the current sorghum contigs span 35 of 130 physical gaps in a recent rice pseudomolecule (14, 7, and 14 by SP, SB, and both; Table 5, which is published as supporting information on the PNAS web site). The sorghum–rice comparison, together with hybridization data for other taxa (Table 1), permits us to begin to identify probable euchromatic regions of other genomes. For maize, these data merely add to a detailed physical map (Cone et al., 2002), but for sugarcane, this comprises a previously undescribed 132 resource. The evolutionary lifespan of euchromatic regions remains unknown; however, hybridization of many of our probes to BACs from banana, separated from the cereals by an estimated 140 million years (Paterson et al., 2004), sets the stage for further study. Limitations of Microsynteny Approaches: Colinearity along recombinational (genetic) maps of related taxa masks much rearrangement in recombinationally inert regions, which contribute little to such maps. Model genomes may be of little assistance in the study of nonrecombinogenic regions of other genomes (Akhunov et al., 2003) . De novo characterization of these regions may be needed for each species, requiring new methods to deal with their highly repetitive nature. Synthesis. Euchromatic and heterochromatic regions, as defined cytologically (19, 24, 25), appear to represent evolutionarily different components of genomes that can be independently detected by several comparative or in silico approaches. The structuring of cereal (and presumably other) genomes into these components has many implications for the evolution of individual genes and groups of genes. Differences in chromosomal context may resolve disparate conclusions regarding the degree to which colinearity and synteny persist over time, that have been drawn from comparative analyses of sets of individual BAC_contig sequences. Nonrecombinogenic domains may nurture the evolution of physically dispersed but genetically tightly linked coadapted gene complexes or supergenes that are predicted to be favored by domestication and have been observed in several cases (42). The genomic context of a gene may influence the degree to which information about gene function from model organisms extrapolates to less characterized organisms.

3.7.6. Summary molecular taxonomy of Sorghum There is a plethora of literature published on molecular genetic analysis on Sorghum, here only a few are commented, more can be found in the bibliography on Sorghum, related to agronomic questions. There are some 750 publications included in this selected bibliography. Progress in molecular knowledge is considerable, and growing rapidly in the last few years, helping to understand taxonomic relationships, but also providing basic knowledge on useful QTLs, possible resources of new resistance genes for future crop protection. A screening in the Web of Science on all publications related to Sorghum reveals some 14000 publications, the bulk coming from publications about basic research of the molecular biology of Sorghum. More publications are discussed in the chapter 1.4. on Sorghum breeding.

3.7.7. Bibliographic references on molecular taxonomy of Sorghum http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Molecular-Taxonomy-20060806.doc

133

3.8. Summary of the Sorghum bicolor taxonomy

3.8.1. The basics, an oversimplified summary 5 basic races of cultivated Sorghum are recognized by (Harlan & De Wet, 1972):

Sorghum, Kafir, Guinea, Caudatum and Durra:

 Sorghum: open inflorescences and long, clasping glumes, usually enclosing the grain at maturity, Africa and Asia.

 Kafir: Symmetrical and nearly spherical grains with glumes shorter than the grain, Africa south of the equator.

 Guinea: Easily recognized by its long and obliquely twisted, gaping glumes revealing grains at maturity. Predominantly in West Africa.

 Caudatum: Grains asymmetrical with a turtleback, pointed beak and short glumes. Throughout Central Africa, of recent origin.

 Durra: Grains obovate, wedge-shaped at base, but slightly broad above the middle. Fringe of Sahara, West Africa, throughout arid Asia.

Natural intercrossings among these five basic races gave rise to ten intermediate or hybrid races, which are found in geographical areas overlapping the basic races and also at experimental stations.

From the breeding point of view, the races Kafir, Caudatum and Durra contain genes with important contributions to yield, they have been exploited in crop improvement programs across the globe.

134

Fig. 53 The five major races of grain Sorghum as described above, from: (Bantilan et al., 2004)

135

Table 1 from (De Wet & Huckabay, 1967) Classification of Sorghum bicolor

136

Fig. 54 Classification of S. bicolor, Table 1 from (De Wet & Huckabay, 1967)

3.8.2. Some helpful summary illustrations for the races of S. bicolor

Fig. 55 Five basic spikelet types of cultivated Sorghum; wild type and shattercane not shown, from (Harlan & De Wet, 1972):

Fig. 56 Head types of cultivated Sorghum. Type 1 is reserved for wild races and is considerably more diffuse than type 2. From (Harlan & De Wet, 1972) Sorghum and guinea: types 2-4 kafir and durra: types 5-7 caudatum: wide range of types broomcorn and half-broomcorn types: 8-9 from (Harlan & De Wet, 1972): 137

3.8.3. Simplified Key for the subspecific taxa of S. bicolor The subspecific taxa of S. bicolor are distinguished, one from an other, by the following key characters after (De Wet & Huckabay, 1967)

This race is based on S. caudatum including S. nigricans of subseries Caffra. Caudatum Sorghums have characteristic asymmetric turtle-backed grains that are flat or even concave on one side and distinctly curved on the opposite side (bulging on the side where the embryo is visible). The grains are usually exposed between the shorter glumes at maturity. The inflorescences range in shape from compact to open.

1) Plants annual or weekly perennial, without rhizomes or at most with short rhizomatous-like structures; inflorescence variable subsp. bicolor.

2) Racemes tough; mature sessile spikelets persistent; grains often exposed by the gaping glumes; inflorescence compact to somewhat loose; cultivated ………………………………var. bicolor.

3) Grains enclosed by the glumes, although often visible between the glumes at maturity; inflorescence usually rather loose; ………………………………………………..race bicolor.

3) Grains usually exposed by the gaping glumes, often longer than and extruding from them.

4) Lower glume of sessile spikelets either transversely wrinkled or depressed at the middle, or with a strongly nerved herbaceous tip when in flower; inflorescence usually compact;……………………………………………………….race durra.

4) Sessile spikelets with the glumes not as above.

5) Grains are characteristically asymmetrically turtle-backed, that are a flat or even concave on one side and distinctly curved on the opposite side (bulging on the side whre the embryo is visible:…………….race caudatum

5) Grains not so

6) Inflorescence usually somewhat loose; sessile spikelets about twice as long as broad when in flower; …………………………….....race guinea.

6) Inflorescence contracted or at least compact, rarely somewhat loose; spikelets often almost as broad as long when in flower…...... race kafir.

2) Racemes articulating at maturity; sessile spikelets deciduous; grain completely enclosed by the larger glumes; inflorescence loose and usually open; wild or weedy.

7) Leaves 2-7 cm wide, up to 75 cm long; inflorescence loose and broad; sessile spikelets usually with a slender, 5-10 mm long awn ……………………………..var. arundinaceum

7) Leaves 0.5-3cm wide, up to 70 cm long, inflorescende variable

8) Leaves mostly about 35 cm long; inflorescence usually somewhat narrow, sessile spikelets with often with a stout, 10-30 mm long awin……………var. aethiopicum. 138

8) Leaves often up to 70 cm long; inflorescence mostly very loose and broad; sessile spikelets often with a slender, 10-20 mm long awn;……………var. verticilliflorum.

1) Plants perennial with well developed and extensive rhizomes; racemes fragile; sessile spikelets deciduous at maturity; grains enclosed by the longer glumes; inflorescence loose and usually open;

wild or weedy;………………………………………………………………………………subsp. halepense.

Fig. 57 Diagram of a spikelet pair showing floret positions scored for sex expression. 1: sessile spikelet , proximal floret; 2: sessile spikelet, distal floret; 3: pedicellate spikelet, proximal floret; 4: pedicellate spikelet, distal floret.

139

Morphology and architecture of the inflorescence of Sorghum, common in variations to all Andropogoneae, a tribe (group of genera) within the Poaceae, to which Sorghum belongs.

3.8.4. Illustrations of the main traits of Sorghum bicolor Taken from (Doggett, 1988) repeated here from the chapter on S. bicolor subsp. bicolor, to facilitate the use of the above simplified key. 140

Fig. 58 The inflorescence and spikelets of Sorghum bicolor var. bicolor. 1: Part of panicle: a: internode of rhachis, b: node with branches, c: branch with several racemes. 2: Raceme: a: node, b: internode, c: sessile spikelet, d: pedicel, e: pedicelled spikelet, f: terminal pedicelled spikelets, g: awn. 3: Upper glume: a: keel, b: incurved margin. 4: Lower glumes: a: keel, b: keel- wing, c: minute tooth terminating keel. 5: Lower gemma: a: nerves. 6: Upper lemma: a: nerves, b: awn. 7: Palea. 8: Lodicules. 9: Flower: a: ovary, b: stigmas, c: anthers. 10: Grain: a: hilum. 11: Grain: a: embryo mark, b: lateral lines (1x 2/4, 4x 2-9, 5x 10- 11 (Doggett, 1988)

141

Fig. 59 S. bicolor (Linn.) Moench var. bicolor (Pers.) Snowden (Snowden, 1936) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 4: Mature fruiting spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 6: Mature fruiting spikelet with two pedicelled spikelets. 7: Grain. 8: Longitudinal section of grain. 9: Transverse section of grain. (Doggett, 1988)

Fig. 60 Sorghum conspicuum var. conspicuum (Snowden) race guinea (Snowden, 1936) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 4: Mature fruiting spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 6: Mature fruiting spikelet with two pedicelled spikelets. 7: Grain. 8: Longitudinal section of grain. 9:Transverse section of grain. (Doggett, 1988)

142

Fig. 61 Sorghum durra (Forsk.) Stapf var. aegyptiacum (Koern.) Snowden (Snowden, 1936) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 4: Mature fruiting spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 6: Mature fruiting spikelet with two pedicelled spikelets. 7: Grain. 8: Longitudinal section of grain. 9:Transverse section of grain. (Doggett, 1988)

Fig. 62 Sorghum caudatum Stapf var. caudatum (Hack.) Snowden (Snowden, 1936) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 4: Mature fruiting spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 6: Mature fruiting spikelet with two pedicelled spikelets. 7: Grain. 8: Longitudinal section of grain. 9:Transverse section of grain. (Doggett, 1988)

143

Fig. 63 Sorghum caffrorum Beauv. var. albofuscum (Koern.) Snowden (Snowden, 1936) 1: Part of raceme. 2: Lower glume of sessile spikelet. 3: Upper glume of sessile spikelet. 4: Mature fruiting spikelet. 5: Mature fruiting spikelet with one pedicelled spikelet. 6: Mature fruiting spikelet with two pedicelled spikelets. 7: Grain. 8: Longitudinal section of grain. 9:Transverse section of grain. (Doggett, 1988)

For more literature and discussion see chapter on evolutionary dynamics and about recent results on numerical taxonomy in chapters numerical taxonomy and molecular taxonomy.

3.8.5. Recent taxonomic treatment of Sorghum bicolor (L.) Moench

(Wiersema & Dahlberg, 2007)

144

4. Biogeography of Sorghum

4.1. Distribution of the genus Sorghum

4.1.1. Word wide distribution of annual cultivated Sorghum The distribution of the genus Sorghum as delimited by (Garber, 1954) in the broad sense including the subgenera has clearly a worldwide distribution. See the description of various occurrences in (Doggett, 1988) and his subtle debate about the influence of the continental drift, a topic to be followed up if we want to understand the global distribution of the genus Sorghum. The distribution of the Section Sorghum overlaps with that of Sorghastrum, Heterosorghum and Para-Sorghum.

The genus Sorghum itself is strictly Old-World in its original distribution, extending almost continuously from Southern Africa to subtropical Australia. The cultivar-weed complex of Sorghum reached Australia and the New World only after the colonization of the Europeans. There is no doubt that today Sorghum has a worldwide distribution and is one of the worlds most important staple crops. A global map is taken from the website of the University of Hohenheim, a description of the breeding program on the website and comments in the chapter on Sorghum breeding. http://www.uni-hohenheim.de/~ipspwww/350b/index.html origin of Sorghum: red, distribution: green

Fig. 64 Global map on annual Sorghum cultivated around the world. Source: University of Hohenheim, http://www.uni- hohenheim.de/~ipspwww/350b/index.html

Three maps on the global distribution of annual Sorghum cultivars from http://www.tropicalforages.info/key/Forages/Media/Html/Sorghum_(annual).htm 145

Fig. 65 Asian and Australian distribution of cultivated annual Sorghum. From the website of http://www.tropicalforages.info/key/Forages/Media/Html/Sorghum_(annual).htm

Fig. 66 Central and South American Distribution of cultivated annual Sorghum, source: http://www.tropicalforages.info/key/Forages/Media/Html/Sorghum_(annual).htm 146

Fig. 67 African distribution of annual cultivated Sorghum. Source: http://www.tropicalforages.info/key/Forages/Media/Html/Sorghum_(annual).htm

Fig. 68 Distribution of cultivated Sorghum races, from (Deu et al., 2003) 147

4.1.2. Sorghum, geographical distribution of the staple crop in Africa

Fig. 69 Sorghum the leading crop of Africa as a whole and one of the four outstanding cereal grains in the world (along with maize, rice, and wheat), was first brought under cultivation by the Negroes of the western Sudan, probably in the fifth millennium before the Christian era. Of its numerous subspecies, Guinea corn (var. guineense) is especially prominent in the region of origin and in the central Sudan; Durra (var. durra), developed in Ethiopia more than a thousand years later, assumes the leading position in East Africa; and Kafir corn (var. caffrorum), presumably evolved fairly early in the Christian era, holds first place from southern Tanganyika to Natal and Bechuanaland. Still other varieties have been developed subsequently in countries to which the plant has spread, notably Malaya, China, and the United States. In Africa, the only introduced crop to which Sorghum appears to have lost substantial ground is maize, particularly in the south. Original caption from (Murdock, 1960), the caveats see the comments in this study on the race Durra cited from (Stemler et al., 1975a)

148

4.1.3. Distribution of Sorghum bicolor taxa and Sorghum halepense Its members are cultivars and weeds, often encountered in man made habitats (Harlan & De Wet, 1972, 1973) Two morphologically distinct complexes, a Mediterranean and a tropical one, were demonstrated by (De Wet & Huckabay, 1967). The Mediterranean ecotype is represented by small plants with narrow leaves and 2n = 40 chromosomes, extending eastward through Asia Minor to the mountains of West Pakistan, where they are replaced by the more robust tropical ecotype with broad leaves, it might have resulted from hybridization with cultivated Sorghums.

4.1.3.1. Wild Sorghums in Europe and Africa

Fig. 70 Distribution of S. bicolor, 13: S. halepense 14: S. bicolor var. verticilliflorum (arundinaceum), 15: S. bicolor var. aethiopicum. From (De Wet & Huckabay, 1967)

4.1.3.1.1. Sorghum halepense Sorghum halepense is widely distributed in the Mediterranean region and extends over India to the islands of South East Asia. There are no records from South Africa. 149

S. halepense is native of southern Eurasia east to India (Monaghan, 1979) and has been introduced as a highly estimated forage grass in the United States at the beginning of the 19th century. Slender plants with relatively small inflorescences and narrow leaf blades occupy the western range of the species. These specimens are commonly classified as S. halepense while more robust plants with large inflorescences and broader leaf-blades, and which occupy the eastern half of the species range are usually recognized as S. miliaceum. The species has also been introduced as a weed to all warm temperate regions of the world.

In the Americas it has introgressed with grain Sorghum to produce the widely distributed Johnsongrass (Celarier, 1958). In Argentina, derivatives of such introgression were described as S. almum Parodi, Columbusgrass (Parodi, 1943). Sorghum almum has also been used for a while in Australia, but it is now replaced there with a triple hybrid S. bicolor x halepense x verticilliflorum (Zeller, 2000). (Chen et al., 1993) found out that the cytoplasmatic sterile plants contain a slightly smaller 3.7 kb HindIII chloroplast DNA fragment, whereas the fertile plants from S. versicolor, S. almum, S. halepense, and Sorghastrum nutans (Yellow Indiangrass) also have the 3.8 kb fragment, and CMS lines of those taxa studied containing Al, A2 and A3 cytoplasms have the 3.7 kb fragment. (Morden et al., 1990) provide a differentiated discussion on the origin of Sorghum halepense, call for further studies, but conclude that S. bicolor (in the broad sense) is one of the parents of Sorghum halepense.

In fact, Johnsongrass in all its derivatives forms with cultivated Sorghums a complex with feral populations, where hybridization is going forth and back and the resulting populations are often difficult to determine to their origin. This is extensively discussed in the chapter on evolutionary dynamics of the genus Sorghum below.

According to (Ejeta & Grenier, 2005; Warwick & Stewart Jr., 2005) the existence of intermediary forms in most Sorghum growing areas of the warm temperate zones in the Americas offers an empirical evidence for noxious weedy forms arising from continued introgression (exo-ferality) among different Sorghum types.

150

4.1.3.1.2. Distribution of the weedy taxa of Sorghum in Africa

Fig. 71 Distribution of the series spontanea. (Sorghum volelianum should read Sorghum vogelianum) From (De Wet et al., 1970)

151

Fig. 72 Distribution of the wild varieties of Sorghum bicolor, original caption from (De Wet et al., 1970) caption see above. See the full list of the taxa in the previous figure. The wild varieties in this map comprise Sorghum race aethiopicum, arundinaceum, verticilliflorum and virgatum.

152

The weedy taxa of Spontanea reproduced in (De Wet et al., 1970) match almost exactly, by crossing various wild members of Spontanea with different races of cultivated Sorghums. The wild taxa are widely distributed in Africa. Sorghum arundinaceum and Sorghum vogelianum are confined, almost exclusively, to the wet, tropical forest of the Guinea coast and the Congo. No combination of morphological traits could be found that would consistently distinguish between these two taxa. They are robust perennials with 10-20 feet tall, erect culms. Well-developed leaf blades are about 7 cm wide and 50-70 cm long. Their inflorescences are open, often over 50 cm long, with long and flexuous lower branches that are usually undivided for 5-10 cm from the base. Racemes are very fragile, with 2-8 spikelet pairs. The sessile spikelets are 6-10 mm long, lanceolate, often awned, and densely hairy to glabrous on the back.

4.1.3.1.2.1. Sorghum bicolor var. arundinaceum (in many publications now under S. verticilliflorum) The common weedy Sorghums of tropical and subtropical West Africa running under the name of S. arundinaceum belong now taxonomically to S. bicolor var. verticilliflorum , extending from Sierra Leone along the moist belt surrounding the coastal tropical forest to about 15° south latitude. These Sorghums are commonly encountered around waste places, invade cultivated fields, and seem to be naturally adapted to stream banks.

4.1.3.1.2.2. Sorghum bicolor var. aethiopicum In the drier inland regions, extending north of the equator from northern Nigeria eastward to Somaliland and along the Nile valley to Cairo, variety aethiopicum is the common weedy Sorghum (see map 15) growing in the same habitats as var. arundinaceum, with which it overlaps at the western edge of the distribution. There, extensive hybridization between the two varieties are recorded, particularly in Northern Nigeria.

4.1.3.1.2.3. Sorghum bicolor var. verticilliflorum Variety verticilliflorum is the widest distributed feral complex, with a broad climatic plasticity, extending almost continuously east of 20° east longitude from the South African coast to 10° north latitude (see two maps above and below). It overlaps along its northern and northwestern borders with var. aethiopicum and var. arundinaceum, again hybridizing extensively. It grows over extensive areas of tall grass savannah in the Sudan (Wood & Lenne, 2001)

4.1.3.1.2.4. Sorghum bicolor var. virgatum Has been collected only along the valley of the Nile, from Cairo south to central Sudan and along irrigation ditches in northeastern Chad. (De Wet et al., 1970)

153

Fig. 73 Distribution of S. bicolor 16: subsp. bicolor var. verticilliflorum, 17: subsp. bicolor race guinea, 18: subsp. bicolor var. bicolor race kafir. From (De Wet & Huckabay, 1967) The cultivated Sorghums are even more variable than the feral complexes through continuous selection and extensive (local!) exchange of germplasm, including hybrid plants. Still it is possible to geographically distinguish consistently between four cultivated complexes:

154

4.1.3.2. Distribution of cultivated races of S. bicolor in Africa

Fig. 74 Distribution of cultivated Sorghums in Africa, from (De Wet et al., 1972)

4.1.3.2.1. S. bicolor race guinea in Africa The commonly cultivated Sorghum of Tropical West Africa is 'guinea corn' (see map 17 above). It is widely cultivated in areas with over 1000mm of annual rain fall (Johnson, 1958) Along the eastern edge of guinea Sorghum cultivation, hybridization with members of the race kafir gave rise to the complex that extends along Eastern Africa south to Zululand. 155

4.1.3.2.2. S. bicolor race bicolor, kafir and durra in Africa Race kafir is the dominant cultivated Sorghum south of 5° north latitude and east of 20° east longitude (see map 18). It is also widely cultivated in northern Nigeria, and gene flow from race kafir into guinea Sorghum is evident. Distribution and morphology suggest close relationships with variety verticilliflorum, and durra Sorghums might have originated out of selections from race kafir.

Fig. 75 Distribution of S. bicolor. 19: subsp. bicolor var Sorghum race durra; 20: subsp. bicolor var Sorghum race bicolor. 21: Distribution of rice (horizontal lines), root crops (vertical lines), and Sorghum (solid circles). From (De Wet & Huckabay, 1967)

156

4.1.3.2.3. Race bicolor This race probably represents relics of ancient cultivated kinds. It shows the most ancestral spikelet morphology, the race has usually very small grains that are fully enclosed by the long, clasping glumes which usually enclose the grains at maturity. The inflorescence is usually open, the racemes baring grains are sitting at long panicle branches. They resemble spontaneous weedy Sorghums, but they lack the ability of natural seed dispersal. (they do not shatter). They are low yielding as cereals, but they grow nearly always besides Sorghum cultivars. They are rarely an important crop, except in Ethiopia West of the Rift Valley. Most bicolors are now grown for their sweet stalks or for production of beer or dye.

(deOliveira et al., 1996) tested three different molecular marker technologies to determine the relatedness of 84 different lines of Sorghum. Both racial characterization and geographical origin were found to be correlated with relatedness. In some cases, the region of origin was the more significant factor, where samples of different races from the same locality were more closely related than were samples of the same race from different localities. Wild Sorghums were shown to have few novel alleles, suggesting that they would be poor sources of germplasm diversity. The results also indicated that Chinese Sorghums are a narrow and distinctive group that is most closely related to race bicolor.

4.1.3.2.4. Race kafir This race is based on subseries Caffra of Snowden with the exclusion of S. caudatum and S. nigricans. It is characterized by more or less compact inflorescences that are often cylindrical in outline. Sessile spikelets are typically elliptic with the glumes, at maturity, tightly clasping the usually much longer broadly ellipsoid grain. The part of the grain which extends beyond the clasping glumes is symmetrical and round. Kafir Sorghums are found in parts of Africa south of the equator, mainly in areas with one well-defined rainy season such as Natal, Transvaal, and Orange Free State (Snowden, 1936).

4.1.4. Summary distribution of Sorghum The genus Sorghum is strictly Old-World in its original distribution, extending almost continuously from Southern Africa to subtropical Australia. The cultivated-weed complex of Sorghum reached Australia and the New World only after the colonization of the Europeans. The cultivated Sorghums are even more variable than the feral complexes through continuous selection and extensive exchange of germplasm, including hybrid plants. Still it is possible to distinguish between four distinct cultivated complexes.

4.1.5. Bibliography distribution of Sorghum http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Biodiversity-20060806.pdf

4.2. Centers of biodiversity and centers of crop origin, the generalities Centers of Biodiversity and centers of crop origin are not necessarily the same, as shown in the maps below: 157

4.2.1. Metric map of global biodiversity

Fig. 76 A metric map of biodiversity, based on counting per area: It shows surprising results, which are not in all aspects matching other world map views such as centers of biodiversity or centers of crop diversity. Source: http://www.nhm.ac.uk/research-curation/projects/worldmap/ Wilhelm Bartlott, Bonn, Germany (Kier et al., 2005; Lughadha et al., 2005)

There are many global maps on biodiversity existing, most of them have been done by compilation, only a very few build on metric data, like the one conceived by Wilhelm Barthlott, see (Kier et al., 2005; Lughadha et al., 2005). Despite this unusual data acquiring the picture presents near to expectation of the present day knowledge: Sub-Saharan Africa, Central South Asia and Central America show clearly the highest concentration of biodiversity.

4.2.2. Hot spots of conservation and centers of biodiversity The second map shows the hot spots of plant conservation, As many as 44% of all species of vascular plants and 35% of all species in four vertebrate groups are concentrated to 25 hotspots comprising only 1.4% of the land surface of the Earth. This opens the way for a `silver bullet' strategy on the part of conservation planners, focusing on these hotspots in proportion to their share of the world's species at risk. It should be noted that many of the hot spots match with the centers of crop diversity, except for Africa.

158

Fig. 77 Centers of 25 hotspots of biodiversity and conservation (Myers, 2000) There is a generally good overlap of the hot spot areas of Myers with a new survey done by (Kuper et al., 2004) in Africa, although there are differences, mainly due to better data quality.

4.2.3. Centers of crop biodiversity Alphonse de Candolle - in 1882 based his ideas on geographical centers of cultivars on floras, historical writings, archaeological information, ethnological data, linguistics etc. N.I.Vavilov added genetics, chromosome studies, and anatomical data. He was interested in the presence of wild ancestors of the crops. Vavilov concluded that the most likely areas of origin are there where the crop was cultivated and their wild ancestors grew in vicinity (Vavilov, 1926, 1951; Vavilov, 1987; Vavilov, 1940). Furthermore, the areas of origin should include a high amount of variation. Vavilov decided that on a global scale there were 8 major areas that met these requirements. These were: Mexico and Central America, the Central Andes, Abyssinia, the Mediterranean, India (Middle East), SW Asia, China, and SE Asia. Later, he added more characters until he got up to about 20. At this point the concept of centers of crop diversity got definitely blurred.

159

Fig. 78 The original eight centers of crop diversity according to Vavilov, N.I. (Hawkes, 1983; Hawkes, 1990, 1991, 1999; Hawkes & Harris, 1990; Vavilov, 1987; Vavilov, 1940; Williams, 1990)

The centers of diversity which Vavilov described were not discrete but overlapped for a number of crops, as regions which have concentrations of variation assessed in terms of recognizable botanical varieties and races. But he also included a complex of properties which include physiological and ecotype characters (Williams, 1990). This is why the concept of the biodiversity centers underwent later many amendments and enlargements, which resulted among others in the map of (Zeven & Zhukovsky, 1975).

160

Fig. 79 P.M. Zhukovsky's alterations (solid lines) and additions (broken lines) to Vavilovs original concept of crop diversity, from (Zeven & Zhukovsky, 1975)

4.2.4. Centers of crop origin The centers of crop origin are still under discussion and change (Harlan, 1971). It is clear from the comparison of the distribution of Sorghum species and races in Africa that there is no fixed opinion in the biogeography of crop and biodiversity center of African Sorghums. S. bicolor subsp. bicolor shows a rather widely W-E - extended area of origin in sub-Saharan Africa.

161

Fig. 80 Centers and noncenters of agricultural origins (A1, Near East center; A2, African noncenter; B1, North Chinese center; B2, Southeast Asian and South Pacific noncenter; C1, Mesoamerican center; C2, South American noncenter, from (Harlan, 1971)

Harlan proposes the theory that agriculture originated independently in three different areas and that, in each case, there was a system composed of a center of origin and a noncenter, in which activities of domestication were dispersed over a span of 5000 to 10000 kilometers. One system includes a definable Near East center and a noncenter in Africa; another center includes a North Chinese center and a noncenter in Southeast Asia and the South Pacific; the third system includes a Mesoamerican center and a South American noncenter. There are suggestions that, in each case, the center and the noncenter interact with each other. Crops did not necessarily originate in the centers of their highest diversity (or in any conventional concept of the term), nor did agriculture necessarily develop in a geographical centre. Harlan developed these ideas further on in (Harlan, 1992b), the case of Sorghum is given in chapter 4.3.

4.2.5. Summary Centers of Biodiversity and Crop Biodiversity are not the same, although sometimes they overlap. Centers of crop origin have changed in concept and are still in debate, since in many cases reliable archaeobotanical data still lack. The best solution is the one from (Harlan, 1971) on Centers and Non-Centers, see map above: (A1, Near East center; A2, African noncenter; B1, North Chinese center; B2, Southeast Asian and South Pacific noncenter; C1, Mesoamerican center; C2, South American noncenter. 162

4.2. Centers of origin of the genus Sorghum according to present day agriculture and landrace distribution Introductory remarks: This chapter on the Centers of Origin can only be fully understood if one takes into account the chapter on reproduction and evolutionary dynamics below, but it is on the other hand so strongly related to biogeography, that the basics are given here, including some few examples of landrace distribution studies, which will be dealt with extensively in the following chapters on evolutionary dynamics and gene flow.

4.2.1. One or several centers of origin for Sorghum ?

Fig. 81 Probable areas of domestication of selected African crops: 1, Brachiaria deflexa; 2, Digitaria exilis and Digitaria iburua; 3, Oryza glaberrima; 4, Dioscorea rotundata; 5, Musa ensete and Guizotia abyssinica; 6, Eragrostis tef; Voandzeia and Kerstingiella; 8, S. bicolor; 9, Pennisetum americanum; 10, Eleusine coracana, from (Harlan, 1971) 163

Despite this map, (Harlan, 1971) states, that it is not possible to clearly locate a center for Sorghum domestication on botanical evidence alone and supports thus the view of (De Wet & Huckabay, 1967). The wild races are widespread and often extremely abundant. Harlans understanding is based on patterns of variation and genetic interaction among wild species, weedy relatives and cultivated races which all would suggest as a center of origin a wide zone in the broad-leaved savanna belt that stretches from about lake Chad to eastern central Sudan. Vast amounts of wild Sorghum are found along the Sudan-Ethiopia border, but there is no indication that the area was ever farmed before government settlement projects were established. Variations in Sorghum do not suggest that its homeland is Ethiopia; by far the bulk of Ethiopian Sorghums are durras, which are the most specialized and derived cultivated Sorghum. Still, pattern of domestication and the spread of Sorghum lets (Ejeta & Grenier, 2005) come to the conclusion, that Sorghum must have spread from a region including parts of Sudan and Ethiopia.

(Harlan, 1992b) gave a latest overview on the centers of biodiversity of Sorghum for Africa: Compare for more details chapter on evolutionary dynamics of the genus Sorghum.

4.2.2. Example from Somalia: Distribution and loss of landraces (Manzelli et al., 2005) give insight in an example from Somalia about the biodiversity and distribution of S. bicolor: They point to the problem of vanishing landraces and their invaluable characteristics, having adapted to harsh conditions over centuries. Today this traditional inter-dependence is at risk with catastrophic consequences for local poor resources farmers and maintenance of genetic diversity within Sorghum. Traditional landraces are being relegated to marginal and risk prone areas as they are replaced by improved varieties. This can lead to the loss of local knowledge of traditional landraces and to an erosion of genetic diversity. This should be taken into account by all breeding programs, including those on biofortification within humanitarian projects like Africa-Harvest.

The case of Somalia follows the same trend observed in other developing countries, where the local agriculture, founded on an ancient relationship between ecological characteristics and social needs, is continuously threatened by unpredictable environmental and social changes. Somali farmers report the loss of old traditional landraces, preferred for traditional bread production, such as Abaadiro, Carabi and Masego. Moreover those farmers have reported a decrease in Sorghum production and consumption due to importation of wheat flour and rice from Ethiopia. This is another reason to work closely with local farmers and listen to their experience, although any kind of romantic view of their situation should be avoided, progress by the introduction of better yielding races does not mean that the cultural and scientific inheritance of highly diverse germplasm has to be traded off. Conservation and improvement of Sorghum accessions are, therefore, of practical value. Improvement of the crop will lead to greater stability of a traditionally reliable food source that is known and understood in Northwest Somalia. The breeding programs of ICRISAT are building on basic work with germplasm collections, actually heading towards 40’000 accessions to the present day. What is even more important, is making these treasures available to the breeders in an efficient way. On the level of landraces, there are many inventories based on morphology and genetics revealing a striking biodiversity, often closely growing together in intricate 164 patterns, and preserving their identity, despite of active seed selection and exchange of the local farmers. Building on accession work from (Grenier et al., 2001a; Grenier et al., 2000a; Grenier et al., 2000b; Grenier et al., 2001b), (Mgonja et al., 2006) propose a selective accession and breeding program, which takes advantage of modern genetics and a better biogeographical insight in the genus Sorghum, respecting at the same time the values of traditional knowledge. For more details see the chapter on Sorghum breeding.

A study by (Mann et al., 1985) shows that cultivated Sorghum arose from the wild S. bicolor subsp. verticilliflorum (incl. arundinaceum). There is no evidence suggesting that presently cultivated Sorghums have evolved from the rhizomatous diploid or tetraploid wild species, see the chapter on evolutionary dynamics.

4.2.3. Domestication of Sorghum started in North-East and Central Africa, the classic model This paragraph serves only as a ‘preview’ on the history of Sorghum landraces, it should serve here only to clarify questions on the origin of Sorghum (which is not the same as the center of biodiversity of Sorghum and not the same as the locations of where archaeobotanically documented domestication of Sorghum started. More details and an account on the history of Sorghum in Africa follows in the chapters on the evolutionary dynamics and the history of domestication of the genus. These lines here must be seen in the context of the biogeography of Sorghum as dealt with in this chapter.

After (Doggett, 1988), domestication of Sorghum crops is presumed to have started in the north-east quadrant of Africa (Harlan & De Wet, 1972) some 5,000 years ago - perhaps in the expanse of arable land area recognized in contemporary geography as Ethiopia and Sudan. The great genetic variability of the crop in this geographical area, the wide range of ecological habitats there, and the long history of human selection efforts in the region have given sufficient credence to the theory of the origin and early domestication of Sorghum in eastern Africa. Several routes have been recognized for the later movement of Sorghum into other parts of Africa and beyond. Early movement between 2,000 and 4,000 years ago has taken the crop to west, central, and southern Africa leading to further domestication and appearance of distinct forms (races) in each of these regions. But as we will see in the following chapters on the history of domestication of Sorghum, domestication must have started 8000 years ago.

One of the most precise analysis stems from an extensive study by (Aldrich et al., 1992) on allozyme enzyme distribution, see the discussion on the distribution of allozymes studied in 78 wild accessions (18 countries) and 351 cultivated accessions (52 countries). See the comments in chapter on gene flow from wild to cultivated Sorghums. Here the comments about the origin of cultivated Sorghums by the authors: Their data suggest that cultivated Sorghum was initially domesticated in northeast-central Africa. These results are consistent with those of (Ollitrault et al., 1989), which indicate that cultivated Sorghum originated from a primitive Sorghum race in eastern, central Africa. One can hypothesize that other domestication events occurred independently of the proposed primary domestication in northeast- central Africa. However, if the “wild” Sorghum from northwest or southern Africa had contributed substantially to the genetic composition of the extant crop's gene pool, then the “wild” Sorghum of these regions should show the greatest genetic similarity to the cultivars of the same region, which is not supported by data. This does not rule out small-scale independent domestications in these regions, but it 165 is unlikely that northwest or southern Africa were as active in the domestication of Sorghum as was northeast central Africa. The data of (Aldrich et al., 1992) also support the view of (Harlan & Stemler, 1976) that cultivated Sorghum originally was selected from a complex consisting primarily of races verticilliflorum and/or aethiopicum. In regions where races arundinaceum and virgatum can be identified as biologically distinct entities, i.e., in the genetically 'purest' sections of their range see below, they are genetically distinct from most cultivated collections. Thus, races aethiopicum and arundinaceum probably have not contributed as greatly to the domestication event that produced the majority of cultivated Sorghum as did races aethiopicum and verticilliflorum, although they may have contributed to the character of the extant crop through introgression (Doggett, 1965). It should be noted, however, that seed collections used by the authors could carry alleles introduced from other lines due to improper control of outcrossing during maintenance at the seed bank. Such occurrences might explain the genetic similarities among some of the collections independently of ancestor-descendent relationships, and the results should be viewed in light of this. The same conclusions are drawn in a similar paper by (Aldrich & Doebley, 1992), who studied in the same manner as above the restriction fragment variation in the nuclear and chloroplast genomes of cultivated and wild Sorghum bicolor. These results are discussed in detail in chapter on evolutionary dynamics of the genus Sorghum.

Fig. 82 Graph of the first two components of a principal component analysis based on isozyme allele frequency data from individual accessions of subsp. arundinaceum and subsp. drummondii. from Countries are abbreviated as follows: AF Afghanistan, AL Algeria, AN Angola, BE Benin, BG Bangladesh, BO Botswana, B U Burma, CD Chad, CH China, CM Cameroon, CO Congo, CR Central African Republic, EG Egypt, ET Ethiopia, GH Ghana, GM Gambia, IC Ivory Coast, ID Indonesia, IN India, IQ Iraq, IR Iran, IS Israel, JA Japan, KE Kenya, KO Korea, LB Lebanon, LS Lesotho, MA Mall, ML Malagasky Republic, MW Malawi, MZ Mozambique, NG Niger, NI Nigeria, NP Nepal, PK Pakistan, SA South Africa, SE Senegal, SI Sierra Leone, SL Sri Lanka, SO Somalia, SU Sudan, SW Swaziland, TA Taiwan, TH Thailand, TU Turkey, TZ Tanzania, UG Uganda; UR USSR, UV Upper Volta, YE Yemen, ZA Zaire, ZI Zimbabwe, ZM Zambia. From (Aldrich et al., 1992) 166

Fig. 83 Graph of the first two components of a principal component analysis based on wild and cultivated isozyme allele frequency data pooled by country of origin. Abbreviations for countries of origin are listed in the caption of the Figure above, from (Aldrich et al., 1992)

Fig. 84 Average linkage cluster analysis based on isozyme allele frequency data for wild and cultivated accessions pooled by country of origin using modified Rogers' distance (Wright, 1978), from (Aldrich et al., 1992)

167

4.2.4. Ancestral farmers chose for the earliest domestication efforts crops which grow in monodominant stands.

Taking into account a seminal paper of (Wood & Lenne, 2001): Sorghum itself can be found in monodominant stands. Harlan identified the verticilliflorum race of Sorghum bicolor as the progenitor of cultivated sorghums, and noted that it was found as the chief dominant, in enormous quantities, of the extensive tall-grass savanna of Sudan and Chad (Harlan, 1989). Harlan also noted for Africa: ‘Massive stands of truly wild races of Sorghum can be found widely distributed over the savanna zones’.45 Sorghum was domesticated somewhere along a belt south of the Sahara from Chad to western Ethiopia. 46 The races, aethiopicum and verticilliflorum of Sorghum bicolor are often dominant grasses in the northern savanna of Africa.47 These ‘massive stands’ of annual wild Sorghum provide both an evolutionary and ecological pedigree for monoculture Sorghum cropping. Typically enough, (Ayana et al., 2000a) found a remarkably narrow genetic basis for the variety verticilliflorum in their analysis. But Sorghum does not stand alone with classic monodominant habitats: Wild rice Oryza coarctata is reported in Bengal as growing in simple, oligodiverse pioneer stands of temporarily flooded riverbanks (Prain, 1903); Harlan ascribed Oryza (Harlan, 1989) to monodominant populations and illustrated harvests from dense stands of wild rice in Africa (Oryza barthii, the progenitor of the African cultivated rice, Oryza glaberrima). Oryza barthii was harvested wild on a massive scale and was a local staple crop across Africa from the southern Sudan to the Atlantic. Grain yields of wild rice stands in Africa and Asia could exceed 0.6 tonnes per hectare — an indication of the stand density of wild rice (Evans, 1998). Botanists and plant collectors have repeatedly and emphatically noted the existence of dense stands of wild relatives of wheat (Wood & Lenne, 2001). For example, in the Near East, massive stands of wild wheats cover many square kilometers (Harlan, 1992a). Wild einkorn (Triticum monococcum subsp. boeoticum) in particular tends to form dense stands, and when harvested, its yields per square meter often match those of cultivated wheats under traditional management (Hillmann, 1996). Wild Einkorn occurs in massive stands as high as 2000 meters altitude in south-eastern Turkey and Iran (Harlan & Zohary, 1966). Wild emmer (Triticum turgidum subsp. dicoccoides) grows in massive stands in the northeast of Israel, as an annual component of the steppe-like herbaceous vegetation and in the deciduous oak park forest belt of the Near East (Nevo, 1998). Wild wheat was also recorded to grow in Turkey and Syria in natural, rather pure stands with a density of 300/ m² (Anderson, 1998).

4.2.5. Good arguments that important domestication processes in Sorghum have happened in Southeastern Asia In (Kimber, 2000) on “Origins of Domesticated Sorghum and its Early Diffusion in India and China” there are (p. 45-68) lots of arguments given, that important domestication phases have happened in Neolithc and Bronze periods. The possible routes of migration of early Sorghum races must have followed the ancient trade activities, this made it possible that secondary domestication zones were established in Asia:

168

Fig. 85 Early trade and pilgrimage routes in Asia, compiled by (Kimber, 2000) with data from (Barraclough, 1985; Grutz, 1998; Needham & Bray, 1984; Warmington, 1974)

169

Fig. 86 The origins of agriculture and the expansion of Neolithic agriculture on the Eurasian continent. 1, the area of origin of wheat and barley in Southwest Asia; 2, the distribution of Neolithic agriculture in Europe, West Asia and India; 3, the area of origin of rice in the Middle and Lower Yangtze Basin; 4, the Neolithic distribution of rice in East and Southeast Asia; 5, the area of origin of millet in the middle reaches of the Yellow River; 6, the distribution of millet in the Neolithic; 7, the area of overlap of rice and millet during the Yangshao Cultural period, From (Li et al., 2007)

It is now clear from high-resolution archaeobotanical evidence, genetic studies and cultural chronologies (eg, (Morrell et al., 2003), (Armelagos & Harper, 2005a, b) that agriculture emerged as the predominant form of food production directly from a huntergatherer background in at least two major regions of the Eurasian continent (Bellwood, 2001, 2005) (Figure 86). Bellwood published a world map with an impressive view on multiple origins of crop domestication, due to early migration: The assumption of Bellwood is that the spread of Afroasiatic occurred as a result of actual human movement, not language diffusion alone. There is no significant archaeological evidence for a population movement from Africa into the Levant, whether Mesolithic or Neolithic, at the time in Question. But the controversy goes on (Bellwood et al., 2004), if far from over, and this complex matters over time, language, agriculture and geography needs careful multidistiplinary (better transdisciplinary) approaches.

170

Fig. 87 The distribution of prehistoric agriculture, with some widespread prehistoric archaeological complexes that appear to be associated with early agriculturalist expansion, right part of the fig. from (Bellwood, 2001).

The Fertile Crescent of Southwest Asia was the place of origin of domesticated wheat and barley between about 9500 and 7500 BC (eg, Zohary and Hopf, 2000; Diamond and Bellwood, 2003; Bellwood, 2005; Wilcox, 2005). Earliest agriculture in China was just as revolutionary as that in Southwest Asia, and may have a similar chronology (Underhill, 1997). Rain-fed crops and wetland rice-based agriculture were developed in the Yellow and Yangtze River basins independently by the early Holocene (Zhao, 1998; Shelach, 2000; Crawford, 2006) (Figure 1). Barley is known from the western Yellow River region between 2000 and 800 BC (Zhao, 2005).

171

Below the latest published interpretation of the origin of cultivated Sorghum (Ejeta & Grenier, 2005)

Fig. 88 Pattern of Domestication and Spread of the Genus Sorghum. (Ejeta & Grenier, 2005)

One of the most comprehensive accounts on the origin of Sorghum, taking into account numerous ascpects including domestication, has been published in (Smith & Frederickson, 2000) as a chapter: (Kimber, 2000): It seems to be the race guinea to be the most ancestral one with its wide open panicle. Already and (De Wet & Harlan, 1971)

4.3. Possible origin of Sorghum in the ancient green Sahara There are a number of reasons why the pattern of domestication spreading from the Sudan and Ethiopia proposed by (Ejeta & Grenier, 2005) might be correct for younger periods back to 2000 B.C., but this does not mean that the real origin of Sorghum would fall into the same pattern. It became clear in the last years with an increasing number of archaeological data that in earlier periods of the Holocene the Sahara was definitely more humid and could well have harbored ancient crops, including Sorghum.

172

Before we can dig into the new hypothesis of Sorghum origin in the green Sahara, we must review the history of domestication in relation to human history and climate history:

4.3.1. Archaeobotany of Sorghum We still lack a great deal of knowledge about the history and the origin of our crops, a lot of work needs to be done in the next decades. We will have to build on modern archaeobotanical methods, some of them are described, including literature references, in (Ammann et al., 2005; Jacomet & Kreuz, 1999)

It is known only from a few incidences when S. bicolor was first brought into cultivation. (Murdock, 1960) proposes that this cereal, as well as several other West African crops were domesticated some 7,000 years ago. (Doggett, 1965) suggests that domestication took place somewhere in the northeastern quadrant of Africa some 5,000 years ago. Sorghum must have been domesticated in the savanna zone south of the Sahara, and certainly more than 3,000 B.P. ago. (Harlan & De Wet, 1971; Harlan & Stemler, 1976).

However, conclusive archeological evidence of the early evolutionary history of Sorghum is in many regions lacking, except in a very few cases. Most identifiable remains of cultivated Sorghum in Africa date back only to the early centuries of the Christian era. (Plumley, 1970) discovered a beautifully preserved sheath of durra Sorghum at Qasr Ibrim dating back to Meroitic times. (Clark & Stemler, 1971) identified S. bicolor from Jebel et Tomat in the Sudan (ca. 245 A.D.). Caudatum Sorghum dated to the ninth century A.D., were found at Daima (Connah, 1971). Cultivated Sorghum, probably belonging to race guinea, was abundant at Muramasapwa in Malawi during the ninth century A.D. (Robinson, 1966), and kafir Sorghums are commonly associated with early iron-age Bantu dwellings across southern Africa (Fuller, 1975). Sorghum domestication certainly began earlier than these dates would suggest. During the beginning of the Christian era, modern races were already well established and, except for race durra, were probably widely distributed across their present ranges. Other traces with Sorghum seeds from desertic areas are recorded from plant remains from Nabta Playa (site E-75-6) dated to the late ninth millennium bp consisted of Sorghum and a broad spectrum of wild grasses, in all 20,000 seeds. All of the plants were morphologically wild and grow today in the Sahelian zone, with summer rain (Wasylikowa et al. 1997; Wendorf & Schild 1998; 2001; 2002). It was probably during this time that people in the Khartoum area started to cultivate local wild sorghum (Haaland 1987; 1995; 1999). This was the beginning of a long history and remarkable diversity in different forms of Sorghum based foodstuffs, porridge and beer (Edwards 2003).

(Nicoll, 2004) reviews the various Late Quaternary records that are available from western Egypt and northern Sudan, which includes more than 500 published radiocarbon dates and various sedimentary archives from local landscape components, including palaeolakes, soils, drainages (wadis), and archaeological sites. This palaeoenvironmental compilation frames the spatial and temporal context of local cultural activities when the region was most hospitable similar to 9000-6000 BP; at this time, monsoonal weather influenced the portion of the African continental interior, creating enough convective rainfall for occasional surface water storage. In this part of the modern Sahara, rapid hydroclimatic changes play a key role in geomorphic evolution and resource availability. As 'watering 173 holes' formed and dried up in the Early to Middle Holocene, Neolithic people developed various subsistence strategies, including opportunistic hunting of small animals (e.g. gazelle and hare), and food- related (e.g. wild Sorghum, millet, and legumes) activities: gathering, plant cultivation and livestock- rearing. During its wettest phases during the 'monsoonal maximum,' the area was drought-prone, sustaining a meager steppe shrub desert flora. Further desertification and aeolian deflation during the Middle and Late Holocene fostered technological innovation, migration and settlement, as well as the further development of agrarian communities and complex culture.

Evidence of early Sorghum cultivation in Africa and India is often cited, but the data on which these conclusions were until recently at best circumstantial. Sorghum domestication probably dates back to the advent of agriculture in sub-Saharan Africa (Clark, 1976). (Harlan & Stemler, 1976) suggest that domestication took place before 3,000 B.P. and that cultivated Sorghum reached northwestern India during the latter part of the second millennium B.C. Other African cereals, Pennisetum americanum (L.) Leeke (pearl millet), and possibly Eleusine coracana (L.) Gaertner (finger millet) were well established as crops in the SindPunjab region of India by 1000 B.C. From northwestern India the crop must have been introduced to China. This cereal became an important crop in China only after the Mongol conquest (Hagerty, 1940). Since the fifteenth century Sorghum has also been widely introduced into the New World and during the last several decades has become an important crop in the U.S. and Mexico. (Dahlberg et al., 2001; Zeller, 2000).

Although not generally known, in the early nineties, a discovery pushed back the origin of Sorghum domestication for many thousand years, this opens new perspectives for the history of domestication: Archaeobotanical results allow to trace back Sorghum grains to much earlier times than assumed before: (Wasylikowa et al., 1997; Wendorf et al., 1992). Excavations at an early Holocene archaeological site in southernmost Egypt, 100 km west of Abu Simbel, have yielded hundreds of carbonized seeds of Sorghum and millets, with consistent radiocarbon dates of 8,000 years before present (BP), thus providing the earliest evidence for the use of these plants. They are morphologically wild, but the lipid fraction of the Sorghum grains shows a closer relationship to domesticated than to wild varieties. Whatever their domestication status, the use of these plants 8,000 years ago suggests that the African plant-food complex developed independently of the Levantine wheat and barley complex. This ancient history, going back maybe even to the old Saharan population which existed before the dramatic desertification of huge regions, would also explain the high degree of biodiversity of landraces which does not follow the rules of their present day distribution properly. 174

Fig. 89 Charred plant remains from Site E-75-6. a-c, Grains of Sorghum sp showing the ventral (a) and dorsal (b, c) surfaces, with consistent radiocarbon dates 8000 BP, from Fig. 2, (Wendorf et al., 1992)

4.3.2. History of domestication of Sorghum The genus Sorghum is complex in its taxonomic and evolutionary aspects, and especially the taxonomy of the cultivated Sorghums shows signs of over-classification. There is still a lot of work to be done in order to fully understand the evolutionary dynamics. Most of the scientific papers dealing with the evolutionary dynamics stem from the 70s, in a time of the wake of molecular biology becoming relevant to systematics of specific groups. In the most comprehensive review on Sorghum domestication (Kimber, 2000) raises an interesting new concept on the dynamic processes:

Domesticated grain Sorghums originated from wild members of S. bicolor subsp. verticilliflorum (including S. bicolor subsp. arundinaceum) that were brought into cultivation. (Snowden, 1936), (Porteres, 1962) and (De Wet & Huckabay, 1967) suggested that races arundinaceum, aethiopicum and verticilliflorum could have independently given rise under complex domestication processes to guinea Sorghums, durra Sorghums and kafir Sorghums, respectively.

(Doggett, 1965) and (De Wet & Harlan, 1971), however, demonstrated that the earliest domesticated Sorghums were Sorghum bicolor - like, and that the modern domesticated races were derived from S. bicolors after the initial domestication. Morphological affinities between guinea Sorghums and race arundinaceum, durra Sorghums and race aethiopicum, and between kafir Sorghums and race verticilliflorum can be explained on the basis that gene exchange occurs across weed races belonging to subspecies drummondii. The distribution of related wild and cultivated races coincide in Africa, and hybridization between sympatric wild and cultivated kinds commonly takes place. 175

(Harlan, 1971) proposed that Sorghum was first taken into cultivation along a broad band of the savanna between the Sudan and Nigeria. In this region race verticilliflorum (including arundinaceum) is an abundant wild grass and is collected as a wild cereal in times of scarcity. From this region the cultivation of Sorghum spread into tropical West Africa, the arid northeast and southeastern Africa. Selection for adaptation to a wet tropical habitat produced race guinea (De Wet et al., 1972). This migration probably occurred before 3,000 B.P. (Harlan & Stemler, 1976).

Archeological results allow to trace back Sorghum grains to much earlier times than assumed before: (Wendorf et al., 1992) Excavations at an early Holocene archaeological site in southernmost Egypt, 100 km west of Abu Simbel, have yielded hundreds of carbonized seeds of Sorghum and millets, with consistent radiocarbon dates of 8,000 years before present (BP), thus providing the earliest evidence for the use of these plants. The Sorghum grains found, including all other findings (Wendorf et al., 1992) are members of the natural subdesertic or sahelian flora. The Sorghum and millets correspond in size and structure to modern wild varieties of these plants and thus seem to be parts of an economy based gathering of wild plant foods. The preliminary investigation by infrared spectroscopy of the lipids in the Sorghum grains even suggests the possibility of some cultivation. Five archaeological specimens from four features were compared with specimens (from the herbarium at Kew, London) of three cultivated races of Sorghum bicolor and four wild species of Sorghum: S. arundinaceum, S. aethiopicum, S. verticilliflorum and S. virgatum. The spectrographic results for the archaeological specimens are all very similar, and are closer to the cultivated specimens than to the wild, particularly in the hexane extracts. Considerably more work will be required to investigate this possibility, but it is particularly intriguing as it has been suggested that Sorghum and millet were domesticated somewhere in the African savannah zone (Harlan & Stemler, 1976). The wild relatives of most of the plants of the African savannah complex grow today in the savannah and it has been assumed that they were domesticated somewhere within that zone. There are several theories of where and when this occurred (Chevalier, 1947), but up to the present day there has been no direct archaeological evidence produced. This absence results primarily from the assumption that the domestication of African plants was of a late and secondary development, which began well after wheat and barley agriculture was established in Egypt, about 6,400 years BP. The problem is compounded by the poor preservation of organic (especially plant) materials in the recent savannah. It also has to be said than only modern archaeology commands a holistic approach in excavations taking care also about the often orphaned botanic micro-remains. The future will undoubtedly bring more data to light about the domestication of African cultivars.

After race guinea was well differentiated, it spread to Malawi and later to southern Africa along the mountains of eastern Africa. Migration of S. bicolors south from the West African savanna probably occurred early in the history of Sorghum domestication and race kafir may have originated from race bicolor south of the equator (De Wet et al., 1972).

(Folkertsma et al., 2005) demonstrate in their genetic studies of the race guinea, that accessions from South Asia most closely resembled those from southern and eastern Africa, supporting earlier 176 suggestions that Sorghum germplasm might have reached South Asia via ancient trade routes along the Arabian Sea coasts of eastern Africa, Arabia and South Asia. Stratification of the accessions according to their Snowdenian classification indicated clear genetic variation between margaritiferum, conspicuum and Roxburghii accessions, whereas the gambicum and guineense accessions were genetically similar. Actually, according to the Haaland hypothesis, the wild-type Sorghums were taken to India and domesticated there and from there reintroduced in Africa

However, (Schechter & De Wet, 1975) demonstrated closer genetic affinities between kafir Sorghums and local wild kinds of race verticilliflorum (including arundinaceum) than between race kafir and the other domesticated grain Sorghums. This suggests an independent and maybe much earlier domes- tication of Sorghum in the southeastern African savanna.

Caudatum Sorghums are more or less limited to the original regions of Sorghum domestication. This race was developed parallel to present day Chari-Nile languages (Stemler et al., 1975a). Durra Sorghums probably originated outside Africa from S. bicolors that were introduced to the Sind-Punjab region of northwestern India some 3,000 years ago (De Wet & Huckabay, 1967; Harlan & Stemler, 1976). Race durra was later introduced into North and West Africa during the Islamic expansion across the Sahel and West African savanna, and later also became widely introduced into Ethiopia (Stemler et al., 1975b).

Domestication of Sorghum involved adaptation to the man-made habitat and adaptation beyond the original area of cultivation (De Wet et al., 1970). Racial evolution of grain Sorghums is closely associated with ethnological, ecological and geographical isolation. Variation within races is determined by conscious selection for particular uses, to satisfy the individual needs of cultivators. Modern evolutionary development is accelerated by movement of cultivated kinds across previous isolating barriers. Experimentally produced cultivars are replacing traditional cultivars in many parts of Africa today. For more details see chapter on S. bicolor and chapter on Sorghum breeding. But we should not omit another view on Sorghum landrace breeding produced by one of the most experienced breeders: (Hawkes, 1983) states that the adaptation of the landraces to the environment is the result of selection ‘largely of an unconscious nature’.

More remarks about Sorghum landrace breeding and development in chapter on ‘Centers of biodiversity are less susceptible to biological invasions’

A still valid, but under the present day knowledge simplified scheme is already given in (Doggett, 1988): It is based on publications by (Garber, 1950) for A, (Celarier, 1958) for B. 177

Fig. 90 Ancestral stock of Sorghum after (Garber, 1950), taken from (Doggett, 1988). Compare the revised view of (Spangler et al., 1999; Spangler, 2003) in chapter on Taxonomy of Sorghum in general. 178

Fig. 91 Suggested relationship among the genera of Sorghastrae, after (Celarier, 1958), from (Doggett, 1988).

179

4.3.3. Climate history in relation to the history of African agriculture

Fig. 92 The distribution of the absolute number of species in the western Sahara at about 5,500 B.P.; from Lauer and Frankenberg, 1979. Number of species found in areas of 6,400 km2: 1. > 300 2. 161-300 3. 81-160 4. 41-80 5. 21-40 6. < 21. From (Messerli & Winiger, 1992)

This is also the view of (Kimber, 2000), adding some more arguments coming from archaeobotany. Her comprehensive and extensive account on the complex domestication dynamics in Africa cannot be reproduced here.

180

Fig. 93 Anthropogenic and natural environmental changes which should be differentiated in space and time and from highlands to lowlands. Reversibility means the capability of an ecosystem to restore its original state after a disturbance or a change. Irreversibility means the absence of this capability (Gigon, 1983; Littmann, 1988). C: Climatological conditions S: Soil type and soil formation H: Hydrological conditions G: Geomorphic processes, hazards V: Vegetation cover and diversity, From (Messerli & Winiger, 1992)

181

Fig. 94 The distribution of elephants (upper map) and rhinoceros (lower) in the humid Holocene; from Mauny, 1956. From (Messerli & Winiger, 1992).

182

4.3.4. Recent biogeographical and migration concepts of Sorghum races

Fig. 95 Areas of origin and development of the domesticated races of bicolor and possible migration routes from and between these areas, C & ½ C caudatum; DB, durra and bicolor; K, kafir; G, guinea; GK, guinea and kafir. (Data from (Harlan & Stemler, 1976)

183

Fig. 96 Early human migrations and associated diffusion of S. bicolor races. Wide arrows, postulated early Sorghum routes, narrow arrows, diffusion of iron making technologies; dots, earliest centers of iron making. Data vom Murdock 1959, Harlan et al. 1976, Doggett 1988 and Shillington 1989.

A comprehensive summary on the latest Sorghum migration views has been published by

4.3.6. Summary:

There is no material proof (yet) of any genuine place or country of origin of Sorghum. But on the other hand, the region of Sudan and Ethiopia harbor a series of ancestral landraces, show a high biodiversity and certainly can be taken as the centers of landrace biodiversity of the present time and recent millenia. Data support the original idea of Harlan about the centers and non-centers of crop diversity. Harlan’s understanding is based on patterns of variation and genetic interaction among wild species, weedy relatives and cultivated races which all would suggest as a center of origin a wide zone in the broad-leaved savanna belt that stretches from about lake Chad to eastern central Sudan. Vast amounts of wild Sorghum are found along the Sudan-Ethiopia border, but there is no indication that the area was ever farmed before government settlement projects were established. Variations in Sorghum do not suggest that its homeland is Ethiopia; by far the bulk of Ethiopian Sorghums are durras, which are the most specialized and derived cultivated Sorghum. Still, pattern of domestication and the spread of Sorghum lets (Ejeta & Grenier, 2005) come to the conclusion, that Sorghum must have spread from a region including parts of Sudan and Ethiopia.

184

4.3.6. Bibliography Sorghum centers of origin http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Origin-20060806.pdf

4.4. Centers of biodiversity robust against introgression, landraces are a dynamic system.

4.4.1. Centers of biodiversity are less susceptible to biological invasions Nowadays it is common sense, that centers of biodiversity need special attention, since they are invaluable resources for genetic variation in a time when the most widespread cultivars have, due to modern selection, a genome which is narrowed down to best performance, robustness against diseases etc. (Swanson, 1995; Tanksley & McCouch, 1997) . It is also clear, that centers of biodiversity suffer from species loss, which is well documented (Brooks et al., 2002; Manzelli et al., 2005), for details about Sorghum see chapter 1.2.1. Centers of origin of Sorghum). Derived from such thoughts and facts, it is a widespread view that centers of crop origin should not be touched by modern breeding. With all due respect as shown by the previous publications, this is a rather undifferentiated view, which is enforced by some erroneous views on population ecology, such as that regions of high biodiversity are particularly susceptible to invasive processes, which is wrong: On the contrary, there are scientific studies showing that a high biodiversity means more stability against invasive species, against genetic introgression: ((Tilman et al., 2005; Whitham et al., 1999; Whitham et al., 1994): They found the hybrid zone between Eucalyptus amygdalina and Eucalyptus risdonii to be a center of insect and fungal species richness and abundance. Of 40 taxa examined, 73% were significantly more abundant in the hybrid zone than in pure zones, 25% showed no significant differences, and 2% were most abundant on a pure host species. The average hybrid tree supported 53% more insect and fungal species, and relative abundances were, on average, 4 times greater on hybrids than on either eucalypt species growing in pure stands. Hybrids may act as refugia for rare species: 5 of 40 species were largely restricted to the hybrid zone. Also, 50% of the species coexisted only in the hybrid zone, making for unique species assemblages. Three predictions are made by the authors:

 Intermediate genetic differences between the parental species will result in the greatest genetic variation in the hybrid zone, which in turn will have a positive effect on biodiversity

 Bidirectional introgression enhances species richness on hybrids, whereas F1 sterility and unidirectional introgression limit the accumulation of species on hybrids

 Although susceptible hybrids are likely to support the greatest biodiversity, the impacts of hybridization on keystone species will be crucial in determining the overall effect. 185

The introduction of new predators and pathogens has caused numerous well-documented extinctions of long-term resident species, particularly in spatially restricted environments such as islands and lakes (Henneman & Memmott, 2001; Howarth, 1991; Stokstad, 2001). However, there are surprisingly few instances in which extinctions of resident species can be attributed to competition from new species.

This suggests either that competition driven extinctions take longer to occur than those caused by predation or that biological invasions are much more likely to threaten species through intertrophic than through intratrophic interactions (Davis, 2003).

There is more evidence, that biological invasions, and thus also transgenic hybrids of wild relatives of GMOs with a potentially higher fitness, are depending on a multitude of factors, some now with recent research stepwise identified:

(Von Holle & Simberloff, 2005) established the first experimental study to demonstrate the primacy of propagule pressure as a determinant of habitat invisibility in comparison with other candidate controlling factors. There is more evidence documented that vegetation structure and diversity has influence on invasion dynamics on various vegetation types (Von Holle et al., 2003; Von Holle & Simberloff, 2004; Weltzin et al., 2003).

In a later paper (Fridley et al., 2007) the same author group makes even clearer statements: “Given a particular location that is susceptible to recurrent exotic invasion, native species richness can contribute to invasion resistance by means of neighborhood interactions and should be maintained or restored.”

See also the caption of the figure from (Fridley et al., 2007) where a similar statement is included.

186

Fig. 97 Conceptualized diagram of the invasion paradox. Fine-grained studies, many of which are experimental, often suggest negative correlations between native and exotic species richness but are highly variable. Nearly all broader-grain observational studies indicate positive native–exotic richness correlations. Likely exceptions are comparisons between temperate and tropical biomes, where preliminary data suggest that biodiversity hotspots have very few exotic species. From (Fridley et al., 2007).

The much more dynamic picture on agricultural fields fits well with farming experience, which builds on much faster ecological processes. It is a widespread error of many ecologists not to take into account the ephemeral ecological situation of agricultural plant communities (Ammann et al., 2004). Also according to (Thebault & Loreau, 2005) biodiversity should still act as biological insurance for ecosystem processes, except when mean trophic interaction strength increases strongly with diversity. The conclusion – which needs to be tested against field studies, is that in tropical environments with a natural high biodiversity the interactions between potentially invasive hybrids of transgenic crops and their wild relatives should be buffered through the complexity of the surrounding ecosystems.

This view is also confirmed by (Davis, 2003): Taken together, theory and data suggest that, compared to intertrophic interaction and habitat loss, competition from introduced species is not likely to be a common cause of extinctions in long-term resident species at global, metacommunity and even most community levels. It should also be clear that the simple introduction of transgenes to the wild populations or any kind of preservable landrace would not cause any harm to biodiversity, except if the introduced transgene is changing the population structure due to some considerable change in the competitiveness of the species or race receiving the transgene. With this caveat it is simply antibiotech propaganda if one claims that the introgression with transgenes has per se and automatically something 187 to do with a reduction of biodiversity. On the contrary, GMO crops often can – wisely used, be the source of betterment of biodiversity (Ammann, 2005).

Useful for our thesis produced here is the paper of (Morris et al., 1994). The authors indeed found that 4-8 m of in width may actually increase seed contamination over what would be expected if the intervening ground were instead planted entirely with a trap crop.

Finally, it should also be made clear that the threat of landraces is not only caused by environmental and biological agents, but is often the cause of vanishing traditional knowledge (Gupta et al., 2003).

(Harlan, 1975a) was among the first to state that landraces have come to rely on cultivation for their survival, active breeding and conscious selection are at the heart of landrace definition, although not all researchers share this view: (Hawkes, 1983) states that the adaptation of the landraces to the environment is the result of selection ‘largely of an unconscious nature’. An excellent review about definitions and classifications of landraces is given by (Zeven, 1998), see also (Zeven, 1999, 2000, 2002). An excellent early, but still valid comment and table summing up advantages and disadvantages of ex- situ and in-situ conservation of landraces is given by (Hawkes, 1991): “No doubt other factors might be considered in addition to those mentioned in Table I. However, the fact that in-situ and ex-situ methods both possess advantages and disadvantages renders it imperative to scrutinize each with great care. The general conclusions reached up to now are that each method is complimentary to the other, rather than antagonistic. However, the problems continue to confront us, namely, that whereas ex-situ storage methods have been established satisfactorily for at least two decades (Frankel & Hawkes, 1975) and are those in common use for "orthodox" seeds, those for in-situ conservation are only just beginning to be formulated. This is particularly worrying when we need to decide on conservation strategies for "recalcitrant" species – those whose seeds have no dormancy period and cannot thus be stored under the reduced temperature and humidity that have proved so satisfactory for orthodox species (Roberts, 1975). It is quite clear that much more thought and research initiatives need to be applied to the problem of in-situ conservation. The International Board for Plant Genetic Resources, IBPGR (1985) developed a provisional list of species for ecogeographical surveying and in-situ conservation for fruit trees, forages and a number of other crops, and pointed out the need for "sufficiently large and diverse" populations so as to sustain the levels of allelic frequencies in conserved populations. The publication called for further research, setting out at the same time a number of useful parameters for genetic conservation. However, the publication is understandably reticent on hard data involved in setting up reserves, for the simple reason that hard data do not yet exist.”

Still, it seems that we do not know yet enough to make out of the above thesis a theory which holds up to all scrutiny, as (Ives & Carpenter, 2007) develop with convincing details and an impressive survey of the existing literature on biodiversity stability and ecosystems. Their final remarks are rather inconclusive and call for a case to case perspective: “Finally, a finding common to many empirical studies is that the presence of one or a handful of species, rather than the overall diversity of an ecosystem, is often the determinant of stability against different perturbations. We suspect that, depending on the types of stability and perturbation, different species may play key roles. Predicting which species, however, is unlikely to be aided by general theory or surveys of empirical studies; each ecosystem might have to be studied on a case-by-case basis. In the face of this uncertainty and our ignorance of what the future might bring, the safest policy is to preserve as much diversity as possible.”

188

Fig. 98 Comparison of advantages and disadvantages of in-situ versus ex-situ conservation, from (Hawkes, 1991)

4.4.2. The case of maize landraces in Mexico We have no experience yet with the introduction of transgenic Sorghum in Africa, this is why an example from Mexico about the landraces of Maize is given here, although the character of wild relatives and outcrossing biology will be different; but still there are many things to learn about the case history. A major argument for the view that the influence of gene introduction from modern maize traits cannot destabilize the landraces are derived from results of studies made by (Kato et al., 2005; Kato & Sanchez, 2002). They did not find, despite of evident gene flow between modern maize and its wild relatives, any change in the genome: "Based on these results, Zea diploperennis germplasm can be transferred under controlled conditions and persist through the third generation of backcrossing. However, even though maize and teosinte hybridize when growing in sympatry forming fertile hybrids, no evidence was found for natural introgression occurring between maize and Z. diploperennis."

Overall: The widespread fears that introgression of (trans)genes into landraces within centers of biodiversity are basically unfounded and are not documented even in a single case in Mexico, despite a 189 misleading publication in Nature (Quist & Chapela, 2001). This study was later criticized by the editor of Nature as having not sufficient evidence for the claim of Bt gene having introgressed the Mexican maize landraces. After a lengthy debate in Nature an extensive study has been published in PNAS demonstrating with the evidence of 150’000 samples drawn, that there is zero introgression at present time. (Ortiz-Garcia et al., 2005). These fears of aggressive transgene introgression into landraces are nourished by false views about landraces as a whole: This has been shown by the studies of (Bellon & Berthaud, 2004, 2006; Berthaud, 2001). Although a landrace is generally assumed to be a local “variety” produced over time through selection by farmers, Berthaud contends that in the Oaxaca study area, the landraces do not meet the basic criteria of a variety: that they be distinct, uniform, and stable. On the other hand, the threat of low prized imports from the North could develop into a real threat to Mexican landraces, since it could destroy the local markets. Another threat is also not building on the biology of invasiveness: Possible future intermixing with transgenes could lead to the rejection of harvested grain for marketing reasons (Bellon & Berthaud, 2006). So what are we trying to conserve, if not landraces? “The active flow of genes,” answers Berthaud, “which carries traits that are of value now or may be found to have value in the future.” Sustaining gene flow in farmers’ fields may not require maintaining landraces. By trying to retain landraces in their current form, we may paradoxically doom landrace conservation to failure. “A decade ago,” says Berthaud, “most people’s vision of in situ conservation was to put up a fence, keep the farmers and the variety in a state of suspended animation, and figure that everything would stay as it was. But this will not work.” People have needs that may change with the market—say, an emerging preference for floury rather than flinty kernels - or with the environment. For example, a series of dry years will affect the supply of maize seed and what is preferred for planting (although we have to realize that landraces are often more robust to such perils). The study area of any kind of landrace conservation project is a dynamic environment, with new genes and traits flowing in and out of it, even under the most traditional systems. Another reason that gene flow may be required to maintain diversity is the accumulation of deleterious mutations. Small-scale farmers select their own seed. Often they choose the best ears at harvest and save seed from only a few cobs—a logical approach but one that increases deleterious mutations. As defects accumulate, the variety loses its genetic value. “We know farmers are putting new diversity into the system and in the process losing some of the old alleles and traits,” Berthaud observes, “which raises several other questions. Is this dynamic process in balance? And will the current flow maintain the valued genetic diversity?”

4.4.3. Conclusion: Landraces can only be preserved with active and participative breeding programs We do not have enough knowledge about the landraces of Sorghum yet, but (Berthaud, 2001) makes one thing clear in the case of Maize: The population genetics dynamics has to be interpreted contrary to the views of conservationists, who want to preserve landraces, but do not know enough about agricultural population and ecological dynamics (Ammann et al., 2004). The situation in Mexico has been clearly described by (Bellon & Berthaud, 2006): Strictly speaking, the adoption of improved maize varieties8 has been limited in Mexico, but there is increasing evidence that through their breeding practices small-scale subsistence farmers have incorporated improved varieties into their farming 190 systems. By exposing improved varieties to their conditions and management, continually selecting seed of these varieties for replanting, and in some cases promoting their hybridization with landraces either by design or accident, farmers produce ‘‘creolized’’ varieties . Creolized varieties are appreciated because they combine the advantages of improved varieties and landraces. Transgene flow is presently zero or so extremely low that it escapes even the most sensitive detection methods (Ortiz-Garcia et al., 2005), but if it ever will happen (most probably by farmers breeding and seed exchange), there might eventually develop a marketing problem., as long as a given population is mislead by false statements about GM risks of NGOs.

The case of Sorghum and its hybrids with wild relatives is more complex, but there are no events reported where new Sorghum hybrids ran wild in Africa. A demonstration of how complex the situation is, has been shown by (Hadley, 1958b): The highly sterile 30-chromosome hybrids have extremely vigorous rhizomes, and present a local threat through vegetative reproduction alone. Conversely, the 40-chromosome hybrids, whether self-fertile or male-sterile, were found to be weakly rhizomatous and did not constitute a threat other than being a source of grassy weed seedlings.

This is taken all together a very strong argument to work with farmers right from the beginning in the SuperSorghum project, a demanding job for Africa Harvest – a job, which has been undertaken already with great success in another agricultural project of Africa Harvest: the successful spread of TeeCee Banana seedlings. See also the comments in chapter origin of Sorghum on the loss of landraces in Somalia by (Manzelli et al., 2005): (Alvarez et al., 2005) demonstrated in a preliminary project in sub-Sahelian Cameroon that suggest that sorghum populations managed by the Duupa function like source–sink metapopulations. Fields of older farmers, larger and containing a greater number of varieties, act as sources, whereas fields of younger farmers act as sinks, becoming sources as their owners mature. In each field, seeds for sowing are selected from a small number of plants. The frequent exchange of germplasm among fields may counteract the genetic bottlenecks associated with the small number of genitors within each field. Again, the conclusion is to establish participative breeding programs, well founded with scientific data delivered for years by ICRISAT and the wise use of core seed collections (Mgonja et al., 2006).

4.4.4. Summary robustness of biodiversity centers and conservation of landraces Centers of biodiversity and centers of crop biodiversity usually are rich in plant species. Although it is clear, that centers of biodiversity suffer from species loss due to many different factors, there are enough data in population ecology to show that plant communities with a rich set of species are less susceptible to alien species invasions, except in cases where those species have a considerable advantage in competitiveness. The rule is that in heavily disturbed areas one finds the most dramatic invasions of alien species. Landraces cannot be preserved with a conservative system, since they are dynamic populations, having been subject of spontaneous breeding and seed interchange for centuries. In consequence, it will be important to establish participative breeding programs in order to preserve landraces, since landraces have only a chance of survival, if there is still a market for the products. 191

5. Reproduction of Sorghum

5.1. Inflorescence general The young inflorescence is forced up through the top leaf sheath (the boot) after the uppermost leaf has expanded. If the lower part of the inflorescence is still surrounded by the uppermost leaf sheath, this might cause trouble from pests and mould. In some varieties the peduncle is recurved (goose-neck).

5.2. Panicle structure The inflorescence is a panicle with a central rhachis with primary branches, which give rise to secondary and sometimes tertiary branches, carrying the racemes of spikelets, typical for the whole group of the Andropogoneae: the usually single or two-flowered spikelets are arranged closely together and mimicking a multiflower spikelet. See fig. below and illustrations from (Doggett, 1988) in Chapter 1.1.3 on S. bicolor.

(Craufurd et al., 1993) made an experiment with seven Sorghum lines, flowering from 50 to 87 days after sowing, they were subjected to early drought stress, late stress, and both early and late stress in the field during the dry season in India. Panicle initiation was delayed by 2-25 days and flowering by 1-59 days by the drought stress treatments, the greatest effect being in the treatment subjected to both early and late stress. Stress increased the period between panicle initiation and flowering by retarding the rate of panicle development; when stress was severe panicle development stopped.

In all multivariate morphogenetic analysis the panicle structure plays a major role, see for example (Abdi et al., 2002) and the fig in chapter 1.1.7, head types of cultivated Sorghum, from (Harlan & De Wet, 1972).

192

Fig. 99 Head types of cultivated Sorghum. Type 1 is reserved for wild races and is considerably more diffuse than type 2. From (Harlan & De Wet, 1972) Sorghum and guinea: types 2-4 kafir and durra: types 5-7 caudatum: wide range of types broomcorn and half-broomcorn types: 8-9

5.3. The pedicelled spikelets There is much variation between varieties and species in the pedicelled spikelets. These may be persistent or deciduous, large or small, with long or short pedicels. Often the spikelet consists of only two glumes, but sometimes the lemmas are present, and with other varieties there is a male floret with three anthers which produce good pollen. Only rarely does the pedicelled spikelet have a functional ovary and produce seed. In varieties where this occurs the grains in the pedicelled spikelets are always smaller than those of the sessile spikelets (Doggett, 1988). (Casady & Miller, 1970) showed inheritance of hermaphrodite pedicelled spikelets. The character was found to be controlled by a single pair of alleles, with the hermaphrodite pedicelled spikelet allele recessive to the non-hermaphrodite pedicelled spikelet allele.

A scheme from (Spangler et al., 1999) explains the somehow complex basic structure of the Andropogoneae:

193

Fig. 100 Diagram of a spikelet pair showing floret positions scored for sex expression. 1: sessile spikelet , proximal floret; 2: sessile spikelet, distal floret; 3: pedicellate spikelet, proximal floret; 4: pedicellate spikelet, distal floret.

5.4. Anthesis The Sorghum inflorescence usually begins to flower when the peduncle has completed elongation. The first flower to open is either the terminal or the second flower of the uppermost panicle branch. Blooming continues downwards in a fairly regular manner, and in general the flowers in a horizontal plane across the panicle pen at about the same time. Pedicelled spikelets bloom from 2-4 days after the sessile ones on the same branch. Flowering spreads over 6-9(-15) days, the pollen is shed usually during the night or soon after sunrise. Sorghum flowers open rapidly according to the normal Poaceae scheme with the help of lodiculae spreading the glumes, the stigmas and anthers emerge. Several studies have shown that resistance to Sorghum midge is associated with short and tight glumes, faster rate of grain development, and tannins (Sharma & Hariprasad, 2002). However, some recent studies (Diarisso et al., 1998; Pendleton et al., 1994) suggested that time of flowering is the principal component of resistance to Sorghum midge. Therefore, the authors conducted a series of experiments under laboratory and field conditions on the flowering behavior of a diverse array of midge-resistant and 194 midge-susceptible genotypes to quantify the contribution of time of flowering in genotypic resistance to Sorghum midge. Time of flowering under field and laboratory conditions did not show any differences between midge-resistant and midge-susceptible genotypes. According to the experiments of (Sharma & Hariprasad, 2002) there was no evidence of change in the susceptibility of Sorghum midge-resistant genotypes when infested at different times in relation to time of flowering. Therefore, flowering behavior of Sorghum genotypes seems to play little role in genotypic susceptibility to Sorghum midge. (Musabyimana et al., 1995) studied ergot resistance and the influence of flowering periods: Ergot is an important disease of Sorghum (S. bicolor) in parts of Africa and Asia. Studies were conducted to determine the relationship between flowering biology and ergot infection, and to develop an artificial field-screening technique to identify ergot resistance in Sorghum. Spikelets resisted infection after anthesis, but each day's delay in anthesis after inoculation supported 8.3% more ergot. The screening technique consisted of three components: trimming of panicles to remove pollinated spikelets before inoculation, a single inoculation of trimmed panicles, and panicle bagging for 7-10 days. Inoculated panicles were evaluated by a qualitative visual rating method (on a 1-5 scale) and a quantitative spikelet counting method. Selected accessions from the world collection of Sorghum germplasm were screened at Karama Research Station, Rwanda, for two seasons and 12 ergot-resistant lines were identified. These were also resistant at ICRISAT Centre, India.

195

Fig. 101 Sorghum bicolor (L.) Moench, from website Missouri Plants http://www.missouriplants.com/Grasses/Sorghum_bicolor_page.html

5.5. Pollination There are differences in opinion on the proportion of stigmas which are pollinated from stamens in the same flower or from stamens of other flowers in the same head. Some cross pollination between plants usually occurs and the amount is influenced by such factors as wind direction and panicle type – open heads more liable to cross-pollination than compact heads. The figures quoted for cross pollination vary considerably: Highest from 50% down to normally 5%. The pollen is viable for 3-6 hours and stigmas may be receptive for up to a week after blooming, but they are best pollinated during the first 72 hours. Stigmas are also receptive for a day or two before blooming, (Doggett, 1988) (Maunder & Sharp, 1963) found that the upper quarter of the panicle produces 2.5 to 4 times more outcrossing than the lower parts, the outcrossing rate also dependent of the pollen production. (Tuinstra & Wedel, 2000) developed a pollen viability assay by in vitro germination. Male fertility of crop plants is a function of pollen production and viability. Differences in pollen production can be evaluated by simple observation, but pollen 196 viability is more difficult to quantify. The objective of this study was to develop an in vitro pollen germination assay for S. bicolor (L..) Moench. In experiments evaluating common germination media substrates, large differences in germination were observed in response to changing concentrations of sucrose.

(Lansac et al., 1994) studied pollen viability of Sorghum and came to the conclusion, that pollen longevity is approximately 30 min to 2 h after release This fits also well with the experience of W. Wenzel, personal communication of Sorghum plant breeder W.

Wenzel, Agricultural Research Council, South Africa, 2003 in (Schmidt & Bothma, 2006).

5.6. Mating systems in Sorghum in general Sorghum is largely self pollinated, but wind pollination between plants does occur (McGuire, 2004). Subspecies or varieties of Sorghum with open, grass-like panicles, such as Sudangrass (S. sudanense), have a higher rate of outcrossing than Sorghum bicolor races with compact heads typical of commercial hybrids. Outcrossing also varies by location on the panicle, with much higher rates at the top of the panicle, where flowering initiates (Maunder & Sharp, 1963). (Djè et al., 2004) concluded from their results based on seed samples taken in situ indicated that in cultivation sorghum is predominantly selfing, as was previously suggested from studies under experimental conditions. Differences in selfing rate between the two fields investigated could be explained by distinct morphological characteristics of the inflorescence. The autogamous mating system is at odds with the very low genetic differentiation observed previously among Moroccan landraces and this suggests that gene flow through seed exchanges among farmers is frequently occurring, with morphological differences among landraces maintained by phenotypic selection by farmers. More comments on this and similar studies in the chapter on gene flow from crop to crop.

On pollen sterility compare also under Apomixis below.

5.6.1. Cleistogamy Not much is known about cleistogamy in Sorghum. (Ayyangar & Ponnaiya, 1939) report a case from India and (Bowden & Neve, 1953) report a case from West Africa. Cleistogamy is due to the rolling of the inner glume, so that it clasps the internal flower structures and prevents normal opening of the glumes. Only papery (chartaceous) glumes can exhibit rolling (Merwine et al., 1981).

5.6.2. Apomixis (Marshall & Downes, 1977; Murty, 1993; Murty et al., 1984; Murty et al., 1979; Rao et al., 1978; Rao & Murty, 1973; Reddy et al., 1979; Tang et al., 1980) report Apomixis and sexual reproduction in the genus Sorghum from India. This was contested by (Ravi, 1993), but the situation seems now clarified, the counter evidence being too weak. (Ross & Wilson, 1969) report on a case of polyembryony in Sorghum. . Later, (Elkonin et al., 1995) found proof of apomixis. In his work, he took advantage of tissue culture as a means of genetic modification of plant reproductive systems. Mutations can cause female sterility and revert male sterility. Evidence is demonstrated in their work that aposporic cells can be generated from tissue culture of Sorghum. Cytoembryological investigation has shown the presence of structures similar 197 to aposporous embryo sacs (ESs), which have developed in the ovules along with the sexual ESs. The ovules with aposporous formations have been observed almost in all of the studied plants in several generations, however, their frequency varied (2-53%). The authors propose to develop a system of Cytoplasmatic Male Sterility which for breeding and propagation methods could be reverted. Modern hybrid Sorghum bicolor seed production relies exclusively on cytoplasmic male sterility (CMS) systems and almost all hybrid Sorghum seed is produced using the A1 CMS system. (Moran & Rooney, 2003). However, the reliance on a single CMS system increases the vulnerability of the crop to diseases and stresses that may attack that particular CMS system. Alternative CMS systems have been described and even used on a limited basis for hybrid seed production, but a direct comparison of the agronomic effects of different cytoplasms has not been possible because male-sterile lines with a common genetic background (and different cytoplasm) were not available. The recent development of isocytoplasmic A- lines allows more direct comparison of cytoplasmic effect on agronomic performance. There will be more modern methods applied for the improvement of hybrid Sorghum, a fast spreading Sorghum crop production. See also chapter on Sorghum breeding.

5.6.4. Grain development

Fig. 102 Sorghum bicolor (L.) Moench from website Missouri Plants http://www.missouriplants.com/Grasses/Sorghum_bicolor_page.html The time taken for the grain to attain its maximum dry weight varies with growing conditions, but this is usually 25-55 days after blooming. The moisture content at this stage may vary from 25-35% and 198 seasonal influences are evidently important. More data on grain development from studies in the USA are given in (Doggett, 1988).

(Tuinstra et al., 1998; Tuinstra et al., 1997a; Tuinstra et al., 1996, 1997b) tested the grain development in relation to drought factors: In this study, a set of 98 recombinant inbred (RI) Sorghum lines was developed from across between two genotypes with contrasting drought reactions, TX7078 (pre-flowering-tolerant, post-flowering susceptible) and B35 (pre-flowering susceptible, post-flowering tolerant). The RI population was characterized under drought and non-drought conditions for the inheritance of traits associated with post-flowering drought tolerance and for potentially related components of grain development. Quantitative trait loci (QTL) analysis identified 13 regions of the genome associated with one or more measures of post-flowering drought tolerance. Two QTL were identified with major effects on yield under fully irrigated conditions suggesting that these tolerance loci have pleiotropic effects on yield under non-drought conditions. Loci associated with rate and/or duration of grain development were also identified. QTL analysis indicated many loci that were associated with both rate and duration of grain development. High rate and short duration of grain development were generally associated with larger seed size, but only two of these loci were associated with differences in stability of performance under drought. (Jambunathan et al., 1990; Jambunathan et al., 1991; Navi et al., 2005; Rodriguez- Herrera et al., 2006) concentrated in their studies on grain development and grain mould resistance compared to various factors. (Ogunlela & Eastin, 1985) studied after-effects of elevated night temperature and heat-preconditioning on net carbon-dioxide exchange and grain development in Sorghum bicolor. Changes in protein-fractions and leucine- C- 14 incorporation during sorghum-grain development have been studied by (Johari et al., 1977a, b).

5.6.5. Bibliography reproduction of Sorghum http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Reproduction-20060806.pdf

199

6. Sorghum breeding in relation to its biology

6.1. Introduction, Summary

6.1.1. The scope of this chapter It is not intended in this report to give a full account on Sorghum breeding, a relatively short chapter must be enough here, the remarks are mainly related to the biology and evolution of Sorghum, a simple screening of the web of science reveals over 500 publications, and many from 2000 onwards, on the combination of two keywords: Sorghum and breeding – so – after all, Sorghum research has grown into a major area targeting better crops.

6.1.3. Basic remarks about Sorghum breeding and previous breeding efforts in Sorghum. In the period before ca. 2000 there are a lot of efforts to be registrated on classic breeding targets, maybe best summarized in (Doggett, 1988; Doggett & Majisu, 1972) with a graph from Jensen 1970:

200

Fig. 103 Fig. 7.3 from Jensen 1970 demonstrates without extensive comments how tedious traditional crop breeding really is.

201

Another example from Ethiopia is given by (Doggett, 1988) showing the same scheme:

Fig. 104 The Sorghum breeding program of ESIP in Ethiopia, as described in detail by (Doggett, 1988). The program comprises in any one year five kinds of activities, each one with a different objective: 1. backcross crossing block, 2. Dented seed recurrent selection block, 3. ms3 crossing block, 4. Hybrid crossing block, 5. Pedigree crossing block. The diagram summarizes the interrelationships and flow of ESIPs programs. * DSBN = dented seed breeding method, ** Keremt = main rainy season, June – September, in which the evaluations are made. From (Doggett, 1988) p. 223. But both figures could be also misleading: Those who think that in the glorious age of gene technology things would be much easier and proceedings can be dramatically shortened, are wrong. Certainly, the construction of the first plants with selected genes is easier, often more elegant. Also in the grass family genetic engineering has made things better and easier in the last few years. The selection process can be 202 started without lots of unwanted genes coming in due to hybridization and consecutive selection. But still, the bulk of selection work remains the same, and needs to be processed even more carefully and methodically In essence, the classic breeding scheme on Sorghum concentrated on yield improvement through a sustainable pathway of getting stable tetraploid cultivars: The grain sorghums are diploid, having been developed from the wild African grass sorghums of the Arundinacea (Doggett & Majisu, 1968). The success of the wild tetraploid sorghums, such as Johnsongrass in the Halepensia, and of the wild x cultivated cross Columbus grass (S. almum) suggested that useful tetraploid cultivated grain Sorghums might be developed. There are prospects of obtaining in tetraploids character expressions extending beyond the range found in the diploids. These might be particularly valuable for characters such as grain size and grain protein content, while some useful yield increase might also be expected. The doubling of the chromosome number of good diploid varieties using colchicine is most unlikely to lead to the direct production of good tetraploid varieties (Majisu, 1971), since stable and regular chromosome behaviour at the tetraploid level cannot have been subjected to any favourable selection at the diploid level. Rather, the approach adopted sought to develop new Sorghum populations at the tetraploid level which could then be subjected to intercrossing, selection for regular chromosome behaviour being undertaken indirectly by selection for high fertility, as estimated by good seed set. Therefore, a wide range of Sorghum cultivars was made tetraploid with colchicine by the authors, and intercrossing with selection resulted in much improved seed set (Doggett, 1964b). Secondly, the good relational chromosome balance developed down the years in S. halepense was transferred to the autotetraploid cultivars. This was done by backcrossing from S. almum, itself a derivative of hybridisation between S. halepense and cultivated Sorghum. It proved possible to select from such material tetrapIoids with seed-set levels similar to those of cultivated diploid grain sorghums (Doggett, 1964a). It is likely that their yields would decline under continued selfing, judging from the behaviour of lines drawn from diploid populations (Doggett, 1972). None the less, the achievement of diploid yield levels both by tetraploid bulks and by lines drawn from a tetraploid bulk population, before any great degree of recombination can have occurred, gives grounds for optimism regarding the future development of tetraploid grain Sorghum. It seems that recurrent selection in tetraploid grain sorghum populations in which high recombination levels are obtained by the use of the genetic male-sterile traits, would be well worth continuing, and should lead to the development of high yielding varieties.

6.1.4. Gradual transition from traditional breeding towards genetic engineering It should also be made clear that there has been a gradual transition from classic breeding to molecular engineering over the years, maybe best illustrated by the following example: (Duncan et al., 1995) used In vitro cell selection for somaclonal variation studies for speeding up screening techniques: S. bicolor (L.) Moench is generally quite sensitive to salt and acid (high aluminum) 203 soil stresses, but quite tolerant of drought stress. As with any stress phenomenon, intra-specific variability exists within the genus. In vitro cell selection and somaclonal variation offer a speedier alternative to traditional breeding methodology for generating improved lines for hybrid development. A field selection protocol was developed for the three soil stresses and inter-stress evaluations were conducted in an effort to find multiple, stress-tolerant genotypes. The acid soil-drought stress, super- tolerant selections were located by the R(7) generation when exposed to a combined aluminum-drought stress field environment when the regeneration population (number of regenerated lines from one callus source) was maintained at 15,000 plants or higher. The optimal strategy for the exploitation of somaclonal variation may be through short-term cell culture (< 12 months) with no attempt at in vitro selection.

More details about the transition from traditional to modern breeding methods see in (Ammann, 2004), chapter on evolution of plant breeding p. 22: Thirteen steps mark a gradual shift from indirect genome changes through selection, targeted selection, the introduction of chemical and physical devices to speed up mutation rates - it finally reaches its present day culmination by making use of biochemical devices such as restriction enzymes (Arber & Linn, 1969) which allow for a never before matched precision in DNA splicing, soon afterwards called enthusiastically ‘genetic engineering’.

Some interesting basic remarks on the genomics related to Sorghum are given by (Hamblin et al., 2004; Hamblin et al., 2005) state about the linkage disequilibrium (LD) (http://en.wikipedia.org/wiki/Linkage_disequilibrium) : The extent of LD in sorghum is greater than that in maize, where it is generally low (Tenaillon et al., 2002; Tenaillon et al., 2001; Yamasaki et al., 2005), and less than that in Arabidopsis (Nordborg, 2000; Nordborg et al., 2002; Pagnol et al., 2006). This qualitative observation is consistent with the mixed mating system of sorghum producing intermediate levels of effective recombination. Quantitative analyses, however, based on estimated rates of recombination, mutation, and selfing, show that both r and the ratio of r to u are lower than expected under an equilibrium model. These analyses, as well as the frequency spectrum of polymorphism, suggest that a genome-wide departure from equilibrium underlies this phenomenon. The greater extent of LD in sorghum makes it amenable for association studies using a limited number of markers.

6.1.5. Main purposes of breeding programs in Sorghum Depending on the region of production, the type of Sorghum and the purpose for its production varies widely. (Rooney, 2004). Whether they are breeding varieties or hybrids, the primary focus of Sorghum breeders throughout the world are yield, adaptation and quality. In addition to breeding for these factors, reducing losses due to stress is equally important. Most breeding programs consistently select for tolerance to abiotic stresses (such as drought and low temperatures) and biotic stresses (such as Sorghum midge, grain mold, anthracnose, and charcoal rot). Finally, the integration of molecular genetic technology is enhancing Sorghum improvement by providing a genetic basis for many important traits and through marker-assisted selection. Sorghum improvement in the future will require effective 204 utilization of all the available tools in order to develop Sorghum genotypes suitable for the needs of their producers and end-users. Species outside the Eu-Sorghum section are sources of important genes for Sorghum improvement, including those for insect and disease resistance, but these have not been used because of the failure of these species to cross with Sorghum: See the chapter on general taxonomy of Sorghum.

6.1.6. Hybridization barriers in Sorghum An understanding of the biological nature of the incompatibility system(s) that prevent hybridization and/or seed development is necessary for the successful hybridization and introgression between Sorghum, divergent Sorghum species and taxa outside the genus Sorghum.

(Hodnett et al., 2005) determined the reason(s) for reproductive isolation between Sorghum species. The study utilized 14 alien Sorghum species and established that pollen-pistil incompatibilities are the primary reasons that hybrids with Sorghum are not obtained. The alien pollen tubes showed major inhibition of growth in Sorghum pistils and seldom grew beyond the stigma. Pollen tubes of only three species grew into the ovary of Sorghum. Fertilization and subsequent embryo development were not common. Seeds with developing embryos aborted before maturation, apparently because of a breakdown of the endosperm.

(Morgan et al., 2002) and also (Zeller, 2000) (in German language) gave an account with lots of important citations on the lessons of breeding activity for Sorghum in recent years. The grass-based grain crops have a high degree of synteny in their genomes, thus raising the possibility of using specific information more broadly. Studies of Sorghum bicolor are beginning to link physiological behavior to specific genes and hormone-based regulatory systems in ways that suggest specific strategies for improvement. Findings from several other grasses are adding to the information pool being derived from Sorghum. This information relates to flowering and floral development, maturity and senescence, temperature effects via the biological clock, shade avoidance behavior, apical dominance, shoot elongation, and root development including constitutive aerenchyma formation. These studies, along with others, offer a number of options for conventional plant breeding and genetic transformations to improve grass-based crops and satisfy part of the projected human food needs of coming decades.

(Cox et al., 2002) produced a thoughtful account on Sorghum breeding for the temperate regions: Besides pointing to the beneficial hybridization strategies including Sorghum propinquum and Sorghum halepense with cultivars the authors finish by stating: Because there is some homology between the chromosomes of grain Sorghum and Johnsongrass, multivalent chromosome associations are common at meiosis in interspecific tetraploids. Multivalents, in turn, cause poor seed set because of nondisjunction of chromosomes. (Luo et al., 1992) demonstrated that selection for fertility can be effective in autotetraploid grain Sorghum, which is generally plagued by low seed set. Breeders could cross perennials with the highly fertile tetraploid germplasm that (Luo et al., 1992) have produced, broadening and improving the genetic base of tetraploid perennial Sorghum will require introduction of more agronomically elite germplasm. One rapid method of incorporation would be to pollinate both diploid and induced-tetraploid strains of elite, large-seeded inbred lines with the best tetraploid perennials. From the diploid/tetraploid crosses, 205 breeders can select 40-chromosome hybrids that arise from unreduced gametes (Hadley, 1953). The Land Institute is now taking this approach to develop genetically diverse breeding populations. It is possible that some of the strategies to come back from the annual to perennial crops could also be considered for future African breeding programs.

6.1.7. Sorghum as a model for forage grass improvement A last basic remark: (Paterson et al., 1995b) made it clear that Sorghum as a whole genus is an important model for forage grass improvement. The infamy of ‘Colonel Johnson’ notwithstanding, rhizomes are an important asset to turf and forage grasses that cover vast land areas. The importance of forage in livestock diets, and turf for aesthetic and sporting purposes, is widely recognized. Further, grasses are important in erosion control-failure to recognize this was a partial cause of the "Dust-Bowl" epochs that have periodically crippled United States agriculture. Genes responsible for rhizome development and tillering in Sorghum may at least partly account for these traits in other grasses, in which up-regulation of rhizomatousness might improve agricultural productivity. Sorghum provides a facile model for detailed investigation of genes controlling rhizomatousness, a trait important to productivity, quality, and protection of agro-ecosystems. See the figure in the chapter on Sorghum halepense.

The global economic importance of Sorghum makes it a prime candidate for genetic transformation. However, genetically modified Sorghum has not been commercially produced to date, see for more details in the chapters on transgenic Sorghum.

6.1.8. Breeding program of the University of Hohenheim A further program with breeding efforts is undertaken since years at the University of Hohenheim: The productivity can be increased through the application of hybrid breeding. To speed up the breeding process, new heterotic groups based on the utilization of the genetic variability of the Sudanese germplasm shall be developed. Clustering will be based on microsatellite markers as well as field data . Generally, clustering based on molecular data leads to different results than based on agronomic data. The objectives of the present project are to: 1. to asses the pattern of genetic diversity between and within Sudanese Sorghum landraces, 2. to compare the genetic diversity of Sudanese landraces with that of the world collection of hybrid parents, 3. to relate marker-based distance measures with heterosis and hybrid performance of the selected crosses, 4. to evaluate selected CMS lines for combining ability to Sudanese germplasm and 5. to estimate the fertility restoring ability and the outcrossing rate of the landraces in field experiment in Sudan. The project is conducted in cooperation with the Agricultural Research Corporation in Sudan and is financially generously supported by the Eiselen Foundation in Ulm. Contact persons are Dr. H.K. Parzies ([email protected]), and Tahani Elagib, M.Sc. ([email protected] hohenheim.de), the website can be consulted on http://www.uni-hohenheim.de/~ipspwww/350b/index.html

6.2.1. Summary of modern breeding situation in Sorghum A summary of the modern breeding situation in Sorghum is also given by (Bantilan et al., 2004): Encouraged by the results from natural hybridization and selection, attempts at deliberate crossing were made, wherein it was found that crosses between divergent cultivars exhibited high levels of heterosis (Conner & Karper, 1927). The discovery of the cytoplasmic genetic male sterility by (Stephens & Holland, 1954) based on the milo-kafir system was a milestone in Sorghum breeding and research. It is widely used in the commercial exploitation of heterosis. Today, more than 30% of the Sorghum area is under 206 hybrids (in the USA!), which have yields about twice that of any local cultivar. Non-mila sources of cytoplasmic male sterility have also been identified (Schertz & Johnson, 1984; Schertz & Ritchey, 1978). However, their commercialization has been hindered by the nonavailability of sufficiently stable restorer lines. (Moran & Rooney, 2003; Tsvetova & Elkonin, 2002). The availability of several single, recessive male sterile genes (ms3, ms7) serving as a genetic male sterility system allows population improvement through recurrent selection procedures involving random mating and selection, see (Mgonja et al., 2006) below. Selected literature on hybrid Sorghums shows the importance of the development: (Ahmed et al., 1993, 1995; Ahmed & Young, 1969; Ali & Wills, 1983; Arnon & Blum, 1965; Bidari et al., 1978; Borges et al., 1997; Brandner & Straus, 1959; Chougule, 1978; Denman et al., 1984; Denman et al., 1985; Doggett, 1962, 1969; Ercoli et al., 1996; Giri & Bainade, 1981; Hariprakash, 1979; Hariprakasrao, 1978; Haussmann et al., 2000a; Hawkins et al., 1986; Hirpara et al., 1992; Hookstra & Ross, 1982; Iptas & Brohi, 2003; Kandasamy & Subramanian, 1979; Kao et al., 1987; Kirby et al., 1983; Krishnasamy & Ramaswamy, 1985; Kumaravadivel & Rangasamy, 1994; Marier, 1985; Moran & Rooney, 2003; Neto et al., 2004; Ogunlela, 1983; Orak, 2001; Raghuwanshi et al., 1988; Sharma, 1977; Singh et al., 1986; Singh et al., 1983; Singh et al., 1987; Sonune & Kalra, 1984; Srivasta.Sp & Singh, 1970; Stephens & Holland, 1954; Turkhede & Prasad, 1978; Waquil et al., 1986).

(Ahmed et al., 2000): In spite of substantial introduction of new Sorghum and millet cultivars in semiarid Sub-Saharan Africa, there has been minimum aggregate impact on yields in contrast with other crops, such as cotton and maize (ICRISAT & FAO, 1996). Only where inorganic fertilizers and improved water retention or irrigation were combined with new cultivars were there large yield increases.

207

Fig. 105 Global trends in sorghum production, 1979-94. From (ICRISAT & FAO, 1996) The situation seems to have been enhanced for Western Africa, but not in Southern and Eastern Africa:

Fig. 106 Development of Yield per 1000 ha in Western and Central Africa, Southern Africa and Eastern Africa. From (Bantilan et al., 2004)

Given the low soil fertility and irregular rainfall in semiarid regions, both increased water availability and higher levels of principal nutrients apparently will be necessary for substantial yield increase. The cultivar-alone strategy is unlikely to have a significant sustainable yield effect and therefore reduce poverty in semiarid Sub-Saharan Africa.

A reason why it did not work in earlier development years and a way forward is given by (Ahnert et al., 1996): The amount of genetic diversity in parental lines of commercial Sorghum bicolor hybrids is unknown, yet such comprehensive knowledge could improve the effectiveness of future hybrid development programs. Restriction fragment length polymorphisms (RFLPs) and pedigree data were used to investigate the genetic relationships in a group of 58 fertility restorer (R) and 47 sterility maintainer (B) elite Sorghum inbred lines. The objectives of this study were to assess the level of genetic variation for RFLPs in these Lines, estimate genetic similarity (GS) based on RFLPs and pedigree information for R- and B-Lines, and examine the agreement between RFLP-based GS and coancestry coefficient (f) for related (f > 0) pairs of inbreds. The average for R-lines was 0.08 and for B-lines 0.07. Cluster analysis of GS estimates from the entire set (105) of inbreds revealed separate groups for R- and B-lines in agreement with parental types, pedigree information, and the classification system used by breeders. R-lines clustered into two main groups, one derived mainly from Feterita and the other from Zera- zera. B-lines were grouped into different sub-clusters. GS and f were positively correlated for R-lines (r = 0.46) and for B-lines (r = 0.43), suggesting that RFLP data may help quantify the degree of relatedness in elite Sorghum germplasm.

208

(Bawazir & Idle, 1989) propose as breeding targets also drought resistance and root morphology. (Biradar & Borikar, 1985) see a future in the genetic analysis of shootfly resistance in relation to growth stages in Sorghum. A lot of efforts have been undertaken in the last decades in breeding modern traits of Sorghum with traditional and molecular methods (Blum et al., 1991; Dahlberg et al., 2001; Frederiksen & Rosenow, 1980; Gebrekidan, 1985; House, 1984; Jordan et al., 2004; Kresovich et al., 1986; Mattei, 1980; Miller, 1980; Mohammad et al., 1993; Yoshida, 2002; Zeller, 2000).

In Africa the potential to hybridize from a cultivated diploid species to a weedy tetraploid, and the production of a fully fertile diploid progeny from that cross, is significant for breeding, (Dweikat, 2005). Valuable agronomic traits are often present but inaccessible in the wild relatives of cultivated crop species. Utilization of wild germplasm depends on the production of fertile interspecific hybrids. Several unsuccessful attempts have been made to hybridize cultivated Sorghum with its wild relatives to broaden its genetic base and enhance agronomic value. The successful approach used by (Dweikat, 2005) employed the nuclear male sterility gene ms3 to generate a diploid fertile hybrid between the diploid cultivated S. bicolor and its weedy tetraploid wild relative Johnsongrass, Sorghum halepense. Eight Sorghum plants were selected from a Nebraska stiff stalk collection that contains the male sterility gene ms3 and were used as the female parent. About 36,000 florets of male sterile Sorghum were pollinated with Johnsongrass pollen to produce an average of one well-developed and 180 severely shriveled seed/18,000 crosses. The well-developed seed gave rise to a self- fertile diploid, while none of the shriveled seed were able to germinate. Preliminary results showed that several desirable traits from Johnsongrass, including resistance to greenbug and chinch bug and adaptability to cold temperatures, were expressed in the resulting progenies. These observations suggest that speciation within the genus Sorghum, giving rise to widely divergent phenotypes, is effected largely by ploidy-maintained crossing barriers but apparently not by extensive genomic divergence.

In conclusion, Johnsongrass, considered one of the 10 most troublesome and biologically very successful and vigorous weeds, possesses several extremely desirable plant traits including tolerance to pests, cold and drought. In earlier studies, interspecific crosses of Johnsongrass and cultivated Sorghum and the reciprocal produced mostly sterile triploid F1 plants (Casady & Anderson, 1952; Piper & Kulakow, 1994; Sangduen & Hanna, 1984) or both triploid and fertile tetraploid F1 (Hadley, 1953, 1958b; Hadley & Mahan, 1956) due to unreduced gametes produced by the diploid Sorghum.

(Chen et al., 1993) showed that fertile lines of S. bicolor differ from cytoplasmic male sterile (CMS) lines by the presence of a 3.8 kb HindIII chloroplast DNA fragment in the former and a smaller (3.7 kb) fragment in the latter. DNA/DNA hybridization studies showed that these two fragments are homologous. Fertile plants from S. versicolor, S. almum, S. halepense, and Sorghastrum nutans (Yellow Indiangrass) also have the 3.8 kb fragment, and CMS lines studied containing Al, A2 and A3 cytoplasms have the 3.7 kb fragment.

Several papers published in 2006 mark significant progress in basic preparative work for future breeding: A comprehensive chromosome mapping and a very useful chromosome nomenclature for the use in the breeding community is now established: (Childs et al., 2001; Kim et al., 2002; Kim et al., 2005a; Kim et al., 2005b, c; Zwick et al., 2000). (Feltus et al., 2006) and his team made progress in mapping QTLs and detected 100 common markers, the interspecific genetic maps show a high degree of colinearity, which will make interchange of useful genes much easier in the future. 209

6.2.2. The present day breeding programs of ICRISAT The activity of ICRISAT in favor of core collections will be an important basis for future breeding efforts. The core collection of ICRISAT is indispensable, it has been established by (Grenier et al., 2001a; Grenier et al., 2000a; Grenier et al., 2004; Grenier et al., 2000b; Grenier et al., 2001b) and many other collaborators, see also (Rao Prasada et al., 1989), the sample Number has now grown to 36000.

This calls for better management methods in order to make efficient use of this invaluable seed collection. In a first attempt different sampling methods have been compared: (1) Random sampling, (2) agronomic characteristics and (3) taxonomic knowledge based sampling: Also a genetic analysis of the same subsets did not reveal clear differences, and in all samples there remained a high proportion of rare alleles. In contrast to those results published 2000, the second analysis (2001) revealed good results of clear-cut differences:

Since 1972, the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) has maintained a large collection of Sorghum in India. The collection size has continuously increased, and the total number of accessions at present conserved in the gene bank has reached over 42000 accessions (Dahlberg et al., 2002). The need to help management was considered, and this study was conducted to establish core collections. This Sorghum collection was earlier stratified into four clusters according to the photoperiod sensitivity. Then, considering the core collection strategy, the authors used three random sampling procedures to determine the specific accessions to be included in the core [i.e., a constant portion (Core C), a proportional (Core P), and a proportional to the logarithm (Core L)] of the photoperiod group size sampling strategy. Both the Core C and L were significantly different from the landrace collection with better representation of the smallest groups, such as landraces insensitive to photoperiod. Despite differences between the three core collections, estimates of global diversity through the Shannon-Weaver Diversity Indices were of the same magnitude as the landrace collection. When compared, the Core C and L were significantly different. Core L sampled better for the characters, the race, and the latitudinal classes that were related to the photoperiod-sensitive landraces. Thus, for establishing a core collection with the widest range of adaptation to photoperiod, the authors propose the use of a logarithmic sampling strategy, which identifies a broadly adapted set of genotypes.

This was also the goal of another sampling study by (Dahlberg et al., 2002): In their comparative assessment of variation among Sorghum germplasm accessions they used seed morphology and RAPD measurements. The results indicate, that no one of the three approaches to develop clusters by means of agronomic descriptors closely 210 approximated the groupings produced by RAPD markers. Test 2, standardization of data by Z-scores and cluster analysis using the complete set of data, provided the highest similarity score for the race bicolor, but other texts were best for other races. Overall, test 2 yielded the highest similarity score. These results underscored the need for further research in the evaluation of techniques used to develop core collections and their validity.

6.2.3. More efficient testing methods for seed collections proposed Progress will be fostered by applying a more efficient testing method for new traits: (Mgonja et al., 2006). The authors applied sequential retrospective (SeqRet) pattern analysis to stratify Sorghum variety testing sites according to their similarity for yield discrimination among genotypes using historical grain yield data from 147 multi-environment trials (METs).

Fig. 107 Dendrogram for cumulative classification of 38 sites based on grain yield per hectare of Sorghum varieties planted during 1987/1988–1992/1993 and 1999/2000 using weighted environment-standardized squared Euclidean distance as dissimilarity measure and incremental sum of squares as clustering strategy. Site codes in Table 1. Italics indicate sites added to site groups based on nearest centroid criterion. The authors conclude: In the face of continually diminishing resources, there is an increasing need to limit the number of testing sites to a level that may allow reliable identification of promising genotypes for maximizing crop productivity in targeted agro-ecological zones of interest. A major step to achieve this goal is to stratify the testing sites into groups in a way that minimizes the within-group G _ E interaction. The application of SeqRet pattern analysis methodology on grain yield data from historical 211

METs stratified the 38 SADC region Sorghum variety testing sites, representing 35 distinct locations (Table 1), into six major groups:

Group 1: Lucydale, Kadoma, Chimoio; Group 2: Hombolo, Bigbend, Chiredzi, Chokwe, Gwebi, Maseru, Mt Makulu, Panmure, Naliendele, Nampula; Group 3: Matopos, Makoholi, Luve, Malkerns, Nhlangano, Tumbi; Group 4: Sebele, Kasintula, Pandamatenga, Good Hope, Mahanene, Mashare, Okashana, Ukiriguru, Umbeluzi; Group 5: Ngabu, Muzarabani, Luanda, Makoka, Lusitu; and Group 6: Ilonga, Golden Valley.

This grouping provides an objective basis to choose a small representative subset of testing sites for future targeting of Sorghum germplasm in the SADC region to increase breeding efficiency and maximize selection gains in pursuance of breeding for specific adaptation. This will also significantly contribute to faster release of new improved varieties and in increasing the efficiency of supply and delivery of seed for relief purposes. The results suggest that future Sorghum variety testing could be restricted to a few representative sites selected from within each of the six identified site-groups.

Farmers have worked over centuries with this variability, using their indigenous knowledge in achieving selection, improvement and utilization. (Doggett, 1965). Plant breeding involving hybridization and selection following testing got firmly established after the rediscovery of Mendels laws already at the beginning of the 20th century. However, the modern Sorghum breeding based on these principles started only in the late 1930s.

6.2.4. Breeding strategy of ICRISAT The following figure is quite useful and summarizes the breeding strategy of the future, it is proposed by ICRISAT in (Bantilan et al., 2004)

212

Fig. 108 ICRISAT’s Sorghum breeding strategy from 1972 onwards (Bantilan et al., 2004)

6.2.5. Regionalization of breeding strategy of ICRISAT In the following figures these ICRISAT strategies are regionalized and show a clear dependency to the main climate characteristics. It is good to know that more and more the research and development of some of the major grass crop genera is improving also on an organizational level. A lot of new research is facilitating the establishment of model systems in grass crop breeding, demonstrated here by a few selected citations: (Ahn et al., 1993; Ahn & Tanksley, 1993; Arber, 2002; Feuillet & Keller, 2002; Hughes, 2006; Ito et al., 2005; Kilian et al., 1995; Nature Genomics Gateway, 2003 ff; Poletti & Sautter, 2005; Young, 1999). It is important, that ICRISAT will be able to follow up the breeding program with modern methodologies set up in (Bantilan et al., 2004; van Jaarsveld et al., 2006).

213

Fig. 109 Sorghum research domains in Africa and India (Bantilan et al., 2004)

214

Fig. 110 Average yield and yield gain in Sorghum in different countries (Bantilan et al., 2004) 215

6.3. Development of modern Sorghum breeding

6.3.1. The first Sorghum transformations (Hagio et al., 1991) achieved a first transformation of Sorghum. Their report establishes the biolistic method as a useful route for delivery of DNA into the difficult-to-transform monocotyledonous plant species and represents the first stable transformation of Sorghum. Cells from a suspension culture of Sorghum vulgare have been transformed to either hygromycin or kanamycin resistance following uptake of pBC1 or pNGI plasmids, respectively, introduced on DNA- coated high velocity microprojectiles. Hygromycin- and kanamycin-resistant transformants contained hygromycin B phosphotransferase- and neomycin phosphotransferase-hybridizing restriction fragments of the expected size, respectively. A second introduced, but unselected for, reporter uidA gene which encodes beta-glucuronidase activity was also detected by DNA gel blot analysis in these transformants and shown to be expressed at low levels in two of the ten transformants analyzed. Transcripts from the introduced foreign genes accumulated to detectable levels in only these two transformants, both of which had a high copy number of genes integrated into their genome.

(Battraw & Hall, 1991) also succeeded at an early stage to obtain transgenic Sorghums: Parameters influencing the stable transformation of S. bicolor protoplasts with a chimeric neomycin phosphotransferase II (NPT II) gene by electroporation were investigated. The mean number of kanamycin-resistant calli produced increased in direct proportion to the concentration of DNA used for transformation. Linearization of the plasmid doubled the mean number of kanamycin-resistant calli produced, while the addition of carrier DNA had no effect. The copy number (1-4) of integrated genes was low compared with that frequently reported for PEG-mediated transformation. Two strategies for transforming protoplasts with a nonselectable, beta-glucuronidase (GUS) gene were compared. One utilized a plasmid containing a CaMV 35S-NPT II gene covalently linked to a CaMV 35S-GUS gene, and the other strategy utilized the two genes on separate plasmids. DNA from all 77 kanamycin-resistant calli analyzed contained restriction fragments hybridizing to the NPT II probe; approximately 70% of the clones from all transformation treatments contained a 1.7-kb EcoRI/HindIII restriction fragment corresponding to the full-length gene. Of the kanamycin-resistant calli, 38-63% (depending on the transformation treatment) contained GUS-hybridizing fragments, and 8-19% contained the full-length gene. The addition of NPT II and GUS genes on a single plasmid or on separate plasmids did not appear to lead to an appreciable difference in the frequency of cointegration of these genes, although an increased proportion of the plasmid bearing the nonselectable (GUS) gene appeared to favor its cointegration.

6.3.2. Difficulties of the transformation of monototylenonous plants There are many reasons, why in Monocotyledons transformation of the major cultivars met major difficulties in the beginning. After all, the first biolistic transformation was done relatively late (Potrykus et al., 1985) and even after this breakthrough the technology had to overcome many additional hurdles. (Hagio, 1998; Jeoung et al., 2002; Sai et al., 2006; Sticklen & Oraby, 2005; Tadesse et al., 2003)

216

Even recently some obstacles have been recognized by (Emani et al., 2002): In the light of results described thus far, it is interesting but not surprising that azaC treatment during the selection process following the transformation resulted in increased number of gusA expressing calli. This suggests that during the early stages of this investigation the reason for our inability to observe stable gusA gene expression was transgene silencing due to methylation. Treatment of the cell cultures with azaC proved to be an efficient way to obtain an increased number of reporter gene-expressing Sorghum cells. However, the feasibility of such a treatment in enhancing the expression of selectable marker gene resulting in the recovery of higher numbers of transgenic Sorghum plants which continue to express the transgenes needs to be investigated further. (Emani et al., 2002) clearly demonstrate that methylation- based transgene silencing is a serious problem in Sorghum. It may also help explain the very poor transformation efficiencies obtained in previous Sorghum transformation studies. The results presented in this report may also aid in designing more effective strategies to increase the efficiency of Sorghum transformation.

Also (Waniska et al., 2001) found Sorghum to be recalcitrant to transformation; however, a gene coding for a rice chitinase was incorporated into an elite Sorghum inbred line, Tx430, by a biolistic transformation protocol. Six primary transgenic plants were obtained and were shown to contain the transgene by Southern blot analysis (Zhu et al., 1998). The chitinase transgene was inherited in a Mendelian fashion by progeny from these primary transgenic plants and was expressed in the T1 and T2 generations.

One of the real problems in many transformation experiments with Sorghum is gene silencing, which was studied by (Emani et al., 2002): They developed a better physiological understanding about the conditions necessary for gene silencing and reactivation. these results suggest that methylation-based silencing is frequent in Sorghum and probably responsible for several cases of transgene inactivation reported earlier for this crop.

6.3.3. Progress in Sorghum transformation methods (Casas et al., 1997; Casas et al., 1993) obtained transgenic Sorghum plants (S. bicolor L. Moench), by microprojectile-mediated DNA delivery to explants derived from immature inflorescences. Explants were precultured on medium supplemented with kinetin, and sucrose prior to bombardment. Bialaphos selection pressure was imposed 2 wk after bombardment and maintained throughout all the culture stages leading to plant regeneration. More than 2500 explants from 1.5 to 3.0 cm inflorescences were bombarded and subjected to bialaphos selection, Out of more than 190 regenerated plants, 5 were determined to be Ignite/Basta resistant. Southern analyses confirmed the likelihood that the 5 herbicide resistant plants derived from two independent transformation events. The bar gene was inherited by and functionally expressed in T-1 progeny. However, no beta-glucuronidase (GUS) activity could be detected in T-1 plants that contained uidA restriction fragments. Histological analyses indicated that in the absence of bialaphos morphogenesis was primarily via embryogenesis while organogenesis was more predominant in callus maintained with herbicide selection.

217

(Zhao et al., 2000) used Agrobacterium tumefaciens for the transformation of Sorghum. Immature embryos of a public (P898012) and a commercial line (PHI391) of Sorghum were used as the target explants. The Agrobacterium strain used was LBA4404 carrying a 'Super-binary' vector with a bar gene as a selectable marker for herbicide resistance in the plant cells. Statistical analysis showed that the source of the embryos had a very significant impact on transformation efficiency, with field-grown embryos producing a higher transformation frequency than greenhouse-grown embryos. Southern blot analysis of DNA from leaf tissues of T-0 plants confirmed the integration of the T-DNA into the Sorghum genome. Mendelian segregation in the T-1 generation was confirmed by herbicide resistance screening. This first successful use of Agrobacterium showed higher transformation yields of stable transformations than other methods reported previously.

(Able et al., 2001) developed optimized particle inflow gun (PIG) parameters for producing transgenic Sorghum bicolor. Both transient and stable expression were examined when determining these parameters. The uidA reporter gene (GUS) encoding beta -glucuronidase was used in transient experiments and the green fluorescent protein (GFP) used to monitor stable expression. Initially, optimization was conducted using leaf segments, as the generation of Sorghum callus in sufficiently large quantities is time-consuming. Following leaf optimization, experiments were conducted using callus, identifying a high similarity between the two tissue types (r(s) = 0.83). High levels of GUS expression were observed in both leaf and callus material when most distant from the DNA expulsion point, and using a pressure greater than 1800 kPa. Three promoters (Ubiquitin, Actin1 and CaMV 35S) were evaluated over a 72-h period using GUS as the reporter gene. Using this optimized protocol, several plants were regenerated after having been bombarded with the pAHC20 construct (containing the bar gene), with molecular evidence confirming integration.

(Jeoung et al., 2002) also worked successfully on the optimization of Sorghum transformation: Early and reliable detection of plant transformation events is essential for establishing efficient transformation protocols. They have compared the effectiveness of using the gene encoding a green fluorescent protein (GFP) and a beta-glucuronidase (gus) as reporter genes for early detection of transgene expression in explants subjected to biolistic bombardment and Agrobacterium-mediated transformation. The results indicate that gfp gene is superior to gas gene in following transgene expression in transiently transformed materials in both methods of transformation. Using GFP as the screenable marker, the authors have optimized Sorghum transformation with respect to the conditions for transformation, type of explants, promoters, and inbreds. These optimized conditions have been used to obtain stably transformed explants for subsequent regeneration

(Tadesse et al., 2003) optimized the conditions for microparticle bombardment for four explant types of Sorghum ( S. bicolor (L.) Moench) based on transient expression of the uidA reporter gene. After selection on geneticin, fertile transgenic Sorghum plants were regenerated from immature embryos as well as from shoot tips. Stable integration and Mendelian inheritance of the neo selectable marker gene was demonstrated in all transgenic plants.

(Devi & Sticklen, 2003) transferred a chitinase gene from American elm, and the bar selectable gene to Sorghum by particle bombardment. Eleven putative transgenic plants regenerated on selection medium 218 were transferred to sterile soil mix and maintained in the containment glasshouse. The presence of chitinase and bar genes in five T0 transgenic plants was confirmed by PCR and Southern blot analysis. Studies on the inheritance pattern of herbicide resistance and fungal resistance will follow.

(Carvalho et al., 2004): The results presented in this work support the hypothesis that Agrobacterium- mediated transformation of Sorghum is feasible, analogous to what has been demonstrated for other cereals such as rice, maize, barley and wheat. The four factors which most influenced transformation were: (1) the sensitivity of immature Sorghum embryos to Agrobacterium infection, (2) the growth conditions of the donor plant, (3) type of explant and (4) co-cultivation medium. A major problem during the development of this protocol was a necrotic response which developed in explants after co-cultivation. Although promising, the overall transformation efficiency of the protocol is still low and further optimization will require particular attention to be given to the number of Agrobacterium in the inoculum and the selection of Sorghum genotypes and explants less sensitive to Agrobacterium infection.

In the same year, (Cheng et al., 2004) summarize the present day situation on the methodology of Agrobacterium mediated transformation with a view on Sorghum breeding: Efficient transformation protocols for agronomically important cereal crops such as rice, wheat, maize, barley, and sorghum have been developed and transformation for some of these species has become routine. Many factors influencing Agrobacterium-mediated transformation of monocotyledonous plants have been investigated and elucidated. These factors include plant genotype, explant type, Agrobacterium strain, and binary vector. In addition, a wide variety of inoculation and co-culture, conditions have been shown to be important for the transformation of monocots. For example, antinecrotic treatments using antioxidants and bactericides, osmotic treatments, desiccation of explants before or after Agrobacterium infection, and inoculation and co-culture medium compositions have influenced the ability to recover transgenic monocots. The plant selectable markers used and the promoters driving these marker genes have also been recognized as important factors influencing stable transformation frequency. Extension of transformation protocols to elite genotypes and to more readily available explants in agronomically important crop species will be the challenge of the future. Further evaluation of genes stimulating plant cell division or T-DNA integration, and genes increasing competency of plant cells to Agrobacterium, may increase transformation efficiency in various systems. Understanding mechanisms by which treatments such as desiccation and antioxidants impact T-DNA delivery and stable transformation will facilitate development of efficient transformation systems.

(Girijashankar et al., 2005) produced transgenic Sorghum plants expressing a synthetic cry1Ac gene from Bacillus thuringiensis (Bt) under the control of a wound-inducible promoter from the maize protease inhibitor gene (mpiC1) via particle bombardment of shoot apices.

(Gao et al., 2005a) transformed grain Sorghum (S. bicolor (L.) Moench) with a visual reporter gene (gfp) and a target gene (tlp), three genotypes (two inbreds, Tx 430 and C401, and a commercial hybrid, Pioneer 8505) were used. They obtained a total of 1011 fertile transgenic plants from 61 independent callus lines, which were produced from 2463 zygotic immature embryos via Agrobacterium-mediated transformation. A total of 320 plants showing GFP expression, derived from 45 calli, were selected and 219 analyzed by Southern Blot analysis. There was a 100% correlation between the GFP expression and the presence of the target gene, tlp, in these plants. The progeny also showed different copy numbers of transgenics. The authors describe the successful use of GFP screening for efficient production of stably transformed Sorghum plants without using antibiotics or herbicides as selection agents

(Gao et al., 2005b) Used a dual-marker plasmid containing the selectable marker gene, manA, and the reporter gene, sgfp, used to transform immature Sorghum embryos by employing an Agrobacterium- mediated system. Both genes were under the control of the ubi1 promoter in a binary vector pPZP201. The Escherichia coli phosphomannose isomerase (PMI) gene, pmi, was used as the selectable marker gene and mannose was used as the selective agent Optimization of this selection system for Sorghum transformation provides an efficient way to produce transgenic plants without using antibiotic or herbicidal agents as selectable markers, and our results showed that the transformation efficiency reached 2.88% for Pioneer 8505 and 3.30% for C401, both values higher than in previously published reports.

(Sai et al., 2006; Sticklen & Oraby, 2005) have developed a working method for invitro cell cultures, often an indispensable tool for plant biotechnology and propagation.

(Sticklen & Oraby, 2005) and (Sai et al., 2006) avocate a new method for more reliable explants in the transformation technique: Immature zygotic embryo has been the widely used explant Source to develop embryogenic callus lines, cell suspensions and protoplasts for transformation of cereal crops including maize, wheat, rice, oat, barley, Sorghum, and millet. However, the lack of competence of immature embryos in certain elite lines is still a barrier to routine production of transgenic cereal crops in certain commercial cultivars. In addition, a great deal of effort is required to produce immature embryos, manipulate cultures of immature embryos or their cell suspensions, and cryopreserve cultures for further use. In addition, undifferentiated cells may have reduced regenerability after a few months of in vitro culture. Alternative explants and regeneration systems for efficient transformation of cereal crops are needed to avoid or reduce the above limitations. During the past decade, scientists have successfully manipulated the shoot apical meristems from seedlings of maize, oat, Sorghum, millet, wheat, and barley in all effort. to develop a less genotype-dependent and efficient cereal regeneration system that can be maintained in vitro for long periods of time without the need for cryopreservation. Furthermore, apical meristem regeneration systems were used to stably transform maize, wheat, rice, oat, barley, Sorghum, and millet.

6.3.4. Situation in the year 2004 in transformed Sorghum cultivars To the present day there are several transgenic Sorghums existing in the laboratories, none of them is already in field testing stage.

220

Fig. 111 Characteristics of Sorghum research domains, from (Bantilan et al., 2004)

6.3.5. Present day situation in transgenic Sorghums

(Franks et al., 2006a) work on an extension of the cultivation area with new cold tolerant traits. Early season cold tolerance in grain Sorghum [S. bicolor (L.) Moench] is a desirable trait for extending its production range and minimizing risks associated with early spring plantings. Chinese accessions from this working group would serve as a source of favorable genes primarily for tolerance to low temperatures during the germination and emergence phase of growth in the breeding of cold tolerance Sorghum lines.

(Dingkuhn et al., 2006) identify systems caught in the agricultural transition from subsistence to intensified, market-oriented production as the most important target for crop improvement, and provide examples of new breeding objectives for cowpea, Sorghum and upland rice. In each of these cases, breeders, with the help of physiologists, have developed innovative plant-type concepts that combine improved yield potential and input responsiveness with specific traditional crop characteristics that remain essential during the agricultural transition. Sorghum breeders who had previously deselected photoperiod sensitivity are now reinserting sensitivity into plants having "modern" architecture, in order to allow for flexible sowing dates while maintaining an agro-ecologically optimal time of flowering near the end of the wet season. The ecophysiological basis of these plant types, their place in current and future cropping systems, as well as the problem of under-funding for their realization, are discussed. Asia's Green Revolution of the 1960s and 1970s has largely bypassed West Africa, and "modem" (high- yielding, input responsive) germplasm for staple crops has found comparatively little adoption, except for systems that are have good access to markets and sufficient water resources. It is unlikely, however, 221 that breeding objectives conserving traditional crop characteristics as found in extensive systems would have been more successful.

See also some citations at the end of the chapter 4.3.3. on progress in Sorghum transformation methods.

6.3.6. Linkage maps of Sorghum (Chittenden et al., 1994) The first ''complete'' genetic linkage map of Sorghum section Sorghum is described, comprised of ten linkage groups putatively corresponding to the ten gametic chromosomes of S. bicolor and S. propinquum. The map includes 276 RFLP loci, predominately detected by PstI-digested S. bicolor genomic probes, segregating in 56 F-2 progeny of a cross between S. bicolor and S. propinquum. Although prior cytological evidence suggests that the genomes of these species are largely homosequential, a high level of molecular divergence is evidenced by the abundant RFLP and RAPD polymorphisms, the marked deviations from Mendelian segregation in many regions of the genome, and several species-specific DNA probes. The remarkable level of DNA polymorphism between these species will facilitate development of a high-density genetic map. The S. bicolor x propinquum RFLP map provides valuable information regarding both transmission genetics and evolutionary genetics, and has later helped in mapping QTLs associated with the divergence of grain Sorghums from grass like relatives. Moreover, a detailed map of DNA markers is a valuable link between the disciplines of classical plant breeding and molecular biology, with many potential applications for the improvement of Sorghum and related species, related from marker assisted selection to map-based cloning.

(Crasta et al., 1999; Xu et al., 2000) made progress with analyzing and localizing QTLs in order to prepare the field for drought resistant trait breeding: The identification of genetic factors underlying the complex responses of plants to drought stress provides a solid basis for improving drought resistance. The stay- green character in Sorghum (S. bicolor L. Moench) is a post-flowering drought resistance trait, which makes plants resistant to premature senescence under drought stress during the grain filling stage. The objective of this study was to identify quantitative trait loci (QTLs) that control premature senescence and maturity traits, and to investigate their association under post-flowering drought stress in grain Sorghum. A genetic linkage map was developed using a set of recombinant inbred lines (RILs) obtained from the cross B35 x Tx430, which were scored for 142 restriction fragment length polymorphism (RFLP) markers. The RILs and their parental lines were evaluated for post-flowering drought resistance and maturity in four environments. The molecular genetic analysis of the QTLs influencing stay-green and maturity, together with the association between these two inversely related traits, provides a basis for further study of the underlying physiological mechanisms and demonstrates the possibility of improving drought resistance in plants by pyramiding the favorable QTL.

(Childs et al., 2001): Sorghum is an important crop for breeders of cereals, because it manages to grow under different kinds of environmental stress and contains well adapted genome parts, and the genome is rather small. The modified cDNA selection procedure described will be useful for genome-wide gene discovery and EST mapping in Sorghum, and for comparative genomics of Sorghum, rice, maize and other grasses. The authors produced a promising partial transcript map along Sorghum linkage group B. 222

(Draye et al., 2001) worked towards the Integration of Comparative Genetic, Physical, Diversity, and Cytomolecular Maps for Grasses and Grains, Using the Sorghum Genome as a Foundation. The small genome of Sorghum (S. bicolor L. Moench.) provides an important template for study of closely related large-genome crops such as maize (Zea mays) and sugarcane (Saccharum spp.), and is a logical complement to distantly related rice (Oryza sativa) as a “grass genome model.” Using a high-density RFLP map as a framework, a robust physical map of Sorghum is being assembled by integrating hybridization and fingerprint data with comparative data from related taxa such as rice and using new methods to resolve genomic duplications into locus-specific groups. By taking advantage of allelic variation revealed by heterologous probes, the positions of corresponding loci on the wheat (Triticum aestivum), rice, maize, sugarcane, and Arabidopsis genomes are being interpolated on the Sorghum physical map. The authors seek to provide a detailed picture of the structure, function, and evolution of the genome of Sorghum and its relatives, together with molecular tools such as locus-specific sequence- tagged site DNA markers and bacterial artificial chromosome contigs (a set of overlapping DNA segments derived from a single genetic source, from contiguous) that will have enduring value for many aspects of genome analysis.

(Dahlberg et al., 2001) Ergot or sugary disease of Sorghum has become an important constraint in North and South American countries that rely on F-1 hybrid seeds for high productivity. The objective of this research was to determine the vulnerability of various germplasm sources and publicly bred Sorghum lines to sugary disease (Claviceps africana) in the United States. The combination of flower characteristics associated with resistance were least exposure time of stigma to inoculum before pollination, rapid stigma drying after pollination, and small stigma. An Ethiopian male-fertile germplasm accession, IS 8525, had good levels of resistance. Its A(3) male-sterile hybrid had the highest level of resistance in the male-sterile background. IS 8525 should be exploited in host-plant resistance strategies.

(Uptmoor et al., 2003) studied the genetic characteristics of South African landraces of S. bicolor and came to the following conclusions: The results on genetic relatedness and genetic diversity within Sorghum accessions grown in Southern Africa revealed a clear separation between landraces and breeding varieties, but a similar level of genetic diversity within both subgroups. This information in connection with knowledge on agronomic traits (Wenzel et al., 2001a; Wenzel et al., 2001b) may have an impact on Sorghum breeding in this region. One breeding strategy may be to choose well-adapted parents that posses many random genetic differences in the hope of an increased number of transgressive recombinants (Graner et al., 1994; Tinker et al., 1993). In comparison to pedigree data, detailed genetic similarity estimates based on molecular markers are better suited for this selection of parental genotypes since additional information about the number of segregating loci is provided. Besides this, these molecular markers facilitate an unequivocal identification of respective germplasms for conservation purposes (Karp & Isaac, 1998; Karp et al., 1997).

(Folkertsma et al., 2005) deliver an important ICRISAT – funded study of the genetic diversity of Guinea- race of Sorghum, S. bicolor (L.) Moench. The Guinea-race is a predominantly inbreeding, diploid cereal 223 crop. It originated from West Africa and appears to have spread throughout Africa and South Asia, where it is now the dominant Sorghum race, via ancient trade routes. To elucidate the genetic diversity and differentiation among Guinea-race Sorghum landraces, the authors selected 100 accessions from the ICRISAT Sorghum Guinea-race Core Collection and genotyped these using 21 simple sequence repeat (SSR) markers. The 21 SSR markers revealed a total of 123 alleles with an average Dice similarity coefficient of 0.37 across 4,950 pairs of accessions, with nearly 50% of the alleles being rare among the accessions analysed. Stratification of the accessions into 11 countries and five eco-regional groups confirmed earlier reports on the spread of guinea race Sorghum across Africa and South Asia: most of the variation was found among the accessions from semi-arid and Sahelian Africa and the least among accessions from South Asia. In addition, accessions from South Asia most closely resembled those from southern and eastern Africa, supporting earlier suggestions that Sorghum germplasm might have reached South Asia via ancient trade routes along the Arabian Sea coasts of eastern Africa, Arabia and South Asia. Stratification of the accessions according to their Snowden classification indicated clear genetic variation between margaritiferum, conspicuum and roxburghii accessions, whereas the gambicum and guineense accessions were genetically similar. The implications of these findings for Sorghum Guinea-race plant breeding activities are discussed: The main objective was to analyze the pattern of genetic diversity within guinea race landraces to facilitate their use in breeding programs for West and Central Africa. The relatively large proportion of rare alleles (allele frequency <5%) and overall divergence among guinea race Sorghum accessions indicates opportunities to select divergent parents with enough genotypic and phenotypic variation to allow mapping of genes and quantitative trait loci and for marker-assisted introgression of traits into elite breeding lines within this race itself. The authors also suggest These results suggest that guinea hybrid breeding work could develop heterotic pools on the basis of ecogeographic/morphological characteristics so as to maximize genetic distance and heterosis. For example, the guineense and gambicum accessions from West Africa could serve as one pool, and the conspicuums from humid West-, East-, and southern Africa could serve as a separate pool.

(Oliver et al., 2005) finished a study on the gene level. It may be possible to add value to crop and animal systems by enhancing the digestibility of the stover residue by the use of brown midrib (bmr) genes if grain yields can be maintained. The objectives of this study were to evaluate the effect of bmr-6 and bmr-12 genes on grain yield of Sorghum and to evaluate the effect of the bmr genes on stover yield and quality in these genetic backgrounds: 'Wheatland', 'Redlan', RTx430, Tx623, Tx630, Tx631, and the hybrid A Wheatland X RTx430. Plant height, maturity, grain yield and test weight, stover neutral detergent fiber (NDF), acid detergent fiber (ADF), acid detergent lignin (ADL), and in vitro NDF digestibility (IVNDFD) were measured in split-plot experiments replicated four times in each of four environments with lines being whole-plots and genotypes being subplots. Although the variable expression of bmr-12 and bmr-6 in different lines indicates that selection of compatible genetic backgrounds will be critical in determining the realized impact on value, the results are promising and merit further selection work.

(Feltus et al., 2006) worked out maps of QTLs for all chromosomes of Sorghum, specifically he concentrated on ‘staygreen’ loci which will be important for drought resistance breeding. As a whole, they have detected overlap of eleven QTL 224 pairs, and at least four of these QTL correspondences are unlikely to be due to chance. However, it should be noted that there could be an increased tendency to correspondence of QTLs if the underlying chromatin is located in a region of low recombination or high genedensity (Noor et al. 2001). Low recombination rates have a tendency to combine the effects of a number of small-effect genetic factors into apparent large- effect QTL. This effect can be offset by low gene density, but it can be enhanced by high gene density. Low recombination rates are especially common in centromeric and telomeric regions. The hypergeometric probability function does not take these biases into account, and corrections must wait until a Sorghum recombination/gene-density map is developed. Collectively, these data provide strong support for the transferability of molecular tools between interspecific and intraspecific crosses, and begin the process of separating QTLs that contribute to morphological and physiological divergence among Sorghum species from those that contribute to diversity within the cultigen (S. Sorghum). Such a ‘categorization’ of QTLs is of value for setting priorities in ongoing studies of botanical diversity, and crop improvement, respectively.

6.4. Present and future breeding efforts

6.4.1. Biofortification of Sorghum crops

Biofortification has been an important breeding target also in earlier periods: Already (Doggett, 1981) made it clear that higher lysine levels would be desirable as breeding goal – but difficult to obtain: Sorghum is a crop greatly in need of improved protein quality (Hulse et al., 1980) state: Recognizing the inferior nutritional quality of normal Sorghum protein, the inhibition of protein by polyphenols, together with nitrogen losses that occur during domestic and small industrial processing, and probably, in storage, it is recommended that research be continued to stabilize a higher than average lysine in combination with an average (c 10% N x 5.7) protein content. “One cannot make bricks without straw: we have not yet found the straw.” In Sorghum grain, the increased protein resulting from better nutrition of the plant is largely prolamine, of low nutritive value. Protein levels are low under conditions in which nitrogen is limiting at grain filling. (Riley, 1980) has made a thorough study of the protein situation is Sorghum: the progress made in trying to use two high-lysine genes from Ethiopia, and the mutant in P-721 from Purdue. There were few grounds for encouragement. Plump seeds were unobtainable with the Ethiopian source in crosses: lysine content transferred better from P721, but yields were poor. See also the early studies of (Jambunathan et al., 1975; Oswalt, 1973)

This view is also reinforced by the National Academy in Washington (Ruskin et al., 1996). In a special chapter 7 p. 127 dedicated to the ‘lost crop’ of Sorghum the authors call for an effort to lift such an important crop to a much better nutritional level:

Biofortification is now belonging to several breeding programs of major crops worldwide. Information can be downloaded mainly from two websites: http://www.harvestplus.org/ and http://www.superSorghum.org 225

(Austin et al., 1972; Cavins et al., 1972; Collins & Pickett, 1972a, b; Deosthal.Yg & Mohan, 1970; Doesthal.Yg et al., 1970; Naik, 1968; Tanksley & Escobosa, 1971; Virupaka.Tk & Sastry, 1967) studied the relation of lysine content to protein levels in Sorghum. (Singh & Axtell, 1973) discovered a high lysine mutant gene (Hl) that improves protein quality and biological value of grain-Sorghum. (Sullins et al., 1975) studied the endosperm structure of high lysine Sorghum. (Singh, 1976) published modified vitreous endosperm recombinants from crosses of normal and high lysine Sorghum. (Copelin et al., 1978; Copelin et al., 1976) studied the availability of lysine in Sorghum. (Ejeta & Axtell, 1987) published a study of high lysine traits in Sorghum: they produced data on Protein and Lysine Levels in developing kernels of normal and high-lysine Sorghum. (Vanscoyoc et al., 1988) studied kernel characteristics and protein-fraction changes during seed development of high-lysine and normal Sorghums. (Monyo et al., 1988) report on combining ability of high lysine Sorghum lines derived from P-721 Opaque. (Jambunathan et al., 1983) write about rapid methods for estimating protein and lysine in Sorghum.

Classical ''high-lysine'' Sorghum lines are according to (Vernaillen et al., 1993) characterized by smaller seeds than average, due to a decrease in prolamin synthesis and a subsequent decrease in yield. To exploit the natural variation in lysine content and to identify ecotypes with a seed lysine content higher than average, characterized by plump seeds, a method was developed based on root-growth inhibition of seeds growing on a medium containing aminoethylcysteine (AEC), a lysine analogue. By using a collection of Sorghum mutants and ecotypes a correlation coefficient of 0.926 between root length and lysine content was established. This method, which uses the root length of plants growing on AEC to indicate which lines have a potential elevated lysine content, can be applied for the screening of Sorghum germplasm. Since this is a non-destructive method it can also be used at the individual seed level, for example for screening progenies of regenerated plants from in vitro culture to exploit the somaclonal variation.

(Weaver et al., 1998) report about the discovery of grain Sorghum germ plasm with high uncooked and cooked in vitro protein digestibilities: Grain Sorghum has been documented to have low protein digestibility relative to other cereal grains. Low protein digestibility of Sorghum is most pronounced in cooked foods and is ranked slightly lower than corn as a feed grain. In this article, Sorghum germ plasm is identified that has substantially higher uncooked and cooked flour in vitro protein digestibility than normal cultivar S. bicolor lines were found within a high-lysine population derived from the mutant P721Q containing approximatly 10-15% higher uncooked and approximate to 25% higher cooked protein digestibilities using a pepsin assay. Highly digestible Sorghum grain showed little reduction in digestibility after cooking, compared to the large reduction that is typical of normal Sorghum cultivars. Using the three-enzyme pH-stat method, the authors showed that the highly digestible lines had the same degree of peptide bond hydrolysis in approximate to 5 min, as was found in 60 min in the normal cultivar, P721N. Differences in protein digestibility were related to enzyme susceptibility of the major storage prolamin, alpha-kafirin, that comprises approximate to 50-60% of the total Sorghum grain protein. Using the enzyme-linked immunosorbent assay (ELISA) technique to track the pepsin digestion of a-kafirin, the highly digestible 226 lines had approximate to 90-95% alpha-kafirin digested in 60 min compared to 45-60% for two normal cultivars. More digestible protein from Sorghum grain, that additionally is high in lysine content and has a fairly hard endosperm, could be of important benefit to populations who lack adequate protein in their diets, and may, pending further studies, prove to increase the value of Sorghum as a feed grain.

(O'Brien, 1999) tried to disentangle genotype and environment effects on feed grain quality in an extensive review covering most cereals. They found that the most thoroughly researched grain has been Sorghum, which is principally grown in developed countries for feeding to livestock, where some studies have been conducted to define the extent of genotype, location, and genotype x environment interaction effects. Scope exists to enhance the nutritive value of Sorghum by breeding through modification of endosperm composition, tannin content, and improved protein digestibility.

(Reddy & Jacobs, 2002) concentrated on Sorghum lysine-rich cultivar verification by SDS-PAGE and Southern blot. The aim was to assess the genetic variability among lysine-rich cultivars of Sorghum and to compare with low lysine cultivar White Martin and a chemically induced high lysine mutant P721O. The lysine-rich cultivars contain approximately 1.5 to 2 times more lysine when compared to low-lysine cultivar. The detected variability among kafirins both in SDS-PAGE and Southern blot could be effectively used as markers in selection of lysine-rich cultivars for further use in breeding program.

(Fontaine et al., 2002) developed NIRS calibrations for the accurate and fast prediction of the total contents of methionine, cystine, lysine, threonine, tryptophan, and other essential amino acids, protein, and moisture in the most important cereals and brans or middlings for animal feed production. More than 1100 samples of global origin collected over five years were analyzed for amino acids following the official methods of the United States and European Union. (Hassen et al., 1986) report on tryptophan levels in normal and high-lysine Sorghums.

(Wu et al., 2003) published an important paper on the enrichment of cereal protein lysine content by altered tRNA(lys) coding during protein synthesis. The world's major crops are deficient in lysine and several other amino acids essential for human and animal nutrition. Increasing the content of these amino acids in cereals, our major source of dietary energy, can help feed a global population whose reliance upon dietary protein is growing faster than crop yields The authors document the heritable expression in rice, the world's major cereal crop, of tRNA(lys) species that introduce lysine at alternative codons during protein synthesis, resulting in a significant enrichment of the lysine content of proteins in rice seeds without changing the types or quantities of the seed storage proteins.

227

Fig. 112 Gene constructs used for rice (O. sativa) transformation. tRNAlys (CUA), A. thaliana tRNAlys gene with anticodon sequence altered from CTT to CTA; UBI, maize ubiquitin promoter; GUS, b-glucuronidase gene; LUCam, firefly luciferase gene with TAG codon at position 206; BAR, Phosphinothricin acetyltransferase(PAT) gene; GST, Glutathione S-transferase gene; NOS, terminator of the nopaline synthase gene; HA, Haemagglutinin epitope coding sequence. From (Wu et al., 2003) (Slingerland et al., 2006) present preliminary results reporting from a project done in a joint venture with the agricultural research center of Wageningen and Ouagadougou, Burkina Faso, suggesting that a feasible chain solution consists of breeding for high Fe and moderate phytic acid contents and using soil organic amendments and P fertilization to increase yields is possible, but that this needs to be followed by improved food processing to remove phytic acid. Further research on timing of application of phosphate, Fe fertilizer and soil organic amendments is needed to improve phytic acid-Fe molar ratios in the grain. Research on the exact distribution of Fe, phosphate, phytic acid and tannins within the Sorghum grain is needed to enable the development of more effective combinations of food processing methods aiming for more favorable phytic acid-Fe molar ratios in Sorghum-based food. See also the comments of (Dicko et al., 2006) below. Here a few more details are shown with the aim to demonstrate the full complexity of the biofortification task: in this thoughtful paper, the authors offer a holistic view how to approach biofortification: In a first attempt, they demonstrate that biofortification is just one of three ways to solve the existing malnutrition problem (here with the example of iron deficiency):

228

Fig. 113 The three steps determining the effective supply of nutrients, the four current strategies (a–d) and a fifth new strategy impacting on these steps. (Source: C.E. West, unpublished). From (Slingerland et al., 2006)

Three strategies are currently in practice to improve Fe supply: (a) dietary diversification, (b) supplementation (Bothwell, 2000; Bothwell & MacPhail, 2004), and (c) fortification (Merx et al., 1996). They interfere at different levels in the above figure. Dietary diversification aims to increase the availability and intake of food that is rich in bioavailable Fe. Supplementation provides Fe sources in the form of pills and suchlike, additionally to daily diets. Fortification enhances levels of bioavailable Fe in frequently consumed foods either by adding Fe or by decreasing the effects of inhibitors. A fourth, more recent strategy aims to increase Fe in plant-based foods by breeding for micronutrient content (biofortification). They all contribute to solving the problem, yet they require certain conditions to be successful.

The food chain approach according to (Slingerland et al., 2006): The food chain approach is represented by the figure below, which forms the framework for the program. The complete program focuses on improving Fe and Zn supply through staple foods in West Africa and China. This paper mainly reports on Fe and phytic acid in sorghum staple food in West Africa although some results on zinc will be mentioned in the agronomic part. The boxes in the middle of Figure 2 form a chain representing the flow of Fe, Zn and phytic acid from natural resources (soil) to the human body (health). Enhancers are only considered in the domain of dietary composition. Breeding, agronomic practices, storage, food preparation and processing, and dietary composition are possible interventions. Plant and soil factors influence the availability of Fe for uptake by plants. Actual uptake can further be stimulated by soil management measures. Varieties differ in their ability to take up micronutrients from the soil or to transport micronutrients and phytic acid to the grain. Knowledge of plant physiological processes and assessment of ranges of difference in Fe and phytic acid contents in the grain are essential for developing a breeding strategy aiming at varieties with favorable grain phytic acid-Fe ratios. Assessment of post-harvest activities is needed for proposing food processing methods that increase the contents of desirable micronutrients and remove or de-activate anti-nutritional factors, at the same time making attractive and digestible foods. The inner circle of the figure below concerns decision-making and resource allocation by producers, processors and consumers of staple foods. In West Africa all activities regarding staple foods generally 229 take place at household level. The household and its actors are therefore important units of research and implementation.

Fig. 114 Analytical framework for the research program: food chain approach and context. From (Slingerland et al., 2006)

The outer circle of the figure represents the ‘environment’. In West Africa the biophysical environment consists of poor soils and of rainfall patterns that are highly irregular both in time and space, making high and sustainable food production difficult. The economic environment is characterized by lack of infrastructure (roads, irrigation facilities and financial institutions), lack of affordable inputs (fertilizers, pesticides), low purchasing power of both producers and consumers and thus lack of attractive markets and large price fluctuations between seasons and between years. Such an environment is characterized by high transaction costs. The social environment consists of a majority of poorly educated people with little clout towards governments and industries. The ‘donor’ environment consists of many developing organizations that work on all kinds of topics, such as the Helen Keller Foundation on home gardening, UNICEF in micronutrient 230 supplementation programs and the International Fertilizer Development Center (IFDC) working on nutrient management to increase crop yields, to name just a few. National policy documents show that the policy environment is aware of the micronutrient problem. However, national governments and their representatives at decentralized levels lack sufficient financial means, scientific insights or political power to co-ordinate these donor-initiated efforts or to choose from proposed interventions. One role of the research program is to support decision-making by providing insight into the impacts of the different interventions and their adequacy in different environments. The program will communicate with donor organizations and policy makers through scientific papers and contributions to regional and international conferences. Local farmers and food processors are included in the research and participate in discussions on the hypotheses, set-up and outcomes of parts of the research. These discussions will lead to the formulation of locally applicable recommendations for best practices.

Preliminary results show that specific management practices with various fertilizers are influencing besides yield itself the content of Zink, Phytic Acid and its ratio:

231

Fig. 115 Effect of Zn and P fertilizer on the relation between sorghum grain yield and (A) Zn in the grain, (B) phytic acid in the grain, and (C) the phytic acid / Zn molar ratio. Results from the Somyaga field experiment 2003. (Source: K. Traore, unpublished results) A problem is that literature does not provide information per genotype for the four components at the same time: Fe, Zn, polyphenol (or tannin) and phytic acid. Information is incomplete and scattered and the genetic component is hardly ever separated from environmental and management effects. Therefore the program decided to screen the sorghum core collections of ICRISAT and CIRAD. Based on these data the program will select sorghum varieties with genetic potential for high Fe and moderate phytic acid contents (e.g. variety 2 in Figure 4) or for moderate Fe contents and low phytic acid (e.g. variety 3 in Figure below). 232

Fig. 116 Zinc, iron and phytic acid contents of the grain of six contrasting sorghum varieties selected from the CIRAD sorghum collection at their field station at Samanke, Mali, in the 2002–2003 season. (Source: M.A. Slingerland, unpublished results).

Literature and the sorghum screening experiment showed that there is scope for breeding to increase Fe content and decrease the phytic acid-Fe molar ratio. Breeding for zero phytic acid or zero tannins, supposedly having negative agronomic side-effects, is not necessary. Within the program, experimental work has started on gathering information on the exact physiological pathways and roles of anti- nutritional factors, Fe and Zn in the plant. The experiments include measuring and observing phenological characteristics of existing varieties. Modern molecular techniques such as amplified fragment-length polymorphism (AFLP) analysis and specific molecular statistics will be used to link the Fe, Zn and phytic acid contents of the grain to the uptake, transport and translocation processes and to the potentially responsible genes. Further experiments will verify the resulting linkages. Fieldwork is being performed to verify agronomic trade-offs in yield or resistance to birds and insects. The work is done on millet in parallel with sorghum, developing an efficient integrated breeding strategy aiming at 233 high Fe and Zn and low phytic acid contents in both crops. So far only preliminary results have been obtained, not allowing for any reporting at this stage. Interestingly enough, the programmes are conducted under the rules of participatory project structures, for which the group has published earlier an important paper on its basic approach, called ‘the unifying power of sustainable development’, which leads towards balanced choices between ‘People, Planet and Profit’ in agricultural production chains and rural land use (Slingerland et al., 2003).

(Tesso et al., 2006) report on the development of high-protein digestibility (HPD)/high-lysine (hl) Sorghum mutant germplasm with good grain quality (i.e., hard endosperm texture). It has been a major research objective at Purdue University. Progress toward achieving this objective, however, has been slow due to challenges posed by a combination of genetic and environmental factors. In this article, the authors report on the identification of a Sorghum grain phenotype with a unique modified endosperm texture that has near-normal hardness and possesses superior nutritional quality traits of high digestibility and enhanced lysine content. These modified endosperm lines were identified among F-6 families developed from crosses between hard endosperm, normal nutritional quality Sorghum lines, and improved HPD/hl Sorghum mutant P721Q-derived lines. A novel vitreous endosperm formation originated in the central portion of the kernel endosperm with opaque portions appearing both centrally and peripherally surrounding the vitreous portion. Kernels exhibiting modification showed a range of vitreous content from a slight interior section to one that filled out to the kernel periphery. Microstructure of the vitreous endosperm fraction was dramatically different from that of vitreous normal kernels in Sorghum and in other cereals, in that polygonal starch granules were densely packed but without the typically associated continuous protein matrix. The authors speculate that, due to the lack of protein matrix, such vitreous endosperm may have more available starch for animal nutrition, and possibly have improved wet-milling and dry-grind ethanol processing properties. The new modified endosperm selections produce a range that approaches the density of the vitreous parent, and have lysine content and protein digestibility comparable to the HPD/hl opaque mutant parent. Endosperm texture is an important quality parameter in grain crops. This paper shows clearly, that it is not only the quantitavels enhanced protein content, but also the structure within the cells having demonstrable influence on the digestability of proteins.. Superior nutritional quality depends on a multitude of factors, including environmental and genetic ones. Figure

234

Fig. 117 Fig 2" Light micrographs of thin section of protein-stained developing sorghum grain (30 days after half- bloom/pollination): Modified HPD/hl mutant endosperm (A); normal vitreous endosperm (B). Note the lack of protein (dark- stained lines) surrounding starch granules in the section of the modified mutant endosperm with high starch packing (white arrows) contrary to the complete protein matrix surrounding starch granules in the normal vitreous endosperm (black arrows).

235

(Dicko et al., 2006) state in a joint project between the university of Ouagadougou and the agricultural research center of Wageningen that Sorghum is a staple food grain in many semi-arid and tropic areas of the world, notably in Sub-Saharan Africa because of its good adaptation to hard environments and its good yield of production. Among important biochemical components for Sorghum processing are levels of starch (amylose and amylopectin) and starch depolymerizing enzymes. Current research focuses on identifying varieties meeting specific agricultural and food requirements from the great biodiversity of Sorghums to insure food security. Results show that some Sorghums are rich sources of micronutrients (minerals and vitamins) and macronutrients (carbohydrates, proteins and fat). Sorghum has a resistant starch, which makes it interesting for obese and diabetic people. In addition, Sorghum may be an alternative food for people who are allergic to gluten. Malts of some Sorghum varieties display alpha- amylase and beta-amylase activities comparable to those of barley, making them useful for various agro- industrial foods. The feature of Sorghum as a food in developing as well as in developed countries is discussed. A particular emphasis is made on the impact of starch and starch degrading enzymes in the use of Sorghum for some African foods, e.g. "to", thin porridges for infants, granulated foods "couscous", local beer "dolo", as well agro-industrial foods such as lager beer and bread. (see also (Slingerland et al., 2006) commented above.

An important effort is now made with the biofortification of Sorghum cultivars within the framework of a Africa Harvest research and development project financed mainly by the Bill and Melinda Gates Project (www.SuperSorghum.org), again with the collaboration of ICRISAT. It is the only Gates funded project which is based nearly entirely in the developing countries. One of the main targets is to heighten the level of lysine in the grain, so as to reach the normal levels encountered in cultivated maize. The bibliography of lysine related scientific papers counts over 170 numbers, the majority of them targeting improved Sorghum for feed. http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Lysine-WOS-20060804.pdf

Many other high-lysine feed studies have been conducted, here just three examples:

(Lekule & Kyvsgaard, 2003) comment on pig feeding in small farms in Africa, where some nutrient components should be enhanced: a diet consisting of e.g. maize or Sorghum, which are some of the feeds available on the small farms, will only provide approximately 30% of the pigs requirements of lysine and methionine, which are the most limiting amino-acids in pig feeds.

(Kumar et al., 2005) published a study on the response of broiler chicks fed high tannin Sorghum diets supplemented with DL-methionine. It was concluded that red Sorghum could replace maize in toto in color broiler chicken diet but economically it can be used only up to 33% in diet replacing 50% of maize. Addition of methionine did not prove beneficial.

(Imik et al., 2006) report on effects of additives on laying performance, metabolic profile, and egg quality of hens fed a high level of Sorghum (Sorghum vulgare) during the peak laying period, including data on high lysine supplements.

236

In a latest review several authors gave insight in Sorghum improvement, mainly concentrating on grain quality and certain resistance aspects, the reviewed literature complementing the one given in this chapter. (Belton et al., 2006; Chandrashekar & Satyanarayana, 2006; Dykes & Rooney, 2006; O'Kennedy et al., 2006; Salinas et al., 2006; Taylor & Shewry, 2006; Taylor et al., 2006)

6.4.2. Sorghum breeding for biofuel production

According to (Zhan et al., 2006) Sorghum (Sorghum bicolor (L.) Moench) is a starch-rich grain crop similar to maize (Zea mays L.), it could be transformed into an ideal biofuel cropl. It is true that Sorghum has been underutilized for biobased products and bioenergy. This study is designed to investigate the effects of supercritical-fluid-extrusion (SCFX) of Sorghum on ethanol production. Morphology, chemical composition, and thermal properties of extruded Sorghum are characterized. Analysis of extruded Sorghum showed increased measurable starch content, free sugar content, and high levels of gelatinized starch. SCFX cooked and non-extruded Sorghum is further liquified, saccharified, and fermented to ethanol by using Saccharomyces cervisiae. The ethanol yield increased as Sorghum concentration increased from 20 to 40% for both extruded and non-extruded Sorghum. Ethanol yields from SCFX cooked Sorghum were significantly greater than that from non- extruded Sorghum (>5%).

237

Fig. 118 Microstructures of Sorghum: (A) non-extruded Sorghum and (B) supercritical-fluid-extrusion cooked Sorghum after (Zhan et al., 2006)

The extrusion process is an effective pretreatment method of Sorghum for fermentation. Chemical changes (such as thermal degradation, depolymerization of starch, dietary fiber and proteins, and recombination of depolymerized fragments), and physicochemical changes (such as destruction of native starch and protein structures) could occur during the extrusion process (Camire, 1998).

6.4.3. Future Breeding Opportunities in general There is still a lot of work to be done, in all domains of research and development, the map on Sorghum yields shows this clearly on a world wide scale. Yield gain has been achieved in many developed countries on a large scale, in Africa only in Egypt. In all other African countries there is still a lot of work to be done. There are more than 350 papers dealing with molecular analysis of the Sorghum genome related to agronomic questions, this should first be taken into account in a thorough review paper (see the bibliographic account on agronomic papers of the genus Sorghum).

238

Future breeding opportunities are described in an extensive review from (Morgan et al., 2002) in Crop Science and by (Zeller, 2000) in the Bodenkultur, in German language. Population trends predict increasing food needs while progress in developmental and genomic plant sciences offer new opportunities for crop improvements. The grass-based grain crops have a high degree of synteny in their genomes, thus raising the possibility of using specific information more broadly. Studies of S. bicolor (L.) Moench are beginning to link physiological behavior to specific genes and hormone-based regulatory systems in ways that suggest specific strategies for improvement. Findings from several other grasses are adding to the pool of information being derived from Sorghum. This information relates to flowering and floral development, maturity and senescence, temperature effects via the biological clock, shade avoidance behavior, apical dominance, shoot elongation, and root development including constitutive aerenchyma formation. These studies, along with others, offer a number of options for conventional plant breeding and genetic transformations to improve grass-based crops and satisfy part of the projected human food needs of coming decades.

It is clear to the reviewer that genomics will be of great importance in the process of developing better traits of Sorghum: (Bennetzen & Freeling, 1993) already made it clear that grass genomics could greatly help here, according to the authors it is less a matter of scientific insight, more a problem of organization: Taken all the present day knowledge together, it is very helpful to know the considerable complementarity of cereal genome studies, they should with time result in a program of grass research that is free from the limitations of any single system. As a group, the 25’000 species of grasses have virtually every characteristic that one could wish for in a model organism: excellent cytogenetics (wheat and maize), unequalled mutant collections (those for barley, maize and wheat are especially impressive), well-developed transposon mutagenesis and tagging (maize), small genome size (rice and Sorghum), acceptable transformation technologies, and a plentiful supply of mapped DNA markers (as a combined figure, from all the cereals, in the number of thousands). Important is the diversity of form and behavior represented in grass species, see also (Tyagi et al., 1999) for a review of the situation in rice.

Instead of summarizing all the many breeding and selection attempts for more resistant Sorghum traits, here just a few examples. (An extensive review should be done for the second phase of the SuperSorghum project). An example on selective breeding for Striga resistance from (Rich et al., 2004) is given. It demonstrates concretely the potential of Sorghum landraces as precious resources of useful resistant genes. Witch weeds (Striga spp.) are noxious parasitic weeds that cause considerable crop damage in the semiarid tropics. Genetic control of Striga is effective, although sources of resistance are limited in most crops. Useful resistance sources have been obtained in Sorghum bicolor, an important host crop that has coevolved with the parasite. Fifty-five wild accessions within the primary gene pool of Sorghum and 20 Sorghum cultivars were screened for resistance to Striga asiatica L. Kuntze in the laboratory. Wild Sorghums assayed included S. almum Parodi, S. bicolor subsp. drummondii (Steud.) De Wet, (race drummondii and race hewisonni), S. bicolor subsp. verticilliflorum (Steud.) Piper with races aethiopicum, arundinaceum, verticilliflorum, and virgatum; S. halepense (L.) Pers.; S. miliaceum; S. rhizomatores; S. sorghastrum; and S. usambarense. Wild Sorghum accessions varied in their effects on S. asiatica at the pre-attachment level of association. Potential striga-resistance mechanisms of low germination stimulant production, germination inhibition, and low haustorial initiation activity were observed in this collection 239 of Sorghums. Some of these potential striga-resistance mechanisms, reported by (Rich et al., 2004) first time, appear to be unique to wild Sorghums. The results described in this study offer the possibility of introgressing valuable resistance genes from wild to cultivated Sorghum. There is no space here to give an extensive account on Sorghum breeding related to disease and pest resistance. (Leslie, 2003) give a more recent account on breeding activities related to Sorghum diseases and pests.

In another field of breeding, distant intergeneric hybridization, the interest has recently grown: Of late, there has been renewed interest in the utilization of related genera in sugarcane varietal improvement as some of them are considered potential sources for important characters in modern Sugar cane breeding like earliness from Sorghum (Nair et al., 2006). Intergeneric crosses are difficult to perform and identification of genuine hybrids among the progeny is truly difficult. With the help of RAPD markers one could precisely identify the true hybrids of Saccharum officinarum x S. bicolor. Although the study has been undertaken for the benefit of modern sugar cane, there is breeding potential also for Sorghum.

(Price et al., 2005b) published a breakthrough study in obtaining a viable hybrid between S. bicolor and Sorghum macrospermum, the first one between S. bicolor and a taxon outside Eu-Sorghum. Although exotic germplasm is extensively used in Sorghum improvement programs, Sorghum species classified in sections other than Eu- Sorghum have not been utilized as germplasm because of strong reproductive barriers involving pollen - pistil incompatibilities. S. macrospermum is of particular interest to Sorghum breeders because of its close phylogenetic relationship and cytogenetic similarities to S. bicolor and its resistance to important Sorghum pests and pathogens, such as Sorghum midge and Sorghum downy mildew. They produced a vegetatively vigorous interspecific hybrid from a cross between a cytoplasmic male- sterile S. bicolor plant and S. macrospermum by using embryo rescue and in vitro culture techniques.

6.5. Bibliography Sorghum breeding http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Breeding-20061016.pdf

6.6. Summary Sorghum breeding It is not intended in this report to give a full account on Sorghum breeding, a short chapter must be enough here, the remarks are mainly related to the biology and Evolution of Sorghum, a simple screening of the web of science reveals over 500 publications, and many from 2000 onwards, on the combination of two keywords: Sorghum and breeding – so – after all, Sorghum research has grown into a major area targeting better crops.

Depending on the region of production, the type of Sorghum and the purpose for its production varies widely. (Rooney, 2004). Whether they are breeding varieties or hybrids, the primary focus of Sorghum breeders throughout the world are yield, adaptation and quality. In addition to breeding for these factors, reducing losses due to stress is equally important. Most breeding programs consistently select for tolerance to abiotic stresses (such as drought and low temperatures) and biotic stresses (such as Sorghum midge, grain mold, anthracnose, and charcoal rot). Finally, the integration of molecular genetic technology is enhancing Sorghum improvement by providing a genetic basis for many important traits and through marker-assisted selection. Sorghum improvement in the future will require effective utilization of all the available tools in order to develop Sorghum genotypes suitable for the needs of their producers and end-users. 240

Species outside the Eu-Sorghum section are sources of important genes for Sorghum improvement, including those for insect and disease resistance, but these have not been used because of the failure of these species to cross with Sorghum. An understanding of the biological nature of the incompatibility system(s) that prevent hybridization and/or seed development is necessary for the successful hybridization and introgression between Sorghum and divergent Sorghum species. (Hodnett et al., 2005) determined the reason(s) for reproductive isolation between Sorghum species. The study utilized 14 alien Sorghum species and established that pollen-pistil incompatibilities are the primary reasons that hybrids with Sorghum are not obtained. The alien pollen tubes showed major inhibition of growth in Sorghum pistils and seldom grew beyond the stigma. Pollen tubes of only three species grew into the ovary of Sorghum. Fertilization and subsequent embryo development were not common. Seeds with developing embryos aborted before maturation, apparently because of breakdown of the endosperm. (Morgan et al., 2002) and also (Zeller, 2000) in German gave an account with lots of important citations on the lessons of breeding activity for Sorghum in recent years. The grass-based grain crops have a high degree of synteny in their genomes, thus raising the possibility of using specific information more broadly. Studies of S. bicolor (L.) Moench are beginning to link physiological behavior to specific genes and hormone-based regulatory systems in ways that suggest specific strategies for improvement. Findings from several other grasses are adding to the pool of information being derived from Sorghum. This information relates to flowering and floral development, maturity and senescence, temperature effects via the biological clock, shade avoidance behavior, apical dominance, shoot elongation, and root development including constitutive aerenchyma formation. These studies, along with others, offer a number of options for conventional plant breeding and genetic transformations to improve grass-based crops and satisfy part of the projected human food needs of coming decades.

The global economic importance of Sorghum makes it a prime candidate for genetic transformation. However, genetically modified Sorghum has not been commercially produced to date, see for more details chapter 2.2.6.

The various breeding programs A summary of the modern breeding situation in Sorghum is also given by (Bantilan et al., 2004): Encouraged by the results from natural hybridization and selection, attempts at deliberate crossing were made, wherein it was found that crosses between divergent cultivars exhibited high levels of heterosis (Conner & Karper, 1927). The discovery of the cytoplasmic genetic male sterility by (Stephens & Holland, 1954) based on the milo-kafir system was a milestone in Sorghum breeding and research. It is widely used in the commercial exploitation of heterosis. Today, more than 30% of the Sorghum area is under hybrids (in the USA!), which have yields about twice that of any local cultivar. Non-mila sources of cytoplasmic male sterility have also been identified (Schertz & Johnson, 1984; Schertz & Ritchey, 1978). However, their commercialization has been hindered by the nonavailability of sufficiently stable restorer lines. (Moran & Rooney, 2003; Tsvetova & Elkonin, 2002). The availability of several single, recessive male sterile genes (ms3, ms7) serving as a genetic male sterility system allows population improvement through recurrent selection procedures involving random mating and selection, see (Mgonja et al., 2006).

Biofortification is now belonging to several breeding programs: (Slingerland et al., 2006) The preliminary results suggest that a feasible chain solution consists of breeding for high Fe and moderate phytic acid contents and using soil organic amendments and P fertilization to increase yields but that this needs to be followed by improved food processing to remove phytic acid. Further research on timing of application of phosphate, Fe fertilizer and soil organic amendments is needed to improve phytic acid-Fe 241 molar ratios in the grain. Research on the exact distribution of Fe, phosphate, phytic acid and tannins within the Sorghum grain is needed to enable the development of more effective combinations of food processing methods aiming for more favorable phytic acid-Fe molar ratios in Sorghum-based food.

(Dicko et al., 2006) State in a joint project between the university of Ouagadougou and the agricultural research center of Wageningen that Sorghum is a staple food grain in many semi-arid and tropic areas of the world, notably in Sub-Saharan Africa because of its good adaptation to hard environments and its good yield of production. Among important biochemical components for Sorghum processing are levels of starch (amylose and amylopectin) and starch depolymerizing enzymes. Current research focus on identifying varieties meeting specific agricultural and food requirements from the great biodiversity of Sorghums to insure food security.

An important effort is now made with the biofortification of Sorghum cultivars within the framework of a Africa Harvest research and development project financed mainly by the Bill and Melinda Gates Project (www.SuperSorghum.org), again with the collaboration of ICRISAT.

There is still a lot of work to be done, in all domains of research and development, the map on Sorghum yields (see fig. from (Bantilan et al., 2004) in this chapter shows this clearly on a world wide scale. Yield gain has been achieved in many developed countries on a large scale, in Africa only in Egypt. In all other African countries there is still a lot of work to be done.

242

7. Evolution and gene flow of Sorghum

Fig. 119 Bantilan, M., Deb, U., Gowda, C., Reddy, B., Obilana, A., & Evenson, R., eds. (2004) Sorghum Genetic Enhancement: Research Process, Dissemination and Impacts, pp 312, ICRISAT,Patancheru, Andhra Pradesh, India, plate p. 224, frontispiece to chapter ‘Impact of Improved Sorghum Cultivars on Genetic Diversity and Yield Stability’. 243

7.1. Evolutionary dynamics and history of Sorghum domestication

7.1.1. Crop – Wild complex, ferality, the basics, related to Sorghum

Generally, the situation of interfertile crop-weed relationships can be described as geographically and historically complex systems with feral populations as described by (Gladis, 1966; Gressel, 2005a; Hammer et al., 2003),

Fig. 120 Plants under human influence: The situation is very complex, if the time axis is fully followed as has been shown by (Gladis, 1966; Hammer et al., 2003), this is demonstrated in a single graph with all details: Feral crops are (or can be) of multiple origin. It becomes clear that ferality is of multiple origins. The pathways between crop plants, weeds and wild plants are manifold, and for sure in many cases it is almost impossible to trace the origin of ferality or to trace back the original and wild species from which crops have been derived. More about crop ferality and voluntarism is summarized by many authors in a volume edited by (Gressel, 2005a).

244

Fig. 121 The original gene pool concept, established by Harlan and de Wet and modified by (Hammer et al., 2003). GP 1 The biological species, including wild, weedy and cultivated races. GP 2 All species that can be crossed with GP 1, with some fertility in individuals of the F1 generation; gene transfer is possible but may be difficult. GP 3 Hybrids with GP 1 do not occur in nature; they are anomalous, lethal, or completely sterile; gene transfer is not possible without applying radical techniques. GP 4 Any synthetic strains with nucleic acid, i.e., DNA or RNA, frequencies that do not occur in nature. 245

Fig. 122 (Ejeta & Grenier, 2005): The adaptation to the gene pool of Sorghum based on classification of a 'biological species' as per (Harlan & De Wet, 1971, 1972) . S. bicolor and propinquum: the biological species

According to (Harlan & De Wet, 1971), S. bicolor and S. propinquum are found in the primary gene pool (GP-1) of Sorghum, which contains biological species of both the cultivated and spontaneous (wild and/or weedy) races. The species in this primary gene pool intercross readily and produce fertile hybrids. S. halepense is found in the secondary gene pool (GP-2) which includes all species that can be crossed with GP-1 with at least some fertility in the hybrids. This suggests that gene transfer between these two gene pools is possible, though it may be difficult in some situations. The other sections of the Sorghum taxa are encompassed in the tertiary gene pool (GP-3) where hybrids with GP-1 are anomalous, lethal or nearly completely sterile, and gene transfer is not readily possible or requires radical techniques (Hadley, 1953) suggests that Johnsongrass arose as a cross between 20-chromosome (2n) species whose chromosome complements were similar but not identical. This explains the fact that hybrids between Sorghum cultivars and Johnsongrass are only partially fertile, resulting in individuals with 30 and 40 chromosomes, some actually producing some viable seeds. Crossing experiments of (Sangduen & Hanna, 1984) show different results, which might be due to differences in populations used for the experiments: Crossabilities ranged from 71% to 83% when S. halepense was used as a female parent, but only ranged from 0 to 30% in case of S. bicolor used as a female parent. Interspecific hybrids were vigorous, leafy and resembled more closely S. halepense. (Sangduen & Hanna, 1984) suggest to study in more detail results of hybridizing experiments so that wild Sorghum species could become a valuable source for modern Sorghum breeding.

246

Fig. 123 Example of an organismoid or a hypothetically designed crop with a genome composed of different gene pools and synthetic genes, for the explanation of this complicated matter, see (Gladis & Hammer, 2000). As complex as this situation is depicted, it might well come close to reality also with specific cases of Sorghum.

7.1.2. Wild and weedy Sorghums in Africa There are four notable wild races which are probably derived from S. bicolor subsp. verticilliflorum including S. arundinaceum, the progenitor of many of the cultivated Sorghum races. (Aldrich & Doebley, 1992; Aldrich et al., 1992; Harlan & De Wet, 1971) and many others. This is also confirmed by the experience of (Ejeta & Grenier, 2005), they noted (with remarks added from other publications)

Race aethiopicum formed part of the grass vegetation across the drier parts of the west African savannah, extending from western Ethiopia to Mauritania. This race can be locally very abundant appearing in large tracts of land and occasionally invades cultivated fields.

247

Race arundinaceum is a forest grass widely distributed in moist tropical forests of the Guinea coast and the Congo. It appears as a common grass along stream banks and forest paths, and occasionally establishes in cultivated fields.

Race verticilliflorum (including arundinaceum) is a common grass across most of the African savannah, between Sudan and Nigeria, found in weedy patches along roadsides, and often found established in cultivated fields. It is the most widely distributed wild Sorghum, having been introduced to tropical Australia, parts of India and the New World (De Wet, 1978). This race grades into race arundinaceum which is distributed across tropical Africa as a forest grass. In the broad leaf savanna race arundinaceum is often difficult to distinguish from race verticilliflorum except by inflorescences in which the branches become pendulous at maturity. Along with race aethiopicum, it is found most everywhere Sorghum is grown in Sudan.

Race virgatum is a desert grass distributed along irrigation ditches and stream banks in central Sudan and extends along the length of the Nile northwards to Cairo. Distribution map in the chapter on the biogeography of wild Sorghum species fig. 62 from (De Wet et al., 1970).

Weedy Sorghums exist parallel to the wild Sorghums, as either the perennial rhizomatous forms derived from Sorghum propinquum or the annual grassy weeds that resulted from hybridization between cultivated and wild Sorghums within the S. bicolor species.

7.1.3. Call for more hybridizing experiments and morphological analysis There needs still a lot of work to be done on Sorghum hybridization, especially also on the morphological level. A good example is the difficulty when one should distinguish Johnsongrass and Sudan grass morphologically: Good descriptions of the morphological differences between Johnson grass and Sudan grass are published by (Long, 1930): But it shows that there are no major morphological differences between the spikelets of Johnson grass and Sudan grass, except maybe the indumentum and size of the sterile flower pair, which remains to be studied. See the fig below:

248

Fig. 124 Spikelets and flowers of Johnson and Sudan grass. 1: terminal group of Sudan grass having one sessile and two pedicelled spikelets. 2: same for Johnson grass. 3: pair of spikelets of Johnson grass. 4: flower of sessile spikelet of Johnson grass with abaxial lodicules (lod). 5: staminate flower of pedicelled spikelet showing aborted pistil (p). (Long, 1930)

249

7.1.4. Evolutionary dynamics on the genomic level

The cereals differ considerably in haploid nuclear genome size of most genera in Poaceace, varying from _450 Mb in rice to _5,000 Mb in barley (Bennett & Leitch, 1998). Comparative genetic maps of Sorghum and maize (Hulbert et al., 1990), and of more distantly related cereals, revealed significant genome conservation among the cereals despite their differences in nuclear DNA content. The discovery of conserved gene content and order (i.e., colinearity) in grasses has opened new frontiers for gene discovery, gene isolation, and the characterization of gene function in an evolutionary context (Bennetzen & Freeling, 1993; Berhan et al., 1993)..

Fig. 125 Schematic diagram of the simplest evolutionary model for events that took place in this orthologous region of the rice, Sorghum, and maize genomes after their divergence from a common ancestor. Arrows indicate genes, whereas gray bars in the two maize segments show truncated gene fragments. The genetic composition of the common ancestor of the three species is proposed, with 11 genes in this region, and is structurally very similar to the rice region. (Ilic et al., 2003) (Ilic et al., 2003) reveal a complex history of the rice, maize and Sorghum genomes: Extensive sequence conservation between rice and Sorghum allows prediction of an ancestral grass genome segment. Analysis of gene content and order in orthologous rice and Sorghum regions, with 11 genes shared between the two, revealed a high degree of sequence conservation. These studies shed a new light on local sequence evolution in the maize, Sorghum, and rice genomes, revealing very different evolutionary dynamics in this region in these three lineages. Additional investigations are needed of a larger number of orthologous gene segments within maize. Particular emphasis should be placed on functional analyses of the orthologous genes, with the goal of relating changes in evolved genome structure to conserved or altered functions of orthologous loci in maize, Sorghum, and rice. For more comments and citations see the review of (Rabinowicz & Bennetzen, 2006).

(Peng et al., 1999) A restriction fragment length polymorphism (RFLP) linkage map of S. bicolor (L.) Moench was constructed in a population of 137 F6-8 recombinant inbred lines using Sorghum, maize, oat, barley and rice DNA clones. The distribution of 250 these loci does not provide support for the hypothesis that S. bicolor (L.) Moench is of tetraploid origin. Comparison of the map with RFLP maps of maize, rice, and oat produced evidence for Sorghum-maize LG rearrangements and homoeologies not reported previously, including evidence that: (1) a segment of maize 5L and a segment of 5S may be homoeologous to Sorghum.

(Aldrich & Doebley, 1992; Aldrich et al., 1992; Cui et al., 1995) give a valid summary of the evolution of Sorghum as a crop: Their own RFLP data support previous concepts of Sorghum evolution; namely, that multiple origins, diverse environments and human involvement have contributed to the existence of different types of wild and cultivated Sorghum. Outcrossing has led to gene introgression and gene flow among the natural populations. Then polymorphic subpopulations develop, and disruptive selection starts. Intermediate types may exist for a period, but differentiation continues until a number of distinct, separate and adaptive populations are formed. In summary, the population structure of modern Sorghums seems to fit well into Wright's "shifting balance" theory of adaptation, which assumes that genetic drift and selection operating on subpopulations leads to a number of genotypes occupying different adaptive peaks, even though gene flow can occur between the subpopulations (Wright, 1931, 1965). Wright's theory has been widely accepted to explain plant evolution and speciation (Hartl & Clarc, 1989), including applications to the evolution of Sorghum by disruptive selection, as proposed by (Doggett, 1988; Doggett & Majisu, 1968): Cultivated Sorghums were developed from a wild Sorghum which was itself the progenitor of the wild diploid Sorghums (Sorghum verticilliflorum including arundinacea).

For the PAUP dendrogram constructed by (Cui et al., 1995) see the figure in chapter on molecular taxonomic analysis in the genus Sorghum, there are more details presented there about the general view in evolution on the genus Sorghum.

Another view on the evolution of cultivated Sorghums is commented in (Price et al., 2005a) Molecular cytogenetic evidence strongly supports a polyploid origin of sorghum. (Gomez et al., 1998; Gomez et al., 1997) and (Zwick et al., 2000) detected by fluorescent in situ hybridization (FISH) a <280-bp tandemly repeated DNA sequence, CEN38, that differentially resided around the centromere of ten of the 20 chromosomes of S. bicolor. (Gomez et al., 1998) proposed that the ten chromosomes displaying strong FISH signals composed one subgenome of a tetraploid, whereas the other ten chromosomes with little or no FISH signal represented another subgenome.

The most conclusive study, fitting well to the views of Price and the cited authors above, has been conducted by (Hoangtang et al., 1991): They propose to dismiss two of the evolutionary origin hypothesis previously supported by various authors, namely that there is (1) an ancestral species within Parasorghum existing (Sorghum versicolor, n=5), or that (2) a hypothetical hybrid with Sorghum versicolor gave rise by doubling the chromosome number to an ancestral Sorghum bicolor. Both versions are dismissed by the authors because of respective incompatibility problems to be highly improbable. A third hypothesis is supported: 2 Species in the two sections Para-Sorghum and Sorghum represent two separate evolutionary lines of the genus, sharing only a common ancestor in a distant past. Previous reports do not substantiate the first two hypothesis because there is no logical explanation for the significant differences in chromosome sizes (Celarier, 1958; Garber, 1950; Gu et al., 1984). Moreover, the genome size of S. versicolor is about 2.5 times larger than that of S. bicolor (Laurie & Bennett, 1985). Although the second hypothesis is possible, it is unlikely because it requires at least two independent processes (i.e., a diminution in chromosome size and an increase in chromosome number) to occur simultaneously or consecutively. The molecular studies presented by (Hoangtang et al., 1991) favor the third hypothesis. 251

7.1.5. Genetic relationships of landraces of Sorghum (Aldrich & Doebley, 1992; Aldrich et al., 1992) conclude from extensive data on allozyme and restriction enzyme patterns of numerous the level of S. bicolor and its wild relatives about the ancestry of the cultivated Sorghums. Their data support the theory that cultivated Sorghum was derived from S. bicolor subsp. arundinacea for the following reasons: cultivated Sorghum and some forms of subsp. arundinaceum are closely related on the basis of nuclear and chloroplast restriction sites, and the cultivars generally contain a subset of the alleles found in subsp. arundinaceum. Studies of wild Sorghum indicate that the wild races are not well-differentiated genetically from one another, although geographic sections of their range were found to be distinct, Genetic relationships among individual accessions of wild and cultivated Sorghum were estimated by principal component analysis and average linkage cluster analysis. The wild and cultivated Sorghum are separated along the first axis (16% of the variation) of the principal component plot. Most wild collections also fail to cluster with the cultivars in the dendrogram. This indicates that subsp. arundinaceum and subsp. bicolor represent fairly distinct germ plasm pools, as previously shown with isozyme analysis. The portion of the wild gene pool that is genetically most alike the cultivars is from Northeast and Central Africa. Wild Sorghum races from these regions share a high degree of similarity with the cultivars in both their nuclear and chloroplast genomes, and these regions probably represent the area of primary domestication. In contrast, the wild Sorghum races of Northwest and Southern Africa possess nuclear and chloroplast genotypes that show less similarity to the cultivars. Introgression and long-distance seed dispersal also appear to have been factors influencing the distribution of diversity in Sorghum’s primary gene pool. Nuclear and chloroplast markers indicate conflicting genetic relationships between some wild and cultivated collections, suggesting that introgression has occurred between wild and cultivated Sorghum.

252

Fig. 126 Graph of the first two components from a principal component analysis based on RFLP allele frequency data from 56 wild and cultivated Sorghum accessions. Countries of origin are abbreviated as follows: CH China, EG Egypt, ET Ethiopia, IN India, IC Ivory Coast, ICE Kenya, NI Nigeria, SA South Africa, SE Senegal, SU Sudan, UG Uganda. Fig. 1 from (Aldrich & Doebley, 1992)

253

Fig. 127 Average linkage cluster analysis based on RFLP allele frequency data from 56 wild and cultivated Sorghum accessions using modified Rogers' distance (Wright 1978) Furthermore, a lack of correlation between genetic and geographic distances in the cultivated gene pool points to evolutionary processes affecting the spatial distribution of genetic diversity in the cultivated gene pool that may be less prevalent in the wild gene pool where genetic and geographic distances were 254 found to be correlated. Causal factors may include long-distance seed dispersal through agricultural trade and a short evolutionary history of the crop allowing insufficient time for differentiation of its populations.

7.1.6. Summary Domesticated grain Sorghums originated according to earlier authors from wild members of S. bicolor subsp. verticilliflorum (including S. bicolor subsp. arundinaceum) that were brought into cultivation. (Snowden, 1936), (Porteres, 1962) and (De Wet & Huckabay, 1967) suggested that races arundinaceum, aethiopicum and verticilliflorum could have independently given rise under domestication to guinea Sorghums, durra Sorghums and kafir Sorghums, respectively. This view is contested in the light of new molecular and archaeological data.

(Harlan, 1971) proposed that Sorghum was first taken into cultivation along a broad band of the savanna between the Sudan and Nigeria. In this region race verticilliflorum (including arundinaceum) is an abundant wild grass and is collected as a wild cereal in times of scarcity. From this region the cultivation of Sorghum spread into tropical West Africa, the arid northeast and southeastern Africa. Selection for adaptation to a wet tropical habitat produced race guinea (De Wet et al., 1972). This migration probably occurred before 3,000 B.P. (Harlan & Stemler, 1976).

Archeological results allow to trace back Sorghum grains to much earlier times than assumed before: (Wendorf et al., 1992) Excavations at an early Holocene archaeological site in southernmost Egypt, 100 km west of Abu Simbel, have yielded hundreds of carbonized seeds of Sorghum and millets, with consistent radiocarbon dates of 8,000 years before present (BP), thus providing the earliest evidence for the use of these plants.

(Cui et al., 1995) and give a valid summary of the evolution of Sorghum as a crop: Their own RFLP data support previous concepts of Sorghum evolution; namely, that multiple origins, diverse environments and human involvement have contributed to the existence of different types of wild and cultivated Sorghum. Outcrossing has led to gene introgression and gene flow among the natural populations. Then polymorphic subpopulations develop, and disruptive selection starts. Intermediate types may exist for a period, but differentiation continues until a number of distinct, separate and adaptive populations are formed. In summary, the population structure of modern Sorghums seems to fit well into Wright's "shifting balance" theory of adaptation. The latest summary based on a thorough and global genetic analysis comes from (Deu et al., 2006), the summarizing figure given at the end of this chapter. (Gomez et al., 1998) produce data showing that cultivated Sorghums are of polyploid origin.

7.1.7. Bibliography Sorghum landraces http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Landrace-Diversity-20060804.pdf

7.2. Gene flow and hybridization of Sorghum, general remarks

Crop-to-weed hybridization in Sorghum has been supported by circumstantial evidence since (Doggett, 1988), only in recent years do we have morphological and molecular genetics data on the outcrossing 255 rate of African landraces at hand which give us a realistic picture in situ (Djè et al., 2004). The studied landraces in Morocco are predominantly autogamous, which may lead to the conclusion that outcrossing rate towards wild relatives of African Landraces is also very low. This is different from case to case as (Ayana & Bekele, 1998, 1999; Ayana & Bekele, 2000; Ayana et al., 2000a; Ayana et al., 2000b, 2001) can show for some other African regions such as the Sudan and Ethiopia. For further comments see chapter 2.4., gene flow from crop to crop.

7.2.1. A field experiment from the USA shows the difference between (high) potential outcrossing rate and actual (low) hybridization rate Date on crossability potential look alarming: (Sangduen & Hanna, 1984) conducted chromosome and fertility studies on reciprocal crosses under artificial conditions between tetraploid S. bicolor (autotetraploid induced by chemical treatment) and S. halepense. Crossability was higher (71 to 83%) when S. halepense was the female parent - compared to only 0 to 33% hybrids produced when the tetraploid S. bicolor was used as the seed parent. The study also showed that when S. halepense was used as the female parent, the average seed set on both selfed and open pollinated panicles were similar and high (from 82 to 95%). In contrast, when S. bicolor was used as the female parent, the average seed set was only 18% on selfed panicles, and as high as 74% on open pollinated panicles. This looks basically scary. But the agricultural reality is different:

(Arriola & Ellstrand, 1996) created a field experiment in order to determine incidence and rate of spontaneous hybridization between S. bicolor and Sorghum halepense, a rate which is, compared to other crops, very low, but substantial and measurable.

256

Fig. 128 shows the scheme of the field experiment performed by (Arriola & Ellstrand, 1996), the diagram is not to scale.

The results show, that the experimentally determined frequency of spontaneous hybridization itself is not alarming: The data only allow for a minimal statistical interpretation, since the values are very low and the hybrid formation was highly variable. The distribution of the hybrid offspring was highest at distances closest to the crops, at 0 meters the values were not higher than 0.35% and decreased rapidly to values close to zero with growing distance from the crop (Fig. below). Compared to hybridization rates between maize crop and its wild relatives this is negligible. All the realistic outcrossing rates are found to be very low in field experiments: The numbers found by (Arriola & Ellstrand, 1996) hardly exceed 0.1% hybridization rates. And in agricultural reality these rates will actually be even lower, since in the field experiment described above all offspring has been irrigated the same way as the cultivars, which will not be the case in production.

Fig. 129 Left: Percentage of hybrids produced among total progeny sampled. 4: mean rate at each distance class at Moreno Valley Field Station. 5: Mean rate at each distance class at South Coast Field Station. Right: relative pollen flow from the crop. 6. Mean relative pollen flow at each distance from the crop at Moreno Valley Field Station. 7: Mean relative pollen flow at each distance from the crop at South Coast Field Station. From (Arriola & Ellstrand, 1996)

Still, one should not forget about the time and competition factor looking at Sorghum halepense: It is one of the world's most noxious weeds, (Chao et al., 2005), and therefore described by (Arriola & Ellstrand, 1996) as “a paradigm for the potential risks of crop-weed hybridization”. Introduced into the southeastern United States about 200 years ago probably from Mediterranean Europe, S. halepense is a close relative of cultivated S. bicolor and is of Mediterranean origin. Hybrids between grain Sorghum 257

(2n=20) and S. halepense (2n=40) can include highly sterile 30-chromosome and relatively fertile 40- chromosome types (Hadley, 1958a).

Though not widely adopted in the Americas (but widespread in Asia), some authors also describe some shattercanes as (feral) hybrids between Sorghum crop and S. halepense (OTA, 1993). Outcrossing can also involve individuals with different ploidy levels.

Many of those weedy Sorghums are actually also a valuable source for various kinds of resistance against pests (Rich et al., 2004), see chapter on breeding.

But wild Sorghums are also unwelcome carriers of diseases and thus become detrimental to modern Sorghum cultivation. detrimental to agriculture not only in terms of direct cost due to crop losses and chemical treatments, but also because they can serve as alternate hosts for pests and viruses that harm crops. For instance Sorghum ergot caused by Claviceps africana greatly reduces quality of grain in infected fields. Ergot can be a serious disease of seed parents in hybrid Sorghum seed production fields, and the presence of ergot can, therefore, impact international seed markets. C. africana, a very specialized fungus that parasitizes only the flowers of specific grasses, survives in the conidial stage on feral Sorghums and alternate hosts such as S. halepense. Furthermore, once established, C. africana could become endemic on S. halepense thus enhancing the risk. S. halepense is also a host for Colletotrichum graminicola, the anthracnose fungus. Isolates collected from S. halepense are highly pathogenic to Sorghum (Alderman et al., 2004; Muthusubramanian et al., 2005; Pazoutova & Frederickson, 2005; Prom et al., 2005a; Prom et al., 2003; Prom et al., 2005b).

Sorghum almum is a tetraploid grass that arose from a natural cross between the tetraploid S. halepense and the diploid S. bicolor (Doggett, 1988) occurring in South America.

Another weed in Sorghum fields is also described as an allopolyploid perennial hybrid resulting from the cross between S. propinquum and S. bicolor (Ellstrand & Schierenbeck, 2000).

Potential hybridization of cultivated Sorghum (S. bicolor) with Sudangrass (S. sudanense) and its feral relatives (S. almum and S. halepense) was assessed for three congeners commonly growing in natural habitats near Sorghum fields (Arriola, 2002; Arriola & Ellstrand, 1996).

7.2.2. Fitness and many other factors influencing the outcome and persistence of outcrossing events, experience from the USA (Arriola & Ellstrand, 1997, 2002) conducted fitness studies of crop-weed hybrids of Sorghum grown under field conditions to estimate the likelihood of persistence of hybrids in the wild. From their experiments, crop-weed hybrids were determined to be as fit as either of their parent plant for sexual characters as well as for their vegetative traits. However the authors tempered their results in suggesting an evaluation of hybrid fitness in wild conditions, without added irrigation will be reduced. Fitness of hybrids obtained from crosses between grain Sorghum (S. bicolor) and S. halepense was evaluated for potential propagation of the weed-gene in subsequent generations (Hadley, 1958a). The results showed 258 that the highly sterile 30-chromosome hybrids have extremely vigorous rhizomes, and present a threat through vegetative reproduction. Conversely, the 40-chromosome hybrids, whether self-fertile or male- sterile, were found to be weakly rhizomatous and did not constitute a threat other than being a source of grassy weed seedlings.

(Morrell et al., 2005) have also addressed the long-term persistence of Sorghum genes in Johnsongrass populations. The number of cultivar-specific alleles and extensive multilocus patterns of cultivar-specific allelic composition observed at both linked and unlinked loci in the Johnsongrass populations are inconsistent with alternatives to introgression such as convergence, or joint retention of ancestral polymorphisms. Naturalized Johnsongrass populations appear to provide a conduit by which transgenes from Sorghum could become widely disseminated. The lack of association between levels of introgression and the location of QTL that differentiate crop and weedy-wild phenotypes may result in part from the mode of inheritance for many of these traits. For crop-specific traits, including the elimination of inflorescence shattering, reduced plant height, reduced tiller number, regrowth after overwintering, and reduced rhizomatousness, alleles from wild Sorghum propinquum often have dominant or additive effects; alleles from the cultivated form are typically recessive (Lin et al., 1999; Paterson et al., 1995b)

The figure below (Morrell et al., 2005) demonstrates that populations with higher levels of exposure to cultivated Sorghum (NE, TX1) contain larger numbers of cultivar-specific alleles and that those alleles occur at higher frequency than in populations with less exposure (GA, NJ, TX2). The null hypothesis that the proportion of S. halepense individuals with cultivar-specific alleles at each locus is equal in each population, was rejected at P < 0.0001. Caveat: Because S. halepense is capable of asexual reproduction by rhizomes, it is possible that the level of introgression in a population could be overestimated by sampling the same genetic individual multiple times.

259

Fig. 130 Plot of the frequency of occurrence of each cultivar-specific allele (sorted and ranked) in each Johnsongrass population vs. the proportion of individuals in each population carrying each allele. The plot demonstrates that Johnsongrass populations differ both in the proportion of cultivar-specific alleles present in the population and in allele frequency at individual loci. (Morrell et al., 2005)

(Arriola & Ellstrand, 2002): A successful transfer of fitness from cultivated Sorghum into weedy types is best illustrated in the creation of a now famous forage crop, Sudangrass (Sorghum sudanense). This forage grass resulted from the introgression of cultivated Sorghum genes into the weedy germplasm of Sorghum sudanense. It probably evolved during the millennia of genetic exchange among Sorghum species in the Sudan. It combines excellent attributes of cultivated and wild Sorghums that allow effective regeneration of seeds as well as fast crop development and re-growth after repeated cuts as a fodder. Sorghum sudan forages are the most widely commercialized forage crop in the world, perhaps next to alfalfa.

(Pedersen & Toy, 2001) state that, despite of a growing demand of S. bicolor with white seeds and tan plant color there is limited information on the overall agronomic fitness of Sorghum with these characters, A set of experiments was conducted to evaluate the combined effects of plant color and seed color on Sorghum germination, emergence, and agronomic performance, Twenty near-isogenic lines with red seed/tan plant (RT), red seed/purple plant (RP), white seed/tan plant (WT), white seed/purple plant (WP) phenotypes were tested under field and laboratory conditions, Plant color X seed color interactions were not significant. Purple plant color phenotypes had higher cold germination, higher germination after accelerated aging, and greater seedling elongation at 10 d than tan plant color phenotypes. plant color did not influence standard warm germination. No differences in standard warm 260 germination or seed vigor test results were attributable to seed color. Seedling emergence under field conditions was higher for the red seed than the white seed phenotype, Grain yield was higher for the white seed than the red seed phenotype, and higher for the purple plant color than the tan plant color phenotype, Grain test weights from purple plant color lines were higher than those from tan plant color lines. All four phenotypes included relatively high yielding lines. There was considerable overlap between WT, WP, RT, RP lines in yield and other indicators of agronomic performance leading to the conclusion that white seed and tan plant color lines with comparable performance to red seed and purple plant color lines can be selected from segregating breeding populations. Overall, the fitness differences were detectable, and useful for breeding segregation.

(Smeda et al., 2000) showed, that Johnsongrass inheritance of resistance to CHD and AOPP herbicides as a dominant, single gene trait is distinctly different from the recessive trait in wild oat (Warkentin et al., 1988). Heritability of resistance is also different than reported for resistant wild oat in Canada, as well as Italian ryegrass (Lolium multiflorum Lam.), rigid ryegrass (Lolium rigidum G.), and sterile oat (Avena sterilis L.) in Australia. In these species, resistance is conferred by a single, partially dominant nuclear gene (Barr et al., 1992; Betts et al., 1992; Devine & Shimabukuro, 1994). With resistance determined to be a dominant trait transferred by pollen, spread of resistant Johnsongrass from an infested field to surrounding areas will be rapid and difficult to control.

7.3. Gene flow from Sorghum cultivars to wild relatives in Africa

7.3.1. General remarks In Africa, there are several wild Sorghums like Sorghum verticilliflorum (see chapter on Sorghum bicolor) and other wild relatives widespread as weeds in the fields of S. bicolor cultivars, causing many problems. Indirect proof of hybridization is given by numerous genomic studies, again described in other chapters of this study (mainly in chapter on S. bicolor and gene flow from crop to crop). Since Africa is the place of origin of the cultivated Sorghum, we encounter a complex situation:

A case by case view needs to be applied, it will also include local agricultural management tradition. S. bicolor offers an excellent example of the sympatric association and interaction of a crop-wild-weed complex of a species in an agro-ecosystem. The nature of genetic interaction among forms of the taxa and the consequences of these exchanges depend not only on the power of the genes involved but also on several other associated factors. Many of the African landraces are autogamous. Prevalence of wild relatives naturally varies from region to region based on extent of inherent genetic diversity, existence of selective pressures, and the farming systems in the region.

Based primarily on knowledge of the agro-ecosystems and the history of genetic resource management and conservation, (Ejeta & Grenier, 2005) have selected Ethiopia, Sudan, and the United States to demonstrate how different consequences may arise from gene flow between wild and cultivated Sorghums.

Although the empirical evidence they describe from Africa is still in many places circumstantial and only based on numerous detailed genomic analysis, it ultimately needs to be assessed experimentally, which 261 needs to be done in the framework of the differential effects of gene flow in contrast to ecological and demographical factors and to farming practices.

7.3.2. Gene flow scenarios are different from region to region Sudan and Ethiopia are the birth places of the crop and have witnessed the evolution of wild and primitive forms of Sorghum, although empirical data from Ethiopia show that Sorghum durra is the predominant race with a rather narrow genome.

The scenario in the Sudan is different from the Americas and from Ethiopia (Ejeta & Grenier, 2005). Sudan has the largest land area in Africa, yet the population base in low, only a third of the population in Ethiopia. Per-capita holdings of arable land are higher in the Sudan than all other African countries. In northern Sudan, where human settlement has been historically very light, the genetic identity of wild Sorghums may have been further protected by their isolation from human disturbance. This is the explanation for results obtained by (Abu Assar et al., 2005): The objective of their genetic study was to estimate genetic diversity and to obtain information on the genetic relationship among 96 S. bicolor accessions from Sudan, ICRISAT, and Nebraska, USA, Genetic similarity estimates ranged from 0 to 0.91, with a mean of 0.30, together with other data indicating the strong differentiation among the Sorghum materials. In the central clay plains of the country where Sorghum farming is practiced under irrigation and in rotation with other crops, wild Sorghums have also survived as weed in cotton and wheat fields and along the irrigation ditches (Harlan, 1992a).

In both the rainfed and irrigated Sudanese agriculture, see also (Larsson, 1996), genetic exchange between Sorghum and its wild relatives has resulted in formation of two widely recognized forms of crop-wild hybrids:

Aggressive forms of weedy Sorghum bicolor have evolved that are readily identified and recognized by most farmers as feral weeds, and known under a local name of “adar”. This is a form of shattercane that is widely distributed and almost accepted as unavoidable. In spite of continual weeding and selective roguing this weedy S. bicolor has not been easy to eradicate in Sudan.

The second form of intermediate is equally feral, but appears to be more similar to cultivated Sorghum and produces grains that only slowly shatter. Continued introgression of cultivated Sorghum genes into wild forms has resulted in this hybrid form called “kerketita”. Farmers selectively harvest these types and encourage their continued existence, for they rely on them as feed and food depending on the harvest prospect: In bad years, these fast growing intermediates provide the only harvest possible, particularly for fodder.

(Ejeta & Grenier, 2005) have recently started a study in Sudan to investigate the extent of gene introgression among these types and hope to assess the differential fitness of these introgressed intermediates in contrast to the cultivated and wild progenitors. An indirect estimate of gene flow by determining geographical and altitudinal allozyme variation in Ethiopia and Eritrea and its adaptation zones is given by (Ayana & Bekele, 1998, 1999; Ayana & Bekele, 2000; Ayana et al., 2000a; Ayana et al., 262

2000b, 2001) in a series of papers. The level of gene flow was low among accessions, regions of origin and among accessions within adaptation zones, but high among adaptation zones.

Gene flow studies encounter some basic difficulties, since recognizing the nature and origin of shattercanes can be difficult and leads to errors, as stated by (Ayana & Bekele, 2000): Shattercanes (derivatives of wild Sorghum x cultivated Sorghum) are often noted as the most serious weeds on the highlands of Ethiopia, where they are known as kello or sepo (which means the fool in local languages) (Doggett, 1988, 1991; Harlan, 1975b). As described by (Harlan, 1992a), the shattercanes mimic the type of cultivated Sorghum race with which they are associated in Ethiopia. It is interesting to note here that many collectors of Sorghum germplasm confuse shattercane with wild Sorghum because of such camouflaging ability of the shattercanes (Harlan, 1984), see also the paper on weed-crop mimicry by (Barrett, 1983). There were certain accessions that the local farmers identified as kello or sepo from the materials that the authors used for this study. If farmers’ folk taxonomy was right (Teshome et al., 1997), then the confusion holds also true the other way round, confusing the shattercanes with cultivated Sorghum.

All these activities are building on previous research by (Dahlberg et al., 2004; Grenier et al., 2004). There is a lot of research activity to be registered in connection with the Sudan: http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Sudan-20060810.pdf

It should also be noted here that yield loss due to other weeds is well documented: (Tamado et al., 2002) showed in Ethiopia, that there were sites, even very low density of Parthenium hysterophorus (Parthenium weed) resulted in a high yield loss (69%). Owing to differences between sites and years, however, it was not possible to specify meaningfully the threshold densities for weeding.

7.3.3. Summary gene flow within Sorghum cultivars Gene flow from Sorghum cultivars to wild and feral species is well documented in the United States. In Africa and elsewhere experimental studies are still needed, but gene flow is well documented indirectly with numerous genetic screenings. Basically, one should distinguish between the high potential crossability and the low hybridization rate in real agricultural practice. Situation in the USA: Sorghum halepense is a proven conduit between Sorghum cultivars and the wild and feral populations, gene flow has been studied in field experiments. Although the rates are very low compared to maize as an example, there seems to be persistent establishment possible on a long term, and also seed exchange might play a role. In Africa, up to now there are only indirect hints that there is gene flow from wild Sorghums to cultivated ones, and there are only a few reports of spontaneous hybrids between African wild and cultivated Sorghums, which could be due to lack of precise scientific knowledge.

7.3.4. Bibliography on outcrossing Sorghum cultivars to wild relatives http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Outcrossing-20060810.pdf

263

7.4. Gene flow from weedy to cultivated Sorghums

7.4.1. No direct experimental proof of gene flow from weedy to cultivated Sorghums Only a few studies have been undertaken on assessing weed to crop gene flow in Sorghum, all refer to weed-to-crop gene flow only in an indirect manner. There are no field studies which aim at the direct assessment of weed to crop gene flow. Therefore, we still have insufficient information on the rate and the extent of genetic exchange in situ. It is probable, based on the potential of cross pollination and the overlap in natural habitats, that there is a continual transfer of an array of fitness and other genes from weedy types to cultivated Sorghum. (Ejeta & Grenier, 2005)

7.4.2. Indirect proof of low gene flow from wild to cultivated Sorghums Natural gene flow between cultivated and wild and weedy Sorghums in areas where they are sympatric must have led to gene exchange between the cultivated crop and wild relatives, see chapter 2.2. on gene flow from cultivated Sorghum towards its wild relatives. Several studies (Arriola & Ellstrand, 1996; Arriola & Ellstrand, 2002; Sangduen & Hanna, 1984) showed that under natural conditions, crop-to-weed gene exchange is very likely in Sorghum. Success in moving genes between crop and wild relatives depends on several factors including crossability, spontaneous hybridization, fertility, and fitness of the resultant hybrids. Potential hybridization of cultivated Sorghum (S. bicolor) with Sudangrass (S. sudanense) and its feral relatives (S. almum and S. halepense) was assessed for three congeners commonly growing in natural habitats near Sorghum fields (Arriola, 2002; Arriola & Ellstrand, 1996). However, although the potential for gene flow among this group of plants was recognized to be high, no deliberate study has been carried out, except for S. halepense, to characterize the extent of actual crossing events and nature of hybrid progenies among these weedy species under natural conditions such as non-irrigated sites.

(Aldrich et al., 1992) show in a study with the structure of allozyme patterns (and below restriction enzyme patterns), that several alleles occur in both wild and cultivated Sorghum of one region and are absent from Sorghum elsewhere, suggesting local introgression between the wild and cultivated forms. Although the same most common allele was found in the wild and cultivated gene pools at 29 of the 30 loci, phenetic analyses separated the majority of wild collections from the cultivars, indicating that the two gene pools are distinct. Wild Sorghum from northeast and central Africa exhibits greater genetic similarities to the cultivars compared to wild Sorghum of northwest or southern Africa.

This is consistent with the theory that wild Sorghum of northeast-central Africa is ancestral to domesticated Sorghum. Wild Sorghums of race arundinaceum of northwest Africa and race virgatum from Egypt are shown to be genetically distinct from both other forms of wild Sorghum and from the cultivars. Suggestions for genetic conservation are presented by (Aldrich et al., 1992) in light of these data. Comments about their conclusions on the origin of cultivated Sorghum see in chapter on the origin of cultivated Sorghums, including some of the important figures.

(Aldrich & Doebley, 1992) also studied the variation of restriction enzyme patterns besides the allozyme enzyme variation, and found in this more powerful analytic method basically the same results: Greater levels of genetic diversity were found in the wild gene pool of Sorghum than in the cultivated gene pool. 264

A total of 201 different alleles are distributed among the 53 loci examined, with an average of 3.45 alleles per locus in wild Sorghum and 2.28 in cultivated Sorghum. Genetic differences between the wild and cultivated Sorghum can be attributed to the existence of both different most common alleles predominating at a locus and low frequency alleles that are unique to one gene pool or the other. An analysis of restriction fragment variation revealed that the wild and cultivated gene pools in Sorghum share the same most common allele at 41 (77%) of the loci examined. The authors again found some evidence of genetic exchange between wild and cultivated Sorghum species.

But overall (Aldrich & Doebley, 1992) also state, that the genetic relationships between S. bicolor subsp. bicolor and subsp. arundinaceum is not a close one: Genetic relationships among individual accessions of wild and cultivated Sorghum were estimated by principal component analysis and average linkage cluster analysis. The wild and cultivated Sorghum are separated along the first axis (16% of the variation) of the principal component plot. Most wild collections also fail to cluster with the cultivars in the dendrogram. This indicates that subsp. arundinaceum and subsp. bicolor represent fairly distinct germ plasm pools, as previously shown with allozyme analysis of the same principal author.

So in conclusion gene flow between wild and cultivated Sorghums in Africa cannot be a really serious problem: over centuries obviously wild Sorghums and cultivated species maintained their identity well, although there are regional differences, as discussed in the chapter on the origin of Sorghum. They are manifested in the results given by (Aldrich & Doebley, 1992): Some geographic portions of the wild gene pool are genetically more similar to the cultivars than others. Collections of wild Sorghum from Uganda, Sudan, and the Ivory Coast exhibit the highest genetic similarity with the cultivars. Although this latter collection from the Ivory Coast is quite similar to the cultivars, the other 4 wild collections from northwest Africa (Ivory Coast and Nigeria) do not share this relationship and are isolated in the principal component plot at the end opposite the cultivars. Similarly, wild Sorghum from southern Africa is quite distinct from the cultivars. In contrast, the majority of wild collections from northeast Africa (Egypt, Ethiopia, and Sudan) and central Africa (Kenya and Uganda) are fairly similar to the cultivars.

Thus, the majority of wild Sorghum from both southern and northwest Africa share less resemblance with the cultivars of the same region than does Sorghum of central and northeast Africa. Therefore, most of the cultivated Sorghum was probably domesticated from wild progenitors of the northeast and central African regions.

(Hoangtang & Liang, 1988) studied the genomic relationship of S. bicolor and S. halepense, based on chromosomal behavior during meiosis, and concluded that Sorghum is a tetraploid species with a genomic formula AAB1B1, and Johnsongrass is a segmental auto-allo-octoploid AAAAB1B1B2B2. If correct, these observations suggest a close phylogenetic relationship between S. bicolor and S. halepense, and likely a high degree of sequence homology due also to gene flow in both directions.

The latest study is carried out by (Morrell et al., 2005) and confirm the intricate relationships within the crop – weed complex of S. halepense and S. bicolor in the United States. The figure reproduced in the chapter on outcrossing of Sorghum cultivars to wild relatives demonstrates that populations with higher levels of exposure to cultivated Sorghum contain larger numbers of cultivar-specific alleles and that 265 those alleles occur at higher frequency than in populations with less exposure. For more comments on this study see the chapter on outcrossing of Sorghum cultivars towards wild relatives, there allele exchange is documented and extensively discussed.

7.4.3. Summary of gene flow from wild to cultivated Sorghums

Only a few studies have been undertaken on assessing weed to crop gene flow in Sorghum, all refer to weed-to- crop gene flow only in an indirect manner. There are no field studies which aim at the direct assessment of weed to crop gene flow. Therefore, we still have insufficient information on the rate and the extent of genetic exchange in situ. It is probable, based on the potential of cross pollination and the overlap in natural habitats, that there is a continual transfer of an array of fitness and other genes from weedy types to cultivated Sorghum. (Morrell et al., 2005) confirm the intricate relationships within the crop – weed complex of S. halepense and S. bicolor in the United States. The figure reproduced in the chapter on outcrossing of Sorghum cultivars to wild relatives demonstrates that populations with higher levels of exposure to cultivated Sorghum contain larger numbers of cultivar-specific alleles and that those alleles occur at higher frequency than in populations with less exposure. Allele exchange has been shown in numerous cases in Africa. However, this is the result of century old neighborhood of wild and cultivated Sorghum. The very few experimental gene flow studies undertaken show that there is a considerable difference between potential gene flow and actually surviving hybrid plants under field conditions. The problem of gene flow from wild to cultivated Sorghum should, seen in the light of the presented evidence, not been over-estimated.

7.4.4. Bibliography not available on direct evidence of weed to crop gene flow in Sorghum

7.5. Gene flow in Sorghum from crop to crop There are extensive studies done to assess crop to crop gene flow in the genus Sorghum, some examples are given in this study:

7.5.1. The agricultural reality: Gene flow from Sorghum crop to crop is low, individuality of land races have been remained intact over centuries. A picture close to agricultural reality of outcrossing rates of landraces has been developed by (Djè et al., 1998; Djè et al., 1999; Djè et al., 2004), building on population genetic data analysis in the field. The authors assessed the outcrossing rate of Sorghum landraces sampled in situ from two fields under traditional cultivation in north-western Morocco using genotypic data from five microsatellite loci. Assuming a mixed mating model, they estimated outcrossing parameters by two methods that are based on progeny analyses. With both methods, the multilocus estimate of outcrossing rate for the overall sample was in the order of t=0.1 to 0.2, meaning that Sorghum landraces are predominantly autogamous in the study area, but with a significant proportion of outcrossing. (t, the outcrossing rate was computed as a function of tm of n over N (1-α), whereas n is the number of offspring for which a non-maternal allele was found in at least one locus, N represents the total number of offspring, and α is the probability of misidentification of outcrossed progeny.) 266

No relationship was detected between pair wise geographical distances and either morphological or allozyme-based distances between fields. This suggests that if Sorghum seeds are exchanged among farmers, the practice has not been restricted to neighbor farmers. In conclusion, Sorghum landraces from north-western Morocco are characterized by a high level of genetic diversity within individual fields. A field thus constitutes a valuable unit of conservation in itself and this has practical consequences in designing in situ conservation programs for Sorghum. Genetic differentiation between the two fields, as estimated from microsatellite loci was very low. This contrasts with the strong morphological differences that were observed between the landraces cultivated in the two fields. Plants from field 1 had open and loose panicles and the grains were largely covered by the glumes, two characteristics that are typical of an intermediate morphological type durra-Sorghum (Harlan & De Wet, 1972). Plants from field 2 were morphologically typical of the race durra, with very compact panicles and grains that were only partially covered by the glumes. Such discrepancy between morphological classification and differentiation based on neutral genetic markers has been pointed out by (Morden et al., 1989) who showed that genetic variation was more closely associated with geographic origin than racial classification, and this was further confirmed by previous surveys (Djè et al., 1998; Djè et al., 1999) of microsatellite variation within and among worldwide accessions belonging to different morphological races (Dje et al., 2000).

The most important conclusion of (Djè et al., 2004) was: The distribution of outcrossing rate among progeny families showed that 30% of them were completely self-fertilized but some ‘families’ showed substantial outcrossing.

Fig. 131 Distributions of the estimate of outcrossing rate among progeny families sampled from two fields of Sorghum. (Djè et al., 2004)

These results are at odds with the very low genetic differentiation observed previously among Moroccan landraces (Djè et al., 1998; Djè et al., 1999; Zongo et al., 2005) and suggest that morphological differences are maintained despite gene flow through seed exchanges among farmers. In a scatter plot of 54 collections from Morocco the morphological differentiation of landraces is maintained, despite the basic gene flow situation described, the 95% percent confidence ellipses are demonstrating a good separation in the principal component scatter plot (PC-1 against PC-2), see fig. below: 267

Fig. 132 Scatter plots of 54 individuals of Sorghum belonging to six fields on principal components 1 (PC 1) and 2 (PC 2) based of morphological variation. For each field, ninety-five percent confidence ellipses are given. From (Djè et al., 1998) A principal component analysis (PCA) was performed on the whole data set. Four principal components with eigenvalues higher than unity were obtained. Principal component 1 (PC1) accounts for 22% of total variance and expresses mainly variation in peduncle curvature, inflorescence shape, and type of endosperm. PC2 accounts for 18% of total variance and is correlated positively with awn length and negatively correlated with grain color and type of endosperm. PC3 and PC4 account for 16 and 11% of total variance, respectively, with glume hairiness and proportion of grain covering (PC3), and glume color (PC4) as the most important characters. A scatter plot of the 54 individual plants on PC1 and PC2 shows that morphological differentiation among some fields occurs.

(Cui et al., 1995) present molecular taxonomic data which demonstrate gene flow among landraces: Outcrossing has led to gene introgression and gene flow among the natural populations. Then polymorphic subpopulations develop, and disruptive selection starts. Intermediate types may exist for a period, but differentiation continues until a number of distinct, separate and adaptive populations are formed. In summary, the population structure of modern Sorghums seems to fit well into Wright's "shifting balance" theory of adaptation, which assumes that genetic drift and selection operating on subpopulations leads to a number of genotypes occupying different adaptive peaks, even though gene flow can occur between the subpopulations (Wright, 1931, 1965) Wright's theory has been widely accepted to explain plant evolution and speciation (Hartl & Clarc, 1989), including applications to the evolution of Sorghum by disruptive selection, as proposed by (Doggett, 1988; Doggett & Majisu, 1968). The PAUP dendrogram constructed by (Cui et al., 1995) can be found in chapter on molecular taxonomic analysis of the genus Sorghum. 268

(Ayana et al., 2001) sum up their results based on allozyme variation in Sorghum by stating that the two accessions studied show low level of gene flow in the regions of origin, higher in the adaptation zones. Their results are in accordance with previous studies by (Djè et al., 1998; Morden et al., 1990), also regarding the proportion of genetic variation maintained among regions.

(Zongo et al., 2005) give real time and geographically well defined data from Burkina Faso: The authors established the genetic relationship among Sahelian Sorghum, S. bicolor (L.) Moench S.L., landraces from Burkina Faso with the means of electrophoretic analysis for 10 enzymatic systems and 18 loci. Four enzymatic systems (ADH, LAP, MDH, PGD) and five loci revealed polymorphism both within and among landraces. Thirty-eight per cent of the landraces were monomorphic in all the 18 loci. The genotypic frequencies in most of the landraces deviated markedly from Hardy–Weinberg proportions due to a major heterozygote deficit, the landrace being homozygous or a mixture of homozygotes. This study showed that in this Sahelian region there is a high level of diversity between Sorghum landraces and a low rate of outcrossing between them. It showed that for the enzymatic systems studied, alleles spread randomly throughout the territory. Landraces gathered in high clusters. Three main clusters are formed by native cultivars, landraces, growing there for a long time.

Fig. 133 Geographical distribution of enzymatic groups. From (Zongo et al., 2005) fig.5 269

The study shows that most landraces are homozygotes (FIT = 0.956). Two facts can account for this: first, the low rate of outcrossing (FIS = 0.773), and second, the high level of diversification between landraces (FST = 0.807). These facts derive from the combined effect of the mating system and genetic drift. Considered together, the high value of the within-zones F and the low value of the between-zones F are largely due to predominant self-fertilization. A farmer gets seeds through mass selection in a field. Every year, the best panicles in a field are selected and used the following planting season. This farming practice, as (Harlan, 1975b) stated, produces a particular balance of selection pressure and does not prevent the local population from containing some internal variability because the best panicles are selected without considering the genetic purity of the seeds. These panicles are bulked to give the seed for the next season. Again, this is another hint at the best practice in the conservation of landraces, see chapter on biodiversity centers, example on Mexican maize.

There are even examples known in Central Niger and in a village in Cameroon where Monique Deu and her team were able to find all the 5 African Sorghum races exept kafir in sympatry without much gene flow and hybridization (personal communication).

In conclusion, the work of (Zongo et al., 2005), one of a few on large scale studies with real field data, confirms again like the work of (Djè et al., 1998; Djè et al., 1999; Djè et al., 2004; Dje et al., 2000) that we should not over-estimate the impact of gene flow between various Sorghum cultivars. This study showed that in this Sahelian region there is a high level of diversity between Sorghum landraces and a low rate of outcrossing between them. It showed that for the enzymatic systems studied, alleles spread randomly throughout the territory. It found that landraces gathered in high clusters. Three main clusters are formed by native cultivars, landraces, growing there for a long time. The five others are marginal and made of more recent introduced varieties. Even if this study gives important conclusions on Sorghum diversity in the Sahelian region, a more performant method of diversity study, using DNA investigators, must be employed for further investigation to support its results.

The large scale analysis of the core seed collection of ICRISAT also revealed data which allow to conclude for a relatively high long term stability of Sorghum landraces: (Grenier et al., 2004), building on previous work (Grenier et al., 2001a; Grenier et al., 2000a; Grenier et al., 2000b; Grenier et al., 2001b). An analysis of only nine quantitative morpho-agronomic characters revealed a high individuality of the caudatum races collected. This analysis can doubtlessly be refined by using more characters. Distribution of racial types and overall phenotypic diversity among Sorghum landraces in the Sudan is uneven. This suggests that selection and differentiation of types took place along a geographical pattern, perhaps associated with climate and use (Asante, 1995; Doggett, 1988) Thus, (Grenier et al., 2004) corroborated the results commented above through the PCA analysis that showed global phenotypic diversity in the Sudan collection as being differentially distributed among regions, with specific patterns drawn for each geographical area of origin.

270

Fig. 134 Principal component analysis on the total Sudanese landrace collection and for nine quantitative morpho-agronomic characters. Graphical representation of phenotypic diversity in the dimensional space as defined from the two first extracted factors (Varimax procedure). Separated scatter plots for each region and for the unknown source of landraces. (Grenier et al., 2004) (Teshome et al., 1999a; Teshome et al., 1999b; Tunstall et al., 2001) published on the activities of farmers maintaining their own landraces over long periods of time in Ethiopia.

(Abdi et al., 2002) also confirmed that the role of farmer management adds an important dimension to the cause of diversity. This work underlines the necessity of complementary in-situ conservation strategies for Sorghum landraces genetic resources whereby conservation effects are linked to rural development projects that emphasize the preservation of traditional farming systems by relying on and giving attention to the maintenance of biological and genetic diversity in these systems. The adaptive significance and pattern of distribution of specific traits through certain ecological conditions help to choose the site of in-situ conservation, which would be integrated with ex-situ conservation. On-farm conservation by small scale farmers not only influences the patterns of distribution of different traits but also determines the adaptive qualities of combined traits to be maintained for greater yield and dry matter composition. Distinct landraces that have been identified with combinations of traits (genotypes of agronomic and economic quality) need to be included in breeding programs. Recent data demonstrate outcrossing rates from crop to crop from an average 16% to 28%, according to Monique Deu, personal communication. The data are derived from molecular analyis calculated on the basis of factors developed by (Ritland, 2000), who developed a marker-inferred relatedness as a tool for detecting heritability in nature on text organisms other than Sorghum.

There are numerous publications, which underline the above analysis of Sorghum landraces, which maintain over centuries all breeding activities including long distance seed exchange of farmers, despite 271 a sometimes relatively high outcrossing rate - most of the publications deal with different kinds of comparative population genetic analysis:

(Abu Assar et al., 2005; Ahnert et al., 1996; Ayana & Bekele, 1998, 1999; Ayana et al., 2001; Bellon et al., 2003; Berhan et al., 1993; Beta & Corke, 2001; Beta et al., 2001; Bhat et al., 1998; Casa et al., 2005; Cui et al., 1995; deOliveira et al., 1996; Deu et al., 1994; Deu et al., 1995; Deu et al., 2006; Djè et al., 1998; Djè et al., 1999; Djè et al., 2004; Ellstrand & Foster, 1983; Folkertsma et al., 2005; Grenier et al., 2001a; Grenier et al., 2004; Grenier et al., 2000b; Grenier et al., 2001b; Guimaraes et al., 1997; Hammer et al., 2003; Hoangtang et al., 1991; Hultquist et al., 1997; Kane et al., 1992; Kenga et al., 2004; Klein et al., 2000; Lee et al., 1989; Li & Li, 1998; Manzelli et al., 2005; Munjal & Narayan, 1995; Pasquet, 2000; Peng et al., 1999; Sane et al., 1996; Smith et al., 2000; Tao et al., 1993; Teshome et al., 1997; Tuinstra et al., 1997a; Uptmoor et al., 2003; White et al., 1995; Xu et al., 1995; Xu et al., 1994; Xu et al., 2000; Yang et al., 1996; Zongo et al., 2005)

7.5.2. Short term gene flow experiment from Sorghum crop to crop under artificial conditions (Schmidt & Bothma, 2006) investigated the crop-to-crop gene flow in S. bicolor subsp. bicolor (race kafir) depending on the distance between pollen source and pollen recipient, and - on the basis of the experimental data - to estimate proper distances, or buffer zones, to avoid gene flow between Sorghum crops and/or its wild relatives. It is clear from the study, that most data are more of theoretical value and need to be verified by real time and real field experiments collecting molecular and morphological data under production conditions. We should make it clear that this study is giving answers to rather theoretical questions, which is demonstrated also by the basic scheme of the field trial: The results are of high (theoretical) interest, but do not reflect agricultural reality of Sorghum cultivation, they are of great interest for the production of certified seed in the context of Sorghum breeding in Africa.

272

Fig. 135 Plan of the Sorghum gene flow trial layout. Numbers were assigned clockwise to radiating arms, starting with the north-eastern arm. The central block with Redlan B-line (male fertile) was the pollen source, the arms with Redlan A-line (male sterile) were the pollen receptors. Each small square represents a block with Sorghum plants. The first 10 blocks closest to the central field were located 13 m, the blocks farther away were located 28 m from each other. Veld grass were located between the arms not containing wild Sorghum grasses. (Schmidt & Bothma, 2006).

(Schmidt & Bothma, 2006) have determined the maximum pollen flow under the artificial conditions of avoiding self pollination of the receiving field. These estimates and field data thus do not reflect farming reality in Africa, but they are of value for Sorghum breeders, giving maximum impact figures of gene flow in the genus Sorghum.

The average hybridization or outcrossing rate for male sterile plants was 2.54% at 13 m, below 1% at a distance of 26 m or greater, and eventually dropping to 0.06% at 158 m. Outcrossing rates are expected to be even lower for male fertile plants, which were not investigated in this study.

273

Fig. 136 Amount of seeds produced per plant and model of seed production gradient, calculated with a power fit model from three data sets: (i) all data, (ii) 95 percentile, and (iii) 99 percentile. Note the double logarithmic scale. The results were calculated based on field experiments with pollen sterile receiving fields and do not reflect agricultural reality in Africa. From (Schmidt & Bothma, 2006).

Fig. 137 Relative rate of pollen flow at male sterile plants. Values represent the amount of seeds produced 2pr, where r is equal to the distance from the pollen-donating field. These results were obtained under artificial conditions experimenting with pollen sterile receiving fields. They do not reflect agricultural reality in Africa. From (Schmidt & Bothma, 2006).

274

Fig. 138 Distance-dependent hybridization rate of pollinated Sorghum plants. The dark solid line represents the average rate and the grey lines indicate the error interval. It has to be stressed here that the results are obtained with experiments of pollen sterile receiving fields, and do not reflect agricultural reality of Sorghum outcrossing from crop to crop in Africa. From (Schmidt & Bothma, 2006).

Field experiments including receiving fields with artificially made pollen sterile crops are well known also from examples with wheat: Selfing has been eliminated (which is an almost perfect protection against fertilization with alien pollen grains, the alien pollen masses have full access to stigmas still ready to be fertilized). This results in unrealistically high pollination rates from crop to crop. (Devries, 1971, 1972; Dsouza, 1970; Waines & Hegde, 2003). Those results have promptly been abused by Anti-GMO-Activists, building on the ignorance of the public, unfortunately, we can expect the same to happen for the case of Sorghum.

7.5.3. Summary: Gene flow between different varieties of the same crop is almost as old as agriculture itself. Native plant genes as well as genes from transgenics can be dispersed either in seed or pollen, but in the case of the partially autogamous crop Sorghum the outbreeding effects remain low, the result is that landraces of Sorghum in Africa were maintaining their individual genomes over centuries. Several papers deal in all detail with the gene flow from crop to crop, and gene flow has been documented on a morphological and genetic basis, its so minimal, that it should not pose any problem with the introduction of new traits in Africa, provided some mitigation measures are taken, see chapter on consequences and prevention of gene flow.

It has to be stated very clearly, that gene flow with most Sorghum species and cultivars did always happen in the last centuries, but there are many data available to show that landraces have been remained stable, despite of free breeding and selection activity by the farmers. To erect absolute gene flow barriers would mean the genetic stabilization of the landraces and in this way ultimately lead to the end of their existence. This has been shown in the case of the Mexican landraces of maize.

275

Pollen mediated gene flow and the separation distances employed to minimize it typically generate more public interest than that for seed dispersal. For the most part, gene flow takes place within a few meters of the plant, but the up to now published data sets cannot be taken for reality: The case of gene flow with Sorghum cultivars and Johnsongrass cannot be compared to African situations, it is restricted to the Americas. (Arriola & Ellstrand, 1996), and it is a clear case of crop to wild relatives gene flow. The now often cited results of (Schmidt & Bothma, 2006) do not apply to agricultural reality: They are of a more theoretical interest, since the receiving field has been planted with pollen sterile traits, in order to get a maximum rate of outcrossing. In reality it is possible to choose traits with a very high degree of autogamy (self- pollination), a similar situation as in wheat, which has no real outcrossing problems from crop to crop. Distance from a pollen source and cross-pollination frequency with neighboring crops has been identified for major crops and employed in the production of certified seeds. In the case of Sorghum, certified seed production fields are isolated by at least 100 meters from other Sorghum fields, sufficient to maintain genetic purity of at least 99 per cent over many years of production. However, restricting gene flow between GM and non-GM varieties needs to be still better understood, and several methods already developed should be put in practice. These will efficiently help to maintain co-existence as it was practiced for many decades in modern agriculture without transgenic crops. There are a variety of methods available to mitigate the problem of gene flow, this will be dealt with in a separate chapter on the consequences and mitigation of gene flow and hybridization, and the SuperSorghum project will be able to take care of it.

To assess the environmental risk of possible future deployment of transgenic Sorghum, more data on the outward flow of genes through Sorghum pollen to nontarget relatives is needed.

Moreover, available information indicates existence of variation in outcrossing rates among varieties and locations (Cui et al., 1995; Dixon et al., 1990; Djè et al., 1998; Djè et al., 1999; Djè et al., 2004; Doggett, 1988; Pedersen et al., 1998; Zongo et al., 2005). Knowledge on actual outcrossing rates in selected and specific cases will be essential to come up with recommendations for variety maintenance and on-farm seed production.

This study on Sorghum Biology proposes to determine the extent of seed and pollen mediated gene flow in selected experiments using both molecular and morphological methods. Present day morphogenetic analysis of landraces, often intricately growing together, shows, that the individuality of the landraces is maintained despite selection and seed exchange activity of the farmers over a long time.

7.5.5. Bibliographic References on Outcrossing in Sorghum http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Outcrossing-20060810.pdf

276

7.6. Consequences of gene flow and regulation

7.6.1. General remarks: Gene flow and hybridization in agriculture,

7.6.1.1. Conventional crops It is necessary to state some basic facts in the beginning: There is a growing literature of risk assessment papers written by ecologists who forget two basic facts:

They are dealing with agricultural ecosystems, which show basic differences to natural ecosystems in several ways: Species composition in even in the best case much lower, the dynamics of change is dramatically higher (except maybe for pioneer surfaces as periodically flooded riverbanks and glacial forefields) and in most cases, crop rotation systems prevent any kind of sustainable development of species composition (Ammann et al., 2004)

Pollen and seed dispersal are not new phenomena, farmers and seed producers were confronted with these phenomena long before the transgenic era in plant breeding, and up to now they coped up very successfully with these factors: Basically, the same good old rules applied to seed and trait purity can be applied to transgenic crops and production of transgenic seeds – the only difference is that now the professionals have new tools at hand to check purity to a much higher precision. Unfortunately, they also have to cope up with false marketing arguments about produce free of transgenics – the propaganda is absolutely ludicrous in stating that there is still uncertainty about the safety of such food and feed. This is a blatant abuse of peoples fears and has no scientific basis. This said the author wants to make it clear that consumers should have a right to choose, labeling should not be contested, but if the strict labeling rules of Europe are abused for preventing fair trading rules and by denying factually freedom of choice, as it happens in Europe every day, then something is wrong. Coexistence is possible, as shown below and these coexistence arguments again should not be abused by people having fundamental believes against the modern breeding technologies.

There is a basic difference between gene flow per se (with all its important factors as pollen viability, aerodynamics of the pollen flights etc.) and hybridization rate, the actual outcome (again with a complex set of influencing factors such as fertility of the receptor plant, pollen tube compatibilities, seed set etc). Crossability determined in the lab with crossing experiments of various kinds often give much too high figures of potential hybridization. On the other hand, we should take into account the whole set of seed biology factors such as longevity in the soil, also the characteristics of farmers practice in seed storing, exchange and selection, and also in the case of industrial production large scale seed transport and intermixing possibilities.

7.6.1.2. Gene flow of transgenic crops Pollen did not learn to fly with the transgenes, it flew before. In consequence, much of the biosafety measures actually can be learned from traditional and professionally done agriculture: Seed segregation, yield segregation are as old problems to be dealt with as there is a more developed agriculture. It does not make any sense to invent the wheel from scratch again.

277

Having said this, we still have to deal with the novelty of transgenes in cases where the familiarity principle of the OECD cannot be applied (like in the case of many of the Bt crops) (Madsen et al., 2002; Naeem, 2002; Sturgis et al., 2005).

It is clear, that raising the question of the familiarity principle is a merely theoretical issue, because it is not put into place in the present day regulatory international and national biosafety frameworks – and there are still other questions to be asked in the connection with the novelty of the transgenes and its methodology of transfer:

We actually should strictly concentrate on the novelty of genes not in the sense of dealing with a new method of genetic engineering, since the most important elements of genetic engineering follow exactly the same three basic strategies of molecular change as nature does it, as has been explained in detail by (Arber, 2000, 2002, 2003, 2004). The novelty triggering biosafety research and regulation should follow with precision the question about novelty whether the gene product something really new in the context of a given crop and its biology. This, in the strict scientific sense, we should dare to raise as a question – and in the case of SuperSorghum there would be a lot of questions still open in the sense of regulatory and biosafety theory. According to the author it would be ideal to take over the regulatory system of the USA or even better the one from Canada, which is focusing on the novelty of any trait, not about the breeding methods.

But we are dealing soon with a real product in a real world, and therefore these above questions are merely theoretical and cannot be answered in the framework of the SuperSorghum project of Africa Harvest. The regulation is given in each of the African countries where the transgenic traits should be introduced and these regulatory frameworks have to be respected.

Therefore: Of course for the years to come we will have to obey to the rules of the Carthagena Biosafety Protocol (see the original texts and many scientific comments on the latest developments of the negotiations on www.pubresreg.org) and certainly also we must follow the rules of the national regulatory systems which are already in place in Africa.

8. Transgenic Sorghum There are up to date no commercialized transgenic Sorghums in production, although there are many events made ready in the Greenhouse, and also a growing number of projects described in the literature: (Bayley et al., 1992; Casas et al., 1997; Casas et al., 1993; Emani et al., 2002; Forrester, 1994; Franks et al., 2006b; Girijashankar et al., 2005; Haussmann et al., 2000b; Kristensen et al., 2005; Lo et al., 1999; Poletti & Sautter, 2005; Repellin et al., 2001; Sairam et al., 2005) As an example, (Casas et al., 1997; Casas et al., 1993) have transformed a variety with a herbicide (Ignite/Basta) resistance marker gene, but there has been no mention of this plant line in the literature since that time. In the beginning of the development, the reasons for a lack of a modified Sorghum commercialized appeared to be 278 centered on technique. (Subudhi & Nguyen, 2000) report that there has been little success in developing stable gene transfer methods for Sorghum in spite of the reports mentioned above. Up to now, no GM Sorghums have made it to commercialization. Apart from technical difficulties, there is also gene flow to Johnsongrass to be considered, a problem restricted to the Americas and perhaps also to Southern Africa (Arriola & Ellstrand, 2002), although in very small percentages – it may be even less a problem than with transgenic maize, where many field experiments have proven that coexistence can be mastered. With a minimum amount of breeding and management skills, a favorable coexistence situation can be reached also with Sorghum.

8.1. Biofortified Sorghum Whether we do have a real gene flow risk with biofortified Sorghum needs to be assessed first. It will depend considerably on the choice of the landraces, some are more recommendable than others for their inbreeding characteristics. It should be assessed in a strict baseline manner, which means that we need to take into account normal agricultural practice, where gene flow has happened for centuries already and did not cause big problems, despite a considerable genetic diversity: It should be possible to choose traits where nutritional or other values have been considerably altered by traditional breeding methods, in order to get a fair comparison line. Genetic diversity has been studied extensively, consequently it will be important to carefully choose the traits we want to finally launch as fortified traits. (Cui et al., 1995; Draye et al., 2001). For details about biofortification of Sorghum see the chapter on breeding of Sorghum.

8.2. Regulation of transgenic crops In most cases transgenes added to highly domesticated crops are not anticipated to survive in nature without human intervention – and this means in many cases also irrigation in tropical Africa. The influence of lacking human intervention on long term conditions have been shown for potato, maize and sugar beet (Crawley et al., 2001): After 10 years, the transgenic crops have vanished from the test fields. However, partially domesticated crops may constitute an exception for the dependency on humans (Brassica). On the other hand, even in the case of Brassica (Stewart et al., 2003) can demonstrate fitness disadvantages for the transgenic traits. Assessment of the survival of transgenes in nature requires to test the impact of the performance of transgenic plants under field conditions (i.e., tolerance to abiotic stresses or pest resistance, as well as the feral tendency of crops). When transgenic traits are not expected to increase plant fitness in natural habitats, such as herbicide resistance, the risk of its persistence in the wild is lessened. However, when transgenes are transferred into forage crops to confer herbicide resistance, increased weediness may occur at least for a few years by reducing the opportunities to control feral populations (Gepts, 2002; Gressel, 1978, 1984, 1985; Gressel, 1994; Gressel & Segel, 1978). See the extensive discussion of the consequences of gene flow in the chapter on consequences (and its mitigation) of gene flow.

Many domestication genes are presumed to represent a loss rather than a gain of function, as indicated by their recessiveness. An exception is found in Sorghum where alleles of S. propinquum for reduced seed size were mainly (67%) recessive to the corresponding S. bicolor alleles (Paterson et al., 1998). For a genus with a relatively high crossability rate such as Sorghum, dominant mutations may have been easily selected during the domestication process, mostly for a trait such as seed size which is under 279 strong selective pressure. In contrast to domestication genes favored in agro-ecosystems, some transgenes represent gains of function that might release wild relatives from constraints that limit their fitness, but often make crops less adapted to natural ecosystems, (Gepts & Papa, 2003; Stewart, 2005; Stewart et al., 2003). See the detailed discussion in the previous chapter on gene flow and its consequences related to fitness and potential genetic swamping.

Undoubtedly gene flow exists between cropped Sorghum and its wild relatives, and transfer of transgenes from a genetically engineered Sorghum is likely, but should not be exaggerated in its consequences. The prevalence of wild relatives in the vicinity of the transgenic crop, as would be true in centers of crop diversity, increases this likelihood. But as we know from field experiments with Sorghum, one should not mix up potential crossability percentages with the actual production of sustainable progeny with the new transgenes ( that is rate of hybrids actually produced under irrigated and/or non- irrigated field conditions).

For a gene to introgress into a wild relative it is necessary that repeated backcrosses and stabilization of the transgene occur in the new host. The likelihood of transgene introgression into a wild relative will depend on its dominance, absence of association with deleterious crop alleles or traits and the location in the genome. In addition, there are a few data sets and hypothesis that the higher the biodiversity, the less susceptible the communities are for successful and permanent introgressions, see chapter on the centers of origin. The speed in which the transgene introgresses and stabilizes in the population will depend on selective pressure and population size and the sum of mitigation measures taken. Furthermore, the nature of the allele that is introgressing and whether or not the function gained as a result of the introgession contributes to fitness and ferality are key considerations.

Important questions remain to be considered to assess the risk of transgene introgression into the wild gene pool.

These questions should ALWAYS be answered in the context of a fair baseline comparison with traditional Sorghum traits and the transgenic ones to be tested.

What is the actual magnitude of the total hybridization rate (both through pollen drift and seed dispersal and possible tillering and ratooning) in different agro-ecosystems?

What kind of interactive effects of the introduced alleles will we see on management practices?

How do transgenes behave within and outside the cultivation fields (expression, stability, persistence and effects on trophic food chains)?

8.3. Gene flow and its consequences in the Americas Wild and weedy species have a potential to be detrimental elements in traditional, but also in modern agriculture. Agro-ecosystems can be affected by non monitored and monitored spread of weeds and 280 feral populations and by gene flow in which genes from landraces and elite germplasm are moved into the wild relative of the crop and vice versa. The prevalence of such a gene exchange in Sorghum and its potential effect on increasing the evolution of feral forms has only been assessed in field studies in association with Johnsongrass. It is a well known problem, that Shattercanes can cause yield reduction in modern agriculture, this has been shown up to now in the United States, related to the formation of post emergence Shattercanes related to S. bicolor and Sorghum halepense in various other crop systems (Bradley et al., 2000; Defelice, 1990; Deines et al., 2004; Fellows & Roeth, 1992; Hans & Johnson, 2002; Johnson et al., 2000; King & Hagood, 2006; Retta et al., 1996; Retta et al., 1991; Rosalesrobles, 1993).

Progenies from an interspecific cross between Sorghum cultivars and S. propinquum were genotyped with molecular markers and several quantitative trait loci associated to weediness were discovered (Paterson et al., 1995b) and discussed in detail in the chapter on S. bicolor. In the Americas, S. halepense has introgressed with grain Sorghum to produce the widely distributed noxious weed. Derivatives of such introgression were described in Argentina as S. almum, a rhizomatous tetraploid grass that appears as a weed of Sorghum. Enhanced weediness due to genetic exchange between wild and cultivated types is also illustrated by the ubiquitous and well-defined stable intermediates, the shattercanes, particularly in Africa (Ejeta & Grenier, 2005; Harlan, 1992a).

Several cases, among which S. almum from South America was also listed, have been reported where invasive taxa have evolved after intertaxon hybridization (Ellstrand & Schierenbeck, 2000). In some cases, such as in S. almum, invasiveness could possibly be a result of the fitness benefits conferred by heterosis. In evaluating the potential risk of genetic introgression into weedy types, therefore, fitness data is critical and need to be determined.

The other consequence of recurrent gene flow from crops to wild relatives is a so called ‘demographic swamping’, which should occur according to (Ejeta & Grenier, 2005) following two different paths, which both are considered by the author of this report as a mere remote possibility and largely theoretical, in fact it is described with words evoking scary scenarios, which have no justification in the case of cultivars with a very old history of thousands of years: It should be made clear here that the concept of the so called ‘species giving origin’ to maize, rice, Sorghum or wheat is a theoretical one, it is highly improbable that those species are still existing today after all this traffic forth an back between the continents, and the dramatic habitat changes on the regions of origin of many crops.

8.3.1. Outbreeding depression of rare species of origin of crops A case of so-called “outbreeding depression” from detrimental gene flow can potentially lead to reduced fitness of a locally rare species and eventually to its extinction (Ellstrand et al., 1999). But the examples cited indirectly - and then spotted by the author of this study in other citations found in (Ellstrand & Elam, 1993) - are not very convincing: (1) A rare Juglans hindsii in California which might be endangered through Juglans regia. This example actually should rather be mentioned in the second case below (if at all). (2) (Ellstrand & Elam, 1993) cite indirectly also two European examples, Salvia pratensis and Scabiosa columbaria which are wrongly labeled as rare species and certainly not endangered in Central Europe and almost certainly not threatened by outbreeding depression, rather by habitat loss. It is therefore often (not always!) more a theory that domestication genes transferred into a weed might 281 cause reduction in fitness as they might decrease the potential for weediness and lead to maladaptation. Other cases, like wheat, cotton and rice, are very complex syngameons which need further clarification, since it will be subject to extensive genomic studies of many years to come to find out what are the wild progenitors of those widespread cultivars – if they still exist… See for more details: (Ellstrand, 2002). It is again a rather doubtful and very theoretical claim to say that the original species of rice have been genetically eroded or swamped by domesticated species, when we know about the history of domestication and migration over the continents, as described in detail by (Kush, 1997). The author of this study cannot hold back by stating that the case of genetic swamping or erosion is a very doubtful one when promoting it in connection of cultivars with a history of thousands of years, in the case of Sorghum with a highly complex history of at least 8000 years (Wendorf et al., 1992). In the numerous cases of detailed genomic studies of Sorghum landraces in Africa there has not a single case of genetic erosion or swamping been reported, although the above described situation would theoretically be possible, and although we do not yet know enough to exclude the possibility for good. We are confronted with an agricultural reality in Africa which comprises a multitude of ecological factors and elements of farmers management (from ancient times to the present) where we still do not know enough to explain why landraces can survive in such a situation of recurrent gene flow and seed exchange for centuries, which is a fact (see the chapter on evolution and domestication of Sorghum, on gene flow from crop to crop, and also the chapter on the origin of Sorghum).

8.3.2. Loss of genetic integrity of rare species of crop origin The other cause leading to potential demographic swamping is when a locally rare species loses its genetic integrity and becomes assimilated into a locally common species as a result of repeated bouts of hybridization (Ellstrand et al., 1999; Levin et al., 1996). Domesticated species have been implicated in the extinction or increased risk of extinction of wild species for two of the world’s 13 most important crops, namely rice and cotton (Ellstrand, 2002; Ellstrand et al., 1999). Again (Ellstrand, 2002) and (Ejeta & Grenier, 2005): The potential risk of extinction by hybridization depends on patterns of mating, and it is expected that the time to extinction by outbreeding depression and/or swamping will double if assortative mating is imposed over random mating more frequently. In Sorghum, there may have been cases of genetic erosion, but only as a result of habitat change. Despite the fact that cultivated Sorghum grows often adjacent to wild species, genetic swamping is unlikely and does not seem to happen. Nevertheless it will be important to start core seed collection programs all over Africa following the Precautionary Approach of the Cartaghena biosafety protocol with efficient shortcut strategies as suggested by (Mgonja et al., 2006).

Predicted genetic swamping dynamics in the case of Maize are based on blatantly unrealistic and theoretical assumptions which do not match at all with the reproduction biology in Maize (Haygood et al., 2004) as critisized heavily by (Gressel, 2005b). See the detailed discussion in chapter biodiversity and centers of origin.

It is a fact that the potential consequences of gene flow have become more severe in the Americas due to modern agricultural practices. According to many authors cited in (Ejeta & Grenier, 2005) Sorghum and Maize production in the USA has been affected by the widely recognized, competitive and noxious S. halepense. Mechanization favors perennial weed propagation as it stimulates rhizome growth. Because 282 mechanized farming is practiced, farmers in the US inadvertently selected with ploughing and modern cultivation methods for fortified rhizomes in Sorghum halepense that spread readily and enhanced the aggressiveness of the perennial weed. Aggressiveness of weeds can further be enhanced by the introgression of genes from highly improved cultivars as long as these genes confer an advantage. Continuous crossing, however low the frequency, also generates hybrid progeny that are ever more vigorous. These resultant hybrids possess additional fitness benefits conferred on to them via heterosis. (Morrell et al., 2005) have shown that crop-to-weed introgression has impacted allelic composition of Johnsongrass populations with and without recent exposure to cultivated Sorghum. Both artificial crossing and experimental field studies have demonstrated the potential for S. halepense x S. bicolor hybrid formation, but no prior study has addressed the long-term persistence of Sorghum genes in Johnsongrass populations. Lower frequencies of cultivar-specific alleles at smaller numbers of loci are found in Johnsongrass populations from New Jersey and Georgia with no recent exposure to cultivated Sorghum, suggesting that introgressed Sorghum alleles may be dispersed across long distances or can be interpreted as relics of former agricultural practices.

8.4. Gene flow and its consequences in Africa Rather than to follow the scenarios caused by the presence of the noxious weed Sorghum halepense (and Sorghum almum) in the Americas the situation in tropical Africa is a different one: Although there are no direct experimental studies on the gene flow in Africa between Sorghum cultivars and its wild relatives, there are numerous scientific publications on African landraces which give ample molecular detail as a proof that Sorghum landraces have existed for centuries and kept their individuality, mainly thanks to the breeding and selection activities of the farmers, but also because Sorghum has often a helpful reproduction biology: Its landraces are predominantly autogamous traits in many cases - which prevents massive outcrossing.

However, intermediate forms such as the Shattercanes appear also in the agro-ecosystems of Africa where Sorghum cultivation is practiced without sufficient isolation from wild forms and in sympatry with weedy relatives (Ejeta & Grenier, 2005). It has to be stressed that this is not a new situation, and it will be interesting to see and learn how African farmers have coped up with such phenomena over centuries. But we have no reason to romanticize the situation: Up to now, this was often accompanied by considerable yield losses. At the end of the day it is not the introgression itself, it’s the consequences of gene flow which counts. There is some experience about the importance of taking in farmers knowledge on various levels of screening landrace diversity and studying farmers management methods and their influence on agricultural ecology: (Ahmed et al., 2000; Bayu et al., 2006; Jarvis & Hodgkin, 1999; Manzelli et al., 2005; Marley et al., 2004; Mulatu & Belete, 2001; Songa et al., 2002; Tefera, 2004; Teshome et al., 1999b; Wenzel et al., 2001a; Wood & Lenne, 2001; Zongo et al., 2005). More details about Sorghum landraces and their genetic interaction with wild Sorghum are summarized in chapters on S. bicolor, origin of Sorghum and gene flow from crop to crop.

African landraces are according to many authors (Tunstall et al., 2001) threatened through the shift in agriculture: The modern world is placing a range of pressures on wild areas and on traditional agricultural communities, and external interests (often dominated by economic or political issues) strongly impinge. The major external forces relevant to this discussion advocate the introduction of high- 283 yield varieties, accompanied by mechanization and major chemical inputs, as the means to increase total production and economic return. These forces change the nature of the decision-making process dramatically; the farmer is encouraged to grow high-yield varieties in monoculture using inputs of fertilizer and pesticides.

Consequences of recurrent gene flow between cultivated Sorghum and its wild relatives should therefore not be generalized in the case of Africa: Sorghum biodiversity, mating characteristics and diversity of farmers management methods of the many landraces are just too diverse: Some of the African Sorghum varieties are strong inbreeders. And for sure we should not make the mistake of copying the concerns of the American Sorghum breeders with their problems with Sorghum halepense. Sorghum gene flow should always be considered on a case by case basis and in individual countries where the crop is grown.

8.4.1. The field trial of (Schmidt & Bothma, 2006), critical comments in the light of regulatory questions for the SuperSorghum project It is necessary to come back to the latest and most concrete case of a gene flow study from Sorghum crop to crop: All the more that the paper of (Schmidt & Bothma, 2006) will undoubtedly be cited by many in the debate about the regulation of SuperSorghum.

For most of the facts see the chapter on gene flow from Sorghum cultivars to the wild relatives, where the paper is summarized with the most important illustrations.

Crop Sorghum had an average of 4916.8 female flowers per panicle (SE 5 642.9; n 5 12). On the basis of the amount of female flowers per plant (and its standard error) and on the average amount of produced seeds per plant and distance class, the hybridization rate of pollinated flowers per plant was calculated. The hybridization rate was relatively low, not surpassing 3.45%, even at the closest distance to the central field.

The rate was even below 1% for plants at a distance of 39 m or greater and below 0.1% at the two farthest distances at 130 and 158 m (see Fig. 24). This is really astonishing when we consider the following critical discussion below:

284

Fig. 139 Distance-dependent hybridization rate of pollinated Sorghum plants. The dark solid line represents the average rate and the grey lines indicate the error interval. The small peak at 100m can be interpreted as a small patch of Sorghum where the cytoplasmatic sterility failed among the receptor plants. (Schmidt & Bothma, 2006)

This study (Schmidt & Bothma, 2006) shows an observable, sharp decrease of pollination within the first 40 m of the pollen-donating field. Although only a very small percentage of plants became pollinated beyond that distance, pollination continued to be observable at a very low level up to the farthest distance investigated (158 m). In brief, Sorghum gene flow was notable within approximately the first 40 m and very low beyond 40 m. In a similar study on the gene flow of rice, Song et al. (2003, 2004) also observed outcrossing distances of up to 40 m under “normal” weather conditions.

A serious discussion of the results obtained here and the estimations of the maximum pollination distance, however, require considering the following points.

The results indicate that Sorghum airborne pollen flow was significantly influenced by wind direction, which agrees with the viewpoint that weather factors have a strong effect on airborne pollen flow (Jones & Sieglinger, 1951; Song et al., 2004a; Song et al., 2004b). Changes in the pollination gradient can be expected when extreme wind conditions occur. Under normal temperature and humidity conditions, however, pollen has an “inbuilt obsolescence” because longevity is limited. In Sorghum, pollen longevity is approximately 30 min to 2 h after release (Lansac et al., 1994) and personal communication with Sorghum plant breeder W. Wenzel, Agricultural Research Council, South Africa, 2003 to J. Schmidt). Desiccation and reduced viability determine pollen longevity and restrict fertilization of female flowers after this time period (Lansac et al., 1994), even though the pollen itself may be transported several hours or days by the wind. Nonetheless, 2 h in an unusually strong wind could transport the pollen several kilometers and it would still be able to potentially pollinate a female flower. Assuming a uniform distribution of the pollen from the pollen source, the pollen rain would be highly diluted with distance because pollen density is approximately related to dispersion distance (1/r2). In the case of unusually 285 strong winds (singular events), we can expect inhomogeneous or even chaotic dispersion rather than a uniform pattern, see e.g., (Di-Giovanni & Kevan, 1991). Since many of the Sorghum traits used in Africa are to a high degree autogamous, the situation is not too dramatic, even more so in the light of the next argument:

The field trial was conducted with male sterile pollen receptor plants. Thus, the pollen from the central donor field did not have to compete with pollen from the receptor plant, which is normally responsible for approximately 70 to 95% of its pollinated female flowers (Ellstrand & Foster, 1983) (Pedersen et al., 1998) (Djè et al., 1999) (Smith & Frederickson, 2000). Moreover, the flowering period of a single plant is completed in 4 to 7 days, but the female flowers can remain receptive up to 16 days in the absence of pollination (Shertz & Dalton, 1980). Under the conditions of this field trial, the absence of self-pollination and the increased receptive time period of female flowers lead to a much higher outcrossing and gene flow rate than under natural conditions. Even though, the percentage of pollinated female flowers per plant only ranged from 0.04 to 3.45%, values similar to the results from a gene flow study conducted with maize. Therefore, under natural conditions, the outcrossing rates can be expected to be considerably lower than the values obtained here.

The study of (Schmidt & Bothma, 2006) has – as all field studies - its limitations in the short term conditions (wind, weather, flowering period etc.) and also in its specific regional and local geographical conditions. Consequently, the results reflect a mere short time photography of data.

In the field trial of (Schmidt & Bothma, 2006), there was an unusually high number of seeds in a plant 104m from the central field. Considering their findings, it is possible that this plant was fertile (or partly fertile) because of cytoplasmic male sterility instability. The number of plants used in the study (656) also supports this hypothesis. If cytoplasmic male sterility is to be used as a biosafety measure to avoid transgene flow, the involved cytoplasmic male sterility instabilities must be taken into account.

A further point should not be forgotten: The pollen receiving fields were all under irrigation. In nature, the receiving fields with the wild relatives and all the abandoned volunteers will therefore have much less favourable pollination conditions, the percentage of gene flow will again be lower than experimentally fixed.

Consequently, a agronomic ‘reality check’ needs to be done, and the gene flow data of this study cannot be taken for granted, they will most probably stay below 1% under natural and real conditions.

8.4.2. Gene flow in the genus Sorghum has to be assessed case by case A case by case view must include the different biological characteristics of the African Sorghum traits and also local agricultural management traditions.

Sorghum offers an excellent example of the sympatric association and interaction of a crop-wild-weed complex of a species in an agro-ecosystem, in many ways similar to the system of the landraces of Maize 286 in Mexico, see chapter Biodiversity and Centers of Origin in this study. The nature of genetic interaction among forms of the taxa and the consequences of these exchanges depend not only on the power of the genes involved and on the mating characteristics of the traits involved, but also on several other associated factors. Prevalence of wild relatives naturally varies from region to region based on extent of inherent genetic diversity, existence of selective pressures, and the farming systems in the region. Based primarily on knowledge of the agro-ecosystems and the history of genetic resource management and conservation, (Ejeta & Grenier, 2005) have selected Ethiopia, Sudan, and the United States to demonstrate how different consequences may arise from gene flow between wild and cultivated Sorghums. Although the empirical evidence they describe from Africa is still circumstantial and based on indirect molecular evidence and ultimately needs to be assessed experimentally, it offers some basis for recognizing the differential effects of gene flow and more importantly of hybridization rate in contrasting ecological, demographical, and farming practices. Sudan and Ethiopia are the birth places of the present day crop traits and have witnessed the evolution of wild and ancestral forms of Sorghum, although empirical data from Ethiopia show that Sorghum durra is the predominant race with a rather narrow genome and compared to other traits, a rather young domestication history.

The scenario in the Sudan and elsewhere in Africa is different from the Americas (Ejeta & Grenier, 2005): Sudan has the largest land area in Africa, yet the population base in low, only a third of the population in Ethiopia. Per-capita holdings of arable land are higher in the Sudan than all other African countries. In northern Sudan, where human settlement has been historically very light, the genetic identity of wild Sorghums may have been further protected by their isolation from human disturbance.

This is the explanation for results obtained by (Abu Assar et al., 2005): The objective of their study was to estimate genetic diversity and to obtain information on the genetic relationship among 96 Sorghum, S. bicolor accessions from Sudan, ICRISAT, and Nebraska, USA, using 16 simple sequence repeats (SSRs). In total, 117 polymorphic bands were detected with a mean of 7.3 alleles per SSR locus. By this approach each accession is uniquely fingerprinted. In both the rainfed and irrigated Sudanese agriculture (see also (Larsson, 1996)), genetic exchange between Sorghum and its wild relatives has resulted in formation of two widely recognized forms of crop-wild hybrids.

Aggressive forms of weedy S. bicolor have evolved that are readily identified and recognized by most farmers as feral weeds, and known under a local name of “adar”. This is a form of shattercane that is widely distributed and almost accepted as unavoidable. In spite of continual weeding and selective roguing this weedy S. bicolor has not been easy to eradicate in Sudan.

The second form of intermediate is equally feral, but appears to be more similar to cultivated Sorghum and produces grains that only slowly shatter. Continued introgression of cultivated Sorghum genes into wild forms has resulted in this hybrid form called “kerketita”. Farmers selectively harvest these types and encourage their continued existence, for they rely on them as feed and food depending on the harvest prospect: In bad years, these fast growing intermediates provide the only harvest possible, particularly for fodder.

287

(Ejeta & Grenier, 2005) have recently started a study in Sudan to investigate the extent of gene introgression among these types and hope to assess the differential fitness of these introgressed intermediates in contrast to the cultivated and wild progenitors. For more comments see the chapter on gene flow from Sorghum cultivars to their wild relatives.

An indirect estimate of gene flow by determining geographical and altitudinal allozyme variation in Ethiopia and Eritrea and its adaptation zones is given by (Ayana & Bekele, 1998, 1999; Ayana & Bekele, 2000; Ayana et al., 2000a; Ayana et al., 2000b, 2001) in a series of papers. The level of gene flow was low in the accessions, regions of origin and accessions within adaptation zones, but high among adaptation zones.

Gene flow studies in Africa encounter some basic difficulties, since recognizing the nature and origin of Shattercanes can be difficult and leads to errors, as stated by (Ayana & Bekele, 2000): Shattercanes (derivatives of wild Sorghum x cultivated Sorghum) are often noted as the most serious weeds on the highlands of Ethiopia, where they are known as kello or sepo (which means the fool in local languages) (Doggett, 1988, 1991; Harlan, 1975b). As described by (Harlan, 1992a), the shattercanes mimic the type of cultivated Sorghum race with which they are associated in Ethiopia. Examples of crop mimicry have been described in rice cultivation (Barrett, 1983). It is interesting to note here that many collectors of Sorghum germplasm confuse shattercane with wild Sorghum because of such camouflaging ability of the shattercanes (Harlan, 1984). There were certain accessions that the local farmers identified as kello or sepo from the materials that the authors used for this study. If farmers’ folk taxonomy was right (Teshome et al., 1997), then the confusion holds also true the other way round, confusing the shattercanes with cultivated Sorghum. In any case, gene flow occurs between the cultivated Sorghum and its weedy relatives.

All these activities are building on previous research by (Dahlberg et al., 2004; Grenier et al., 2004). There is a lot of research activity to be registered in connection with the Sudan: http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Sudan-20060810.pdf

8.4.3. Summary The situation in the USA and Africa is different regarding gene flow:

In the Americas Sorghum halepense and Sorghum almum pose problems with some gene flow and their weedy character, causing also problems in cultures like maize and soya,

In Africa there is a steady but slow gene flow indirectly documented through genomic analysis. Wild Sorghums are widespread and cause problems for the local farmers, weeding is often not enough and in some years yield reduction is considerable. It will be crucial to learn from the local farmers about their specific problems with wild Sorghums. 288

8.5. Mitigation of gene flow

8.5.1. Coexistence, general remarks For the basics on coexistence there are several papers to be mentioned: (Ammann, 2003; Brookes, 2004; Brookes & Barfoot, 2003, 2004; Byrne & Fromherz, 2003; Eastham & Sweet, 2002; Moschini et al., 2005; Schiemann, 2003). Most of these papers are focussing on Europe, but they provide excellent insight in the basic rules of co-existence:

No zero tolerance: Zero values with intermixing seeds or yield cannot be achieved in most cases (and is not necessary, since the introgressed genes stem from cultivars which have been approved for environmental and food and feed safety anyway). So it is a misleading expression to talk about “contamination”, we should use terms like intermixing, and threshold values instead.

Field size influences results: The impact of gene flow and the result of intermixing tracable genes in the yield is depending on the size of the fields involved: Big fields of many hectares usually do not pose a problem, since intermixing will be homogenized to minute traces well below the threshold values fixed by law.

Safety distances: Coexistence depends on safety distances, which have to be determined in a case by case decision. A multitude of characters of crop reproduction biology will have to be evaluated in order to set up the individual rules of lawful segregation

Biology and production characteristics of Seeds: Seed biology, production (and selection practices and exchange) also has to be involved in coexistence ruling

Time scale: There is, depending on the crop biology, a need to look at several years of crop production because of seed exchange by farmers, volunteer plants, tillering and dormant seeds, all these factors depend on the specific characters of a given crop, soil and regional factors.

Traditional knowledge: In the majority of cases, these rules do not have to be invented and developed from scratch, they can be derived from the practices of traditional and modern farming.

Mitigation needs a case to case approach: Mitigation of gene flow and the insurance of coexistence will depend on a variety of methods, which have to be adapted from case to case.

8.5.2. The choice of inbreeding landraces of Sorghum It will be important to make the right choice of an inbreeding Sorghum landrace, this would be a very elegant and natural way of avoiding unwelcome gene flow. The published literature will have to be scrutinized properly and it is possible that more studies about the correct choice are necessary. According to some authors (Folkertsma et al., 2005), the guinea-race is a predominantly inbreeding, diploid cereal crop, but more research is needed: According to (Dje et al., 2000) the heterozygosity in the 289 guinea populatioins studied is low: Employing SSR markers to analyse the genetic variation among five Guinea-race sorghum accessions, revealed an observed heterozygosity (Hobv) of 0.089, with average expected gene diversity (Hexp) of 0.224. These figures nicely comply with the observations among the 100 Guinea-race sorghum accessions reported by (Folkertsma et al., 2005) and point to the fact that Guinea-race sorghum is predominantly inbreeding, resulting in low levels of observed heterozygosity, but that the gene pool as a whole maintains a high level of allelic variation. It originated from West Africa and appears to have spread throughout Africa and South Asia, where it is now the dominant Sorghum race, via ancient trade routes. On the other hand, there are data available which hint to the fact that at least certain guinea races are also outbreeding (Monique Deu, personal communication).

8.5.3. Safety distances Working with safety distances is often the easiest strategy of choice, this has been shown with many different crops, the safety distances depending on the field size according to experimental data, see below the table of (DEFRA, 2006), based on a multitude of field experiments cited in this extensive report. Safety distances have to be assessed in the case of Africa, they will have to be assessed by field experiments with marker genes and with morphometric methods, details see chapters on proposed field experiments of Africa Harvest SuperSorghum project at the end of the study.

Fig. 140 The measures required to operate below a GM presence threshold lower than 0.9% will be the same as those needed for the 0.9% threshold, except that longer crop separation distances would have to be applied for oilseed rape and maize. Based on Table 6, the aim would be to limit cross-pollination to no more than 0.1%, and on the face of it the NIAB report provides recommended distances for this purpose (Annex C). However, there is a general point here about the difficulty of establishing measures for a 0.1% threshold. The NIAB data is robust for cross-pollination thresholds down to 0.3%, but at the 0.1% level any data has to be treated with caution. Margins of error and issues of statistical sensitivity and uncertainty become much more pronounced at such a very low level. It is possible that appreciably longer distances would be required than those in the NIAB report to be confident of routinely meeting a 0.1% cross-pollination threshold. Defra will consider this further as new scientific information becomes available. Annex C Taken with the comment in footnote No. 126, p. 42 from (DEFRA, 2006)

290

The experiments of (Arriola & Ellstrand, 1996; Arriola & Ellstrand, 1997, 2002) give first hints about safety distances, but they should not be taken for granted and no figures are given here on purpose, since they are derived from the situation from the US with Sorghum halepense and they are building on a rather scanty statistical material, not showing any instructive continuous gene flow gradient data along the field trial distances, and also because the receiving plants were all under irrigation. The results of (Schmidt & Bothma, 2006) can also not be taken for granted, since they base on an artificial situation, since the receiving fields all were containing male sterile plants, producing thus much to high hybridization results.

Nevertheless, the gene flow observed by (Schmidt & Bothma, 2006) fit well into the experiences of Sorghum plant breeders, who - for conventionally bred varieties or hybrids in southern Africa - use an isolation distance of 100 m to achieve less than 1% genetic intermixing (personal communication with ARC-Grain Research Institute Sorghum plant breeder Dr. W. Wenzel and ICRISAT Sorghum plant breeder Dr. E.S. Monyo, 2003 to (Schmidt & Bothma, 2006). Other sources give an isolation distance of 400 to 800 m for hybrid seed production but without additional information on seed purity . (Chopra, 1987; Chopra et al., 1999). Thus, the estimations obtained here agree with the experiences from Sorghum hybrid plant breeding.

8.5.4. Tandem constructs for a sustainable mitigation of gene flow (Al-Ahmad et al., 2006a; Al-Ahmad et al., 2004, 2005, 2006b; Al-Ahmad & Gressel, 2005, 2006; Gressel & Al-Achmad, 2005; Gressel & Al-Ahmad, 2005) propose tandem constructs, which would avoid or at least reduce considerably the spread of transgenes, the concept has been proven to produce stable gene constructs and sustainable mitigation: Transgenic mitigation (TM), where a desired primary gene is tandemly coupled with mitigating genes that are positive or neutral to the crop but deleterious to hybrids and their progeny. This was tested experimentally by the team of Gressel from Rehovot in Israel as a mechanism to mitigate transgene introgression. Dwarfism, which typically increases crop yield while decreasing the ability to compete, was used as a mitigator. A construct of a dominant ahasR (acetohydroxy acid synthase) gene conferring herbicide resistance in tandem with the semidominant mitigator dwarfing Δ gai (gibberellic acid-insensitive) gene was transformed into tobacco (Nicotiana tabacchum). The highest reproductive TM fitness relative to the wild type was 17%. The results demonstrate the suppression of crop–weed hybrids when competing with wild type weeds, or such crops as volunteer weeds, in seasons when the selector (herbicide) is not used. The linked unfitness would be continuously manifested in future generations, keeping the transgene at a low frequency.

8.5.5. Male sterility and apomixis in Sorghum, the prevention of gene flow Another possibility is shown by (Pedersen et al., 2003). The authors come to a rather pragmatic view for the reduction of gene flow with Sorghum: The results from this study support the hypothesis that gene flow through pollen can be severely restricted in Sorghum by adaptation of the technique proposed by (Feil et al., 2003) to A3 cytoplasmic male sterility. The low levels of seed set on selfed A3 F2 individuals (0.04%) as compared with A1 F2 individuals (74%) is a clear indication of greatly reduced levels of viable 291 pollen production from volunteer progeny of seed that escapes harvest in the cropping year. In field scale research and crop production, it is unlikely that all risks associated with new technologies can be completely eliminated. This research provides evidence that the risk of gene flow through Sorghum pollen can be greatly reduced by using, at least in part, the technique described herein.

In the event that transgenic Sorghum is developed to the stage that field testing becomes appropriate, utilization of this technique should be considered in combination with cultural controls including spatial isolation, crop rotation or fallowing in subsequent cropping sea-sons, and herbicides active against Sorghum to reduce risk of gene flow to weedy relatives. Related to yield, the experiments of (Feil et al., 2003) indicate that such associations can bring about grain production as high or even higher than those produced by pure male-fertile maize crops, especially when the male-sterile component is pollinated non-isogenically. The grain yield benefits from cytoplasmic male sterility and xenia as well as the fact that seed of male- sterile varieties can be produced cheaply and reliably in large quantities would facilitate the implementation of the proposed system in agricultural practice. Exploiting this technique has proven successful for the production of high oil maize (Bergquist et al., 1998b) and high grain quality maize (Bergquist et al., 1998a) and may be used to restrict gene flow in genetically modified Sorghum as well.

More details about cytoplasmatic male sterility are given in (McVetty, 1997) (Smith & Frederickson, 2000) (Pedersen et al., 2003). The robustness of cytoplasmic male sterility, however, is an important issue. (Worstell et al., 1984) noted that higher temperatures tend to increase male fertility of the A-lines (male sterile line). In a study by (Pedersen et al., 2003), the probability of a cytoplasmic male sterility plant becoming fertile was estimated to be about 0.39%, or approximately 1 out of 252 plants.

Developers of the first experiments with transgenic Sorghums should also have a close look at the experience with the production of hybrid Sorghum traits: Modern hybrid S. bicolor seed production relies exclusively on cytoplasmic male sterility (CMS) systems and almost all hybrid Sorghum seed is produced using the A1 CMS system. (Moran & Rooney, 2003). However, the reliance on a single CMS system increases the vulnerability of the crop to diseases and stresses that may attack that particular CMS system. Alternative CMS systems have been described and even used on a limited basis for hybrid seed production, but a direct comparison of the agronomic effects of different cytoplasms has not been possible because male-sterile lines with a common genetic background (and different cytoplasm) were not available. The recent development of isocytoplasmic A-lines allows more direct comparison of cytoplasmic effect on agronomic performance. There will be more modern methods applied for the improvement of Hybrid Sorghum, a fast spreading Sorghum crop production.

The discovery of the cytoplasmic genetic male sterility by (Stephens & Holland, 1954) based on the milo- kafir system was a milestone in Sorghum breeding and research. It is widely used in the commercial exploitation of heterosis. Today, more than 30% of the Sorghum area is under hybrids (in the USA!), which have yields about twice that of any local cultivar. Non-mila sources of cytoplasmic male sterility have also been identified (Schertz & Johnson, 1984; Schertz & Ritchey, 1978).

292

However, their commercialization has been hindered by the non-availability of sufficiently stable restorer lines (Moran & Rooney, 2003; Tsvetova & Elkonin, 2002). (Tang & Pring, 2003) demonstrate successful restauration of male-fertile and male-sterile lines carrying either of the two required alleles in a homozygous state, substantiate complementary gene action, and show successful single-gene fertility restauration patterns resulting in approximately 50% viable pollen.

The availability of several single, recessive male sterile genes (ms3, ms7) serving as a genetic male sterility system allows population improvement through recurrent selection procedures involving random mating and selection, see (Mgonja et al., 2006)

It will also be worth wile to have a closer look at the research development of apomictic Sorghums: (Elkonin et al., 1995) found proof of apomixis. In his work, he took advantage of tissue culture as a means of genetic modification of plant reproductive systems. Mutations can cause female sterility and revert male sterility. Evidence is demonstrated in their work that aposporic cells can be generated from tissue culture of Sorghum. Cytoembryological investigation has shown the presence of structures similar to aposporous embryo sacs, which have developed in the ovules along with the sexual embryo sacs. The ovules with aposporous formations have been observed almost in all of the studied plants in several generations, however, their frequency varied (2-53%). The authors propose to develop a system of Cytoplasmatic Male Sterility which for breeding and propagation methods could be reverted.

More details on male sterility and apomixis can be seen in the chapters on Sorghum reproduction and Sorghum breeding.

8.5.6. Gene switching as a prevention method to avoid gene flow Gene switching might develop in a real opportunity, first studies for other purposes show, that mechanisms could be developed for a crop protection technology: (Emani et al., 2002), (Zuo & Chua, 2000) claim that chemical regulation of transgene expression presents a powerful tool for basic research in plant biology and biotechnological applications. Various chemical-inducible systems based on de-repression, activation and inactivation of the target gene have been described. The utility of inducible promoters has been successfully demonstrated by the development of a marker-free transformation system and large scale gene profiling. In addition, field applications appear to be promising through the use of registered agrochemicals (e.g. RH5992) as inducers. most of the systems reported thus far are unsuitable for field applications because of the chemical nature of the inducers. Further work should focus on systems suitable for applications with transgenic crop plants, with particular emphasis on agricultural chemicals (e.g. insecticides and safeners; the latter chemicals are used in agriculture to render crops tolerate to herbicides) that have already been registered for field usage. An additional interest would be to develop multiple-inducible systems to independently regulate several target genes.

(Tang et al., 2004) review genetically engineered transgenic animals and plants which have proven to be extremely useful for analyzing biochemical and developmental processes. Promoters responding to chemical inducers will be powerful tools for basic research in molecular biology and biotechnological applications. Various chemical-inducible systems based on activation and inactivation of the target gene 293 had been developed. The transfer of regulatory elements from prokaryotes, insects, and mammals has opened new avenues to construct chemically-inducible promoters that differ in their ability to regulate the temporal and spatial expression patterns, and this will dramatically increase the application of transgenic technology. The authors provide an overview on promoter activating systems, promoter inactivation systems, inducible gene over-expression, and inducible anti-suppression. Several inducers have been experimentally tested for various maize transgene systems, they are cited in this review. It would therefore be possible to develop promoter systems which respond to chemical inducers, which are in use in modern Sorghum agricultural production. See also earlier important work as an example (Gatz, 1996, 1997, 1998; Gatz & Lenk, 1998).

8.5.7. Summary In order to avoid gene flow, there are several possibilities open for future development and regulation:

Regulation through establishing safety distances

Transgenic mitigation (TM), where a desired primary gene is tandemly coupled with mitigating genes that are positive or neutral to the crop but deleterious to hybrids and their progeny.

Establishing a sustainable system of male sterility, as it was developed for maize

Application of gene switching technologies, which are still in the stage of development

294

9. Geneflow assessment with morphometric methods

Fig. 141 Sorghum verticilliflorum from the National Herbarium of Nairobi, a specimen of a wild taxa included in the species Sorghum bicolor. Herbarium specimens offer lots of morphological characters ready to be included in statistical morphometrics, also they can be precious sources for genetic analysis. Furthermore, authors of new descriptions and new definitions of taxa, based on their taxonomic studies are obliged to deposit in a Herbarium open to public access a type specimen.

9.1. General remarks The following methods should be applied to determine whether gene flow constitutes a biosafety problem: 295

In a first phase it will be advisable to study the potential gene flow with methods using non-transgenic Sorghum traits, later in the course of the SuperSorghum project there will be opportunities for gene flow studies with transgenic traits.

Any kind of gene flow study needs to be organized as a well chosen mix of field studies and an assessment of the present day knowledge about hybrids of Sorghum with its wild relatives. This can best be done with a Dutch-Swiss-Irish method to assess hybrids in scientific plant collections called Herbaria (in Africa major herbaria are located in Nairobi and Pretoria). This methodology has been established by a Dutch research group, a Swiss group and has recently been enhanced by an Irish research group, all summarized and cited in the latest publication on the matter: (Flannery et al., 2005) and (Ammann et al., 2000b). A collaboration between the Irish group and the Berne group has been foreseen, the plan is to organize a follow-up on gene flow research within the EU-research programs.

9.2. The methodology proposed

9.2.1. General situation To be able to evaluate the risk of the field release of a genetically engineered crop it is important to have knowledge on the possibilities of gene flow for a certain crop in a certain region. (de Vries et al., 1992; Frietema, 1996) has introduced gene flow indices that give an indication of the possibilities of a certain plant to successfully hybridize with wild relatives and the impact this may have. These gene flow indices were not specifically developed for transgenic crops, but apply to them just as well. These indices can help in the risk assessment of gene flow of transgenic crops, in particular the risk assessment for field trials. The higher the gene flow indices, the more containment measures one will have to organize, if one wants to prevent outcrossing to wild relatives. For the risk assessment in case of the marketing of a genetically engineered crop these gene flow indices are less relevant, because for this particular risk assessment it is more a black-and-white situation: either plants are able to hybridize with wild relatives (even if chances are low), or they are not. This is because after the commercialization of a particular genetically engineered crop with a high outcrossing rate, and grown on a very large scale, it will be very difficult or even impossible to prevent outcrossing with wild relatives, unless particular measures of mitigation are taken (Al-Ahmad et al., 2006a; Al-Ahmad et al., 2004; Gressel, 1999) Even though the barriers to hybridization exist in nature as described above, many economically important crops show low hybridization barriers. Such low hybridization barriers have been important for agriculture because it has enabled plant breeders to develop new varieties using cross-breeding. Sorghum is here a typical example. See many references on gene flow in (Ammann et al., 2000a), check also the cited literature in other reviews such as (Ellstrand, 2003) and (Eastham & Sweet, 2002).

9.2.2. Early developments of the methodology The earlier literature on the method is basically building upon a Dutch dissertation from the University of Leiden (Frietema, 1996) where the leader of the Berne group (Ammann) has been involved as an expert, and which has then been taken up with research grants from Switzerland resulting in more publications on gene flow matters: (Ammann et al., 2000b, c; Jacot et al., 2004) 296

9.2.2.3. Dutch-Swiss method This 'Dutch-Swiss Method' deals with the agricultural reality, therefore its hybridization reality in the field which is here the target of study. Thus, its not the potential hybridization, but the reality of persistent fertile hybrids which the authors are looking for.

9.2.2.4. Conditions for successful pollination The probability of successful pollination depends on a great number of interrelated factors, such as: - Level of pollen production of (transgenic) plants. - Rate of self- and cross-fertilization of receptor plants. - Rate of dispersion of donor pollen. - Properties of pollinating agents, where plants sometimes use highly specialized insect pollinators and have - highly adapted flowers, resulting in a very effective genetic isolation (mechanical isolation). - The existence of spatial distance between the pollen donor and the wild recipient population (spatial isolation). How large the distance is within which pollination can occur depends on factors like: wind turbulence, speed and direction and/or the flying range of insects and the time during which pollen is viable. - Local density of recipient population. - A difference in flowering season between the crop and the wild population (phenological isolation).

In addition to these factors a number of other factors are sometimes responsible for the fact that even though plants are sexually compatible, they do not hybridize in practice, or that the formed hybrid is not viable (Hadley & Openshaw, 1980). These are: - prevention of fertilization, - hybrid weakness or inviability, - hybrid sterility, and - hybrid breakdown.

9.2.2.5. Gene flow indices The gene flow indices are built up from data for a particular plant on its possibilities for dispersal of pollen, dispersal of diaspora3, and the presence and density of wild relatives in a given region. These data are classified into three different codes: the Dp-code (dispersal of pollen), the Dd-code (dispersal of diaspores) and the Df code (frequency of distribution of wild relatives). Together they are a measure for the possibilities of gene flow of a certain crop in a certain region and a measure of how widespread this gene flow may be. The Dpdf gene flow indices are always applicable for a given region, and this region therefore has to be mentioned in all cases where gene flow indices are used. This system has been in a first attempt refined by (Ammann & Jacot, 2003; Ammann et al., 1996, 2000c; Jacot et al., 2004).

9.2.2.6. Recent improvements of the Dutch-Swiss Method, now called Dutch-Swiss-Irish Method In a recent attempt, an Irish research group has succeeded to employing a composite gene-flow index to numerically quantify a crop’s potential for gene flow (Flannery et al., 2005). The purpose of this strategy is to describe the measures a farmer must adopt to minimize the admixture of GM and non-GM crops. Minimizing pollen / seed-mediated gene flow between GM and non-GM crops is central to successful coexistence. 297

(de Vries et al., 1992; Frietema, 1996) analyzed the potential for spontaneous gene flow by examining three codes of dispersal: dispersal of pollen (Dp), dispersal of diaspores (Dd: fruit and seed) and the frequency of wild relatives (Df). Though the index does not address the potential impact of a transgene on a related crop/wild relative (Conner, 2003), (Ammann et al., 2000a) have highlighted the necessity for a fourth code that would attend to this issue.

9.2.3. Study elements of Dutch-Swiss-Irish method The study elements considered for the model of (Flannery et al., 2005) are summarized in the following table:

Fig. 142 Study elements of the Morphometric method, from (Flannery et al., 2005)

The Dutch-Swiss-Irish method makes some additions: (Flannery et al., 2005) did not take into account that seed mediation might not be the only element of propensity in the case of Johnson Grass (and maybe other wild Sorghum species): Sorghum halepense is known for its extensive vegetative 298 proliferation with shoots. It will be now important for the project to carefully evaluate the detailed reproduction biology of the chosen trait and its close wild relatives like Johnson Grass.

9.2.3.1. Strand CSV (Crop seed-to-volunteer gene flow) Following on from the previous strand, which addressed pollen-mediated crop-to-crop gene flow, is the potential of a crop-derived seed to develop into a volunteer. Seed spillage is an unwanted but assured outcome of harvesting. Testament to this is the issue of managing volunteer populations within a rotation (e.g. recurrence of oilseed rape volunteers from long-lived seed banks), which can provide a conduit for seed-mediated gene flow. In this model, three questions serve to address a crop’s potential for successful seed- mediated gene flow within the confines of an agricultural system: CSV1: Does the crop produce seed during its cultivation? CSV2: Post-harvest Will the seed survive and germinate within the confines of a managed field? Whether seed enters the seed-bank because of seed spillage at sowing, pod shatter or harvesting, their ability to germinate and develop into volunteers affords them the potential to act as future sink populations. CSV3: Will the volunteer develop into a viable individual? By confining this strand to seeds that germinate within the boundary of a crop system, it is possible to differentiate between a crops ability to complete seed-mediated gene flow both inside and outside (see Sect. Strand CSF) the managed environment of the field.

9.2.3.2. Strand CSF (Crop seed-to-feral gene flow) The tendency of a crop to establish a persistent population outside the boundaries of the field (via seed- mediated gene flow) is an important portal of gene flow for all crops and a significant issue when considered in the context of coexistence. Similar to Strand CSV, three questions can be employed to gauge the ability of an errant seed to develop into a feral individual:

CSF1: Does the crop produce seed during its cultivation? CSF2: Following transfer from site of cultivation will wayward seed survive and germinate? CSF3: Will the resulting individuals establish into a viable feral population? The provision of CSF2 and CSF3 is informative for risk assessment purposes as they identify a crop’s ecological viability, measured as an ability to compete for space and nutrients with the adjacent flora. Whereas certain crops (e.g. maize) will not establish a persistent population as they are unable to survive outside of cultivation (Gould, 1968). Certain crop species have the capacity to freely hybridize and backcross with wild and feral populations e.g. ryegrass and oilseed rape and e.g. Johnson Grass.

9.3. The organization of the gene flow study part morphometrics Example of morphometric analysis of gene flow in wheat towards its wild relatives from (Jacot et al., 2004): Scattergram of gene flow between wheat and its wild relatives.

299

Fig. 143 Scattergram Aegilops ovata x Triticum aestivum, based upon unpublished data of Jacot, J. and P. Rufener Al Mazyad. From (Jacot et al., 2004)

9.3.1. Assessing the biosystematics of Sorghum traits and wild relatives, which are relevant for the project a) Screening the relevant literature (now done in the Report Version 10) b) Establishing the schemes of potential gene flow for major traits and its wild relatives: see chapters 1.1

There are some publications describing in detail how morphometric data assessment has been done for the multivariate analysis of Sorghum landraces, here just one example from (Abdi et al., 2002). The following characters have been assessed metrically or qualitatively: Grain form, midrib color, awn, grain plumpness, peduncle shape, endosperm texture, grain color, inflorescence compactness and shape, glume hairiness, grain covering, glume color, stem juiciness, juice flavor, threshability. See table 6 from (Abdi et al., 2002) below for the eigen values of the multivariate analysis.

300

Fig. 144 Dendrogram showing the clustering patterns of the 34 Sorghum (Sorghum bicolor (L.) Moench) landraces in North Shewa and South Welo, Ethiopia (key to legend: for the local name of Sorghum landraces, refer to Table below. From (Abdi et al., 2002)

301

Fig. 145 Table with morphological details, from (Abdi et al., 2002)

Overall, there are numerous publications using morphological variation in describing landraces and in comparison to genetic data: (Abdi et al., 2002; Ayana & Bekele, 1998, 1999; Ayana & Bekele, 2000; Dahlberg et al., 2002; Djè et al., 1998; Hart et al., 2001; Haussmann et al., 1998; Ogunbayo et al., 2005; Okonkwo & Onoenyi, 1998; Olah, 1981; Rami et al., 1998; Rao Prasada & Rao, 1990; Rao et al., 1996; Redfearn et al., 1999; Wu, 1979, 1993)

9.3.3. Fixing the framework of the geneflow study, start with study with preparative a) Screening herbarium or field materials on Sorghum hybrids or field collections, morphometric statistics including multivariate statistics used for numerical taxonomy. b) Establishing in various sites the abundance of hybrids of Sorghum cultivars to the wild relatives. c) Farmers knowledge: Interviews about the feral Sorghum problems: 6 interviews in 3 countries d) Establishing the scheme for the gene flow indices according to the methods described in (Flannery et al., 2005) and (Ammann et al., 1996, 2000a) 302 e) Assessing the gene flow indices for Sorghum hybrids with wild relatives (with herbarium and field morphometric data f) writing up the results, establishing a definite literature list with annotations g) assessing the results and conclusions for the SuperSorghum project, meeting with regulatory experts. 303

10. Cited references

Abdi, A., Bekele, E., Asfaw, Z., & Teshome, A. (2002) Patterns of morphological variation of sorghum (Sorghum bicolor (L.) Moench) landraces in qualitative characters in North Shewa and South Welo, Ethiopia. Hereditas, 137, 3, pp 161-172 ://000182769600001

Able, J.A., Rathus, C., & Godwin, I.D. (2001) The investigation of optimal bombardment parameters for transient and stable transgene expression in sorghum. In Vitro Cellular & Developmental Biology-Plant, 37, 3, pp 341-348 ://000171552000006 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Able-optimal-Bombard-Sorgh- 2001.pdf

Abu Assar, A.H., Uptmoor, R., Abdelmula, A.A., Salih, M., Ordon, F., & Friedt, W. (2005) Genetic variation in sorghum germplasm from Sudan, ICRISAT, and USA assessed by simple sequence repeats (SSRs). Crop Science, 45, 4, pp 1636-1644 ://000230713700058 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/AbuAssar-Variation-Sudan- 2005.pdf

Ahmed, M.M., Masters, W.A., & Sanders, J.H. (1993) Returns from Research in Distorted Economies - Hybrid Sorghum in Sudan. American Journal of Agricultural Economics, 75, 5, pp 1303-1303 ://A1993MW71000151

Ahmed, M.M., Masters, W.A., & Sanders, J.H. (1995) Returns from Research in Economies with Policy Distortions - Hybrid Sorghum in Sudan. Agricultural Economics, 12, 2, pp 183-192 ://A1995RY60700007

Ahmed, M.M., Sanders, J.H., & Nell, W.T. (2000) New sorghum and millet cultivar introduction in Sub-Saharan Africa: impacts and research agenda. Agricultural Systems, 64, 1, pp 55-65 ://000087481900004 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Ahmed-NewSorghum-2000.pdf

Ahmed, S.M. & Young, W.R. (1969) Field Studies on Chemical Control of Stem Borer Chilo Partellus on Hybrid Sorghum in India. Journal of Economic Entomology, 62, 2, pp 478-& ://A1969D005200054

Ahn, S., Anderson, J.A., Sorrells, M.E., & Tanksley, S.D. (1993) Homoeologous Relationships of Rice, Wheat and Maize Chromosomes. Molecular & General Genetics, 241, 5-6, pp 483-490 ://A1993ML71400001

Ahn, S. & Tanksley, S.D. (1993) Comparative Linkage Maps of the Rice and Maize Genomes. Proceedings of the National Academy of Sciences of the United States of America, 90, 17, pp 7980-7984 ://A1993LV64400016

Ahnert, D., Lee, M., Austin, D.F., Livini, C., Woodman, W.L., Openshaw, S.J., Smith, J.S.C., Porter, K., & Dalton, G. (1996) Genetic diversity among elite sorghum inbred lines assessed with DNA markers and pedigree information. Crop Science, 36, 5, pp 1385-1392 ://A1996VK30900049

Akhunov, E.D., Goodyear, A.W., Geng, S., Qi, L.-L., Echalier, B., Gill, B.S., Miftahudin, Gustafson, J.P., Lazo, G., Chao, S., Anderson, O.D., Linkiewicz, A.M., Dubcovsky, J., Rota, M.L., Sorrells, M.E., Zhang, D., Nguyen, H.T., Kalavacharla, V., Hossain, K., Kianian, S.F., Peng, J., Lapitan, N.L.V., Gonzalez-Hernandez, J.L., Anderson, J.A., Choi, D.-W., Close, T.J., Dilbirligi, M., Gill, K.S., Walker-Simmons, M.K., Steber, C., McGuire, P.E., Qualset, C.O., & Dvorak, J. (2003) The Organization and Rate of Evolution of Wheat Genomes Are Correlated With Recombination Rates Along Chromosome Arms. Genome Res. %R 10.1101/gr.808603, 13, 5, pp 753-763 http://www.genome.org/cgi/content/abstract/13/5/753 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit- 1/Akhunov-Evolution-Wheat-2003.pdf

Al-Ahmad, H., Dwyer, J., Moloney, M., & Gressel, J. (2006a) Mitigation of establishment of Brassica napus transgenes in volunteers using a tandem construct containing a selectively unfit gene. Plant Biotechnology Journal, 4, 1, pp 7-21 304

://000234030000003 AND http://www.botanischergarten.ch/Geneflow/AlAchmad-Tandem-napus-2006.pdf

Al-Ahmad, H., Galili, S., & Gressel, J. (2004) Tandem constructs to mitigate transgene persistence: tobacco as a model. Molecular Ecology, 13, 3, pp 697-710 ://000188825700016 AND http://www.botanischergarten.ch/Geneflow/Al-Ahmad-Gressel-Tandem-2004.pdf

Al-Ahmad, H., Galili, S., & Gressel, J. (2005) Poor competitive fitness of transgenically mitigated tobacco in competition with the wild type in a replacement series. Planta, 222, 2, pp 372-385 ://000232498200016 AND http://www.botanischergarten.ch/Geneflow/AlAchmad-Fitness-Tandem-2005.pdf

Al-Ahmad, H., Galili, S., & Gressel, J. (2006b) Infertile interspecific hybrids between transgenically mitigated Nicotiana tabacum and Nicotiana sylvestris did not backcross to N. sylvestris. Plant Science, 170, 5, pp 953-961 ://000236517700006 AND http://www.botanischergarten.ch/Geneflow/AlAchmad-Nobackcross-2006.pdf

Al-Ahmad, H. & Gressel, J. (2005) Transgene containment using cytokinin-reversible male sterility in constitutive, gibberellic acid-insensitive (Delta gai) transgenic tobacco. Journal of Plant Growth Regulation, 24, 1, pp 19-27 ://000231700700004 AND http://www.botanischergarten.ch/Geneflow/AlAchmad-Cytokinin-2005.pdf

Al-Ahmad, H. & Gressel, J. (2006) Mitigation using a tandem construct containing a selectively unfit gene precludes establishment of Brassica napus transgenes in hybrids and backcrosses with weedy Brassica rapa. Plant Biotechnology Journal, 4, 1, pp 23-33 ://000234030000004 AND http://www.botanischergarten.ch/Geneflow/AlAchmad-Tandem-Establishment-2006.pdf

Al-Janabi, S.M., McClelland, M., Petersen, C., & Sobral, B.W.S. (1994) Phylogenetic analysis of organellar DNA sequences in the Andropogoneae: Saccharinae. TAG Theoretical and Applied Genetics, 88, 8, pp 933-944 http://dx.doi.org/10.1007/BF00220799 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Al-Janabi- Andropogoneae-1994.pdf

Alderman, S.C., Halse, R.R., & White, J.F. (2004) A reevaluation of the host range and geographical distribution of Claviceps species in the United States. Plant Disease, 88, 1, pp 63- 81 ://000220313300010

Aldrich, P.R. & Doebley, J. (1992) Restriction Fragment Variation in the Nuclear and Chloroplast Genomes of Cultivated and Wild Sorghum-Bicolor. Theoretical and Applied Genetics, 85, 2-3, pp 293-302 ://A1992JY22200025 ADN http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Aldrich-Restriction-Fragment- Sorgh-bicolor-1992.pdf

Aldrich, P.R., Doebley, J., Schertz, K.F., & Stec, A. (1992) Patterns of Allozyme Variation in Cultivated and Wild Sorghum-Bicolor. Theoretical and Applied Genetics, 85, 4, pp 451-460 ://A1992KE52200011 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Aldrich-Allozymes-Sorghum- 1992.pdf

Ali, M.R. & Wills, R.B.H. (1983) Milling and Cooking of Hybrid Sorghum Grain. Qualitas Plantarum-Plant Foods for Human Nutrition, 32, 1, pp 59-66 ://A1983QF51200007

Alvarez, N., Garine, E., Khasah, C., Dounias, E., Hossaert-McKey, M., & McKey, D. (2005) Farmers' practices, metapopulation dynamics, and conservation of agricultural biodiversity on-farm: a case study of sorghum among the Duupa in sub-sahelian Cameroon. Biological Conservation, 121, 4, pp 533-543 ://000225052500005 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Alvarez-Practice-2005.pdf

Ammann, K. (1986) Die Bedeutung der Herbarien als Arbeitsinstrument der botanischen Taxonomie. Zur Stellung der organismischen Biologie heute. Botanica Helvetica, 96, 1, pp 109-132 ://A1986D456500011 and http://www.botanischergarten.ch/Herbarien/herbarien1.pdf

Ammann, K. (2003) Co-existence between organic, traditional and biotech agriculture Conflict resolution on the basis of discoursive processes and different kinds of knowledge, Klaus Ammann pp 6 Draft Text for the Hearings of the Danish Parliament on Co-existence in Helsingoer, November 2003 Bern (Report) http://www.botanischergarten.ch/Coexistence/Ammann-Hearing-Danmark2004.pdf

305

Ammann, K. (2004) Electronic Source: The impact of agricultural biotechnology on biodiversity, a review, published by: Klaus Ammann http://www.botanischergarten.ch/Biotech-Biodiv/Report-Biodiv-Biotech12.pdf

Ammann, K. (2005) Effects of biotechnology on biodiversity: herbicide-tolerant and insect-resistant GM crops. Trends in Biotechnology, 23, 8, pp 388- 394 ://000231342700005 and http://www.sciencedirect.com/science/article/B6TCW-4GG2HJM- 2/2/b23d0cc8c6846b9f6625162f3351b0ae and http://www.botanischergarten.ch/TIBTECH/Ammann-TIBTECH-Biodiversity-2005.pdf

Ammann, K. (2006) Reconciling Traditional Knowledge with Modern Agriculture: A Guide for Building Bridges. In Intellectual Property Management in Health and Agricultural Innovation a handbook of best practices, Chapter 16.7 (eds A. Krattiger, R.T.L. Mahoney, L. Nelsen, G.A. Thompson, A.B. Bennett, K. Satyanarayana, G.D. Graff, C. Fernandez & S.P. Kowalsky), pp. 1539-1559. MIHR, PIPRA, Oxford, U.K. and Davis, USA chapter 16.7 http://www.botanischergarten.ch/TraditionalKnowledge/Ammann-Traditional-Biotech-2007.pdf free of copyrights AND the exported bibliography with the links: http://www.botanischergarten.ch/TraditionalKnowledge/Exported-Bibliography-links- Ammann-2007.pdf

Ammann, K. (2008) Feature: Integrated farming: Why organic farmers should use transgenic crops. New Biotechnology, 25, 2, pp 101 - 107 http://www.botanischergarten.ch/NewBiotech/Ammann-Integrated-Farming-Organic-2008.publ.pdf AND DOI: http://dx.doi.org/10.1016/j.nbt.2008.08.012

Ammann, K. (2009) Feature: Why farming with high tech methods should integrate elements of organic agriculture. New Biotechnology, 25, 6, pp 378- 388 http://www.sciencedirect.com/science/article/B8JG4-4WKTX50-1/2/1698b7149ed724fd0a49b3ae49f234ab AND http://www.botanischergarten.ch/Organic/Ammann-High-Tech-and-Organic-2009.pdf

Ammann, K. & Jacot, Y. (2003) Vertical Gene Flow. In Methods for Risk Assessment of Transgenic Plants (eds K. Ammann, Jacot, Y. & R. Braun). Birkhäuser, Basel

Ammann, K., Jacot, Y., & Rufener Al Mazyad, P. (1996) Field release of transgenic crops in Switzerland : an ecological assessment of vertical gene flow. In Gentechnisch veränderte krankheits- und schädlingsresistente Nutzpflanzen. Eine Option für die Landwirtschaft ? (eds E. Schulte & O. Käppeli), Vol. 1, 3, pp. 101-157. Schwerpunktprogramm Biotechnologie, Schweiz. Nationalfonds zur Förderung der Wissenschaftlichen Forschung, BATS, Basel http://www.botanischergarten.ch/debate/techdef5a.pdf

Ammann, K., Jacot, Y., & Rufener Al Mazyad, P. (2000a) an Ecological Risk Assessment of Vertical Gene Flow. In Safety of Genetically Engineered Crops (ed R. Custers). Flanders Interuniversity Institute for Biotechnology, Zwijinarde, BE.J. Bury, VIB VIB Publication http://www.vib.be and http://www.botanischergarten.ch/Geneflow/VIBreport.pdf

Ammann, K., Jacot, Y., & Rufener Al Mazyad, P. (2000b) an Ecological Risk Assessment of Vertical Gene Flow. In Safety of Genetically Engineered Crops (ed R. Custers), pp. 60-87. Flanders Interuniversity Institute for Biotechnology, Zwijinarde, BE.J. Bury, VIB VIB Publication http://www.vib.be and http://www.botanischergarten.ch/Geneflow/VIBreport.pdf

Ammann, K., Jacot, Y., & Rufener Al Mazyad, P. (2000c) Weediness in the light of new transgenic crops and their potential hybrids. Journal of Plant Diseases and Protection, Special Issue 2000, pp http://www.botanischergarten.ch/debate/weeds1.pdf

Ammann, K., Jacot, Y., & Rufener Al Mazyad, P. (2005) The ecology and detection of plant ferality in the historic records. In Crop Ferality and Volunteerism (ed J. Gressel), pp. 31-43. Taylor & Francis, Boca Raton http://www.botanischergarten.ch/Feral-def/Feral-MS-5-20040703.pdf extended version

Ammann, K. & van Montagu, M. (2009) Organotransgenesis arrives. New Biotechnology, 25, 6, pp 377-377 http://www.sciencedirect.com/science/article/B8JG4-4W7YY1F-1/2/28740dfca05cf12bb3b973b4f4568ccb AND http://www.botanischergarten.ch/NewBiotech/Ammann-Montagu-Editorial-Oranotransgenesis-arrives-2009.pdf

Ammann, K.i., Wolfenbarger, L., Andow DA. and Hilbeck, A., Nickson, T., Wu, F., Thompson, B., & Ammann, K. (2004) 306

Electronic Source: Biosafety in agriculture: is it justified to compare directly with natural habitats ? (ed ESA Ecological Society of America), Frontiers in Ecology, Forum: GM crops: balancing predictions of promise and peril published by: Ecological Society of America www.frontiersinecology.org and http://www.botanischergarten.ch/Frontiers-Ecology/Ammann-Forum-def1.pdf

Anderson, P.C. (1998) History of harvesting and threshing techniques for cereals in the prehistoric Near East. In The Origins of Agriculture and Crop Domestication (eds A.B. Damania, J. Valkoun, G. Willcox & C.O. Qualset), pp. 145-159. ICARDA, Aleppo

Andrews, D.J., Bramelcox, P.J., & Wilde, G.E. (1993) New Sources of Resistance to Greenbug, Biotype-I, in Sorghum. Crop Science, 33, 1, pp 198-199 ://A1993KT47500035

Arber, W. (2000) Genetic variation: molecular mechanisms and impact on microbial evolution. Fems Microbiology Reviews, 24, 1, pp 1-7 ://000084915900001 AND http://www.botanischergarten.ch/Mutations/Arber-Gen-Variation-FEMS-2000.pdf

Arber, W. (2002) Roots, strategies and prospects of functional genomics. Current Science, 83, 7, pp 826-828 ://000178662800019 and http://www.botanischergarten.ch/Mutations/Arber-Comparison-2002.pdf

Arber, W. (2003) Elements for a theory of molecular evolution. Gene, 317, 1-2, pp 3-11 ://000186667000002 and http://www.botanischergarten.ch/Mutations/Arber-Gene-317-2003.pdf

Arber, W. (2004) Biological evolution: Lessons to be learned from microbial population biology and genetics. Research in Microbiology, 155, 5, pp 297- 300 ://000222736200001 AND http://www.botanischergarten.ch/Mutations/Arber-Evolution-Lessons-2004.pdf

Arber, W. & Linn, S. (1969) DNA Modification and Restriction. Annual Review of Biochemistry, 38, pp 467-& ://A1969D759300017

Armelagos, G.J. & Harper, K.N. (2005a) Genomics at the origins of agriculture, part one. Evolutionary Anthropology, 14, 2, pp 68-77 ://WOS:000228959200005 AND http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Armelagos-Genomics- Origins-1-2005.pdf

Armelagos, G.J. & Harper, K.N. (2005b) Genomics at the origins of agriculture, part two. Evolutionary Anthropology, 14, 3, pp 109-121 ://WOS:000230294700004 AND http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Armelagos-Genomics- Origins-2-2005.pdf

Arnon, I. & Blum, A. (1965) Superiority of a Hybrid Sorghum over Self-Pollinated Varieties at Different Dates and Rates of Planting. Israel Journal of Agricultural Research, 15, 1, pp 3-& ://A19657113000001

Arriola, P.E. (2002) Gene flow and hybrid fitness in the Sorghum bicolor - Sorghum halepense complex, Ohio State University Gene Flow Workshop Ed. pp http://www.biosci.ohio-state.edu/~asnowlab/Proceedings.pdf

Arriola, P.E. & Ellstrand, N.C. (1996) Crop-to-weed gene flow in the genus Sorghum (Poaceae): Spontaneous interspecific hybridization between johnsongrass, Sorghum halepense, and crop sorghum, S. bicolor. Amer. J. Bot., 83, pp 1153-1160 http://www.botanischergarten.ch/Africa-Harvest-Geneflow/Arriola-1996.pdf

Arriola, P.E. & Ellstrand, N.C. (1997) Fitness of interspecific hybrids in the genus Sorghum: Persistence of crop genes in wild populations. Ecological Applications, 7, 2, pp 512-518 ://A1997WW23300013 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Arriola-Fitness-1997.pdf

Arriola, P.E. & Ellstrand, N.C. (2002) Gene flow and hybrid fitness in the Sorghum bicolor - Sorghum halepense complex., Ohio State University Ohio State Univesity Gene Flow Workshop Ed. A. Snow pp 25-31 http://www.biosci.ohio-state.edu/~asnowlab/Proceedings.pdf 307

Asante, S.A. (1995) Sorghum quality and utilization. Afican Crop Science Journal, 3, pp 231-240

Austin, A., Singh, H.D., Rao, N.G.P., & Hanslas, V.K. (1972) Variations in Protein and Lysine Content in Sorghum-Vulgare. Acta Agronomica Academiae Scientiarum Hungaricae, 21, 1 2, pp 81-& ://A1972M320100011

Ayana, A. & Bekele, E. (1998) Geographical patterns of morphological variation in sorghum (Sorghum bicolor (L.) Moench) germplasm from Ethiopia and Eritrea: qualitative characters. Hereditas, 129, 3, pp 195-205 ://000079994600002

Ayana, A. & Bekele, E. (1999) Multivariate analysis of morphological variation in sorghum (Sorghum bicolor (L.) Moench) germplasm from Ethiopia and Eritrea. Genetic Resources and Crop Evolution, 46, 3, pp 273-284 ://000081066600010

Ayana, A. & Bekele, E. (2000) Geographical patterns of morphological variation in sorghum (Sorghum bicolor (L.) Moench) germplasm from Ethiopia and Eritrea: Quantitative characters. Euphytica, 115, 2, pp 91-104 http://www.springerlink.com/openurl.asp?genre=article&id=doi:10.1023/A:1003998313302 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Ayana-Patterns-Euphytica-2000.pdf

Ayana, A., Bekele, E., & Bryngelsson, T. (2000a) Genetic variation in wild sorghum (Sorghum bicolor ssp verticilliflorum (L.) Moench) germplasm from Ethiopia assessed by random amplified polymorphic DNA (RAPD). Hereditas, 132, 3, pp 249-254 ://000089938200009 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Ayana-Verticilliflorum- Varriation-2000.pdf

Ayana, A., Bryngelsson, T., & Bekele, E. (2000b) Genetic variation of Ethiopian and Eritrean sorghum (Sorghum bicolor (L.) Moench) germplasm assessed by random amplified polymorphic DNA (RAPD). Genetic Resources and Crop Evolution, 47, 5, pp 471-482 ://000089346700002

Ayana, A., Bryngelsson, T., & Bekele, E. (2001) Geographic and altitudinal allozyme variation in sorghum (Sorghum bicolor (L.) Moench) landraces from Ethiopia and Eritrea. Hereditas, 135, 1, pp 1-12 ://000175142700001 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Ayana-Allozyme-2001.pdf

Ayyangar, G. & Ponnaiya, B. (1939) Studies in Cleistogamy and its inheritance in sorghum. Current Science, 8, pp 418-419

Bantilan, M., Deb, U., Gowda, C., Reddy, B., Obilana, A., & Evenson, R., eds. (2004) Sorghum Genetic Enhancement: Research Process, Dissemination and Impacts, pp 312, ICRISAT, Patancheru, Andhra Pradesh, India,

Barr, A.R., Mansooji, A.M., Holtum, J.A.M., & Powles, S.B. (1992) The inheritance of herbicide resistance in Avena sterilis ssp. ludoviciana, biotype SAS, Melbourne: Weed Science Society of Victoria 1. Proceedings of the First International Weed Control Congress. Ed. pp 70-75

Barraclough, G. (1985) The Times Atlas of World History, rev. ed. Times Books, London, pp

Barrett, S.C.H. (1983) Crop Mimicry in Weeds. Economic Botany, 37, 3, pp 255-282 ://A1983RC29200001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Barrett-Mimicry-1983.pdf

Battraw, M. & Hall, T.C. (1991) Stable Transformation of Sorghum-Bicolor Protoplasts with Chimeric Neomycin Phosphotransferase-Ii and Beta-Glucuronidase Genes. Theoretical and Applied Genetics, 82, 2, pp 161-168 ://A1991GB09800006 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Battraw-protoplasts-transf- 1991.pdf

Bawazir, A.A.A. & Idle, D.B. (1989) Drought Resistance and Root Morphology in Sorghum. Plant and Soil, 119, 2, pp 217-221 ://A1989AW31500005

308

Bayley, C., Trolinder, N., Ray, C., Morgan, M., Quisenberry, J.E., & Ow, D.W. (1992) Engineering 2,4-D Resistance into Cotton. Theoretical and Applied Genetics, 83, 5, pp 645-649 ://A1992HK21600015

Bayu, W., Rethman, N.F.G., Hammes, P.S., & Alemu, G. (2006) Effects of farmyard manure and inorganic fertilizers on sorghum growth, yield, and nitrogen use in a semi-arid area of ethiopia. Journal of Plant Nutrition, 29, 2, pp 391-407 ://000235346100016

Bellon, M.R. & Berthaud, J. (2004) Transgenic maize and the evolution of landrace diversity in Mexico. The importance of farmers' behavior. Plant Physiology, 134, 3, pp 883-888 ://000220360400001

Bellon, M.R. & Berthaud, J. (2006) Traditional Mexican agricultural systems and the potential impacts of transgenic varieties on maize diversity. Agriculture and Human Values, 23, 1, pp 3-14 ://000236273900002 AND http://www.botanischergarten.ch/Maize/Bellon-Impacts-Landraces-2006.pdf

Bellon, M.R., Berthaud, J., Smale, M., Aguirre, J.A., Taba, S., Aragon, F., Diaz, J., & Castro, H. (2003) Participatory landrace selection for on-farm conservation: An example from the Central Valleys of Oaxaca, Mexico. Genetic Resources and Crop Evolution, 50, 4, pp 401-416 ://000183201600008 AND http://www.botanischergarten.ch/Maize/Bellon-Participatory-Maize-2003.pdf

Bellwood, P. (2001) Early agriculturalist population diasporas? Farming, languages, and genes. Annual Review of Anthropology, 30, pp 181-207 ://WOS:000171900700011 AND http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Bellwood-Early- Agriculturalist-Populations-2001.pdf

Bellwood, P., Ehret, C., keita, S.O.Y., & Newmann, P. (2004) The origins of Afroasiatic - Response. Science, 306, 5702, pp 1681-1681 ://WOS:000225630800018 AND http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Bellwood-Ehret- Controversy-Science-2004.pdf

Belton, P.S., Delgadillo, I., Halford, N.G., & Shewry, P.R. (2006) Kafirin structure and functionality. Journal of Cereal Science, 44, 3, pp 272-286 ://000242699500005 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Belton-Kafirin-2006.pdf

Bennett, M.D. & Leitch, I.J. (1998) Nuclear DNA amount and genome size in angiosperms - Preface. Annals of Botany, 82, pp 1-1 ://000077598900001

Bennetzen, J.L. & Freeling, M. (1993) Grasses as a Single Genetic System - Genome Composition, Collinearity and Compatibility. Trends in Genetics, 9, 8, pp 259-261 ://A1993LN54200001 AND

Bergquist, R.R., Nubel, D.S., & D.L., T. (1998a) Production method for corn with enhanced quality grain traits. U.S. Patent 5 706 603, Date issued: 13 January, 1998, pp

Bergquist, R.R., Nubel, D.S., & Thompson, D.L. (1998b) Production method for high-oil corn grain. U.S. Patent 5 704 160., Date issued: 6 January, pp

Berhan, A.M., Hulbert, S.H., Butler, L.G., & Bennetzen, J.L. (1993) Structure and Evolution of the Genomes of Sorghum-Bicolor and Zea-Mays. Theoretical and Applied Genetics, 86, 5, pp 598-604 ://A1993LK33200011

Berthaud, J. (2001) Maize Diversity in Oaxaca, Mexico: Simple Questions but No Easy Answers, CIMMYT-Report pp 2 (Report) http://www.cimmyt.cgiar.org/whatiscimmyt/AR002001Spa/latinamerica/maize/diversity.htm#Maize

Beta, T. & Corke, H. (2001) Genetic and environmental variation in sorghum starch properties. Journal of Cereal Science, 34, 3, pp 261-268 ://000172194800004

Beta, T., Obilana, A.B., & Corke, H. (2001) Genetic diversity in properties of starch from Zimbabwean sorghum landraces. Cereal Chemistry, 78, 5, pp 583-589 ://000170887300014

309

Betts, K.J., Ehlke, N.J., Wyse, D.L., Gronwald, J.W., & Somers.D.A. (1992) Mechanism of inheritance of diclofop resistance in Italian rygrass (Lolium multiflorum). Weed Science, 40, pp 184-189

Bhat, K.V., Chandel, K.P.S., Shivakumar, T.M., & Sachan, J.K.S. (1998) Isozyme diversity in Indian primitive maize landraces. Journal of Plant Biochemistry and Biotechnology, 7, 1, pp 23-27 ://000073366200004

Bidari, V.B., Satyanarayan, H.V., Hegde, R.K., & Ponnappa, K.M. (1978) Effect of Fungicides against the Seed Rot and Seedling Blight of Hybrid Sorghum Csh-5. Mysore Journal of Agricultural Sciences, 12, 4, pp 587-593 ://A1978GW01400016

Biradar, S.G. & Borikar, S.T. (1985) Genetic-Analysis of Shootfly Resistance in Relation to Growth-Stages in Sorghum. Zeitschrift Fur Pflanzenzuchtung-Journal of Plant Breeding, 95, 2, pp 173-178 ://A1985AMZ0200010

Blum, A., Golan, G., & Mayer, J. (1991) Progress Achieved by Breeding Open-Pollinated Cultivars as Compared with Landraces of Sorghum. Journal of Agricultural Science, 117, pp 307-312 ://A1991GX18100004

Borges, A., Goncalves, L.C., Rodriguez, N.M., Zago, C.P., & Sampaio, I.B.M. (1997) Quality of hybrid sorghum silage with different tannin concentrations and stem moisture contents. Arquivo Brasileiro De Medicina Veterinaria E Zootecnia, 49, 4, pp 441-452 ://A1997YB40600006

Bothwell, T.H. (2000) Iron requirements in pregnancy and strategies to meet them. American Journal of Clinical Nutrition, 72, 1, pp 257S-263S ://000088005800007 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Bothwell-Iron-Strategy-2000.pdf

Bothwell, T.H. & MacPhail, A.P. (2004) The potential role of NaFeEDTA as an iron fortificant. International Journal for Vitamin and Nutrition Research, 74, 6, pp 421-434 ://000227171400004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Bothwell-NaFeEDTA-2004.pdf

Bowden, J. & Neve, R.A. (1953) Sorghum Midge and Resistant Varieties in the Gold Coast. Nature, 172, 4377, pp 551-551 ://A1953UA55900032

Bowers, J.E., Arias, M.A., Asher, R., Avise, J.A., Ball, R.T., Brewer, G.A., Buss, R.W., Chen, A.H., Edwards, T.M., Estill, J.C., Exum, H.E., Goff, V.H., Herrick, K.L., Steele, C.L.J., Karunakaran, S., Lafayette, G.K., Lemke, C., Marler, B.S., Masters, S.L., McMillan, J.M., Nelson, L.K., Newsome, G.A., Nwakanma, C.C., Odeh, R.N., Phelps, C.A., Rarick, E.A., Rogers, C.J., Ryan, S.P., Slaughter, K.A., Soderlund, C.A., Tang, H.B., Wing, R.A., & Paterson, A.H. (2005) Comparative physical mapping links conservation of microsynteny to chromosome structure and recombination in grasses. Proceedings of the National Academy of Sciences of the United States of America, 102, 37, pp 13206-13211 ://000231916300037 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Bowers-Comp-Sorghum- PNAS-2005.pdf

Bradley, P.R., Johnson, W.G., Hart, S.E., Buesinger, M.L., & Massey, R.E. (2000) Economics of weed management in glufosinate-resistant corn (Zea mays L.). Weed Technology, 14, 3, pp 495-501 ://000166693600007

Brandner, L. & Straus, M.A. (1959) Congruence Versus Profitability in the Diffusion of Hybrid Sorghum. Rural Sociology, 24, 4, pp 381-383 ://A1959CAW8700008

Brookes, G. (2004) Co-existence of GM and non GM crops: current experience and key principles, PG Economics Ltd pp 18 Dorchester, UK (Report) http://www.botanischergarten.ch/Coexistence/Brookes-Coexistence-Key-Principles-2004.pdf

Brookes, G. & Barfoot, P. (2003) Co-existence of GM and non GM arable crops: case study of the UK, PG Economics Ltd pp 27 Dorchester, UK (Report) http://www.botanischergarten.ch/Coexistence/Brookes-Coexistence-Casestudy-UK-2003.pdf

Brookes, G. & Barfoot, P. (2004) Co-existence of GM and non GM crops: case study of maize grown in Spain, PG Economics Ltd, pp 13 Dorchester, UK1 (Report) http://www.botanischergarten.ch/Coexistence/Brookes-Coexistence-Casestudy-Spain-2004.pdf

310

Brooks, T.M., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Rylands, A.B., Konstant, W.R., Flick, P., Pilgrim, J., Oldfield, S., Magin, G., & Hilton-Taylor, C. (2002) Habitat Loss and Extinction in the Hotspots of Biodiversity. Conservation Biology, 16, 4, pp 909-923 http://www.blackwell-synergy.com/doi/abs/10.1046/j.1523-1739.2002.00530.x AND http://www.botanischergarten.ch/biodiversity/Brooks-Biodiccenters-Loss-2002.pdf

Byrne, P. & Fromherz, S. (2003) Can GM and Non-GM Crops Coexist? Setting a Precedent in Boulder County, Colorado, USA. Journal of Food, Agriculture & Environment, 1, pp 258-261 http://www.botanischergarten.ch/Coexistence/Byrne-Fromherz-2003.pdf and www.world-food.net

Camire, M.E. (1998) Chemical changes during extrusion cookin. In Processed-induced Chemical Changes in Food (eds F. Shahidi, C.T. Ho & N.V. Chuyen), pp. 109–121. Pergamon Press, New York

Carvalho, C.H.S., Zehr, U.B., Gunaratna, N., Anderson, J., Kononowicz, H.H., Hodges, T.K., & Axtell, J.D. (2004) Agrobacterium-mediated transformation of sorghum: factors that affect transformation efficiency. Genetics and Molecular Biology, 27, 2, pp 259-269 ://000222989100022 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Carvalho-Factors- Transformation-Sorghum-2004.pdf

Casa, A.M., Mitchell, S.E., Hamblin, M.T., Sun, H., Bowers, J.E., Paterson, A.H., Aquadro, C.F., & Kresovich, S. (2005) Diversity and selection in sorghum: simultaneous analyses using simple sequence repeats. Theoretical and Applied Genetics, 111, 1, pp 23-30 ://000230141700004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Casa-Selection-2005.pdf

Casady, A.J. & Anderson, K.L. (1952) Hybridization, Cytological, and Inheritance Studies of a Sorghum Cross - Autotetraploid Sudangrass X (Johnsongrass X 4n Sudangrass). Agronomy Journal, 44, 4, pp 189-194 ://A1952YD73600007 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Casady-Hybrid-1952.pdf

Casady, A.J. & Miller, F.R. (1970) Inheritance of Hermaphrodite Pedicelled Spikelets of Sorghum. Crop Science, 10, 5, pp 612-& ://A1970H686700053

Casas, A.M., Kononowicz, A.K., Haan, T.G., Zhang, L.Y., Tomes, D.T., Bressan, R.A., & Hasegawa, P.M. (1997) Transgenic sorghum plants obtained after microprojectile bombardment of immature inflorescences. In Vitro Cellular & Developmental Biology-Plant, 33, 2, pp 92-100 ://A1997XL31700003 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Casas-transgenic-Sorghum- 1997.pdf

Casas, A.M., Kononowicz, A.K., Zehr, U.B., Tomes, D.T., Axtell, J.D., Butler, L.G., Bressan, R.A., & Hasegawa, P.M. (1993) Transgenic Sorghum Plants Via Microprojectile Bombardment. Proceedings of the National Academy of Sciences of the United States of America, 90, 23, pp 11212-11216 ://A1993MK09400068 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Casas-Transgenic-1993.pdf

Casler, M.D., Pedersen, J.F., & Undersander, D.J. (2003) Forage yield and economic losses associated with the brown-midrib trait in sudangrass. Crop Science, 43, 3, pp 782-789 ://000182540300005

Cavins, J.F., Blessin, C.W., & Inglett, G.E. (1972) Infusion of Grain-Sorghum with Lysine, Methionine, and Tryptophan. Cereal Chemistry, 49, 5, pp 605-& ://A1972N474200014

Celarier, R. (1958) Cytotaxonomic notes on the subsection halepensia of the genus Sorghum. Bull. Torrey Bot. Club, 85, pp 49-62

Chandrashekar, A. & Satyanarayana, K.V. (2006) Disease and pest resistance in grains of sorghum and millets. Journal of Cereal Science, 44, 3, pp 287-304 ://000242699500006 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Chandrashekar-Resistance- 2006.

Chantereau, J., Arnaud, M., Ollitrault, P., Nabayaogo, P., & Noyer, J.L. (1989) Study of Morphophysiological Diversity and Classification of Cultivated Sorghums. Agronomie Tropicale, 44, 3, pp 223-232 ://A1989DL10100009

Chao, W.S., Horvath, D.P., Anderson, J.V., & Foley, M.E. (2005) Potential model weeds to study genomics, ecology, and physiology in the 21st century. Weed Science, 53, 6, pp 929-937 311

://000233835300026

Chen, Z.J., Muthukrishnan, S., Liang, G.H., Schertz, K.F., & Hart, G.E. (1993) A Chloroplast DNA Deletion Located in Rna-Polymerase Gene Rpoc2 in Cms Lines of Sorghum. Molecular & General Genetics, 236, 2- 3, pp 251-259 ://A1993KK98500014

Cheng, M., Lowe, B.A., Spencer, T.M., Ye, X.D., & Armstrong, C.L. (2004) Factors influencing Agrobacterium-mediated transformation of monocotyledonous species. In Vitro Cellular & Developmental Biology-Plant, 40, 1, pp 31-45 ://000220431300003

Chevalier, A. (1947) *Geographie Botanique - Survivances Darbres Et Darbustes Du Desert Nubioarabo-Scindien Dans Le Sahara Central Et Occidental. Comptes Rendus Hebdomadaires Des Seances De L Academie Des Sciences, 225, 10, pp 424-426 ://A1947YG27000002

Childs, K.L., Klein, R.R., Klein, P.E., Morishige, D.T., & Mullet, J.E. (2001) Mapping genes on an integrated sorghum genetic and physical map using cDNA selection technology. Plant Journal, 27, 3, pp 243- 255 ://000170466800007 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Childs-Mapping-Sorghum- 2001.pdf

Chittenden, L.M., Schertz, K.F., Lin, Y.R., Wing, R.A., & Paterson, A.H. (1994) A Detailed Rflp Map of Sorghum-Bicolor X S-Propinquum, Suitable for High-Density Mapping, Suggests Ancestral Duplication of Sorghum Chromosomes or Chromosomal Segments. Theoretical and Applied Genetics, 87, 8, pp 925-933 ://A1994NE95700005 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Chittenden-RFLP-bic-prop- 1994.pdf

Chopra, K.R. (1987) Technical and economic aspects of seed production of hybrid varieties of sorghum. In Hybrid seed production of selected cereals, oil and vegetable crops. (eds W. Feistritzer & A. Kelly), pp. 193–217. FAO, Rome

Chopra, S., Brendel, V., Zhang, J.B., Axtell, J.D., & Peterson, T. (1999) Molecular characterization of a mutable pigmentation phenotype and isolation of the first active transposable element from Sorghum bicolor. Proceedings of the National Academy of Sciences of the United States of America, 96, 26, pp 15330-15335 ://000084375400117

Chougule, J.D. (1978) Effect of Incidences on Control of Sorghum Shoot Fly in Hybrid Sorghum. Indian Journal of Agronomy, 23, 2, pp 170-171 ://A1978GB17200018

Cisse, N. & Ejeta, G. (2003) Genetic variation and relationships among seedling vigor traits in sorghum. Crop Science, 43, 3, pp 824-828 ://000182540300010

Clark, J. (1976) Prehistoric populations and pressures favoring plant domestication in Africa. In Origins of African plant domestication. (eds J.R. Harlan, J.M.J. Dewet & A.B.L. Stemler). Mouton Press, The Hague

Clark, J. & Stemler, A.B.L. (1971) Evidence of agricultural origins in the Nile Valley. Proc. Prehist. Soc., 37, pp 34-79

Clayton, W. & Renvoize, S. (1986) Genera Graminum : grasses of the world. kew bulletin additional series, 13, pp 389

Clayton, W.D. (1981) Evolution and Distribution of Grasses. Annals of the Missouri Botanical Garden, 68, 1, pp 5-14 ://A1981MS42100002 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Clayton-Evolution-Grasses- 1981.pdf

Clayton, W.D., Harman, K.T., & Williamson, H. (2002 ff) Electronic Source: World Grass Species: Descriptions, Identification, and Information Retrieval, http://www.rbgkew.org.uk/herbarium/gramineae/wrldgr.htm

Clayton, W.D. & Renvoize, S.A. (1986, repr. 1999) Genera Graminum: Grasses of the World Kew Publishing, 1986, reprinted 1999, IS: 0 900347 75 X, pp 389pp http://www.kewbooks.com/asps/ShowDetails.asp?id=71 312

Collins, F.C. & Pickett, R.C. (1972a) Combining Ability for Grain Yield, Percent Protein, and G Lysine/100 G Protein in a 9-Parent Diallel of Sorghum-Bicolor (L) Moench. Crop Science, 12, 4, pp 423-& ://A1972N361300007

Collins, F.C. & Pickett, R.C. (1972b) Combining Ability for Yield, Protein, and Lysine in an Incomplete Diallel of Sorghum-Bicolor(L.)Moench. Crop Science, 12, 1, pp 5-& ://A1972L705300002

Cone, K.C., McMullen, M.D., Bi, I.V., Davis, G.L., Yim, Y.S., Gardiner, J.M., Polacco, M.L., Sanchez-Villeda, H., Fang, Z.W., Schroeder, S.G., Havermann, S.A., Bowers, J.E., Paterson, A.H., Soderlund, C.A., Engler, F.W., Wing, R.A., & Coe, E.H. (2002) Genetic, physical, and informatics resources for maize on the road to an integrated map. Plant Physiology, 130, 4, pp 1598-1605 ://000179990100004 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Cone-Integrated-Map-2002.pdf

Connah, G. (1971) Recent Contributions to Bornu Chronology. West African Journal of Archaeology, 1, pp 55-60 ://A1971Y017900004

Conner, A. & Karper, R. (1927) Hybrid vigour in sorghum, Texas A&M University College Station pp Texas Agricultural Experiment Station Bulletin Texas, USA (Report)

Conner, A.J., Travis R. Glare and Jan-Peter Nap (2003) The release of genetically modified crops into the environment. Plant Journal, 33, pp 19 - 46 http://www.botanischergarten.ch/Geneflow/Connor-Management-2003.pdf

Copelin, J.L., Gaskins, C.T., & Tribble, L.F. (1978) Availability of Tryptophan, Lysine and Threonine in Sorghum for Swine. Journal of Animal Science, 46, 1, pp 133-142 ://A1978ER06600017

Copelin, J.L., Sasse, C.E., & Tribble, L.F. (1976) Availability of Lysine in Sorghum. Journal of Animal Science, 42, 1, pp 245-245 ://A1976BD84100046

Cornelissen, J.H.C., Lavorel, S., Garnier, E., Diaz, S., Buchmann, N., Gurvich, D.E., Reich, P.B., ter Steege, H., Morgan, H.D., van der Heijden, M.G.A., Pausas, J.G., & Poorter, H. (2003) A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian Journal of Botany, 51, 4, pp 335-380 ://000184733000001

Cox, T.S., Bender, M., Picone, C., Van Tassel, D.L., Holland, J.B., Brummer, E.C., Zoeller, B.E., Paterson, A.H., & Jackson, W. (2002) Breeding perennial grain crops. Critical Reviews in Plant Sciences, 21, 2, pp 59-91 ://000174854200001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Cox-Perennial-Crops-2002.pdf

Crasta, O.R., Xu, W.W., Rosenow, D.T., Mullet, J., & Nguyen, H.T. (1999) Mapping of post-flowering drought resistance traits in grain sorghum: association between QTLs influencing premature senescence and maturity. Molecular and General Genetics, 262, 3, pp 579-588 ://000083839000022

Craufurd, P.Q., Flower, D.J., & Peacock, J.M. (1993) Effect of Heat and Drought Stress on Sorghum (Sorghum-Bicolor) .1. Panicle Development and Leaf Appearance. Experimental Agriculture, 29, 1, pp 61-76 ://A1993KT38200007

Crawley, M.J., Brown, S.L., Hails, R.S., Kohn, D.D., & Ress, M. (2001) Transgenic crops in natural habitats. Nature, 409, pp 682-683

Cui, Y.X., Xu, G.W., Magill, C.W., Schertz, K.F., & Hart, G.E. (1995) Rflp-Based Assay of Sorghum-Bicolor (L) Moench Genetic Diversity. Theoretical and Applied Genetics, 90, 6, pp 787-796 ://A1995RB06600006 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Cui-bicolor-1995.pdf

Da Silva, J., Honeycutt, R., Burnquist, W., Al-Janabi, S., Sorrells, M., & Tanksley, S. (1995) Saccharum spontaneum L. ‘SES 208’ genetic linkage map combining RFLP- and PCR-based markers. Molecular Breeding, 1, pp 165- 171

Dahlberg, J.A., Bandyopadhyay, R., Rooney, W.L., Odvody, G.N., & Madera-Torres, P. (2001) 313

Evaluation of sorghum germplasm used in US breeding programmes for sources of sugary disease resistance. Plant Pathology, 50, 6, pp 681-689 ://000172806100004

Dahlberg, J.A., Burke, J.J., & Rosenow, D.T. (2004) Development of a sorghum core collection: Refinement and evaluation of a subset from Sudan. Economic Botany, 58, 4, pp 556-567 ://000227297600007 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Dahlberg-Development-Core- Coll-2004.pdf

Dahlberg, J.A., Zhang, X., Hart, G.E., & Mullet, J.E. (2002) Comparative assessment of variation among sorghum germplasm accessions using seed morphology and RAPD measurements. Crop Science, 42, 1, pp 291-296 ://000173296000042 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Dahlberg-Comparative- Assessment-2002.pdf

Damania, A.B., Pecetti, L., Qualset, C.O., & Humeid, B.O. (1996) Diversity and geographic distribution of adaptive traits in Triticum turgidum L (durum group) wheat landraces from Turkey. Genetic Resources and Crop Evolution, 43, 5, pp 409-422 ://A1996VF75500003

Davidse, G. & Turland, N.J. (2001) (1480) Proposal to Reject the Name Holcus saccharatus (Poaceae). Taxon, 50, 2, pp 577-580 http://www.jstor.org/stable/1223906 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Davidse-Reject-Holcus- saccharatus-2001.pdf

Davis, M.A. (2003) Biotic globalization: Does competition from introduced species threaten biodiversity? Bioscience, 53, 5, pp 481-489 ://000182833000011 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Davis-Competition-2003.pdf de Vries, A., Meijden, V., & Brandenburg, W. (1992) Botanical files – a study of the real chances for spontaneous gene flow from cultivated plants of the wild flora of the Netherlands. Gorteria, Supplement 1, pp 100

De Wet, J., Harlan, J., & Kurmarohita, B. (1972) Origin and evolution of guinea sorghums. East African Agricultural and Forestry Journal, 38, pp 114-119 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/DeWet-Guineas-1972.pdf

De Wet, J.M.J. (1978) Systematics and Evolution of Sorghum-Sect Sorghum (Gramineae). American Journal of Botany, 65, 4, pp 477-484 ://A1978EZ09100015 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/de-Wet-Syst-Evol-Sorghum- 1978.pdf

De Wet, J.M.J. & Harlan, J.R. (1971) Origin and Domestication of Sorghum-Bicolor. Economic Botany, 25, 2, pp 128-& ://A1971J685300002

De Wet, J.M.J. & Harlan, J.R. (1975) Weeds and Domesticates - Evolution in Man-Made Habitat. Economic Botany, 29, 2, pp 99-107 ://A1975AN46400001 AND http://www.botanischergarten.ch/Weeds/Dewet-Harlan-Weeds-Domest-1975.pdf

De Wet, J.M.J., Harlan, J.R., & Price, E.G. (1970) Origin of Variability in Spontanea Complex of Sorghum Bicolor. American Journal of Botany, 57, 6, pp 704-& ://A1970G675500013

De Wet, J.M.J. & Huckabay, J.P. (1967) The Origin of Sorghum bicolor II: Distribution and Domestication. Evolution, 21, pp 787-802 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Dewet-Origin-1967.pdf

Defelice, M.S. (1990) Shattercane (Sorghum-Bicolor) Control in Soybeans (Glycine-Max) with Preplant and Postemergence Herbicides. Weed Technology, 4, 4, pp 776-780 ://A1990EY83300014

DEFRA (2006) Electronic Source: Consultation on proposals for managing the coexistence of GM, conventional and organic crops (ed D.o.E. DEFRA, Food and rural Affairs), published by: DEFRA 314

www.defra.gov.uk AND http://www.defra.gov.uk/corporate/consult/gmnongm-coexist/consultdoc.pdf AND http://www.botanischergarten.ch/Coexistence/DEFRA-consultdoc-2006.pdf

Deines, S.R., Dille, J.A., Blinka, E.L., Regehr, D.L., & Staggenborg, S.A. (2004) Common sunflower (Helianthus annuus) and shattercane (Sorghum bicolor) interference in corn. Weed Science, 52, 6, pp 976-983 ://000225150300010

Denman, C.E., McNew, R.W., Merritt, R.G., Reeves, H.E., Belete, K., & Wenger, R.F. (1984) Performance Trials for Hybrid Sorghum and Corn in Oklahoma, 1983. Oklahoma Agricultural Experiment Station Research Report, NP- 84, pp 1-42 ://A1984SK23200001

Denman, C.E., McNew, R.W., Merritt, R.G., Reeves, H.E., Strickland, G.L., Belete, K., & Ekpin, E.U. (1985) Performance Trials for Hybrid Sorghum and Corn in Oklahoma, 1984. Oklahoma Agricultural Experiment Station Research Report, NP- 86, pp 2-38 ://A1985AFM3800001 deOliveira, A.C., Richter, T., & Bennetzen, J.L. (1996) Regional and racial specificities in sorghum germplasm assessed with DNA markers. Genome, 39, 3, pp 579-587 ://A1996UQ16000013 http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=8675002&dopt=Citation AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Oliveira-Regional-DNA-Markers-1996.pdf

Deosthal.Yg & Mohan, V.S. (1970) Locational Differences in Protein, Lysine and Leucine Content of Sorghum Varieties. Indian Journal of Agricultural Sciences, 40, 11, pp 935-& ://A1970H985300001

Deu, M., Gonzalezdeleon, D., Glaszmann, J.C., Degremont, I., Chantereau, J., Lanaud, C., & Hamon, P. (1994) Rflp Diversity in Cultivated Sorghum in Relation to Racial Differentiation. Theoretical and Applied Genetics, 88, 6-7, pp 838-844 ://A1994PB68100032 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Deu-RFLP-Diversity-1994.pdf

Deu, M., Hamon, P., Bonnot, F., & Chantereau, J. (2003) Sorghum. In Genetic diversity of cultivated tropical plants. (eds P. Hamon, M. Seguin, X. Perrier & J. Glaszmann), pp. 307-336. Science Publishers, Inc., and CIRAD

Deu, M., Hamon, P., Chantereau, J., Dufour, P., Dhont, A., & Lanaud, C. (1995) Mitochondrial-DNA Diversity in Wild and Cultivated Sorghum. Genome, 38, 4, pp 635-645 ://A1995RR17400002

Deu, M., Rattunde, F., & Chantereau, J. (2006) A global view of genetic diversity in cultivated sorghums using a core collection. Genome, 49, 2, pp 168-180 ://000236286100009 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Deu-Globalview-Genetics- Sorghum-2006.pdf AND http://www.botanischergarten.ch/Facultyof1000/Deu-F1000-2006.pdf

Devi, P.B. & Sticklen, M.B. (2003) In vitro culture and genetic transformation of sorghum by microprojectile bombardment. Plant Biosystems, 137, 3, pp 249-254 ://000188593900002

Devine, M.D. & Shimabukuro, R.H. (1994) Resistance to acetyl coenzyme A carboxylase inhibiting herbicides. In Herbicide Resistance in Plants. (eds S.B. Powles & J.A.M. Holtum), pp. 141-169. CRC Press, Boca Raton, FL, USA

Devries, A.P. (1971) Flowering Biology of Wheat, Particularly in View of Hybrid Seed Production - Review. Euphytica, 20, 2, pp 152-& ://A1971J654900002

Devries, A.P. (1972) Some Aspects of Cross-Pollination in Wheat (Triticum-Aestivum L) .1. Pollen Concentration in Field as Influenced by Variety, Diurnal Pattern, Weather Conditions and Level as Compared to Height of Pollen Donor. Euphytica, 21, 2, pp 185-& ://A1972N060300004

Di-Giovanni, F. & Kevan, P. (1991) Factors affecting pollen dynamics and its importance to pollen contamination: A review. CAN. J. FOR. RES./J. CAN. RECH. FOR., 21, 8, pp 1155-1170

Diarisso, N.Y., Pendleton, B.B., Teetes, G.L., Peterson, G.C., & Anderson, R.M. (1998) 315

Spikelet flowering time: Cause of sorghum resistance to sorghum midge (Diptera : Cecidomyiidae). Journal of Economic Entomology, 91, 6, pp 1464-1470 ://000077708400035

Dicko, M.H., Gruppen, H., Traore, A.S., Voragen, A.G.J., & van Berkel, W.J.H. (2006) Sorghum grain as human food in Africa: relevance of content of starch and amylase activities. African Journal of Biotechnology, 5, 5, pp 384-395 ://000237013800001

Dillon, S.L., Lawrence, P.K., & Henry, R.J. (2001) The use of ribosomal ITS to determine phylogenetic relationships within Sorghum. Plant Systematics and Evolution, 230, 1-2, pp 97- 110 ://000173035900007 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Dillon-Phylogenetic-Sorghum- 2001.pdf

Dillon, S.L., Lawrence, P.K., Henry, R.J., Ross, L., Price, H.J., & Johnston, J.S. (2004) Sorghum laxiflorum and S-macrospermum, the Australian native species most closely related to the cultivated S-bicolor based on ITS1 and ndhF sequence analysis of 25 Sorghum species. Plant Systematics and Evolution, 249, 3-4, pp 233-246 ://000225206000007 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Dillon-Sequence-25-Sorghum- 2004.pdf

Dingkuhn, M., Singh, B.B., Clerget, B., Chantereau, J., & Sultan, B. (2006) Past, present and future criteria to breed crops for water-limited environments in West Africa. Agricultural Water Management, 80, 1-3, pp 241-261 ://000235558000017

Dixon, A.G.O., Bramelcox, P.J., & Reese, J.C. (1990) Feeding-Behavior of Biotype-E Greenbug (Homoptera, Aphididae) and Its Relationship to Resistance in Sorghum. Journal of Economic Entomology, 83, 1, pp 241-246 ://A1990CM66400040

Djè, Y., Ater, M., Lefebvre, C., & Vekemans, X. (1998) Patterns of morphological and allozyme variation in sorghum landraces of northwestern Morocco. Genetic Resources and Crop Evolution, 45, 6, pp 541-548 ://000077662600007 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Dje-Morpho-Allo-Morocco- 1998.pdf

Djè, Y., Forcioli, D., Ater, M., Lefèbvre, C., & Vekemans, X. (1999) Assessing population genetic structure of sorghum landraces from North-western Morocco using allozyme and microsatellite markers. TAG Theoretical and Applied Genetics, 99, 1 - 2, pp 157-163 http://springerlink.metapress.com/openurl.asp?genre=article&id=doi:10.1007/s001220051220 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Dje-PopGen-Sorghum-1999.pdf

Djè, Y., Heuertz, M., Ater, M., Lefebvre, C., & Vekemans, X. (2004) In situ estimation of outcrossing rate in sorghum landraces using microsatellite markers. Euphytica, 138, 3, pp 205-212 ://000224911700002 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Dje-Outcrossing-Marker- 2004.pdf

Dje, Y., Heuertz, M., Lefebvre, C., & Vekemans, X. (2000) Assessment of genetic diversity within and among germplasm accessions in cultivated sorghum using microsatellite markers. Theoretical and Applied Genetics, 100, 6, pp 918-925 ://000087061100014 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Dje-Diversity-2000.pdf

Doebley, J. & Stec, A. (1991) Genetic-Analysis of the Morphological Differences between Maize and Teosinte. Genetics, 129, 1, pp 285-295 ://A1991GD51200027

Doesthal.Yg, Mohan, V.S., & Rao, K.V. (1970) Varietal Differences in Protein, Lysine, and Leucine Content of Grain Sorghum. Journal of Agricultural and Food Chemistry, 18, 4, pp 644-& ://A1970G822000023

Doggett, H. (1962) Tetraploid Hybrid Sorghum. Nature, 196, 4856, pp 755-& ://A19628800B00025

Doggett, H. (1964a) Fertility Improvement in Autotetraploid Sorghum .2. Sorghum Almum Derivatives. Heredity, 19, 4, pp 543-& 316

://A19642238B00004

Doggett, H. (1964b) Fertility Improvement in Autotetraploid Sorghum .I. Cultivated Autotetraploids. Heredity, 19, 3, pp 403-& ://A19642237B00008

Doggett, H. (1965) The development of cultivated sorghums. In In: Crop plant evolution (ed J. Hutchinson). Cambridge University Press, London

Doggett, H. (1969) Yields of Hybrid Sorghums. Experimental Agriculture, 5, 1, pp 1-& ://A1969C663900001

Doggett, H. (1972) Recurrent Selection in Sorghum Populations. Heredity, 28, FEB, pp 9-& ://A1972L845400002

Doggett, H. (1976) Sorghum: Sorghum bicolor (Gramineae-Andropogoneae). In Evolution of Crop Plants (ed N.W. Simmonds), pp. 112-117. Longman, London

Doggett, H. (1981) The improvement of nutritional quality by genetic means, Tokyo Proceedings of a workshop held at Hyderabad, India, 10-12 November 1981, cosponsored by the United Nations University (UNU), the international Crops Research Institute for the Semi-arid Tropics (ICRISAT), the National Institute of Nutrition (NIN), and the Central Food Technological Research Institute (CFTRI), Session 1: interaction at the production stage Ed. pp http://www.unu.edu/unupress/food/unu01/cap_5.htm AND http://www.unu.edu/unupress/food/unu01/cap_5.htm#The%20improvement%20of%20nutritional%20quality%20by%20gen AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Doggett-Lysine-1981.pdf

Doggett, H. (1988) Sorghum, 2nd Ed. edn. Longman Scientific & Technical, Published in Association with the International Development Research Centre in Canada, copublished in the USA by John Wiley and Sons, Inc., New York, IS: 0-582-46345-9, 0-470-20984-4, pp 512

Doggett, H. (1991) Sorghum history in relation to Ethiopia. In Plant Genetic Resources of Ethiopia (eds J. Engels, J. Hawks & M. Worede), pp. 140-159. Cambridge University Press, Cambridge

Doggett, H. & Majisu, B.N. (1968) Disruptive selection in crop development. Heredity, 23, pp 1-26 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Doggett-Disruptive-1968.pdf

Doggett, H. & Majisu, B.N. (1972) Fertility Improvement in Autotetraploid Sorghum .3. Yields of Cultivated Tetraploids. Euphytica, 21, 1, pp 86-& ://A1972M112500010 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Doggett-Fertility-1972.pdf

Doggett, H. & Parasada, R.K.E. (1995) Sorghum. In Evolution of crop plants 2nd edition (eds J. Smartt & N. Simmonds), pp. 173-180. Chapman, London, UK.

Drabkova, L., Kirschner, J., & Vlcek, C. (2002) Comparison of seven DNA extraction and amplification protocols in historical herbarium specimens of Juncaceae. Plant Molecular Biology Reporter, 20, 2, pp 161-175 ://000177202200006 AND http://www.botanischergarten.ch/Herbarien/Drabovka-DNA-extration-2002.pdf

Draye, X., Lin, Y.R., Qian, X.Y., Bowers, J.E., Burow, G.B., Morrell, P.L., Peterson, D.G., Presting, G.G., Ren, S.X., Wing, R.A., & Paterson, A.H. (2001) Toward integration of comparative genetic, physical, diversity, and cytomolecular maps for grasses and grains, using the sorghum genome as a foundation. Plant Physiology, 125, 3, pp 1325-1341 ://000168179800022 and http://www.botanischergarten.ch/Africa-Harvest-Geneflow/Draye-2001.pdf

Dsouza, L. (1970) Investigations Concerning Suitability of Wheat as Pollen-Donor for Cross-Pollination by Wind as Compared to Rye, Triticale, and Secalotricum. Zeitschrift Fur Pflanzenzuchtung, 63, 3, pp 246-& ://A1970H183500007

Duncan, R.R., Waskom, R.M., & Nabors, M.W. (1995) In-Vitro Screening and Field-Evaluation of Tissue-Culture-Regenerated Sorghum (Sorghum-Bicolor (L) Moench) for Soil Stress Tolerance. Euphytica, 85, 1-3, pp 373-380 317

://A1995TF37600047

Duvall, M.R. & Doebley, J.F. (1990) Restriction Site Variation in the Chloroplast Genome of Sorghum (Poaceae). Systematic Botany, 15, 3, pp 472-480 ://A1990DT08100016 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Duvall-Restric-Chloropl- Sorghum-1990.pdf

Dweikat, I. (2005) A diploid, interspecific, fertile hybrid from cultivated sorghum, Sorghum bicolor, and the common Johnsongrass weed Sorghum halepense. Molecular Breeding, 16, 2, pp 93-101 ://000232919000001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Dweikat-Hybrid-2005.pdf

Dykes, L. & Rooney, L.W. (2006) Sorghum and millet phenols and antioxidants. Journal of Cereal Science, 44, 3, pp 236-251 ://000242699500003 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Dykes-Phenols-2006.pdf

Eastham, C. & Sweet, J. (2002) Genetically modified organisms (GMOs): The significance of gene flow through pollen transfer, European Environment Agency pp 75 Copenhagen (Report) http://reports.eea.eu.int/environmental_issue_report_2002_28/en

Ejeta, G. & Axtell, J. (1987) Protein and Lysine Levels in Developing Kernels of Normal and High-Lysine Sorghum. Cereal Chemistry, 64, 3, pp 137-139 ://A1987H530800001

Ejeta, G. & Grenier, C. (2005) Sorghum and its weedy hybrids. In Crop Ferality and Volunteerism (ed J. Gressel). CRC Press, Boca Raton http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Ejeta-Feral-Chapt8-Sorghum.pdf

Elkonin, L.A., Enaleeva, N.K., Tsvetova, M.I., Belyaeva, E.V., & Ishin, A.G. (1995) Partially Fertile Line with Apospory Obtained from Tissue-Culture of Male-Sterile Plant of Sorghum (Sorghum-Bicolor L Moench). Annals of Botany, 76, 4, pp 359-364 ://A1995RZ64900005 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Elkonin-Apomixis-Sorghum- 1995.pdf

Ellstrand, N.C. (2002) Gene Flow from Transgenic Crops to Wild Relatives: What Have We Learned, What Do We Know, What Do We Need to Know?, The University Plaza Hotel and Conference Center The Ohio State University Columbus, OH Ohio State University Ecological and Agronomic Consequences of Gene Flow from Transgenic Crops to Wild Relatives Ed. A. Snow pp http://www.biosci.ohio-state.edu/~asnowlab/Proceedings.pdf

Ellstrand, N.C. (2003) Dangerous liaisons – when cultivated plants mate with their wild relatives Johns Hopkins University Press, Baltimore MD, pp 244

Ellstrand, N.C. & Elam, D.R. (1993) Population Genetic Consequences of Small Population Size: Implications for Plant Conservation. Annual Review of Ecology and Systematics, 24, 1, pp 217-242 http://arjournals.annualreviews.org/doi/abs/10.1146/annurev.es.24.110193.001245 AND

Ellstrand, N.C. & Foster, K.W. (1983) Impact of population structure on the apparent outcrossing rate of grain sorghum (Sorghum bicolor). TAG Theoretical and Applied Genetics, 66, 3 - 4, pp 323-327 http://springerlink.metapress.com/openurl.asp?genre=article&id=doi:10.1007/BF00251167

Ellstrand, N.C., Prentice, H.C., & Hancock, J.F. (1999) Gene flow and introgression form domesticated plants into their wild relatives. Annual Review of Ecology and Systematics, 30, 1, pp 539-563 http://arjournals.annualreviews.org/loi/ecolsys

Ellstrand, N.C. & Schierenbeck, K.A. (2000) Hybridization as a stimulus for the evolution of invasiveness in plants? PNAS, 97, 13, pp 7043-7050 http://www.pnas.org/cgi/content/abstract/97/13/7043

Emani, C., Sunilkumar, G., & Rathore, K.S. (2002) Transgene silencing and reactivation in sorghum. Plant Science, 162, 2, pp 181-192 ://000173673400001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-LitEmani-GeneSwitching-Sorghum- 2001.pdf

318

Ercoli, L., Mariotti, M., Masoni, A., & Massantini, F. (1996) Effect of temperature and phosphorus fertilization on phosphorus and nitrogen uptake by sorghum. Crop Science, 36, 2, pp 348-354 ://A1996UE90300023

Evans, L.T. (1998) Feeding the Ten Billion: Plants and Population Growth, Cambridge University Press, Cambridge, UK., pp p. 34

FAO (2004) Production Yearbook 2002 No. 56 176 FAO, Rome, pp

Feil, B., Weingartner, U., & Stamp, P. (2003) Controlling the release of pollen from genetically modified maize and increasing its grain yield by growing mixtures of male-sterile and male-fertile plants. Euphytica, 130, 2, pp 163-165 http://www.botanischergarten.ch/Terminator/Feil-Stamp-Euphytica-Maize.pdf

Fellows, G.M. & Roeth, F.W. (1992) Shattercane (Sorghum-Bicolor) Interference in Soybean (Glycine-Max). Weed Science, 40, 1, pp 68-73 ://A1992HM38500013

Feltus, F.A., Hart, G.E., Schertz, K.F., Casa, A.M., Kresovich, S., Abraham, S., Klein, P.E., Brown, P.J., & Paterson, A.H. (2006) Alignment of genetic maps and QTLs between inter- and intra-specific sorghum populations. Theoretical and Applied Genetics, 112, 7, pp 1295-1305 ://000236585600011 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Feltus-Genetic-Maps-QTLs- Sorghum-2006.pdf

Feuillet, C. & Keller, B. (2002) Comparative Genomics in the grass family: Molecular characterization of grass genome structure and evolution. Annals of Botany, 89, 1, pp 3-10 ://000174774200002

Flannery, M.-L., Meade, C., & Mullins, E. (2005) Employing a composite gene-flow index to numerically quantify a crop’s potential for gene flow: an Irish perspective. Environ. Biosafety Res., 4, pp 29-43 http://www.edpsciences.org/ebr AND http://www.botanischergarten.ch/Africa-Harvest-Geneflow/Flannery-ebr0418.pdf

Folkertsma, R.T., Rattunde, H.F.W., Chandra, S., Raju, G.S., & Hash, C.T. (2005) The pattern of genetic diversity of Guinea-race Sorghum bicolor (L.) Moench landraces as revealed with SSR markers. Theoretical and Applied Genetics, 111, 3, pp 399-409 ://000231346200001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Folkertsma-Guinea-Diversity- 2005.pdf

Fontaine, J., Schirmer, B., & Horr, J. (2002) Near-infrared reflectance spectroscopy (NIRS) enables the fast and accurate prediction of essential amino acid contents. 2. Results for wheat, barley, corn, triticale, wheat bran/middlings, rice bran, and sorghum. Journal of Agricultural and Food Chemistry, 50, 14, pp 3902-3911 ://000176550500003

Forrester, N.W. (1994) Resistance Management Options for Conventional Bacillus-Thuringiensis and Transgenic Plants in Australian Summer Field Crops. Biocontrol Science and Technology, 4, 4, pp 549-553 ://A1994QU28100020

Frankel, O.H. & Hawkes, J.G. (1975) Crop Genetic Resources for Today and Tomorrow Cambridge University Press, Cambridge, England, pp

Franks, C.D., Burow, G.B., & Burke, J.J. (2006a) A comparison of US and Chinese sorghum germplasm for early season cold tolerance. Crop Science, 46, 3, pp 1371-1376 ://000237375900047 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Franks-Comp-US-China- Sorghum.2006.pdf

Franks, T.K., Powell, K.S., Choimes, S., Marsh, E., Iocco, P., Sinclair, B.J., Ford, C.M., & van Heeswijck, R. (2006b) Consequences of transferring three sorghum genes for secondary metabolite (cyanogenic glucoside) biosynthesis to grapevine hairy roots. Transgenic Research, 15, 2, pp 181-195 ://000236685800005

Frederiksen, R.A. & Rosenow, D.T. (1980) Breeding for Disease Resistance in Sorghum. Texas Agricultural Experiment Station Miscellaneous Publication, 1451, pp 137-167 ://A1980KS86700013 319

Fridley, J.D., Stachowicz, J.J., Naeem, S., Sax, D.F., Seabloom, E.W., Smith, M.D., Stohlgren, T.J., Tilman, D., & Von Holle, B. (2007) The invasion paradox: Reconciling pattern and process in species invasions. Ecology, 88, 1, pp 3-17 ://WOS:000245668300002 AND http://www.botanischergarten.ch/Invasive/Fridley-Invasion-Paradox-2007.pdf

Frietema, D.V.F. (1996) Cultivated plants and the wild flora. Effect analysis by dispersal codes, University of Leiden, Leiden Thesis, pp 100

Fuller, T. (1975) Bantu Civilization of Southern-Africa - Murphy,Ej. International Journal of African Historical Studies, 8, 4, pp 678-680 ://A1975BX80800012

Gao, Z.S., Jayaraj, J., Muthukrishnan, S., Claflin, L., & Liang, G.H. (2005a) Efficient genetic transformation of Sorghum using a visual screening marker. Genome, 48, 2, pp 321-333 ://000229115700017 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Gao-Efficient-Visual-Marker- 2005.pdf

Gao, Z.S., Xie, X.J., Ling, Y., Muthukrishnan, S., & Liang, G.H. (2005b) Agrobacterium tumefaciens-mediated sorghum transformation using a mannose selection system. Plant Biotechnology Journal, 3, 6, pp 591-599 ://000232537600005

Garber, E.D. (1950) Cytotaxonomic Studies in the Genus Sorghum. Genetics, 35, 1, pp 107-107 ://A1950XW83800034

Garber, E.D. (1954) Cytotaxonomic Studies in the Genus Sorghum .3. The Polyploid Species of the Subgenera Para-Sorghum and Stiposorghum. Botanical Gazette, 115, 4, pp 336-342 ://A1954XW28000004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Garber-Cytotax-Sorghum- 1954.pdf

Gatz, C. (1996) Chemically inducible promoters in transgenic plants. Current Opinion in Biotechnology, 7, 2, pp 168-172 ://A1996UE90500008

Gatz, C. (1997) Chemical control of gene expression. Annual Review of Plant Physiology and Plant Molecular Biology, 48, pp 89-108 ://A1997XD14000005

Gatz, C. (1998) An intoxicating switch for plant transgene expression. Nature Biotechnology, 16, 2, pp 140-140 ://000071831300020

Gatz, C. & Lenk, I. (1998) Promoters that respond to chemical inducers. Trends in Plant Science, 3, 9, pp 352-358 ://000075885000008

Gebrekidan, B. (1985) Breeding Sorghum for Resistance to Insects in Eastern-Africa. Insect Science and Its Application, 6, 3, pp 351-357 ://A1985AVM5900017

Gepts, P. (2002) A comparison between crop domestication, classical plant breeding, and genetic engineering. Crop Sci., 42, pp 1780-1790

Gepts, P. & Papa, R. (2003) Possible effects of (trans)gene flow from crops on the genetic diversity from landraces and wild relatives. Environ. Biosafety Res., 2, pp 89-103 http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Gepts-Possible-Effects-transgeneflow-2003.pdf

Giri, A.N. & Bainade, S.S. (1981) Studies on Intercropping System with Hybrid Sorghum-Csh-6. Indian Journal of Agronomy, 26, 3, pp 351-352 ://A1981NX48700032

Girijashankar, V., Sharma, H.C., Sharma, K.K., Swathisree, V., Prasad, L.S., Bhat, B.V., Royer, M., San Secundo, B., Narasu, M.L., Altosaar, I., & Seetharama, N. (2005) Development of transgenic sorghum for insect resistance against the spotted stem borer (Chilo partellus). Plant Cell Reports, 24, 9, pp 513-522 320

://000232844400003

Gladis, T. (1966) Unkräuter als Genressourcen. Z Pflanzenkr Pflanzenschutz, 15, pp 39–43

Gladis, T. & Hammer, J. (2000) The relevance of plant genetic resources in plant breeding. In FAL agricultural research: animal breeding and animal genetic resources, special issue 228, (ed E.G. Groeneveld, P (eds)), pp. 3–13

Gomez, M.I., Islam-Faridi, M.N., Zwick, M.S., Czeschin, D.G., Hart, G.E., Wing, R.A., Stelly, D.M., & Price, H.J. (1998) Tetraploid nature of sorghum bicolor (L.) Moench. Journal of Heredity, 89, 2, pp 188-190 ://000073364100015 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Gomez-Tetraploid-1988.pdf

Gomez, M.I., IslamFaridi, M.N., Woo, S.S., Schertz, K.F., Czeschin, D., Zwick, M.S., Wing, R.A., Stelly, D.M., & Price, H.J. (1997) FISH of a maize sh2-selected sorghum BAC to chromosomes of Sorghum bicolor. Genome, 40, 4, pp 475-478 ://A1997XR48000009

Gould, F. (1968) Grass Systematics McGraw Hill, New York, pp p. 382

Gourou, P. (1970) l'Afrique Hachette, Paris, France, pp

Grabber, J.H., Jung, G.A., Abrams, S.M., & Howard, D.B. (1992) Digestion Kinetics of Parenchyma and Sclerenchyma Cell-Walls Isolated from Orchardgrass and Switchgrass. Crop Science, 32, 3, pp 806-810 ://A1992JB24000047

Graner, A., Ludwig, W.F., & Melchinger, A.E. (1994) Relationships among European Barley Germplasm .2. Comparison of Rflp and Pedigree Data. Crop Science, 34, 5, pp 1199-1205 ://A1994PF01600010

Grenier, C., Bramel-Cox, P.J., & Hamon, P. (2001a) Core collection of sorghum: I. Stratification based on eco-geographical data. Crop Science, 41, 1, pp 234-240 ://000169469600037 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Grenier-Core-Collection-I- 2001.pdf

Grenier, C., Bramel-Cox, P.J., Noirot, M., Rao, K.E.P., & Hamon, P. (2000a) Assessment of genetic diversity in three subsets constituted from the ICRISAT sorghum collection using random vs non-random sampling procedures A. Using morpho-agronomical and passport data. Theoretical and Applied Genetics, 101, 1-2, pp 190-196 ://000088403800028 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Grenier-Assessment-A- morpho-2000.pdf

Grenier, C., Bramel, P.J., Dahlberg, J.A., El-Ahmadi, A., Mahmoud, M., Peterson, G.C., Rosenow, D.T., & Ejeta, G. (2004) Sorghums of the Sudan: analysis of regional diversity and distribution. Genetic Resources and Crop Evolution, 51, 5, pp 489-500 ://000220869000004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Grenier-Sorghum-Sudan- 2004.pdf

Grenier, C., Deu, M., Kresovich, S., Bramel-Cox, P.J., & Hamon, P. (2000b) Assessment of genetic diversity in three subsets constituted from the ICRISAT sorghum collection using random vs non-random sampling procedures B. Using molecular markers. Theoretical and Applied Genetics, 101, 1-2, pp 197-202 ://000088403800029 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Grenier-Assessment-B-molec- 2000.pdf

Grenier, C., Hamon, P., & Bramel-Cox, P.J. (2001b) Core collection of sorghum: II. Comparison of three random sampling strategies. Crop Science, 41, 1, pp 241-246 ://000169469600038 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Grenier-Core-Collection-II- 2001.pdf

Gressel, J. (1978) Factors Influencing the Selection of Herbicide-Resistant Biotypes of Weeds. Outlook on Agriculture, 9, 6, pp 283-287 ://A1978GC45200005

Gressel, J. (1984) Evolution of Herbicide-Resistant Weeds. Ciba Foundation Symposia, 102, pp 73-87 ://A1984SH88200010

Gressel, J. (1985) 321

Prevention of the Appearance of Herbicide-Resistant Weeds on Roadsides. Phytoparasitica, 13, 3-4, pp 248-248 ://A1985AXX1100040

Gressel, J. (1994) Can wild species become problem weeds because of herbicide resistance? Brachypodium distachyon: a case study. Crop Protection, 13, 8, pp 536-566

Gressel, J. (1999) Tandem constructs: preventing the rise of superweeds. Trends in Biotechnology, 17, 9, pp 361-366 ://000082263400006

Gressel, J., ed. (2005a) Crop Ferality and Volunteerism, CRC Press, Boca Raton,

Gressel, J. (2005b) Problems in qualifying and quantifying assumptions in plant protection models: Resultant simulations can be mistaken by a factor of million. Crop Protection, 24, 11, pp 1007-1015 ://000232249200008 AND http://www.botanischergarten.ch/Modelling/Gressel-Haygood-Million-2005.pdf

Gressel, J. & Al-Achmad, H. (2005) Molecular containment of genes within crops, prevention of gene establishment in volunteer offspring. In Crop ferality and volunteerism. (ed J. Gressel). CRC Press, Boca Raton, Fla, USA http://www.botanischergarten.ch/Geneflow/Gressel-Containment-Ferality-2005.pdf

Gressel, J. & Al-Ahmad, H. (2005) Assessing and managing biological risks of plants used for bioremediation, including risks of transgene flow. Zeitschrift für Naturforschung C-a Journal of Biosciences, 60, 3-4, pp 154-165 ://000230152300002 AND http://www.botanischergarten.ch/Modelling/Gressel-Mitigation-2005.pdf

Gressel, J. & Segel, L.A. (1978) Have Weeds Developed Genetic Resistance to Herbicides in Field. Israel Journal of Botany, 27, 1, pp 45-46 ://A1978FL18200039

Grutz, W. (1998) Oasis of Turquoise and Ravens. Aramco World, 49, 4, pp 16-27 http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Grutz-Oasis-Turquoise-Ravens-1998.PDF

Gu, M.H., Ma, H.T., & Liang, G.H. (1984) Karyotype Analysis of 7 Species in the Genus Sorghum. Journal of Heredity, 75, 3, pp 196-202 ://A1984SU01900009

Guimaraes, C.T., Sills, G.R., & Sobral, B.W.S. (1997) Comparative mapping of Andropogoneae: Saccharum L. (sugarcane) and its relation to sorghum and maize. Proceedings of the National Academy of Sciences of the United States of America, 94, 26, pp 14261-14266 ://000071182800010

Guo, J.H., Skinner, D.Z., & Liang, G.H. (1996) Phylogenetic relationships of sorghum taxa inferred from mitochondrial DNA restriction fragment analysis. Genome, 39, 5, pp 1027- 1034 ://A1996VM93100025 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Guo-Phylog-Sorgh-Shatter- 1996.pdf

Gupta, A.K., Sinha, R., Koradia, D., Patel, R., Parmar, M., Rohit, P., Patel, H., Patel, K., Chand, V.S., James, T.J., Chandan, A., Patel, M., Prakash, T.N., & Vivekanandan, P. (2003) Mobilizing grassroots' technological innovations and traditional knowledge, values and institutions: articulating social and ethical capital. Futures, 35, 9, pp 975-987 ://000186224000006

Hackel, E. (1889) Andropogoneae, Monographiae Phanerogamarum Sumptibus G. Masson, Paris, pp

Hadley, H.H. (1953) Cytological Relationships between Sorghum-Vulgare and S-Halepense. Agronomy Journal, 45, 4, pp 139-143 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Hadley-Cytological-1953.pdf

Hadley, H.H. (1958a) Chromosome numbers, fertility and rhizome expression of hybrids between grain sorghum and johnsongrass. J. Agron., 50, pp 278- 282 322

Hadley, H.H. (1958b) Chromosome numbers, fertility, and rhizome expression of hybrids between grain Sorghum and Johnsongrass. Agronomy Journal, 50, pp 278–282 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Hadley-Chromosome-1958.pdf

Hadley, H.H. & Mahan, J.L. (1956) The cytogenetic behavior of the progeny from a backcross (Sorghum vulgare · S. halepense · S. vulgare). Agronomy Journal, 48, pp 102-106

Hadley, H.H. & Openshaw, S. (1980) Interspecific and intergeneric hybridization. In Hybridization of crop plants (eds W. Fehr & H. Hadley), pp. 133-159. American Society of Agronomy-Crop Science Society of America,, Madison

Hagerty, M. (1940) Comments on writings concerning Chinese sorghums. Harvard J. Asiatic Stud., 5, pp 234-260

Hagio, T. (1998) Optimizing the particle bombardment method for efficient genetic transformation. Jarq-Japan Agricultural Research Quarterly, 32, 4, pp 239-247 ://000076896700002 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Hagio-Improvement-1998.pdf

Hagio, T., Blowers, A.D., & Earle, E.D. (1991) Stable Transformation of Sorghum Cell-Cultures after Bombardment with DNA-Coated Microprojectiles. Plant Cell Reports, 10, 5, pp 260-264 ://A1991GC35200011 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Hagio-Transformation- Sorghum-1991.pdf

Hamblin, M., Mitchell, S., White, G., Gallego, J., Kukatla, R., Wing, R., Paterson, A., & Kresovich, S. (2004) Comparative population genetics of the panicoid grasses: sequence polymorphism, linkage disequilibrium and selection in a diverse sample of Sorghum bicolor. Genetics, 167, pp 471–483

Hamblin, M.T., Fernandez, M.G.S., Casa, A.M., Mitchell, S.E., Paterson, A.H., & Kresovich, S. (2005) Equilibrium processes cannot explain high levels of short- and medium-range linkage disequilibrium in the domesticated grass Sorghum bicolor. Genetics, 171, 3, pp 1247-1256 ://000233967200031 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Hamblin-Equilibrium- Processes-2005.pdf

Hammer, K., Arrowsmith, N., & Gladis, T. (2003) Agrobiodiversity with emphasis on plant genetic resources. Naturwissenschaften, 90, 6, pp 241-250 ://000183825100001 or http://www.botanischergarten.ch/Feral/Hammer-P0DTA6LLA1PE9QET.pdf

Hanna, W.W. (1990) Transfer of Germ Plasm from the Secondary to the Primary Gene Pool in Pennisetum. Theoretical and Applied Genetics, 80, 2, pp 200-204 ://A1990DV10200010 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Hanna-Germplasm- Pennisetum-1990.pdf

Hans, S.R. & Johnson, W.G. (2002) Influence of shattercane Sorghum bicolor (L.) Moench. interference on corn (Zea mays L.) yield and nitrogen accumulation. Weed Technology, 16, 4, pp 787-791 ://000179784600015

Hariprakash, M. (1979) Soil Testing and Plant Analysis Studies on Hybrid Sorghum Csh-1. Mysore Journal of Agricultural Sciences, 13, 2, pp 178-181 ://A1979HS04100011

Hariprakasrao, M. (1978) Effect of Different Levels of Nitrogen, Phosphorus and Potassium on the Growth and Yield of Hybrid Sorghum, Csh-1. Mysore Journal of Agricultural Sciences, 12, 4, pp 566-568 ://A1978GW01400009

Harlan, J. & Zohary, D. (1966) Distribution of wild wheats and barley. Science, 153, pp 1074-1080

Harlan, J.R. (1971) Agricultural Origins - Centers and Noncenters. Science, 174, 4008, pp 468-& ://A1971K638500005 and http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Harlan-Centers-1971.pdf 323

Harlan, J.R. (1975a) Crops and Man American Society for Agronomy, Madison, Wisconsin, pp

Harlan, J.R. (1975b) Geographic Patterns of Variation in Some Cultivated Plants. Journal of Heredity, 66, 4, pp 182-191 ://A1975AQ03100001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Harlan-Patterns-1975.pdf

Harlan, J.R. (1984) Evaluation of wild relatives of crop plants. In Crop Genetic Resources: Conservation and Evaluation, (eds J.W.H. Holden & J.T. Williams), pp. 213–222. IBPGR, London

Harlan, J.R. (1989) Wild-grass harvesting in the Sahara and Sub-Sahara of Africa. In Foraging and Farming: the Evolution of Plant Exploitation (eds D.R. Harris & G.C. Hillman), pp. 79–98, 88–91 and Figures. 5.2–5.3. Unwin Hyman, London

Harlan, J.R. (1992a) Crops and Man, 2nd edition American Society of Agronomy, Madison, Wisconsin, IS: 0-89118-107-5, pp 295

Harlan, J.R. (1992b) Origin and processes of domestication. In Grass evolution and domestication (ed G.P. Chapman), pp. 159-175. Cambridge University Press, Cambridge

Harlan, J.R. & De Wet, J.M.J. (1971) Toward a relational classification of cultivated plants. Taxon, 20, pp 509-517 NEBIS 20070708

Harlan, J.R. & De Wet, J.M.J. (1972) Simplified Classification of Cultivated Sorghum. Crop Science, 12, 2, pp 172-& ://A1972M322900005 and http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Harlan-simplified-1972.pdf

Harlan, J.R. & De Wet, J.M.J. (1973) Sympatric Evolution in Sorghum. Genetics, 74, JUN, pp S109-S109 ://A1973Q333000318 AND http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Harlan-Sympatric-1974.pdf

Harlan, J.R. & Pasquereau.J (1969) Déçrue Agriculture in Mali. Economic Botany, 23, 1, pp 70-& ://A1969C904500008

Harlan, J.R. & Stemler, A.B.L. (1976) Races of Sorghum. In Origins of African plant domestication (ed J.R. In Harlan, de Wet, J.M.J. and Stemler, A.B.L.). Mouton Press, The Hague

Hart, G.E., Schertz, K.F., Peng, Y., & Syed, N.H. (2001) Genetic mapping of Sorghum bicolor (L.) Moench QTLs that control variation in tillering and other morphological characters. Theoretical and Applied Genetics, 103, 8, pp 1232-1242 ://000173261300015

Hartl, D. & Clarc, A. (1989) Principles of population genetics, Sinauer Associates Inc.,, Massachusetts, pp pp 293-323

Hassen, M.M., Mertz, E.T., Kirleis, A.W., Ejeta, G., Axtell, J.D., & Villegas, E. (1986) Tryptophan Levels in Normal and High-Lysine Sorghums. Cereal Chemistry, 63, 2, pp 175-176 ://A1986C081000026

Haussmann, B.I.G., Hess, D.E., Reddy, B.V.S., Welz, H.G., & Geiger, H.H. (2000a) Analysis of resistance to Striga hermonthica in diallel crosses of sorghum. Euphytica, 116, 1, pp 33-40 ://000089501700005

Haussmann, B.I.G., Hess, D.E., Welz, H.G., & Geiger, H.H. (2000b) Improved methodologies for breeding striga-resistant sorghums. Field Crops Research, 66, 3, pp 195-211 ://000087128900001

Haussmann, B.I.G., Obilana, A.B., Blum, A., Ayiecho, P.O., Schipprack, W., & Geiger, H.H. (1998) Hybrid performance of sorghum and its relationship to morphological and physiological traits under variable drought stress in Kenya. Plant Breeding, 117, 3, pp 223-229 ://000075128600005

324

Hawkes, J.G. (1983) The Diversity of Crop Plants Cambridge University Press, Cambridge, U.K., pp

Hawkes, J.G. (1990) VAVILOV,N.I. - THE MAN AND HIS WORK. Biological Journal of the Linnean Society, 39, 1, pp 3-6 ://WOS:A1990CN43400002 AND http://www.botanischergarten.ch/Genomics/Hawkes-Vavilov-Man-Work-1990.pdf

Hawkes, J.G. (1991) International Workshop on Dynamic Insitu Conservation of Wild Relatives of Major Cultivated Plants - Summary of Final Discussion and Recommendations. Israel Journal of Botany, 40, 5-6, pp 529-536 ://A1991JU49800012 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Hawkes-Landraces-Workshop- 1991.pdf

Hawkes, J.G. (1999) The evidence for the extent of N.I. Vavilov's new world Andean centres of cultivated plant origins. Genetic Resources and Crop Evolution, 46, 2, pp 163-168 ://WOS:000079329900009

Hawkes, J.G. & Harris, D.R. (1990) VAVILOV,N.I. CENTENARY SYMPOSIUM - SYMPOSIUM PAPERS. Biological Journal of the Linnean Society, 39, 1, pp 1-1 ://WOS:A1990CN43400001

Hawkins, S.E., Denman, C.E., Weibel, D.E., McNew, R.W., Merritt, R.G., Reeves, H.E., Belete, K., & Strickland, G.L. (1986) 1985 Performance Trials for Hybrid Sorghum and Corn in Oklahoma. Oklahoma Agricultural Experiment Station Research Report, P- 882, pp 1-39 ://A1986D415500001

Haygood, R., Ives, A.R., & Andow, D.A. (2004) Population genetics of transgene containment. Ecology Letters, 7, 3, pp 213-220 ://000189232400007 http://www.botanischergarten.ch/Modelling/Haygood-Andow-Swamping-2003.pdf

Henneman, M.L. & Memmott, J. (2001) Infiltration of a Hawaiian Community by Introduced Biological Control Agents. Science, 293, 5533, pp 1314-1316 http://www.sciencemag.org/cgi/content/abstract/293/5533/1314 and http://www.botanischergarten.ch/BioControl/Hennemann- Science-2001.pdf

Hetterscheid, W.L.A. & Brandenburg, W.A. (1995) Culton Versus Taxon - Conceptual Issues in Cultivated Plant Systematics. Taxon, 44, 2, pp 161-175 ://A1995RA04000001 NEBIS 20070709

Hetterscheid, W.L.A., Van Ettekoven, C., Van den Berg, R.G., & Brandenburg, W.A. (1999) Cultonomy in statutory registration exemplified by Allium L-crops. International Journal of Plant Varieties and Seeds (NIAB), 12, 3, pp 149-160 ://000085747600002 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Hetterscheid-Cultonomy- 1999.pdf

Hetterscheid, W.L.A., VandenBerg, R.G., & Brandenburg, W.A. (1996) An annotated history of the principles of cultivated plant classification. Acta Botanica Neerlandica, 45, 2, pp 123-134 ://A1996UT37300002

Hicks, C., Bean, S.R., Lookhart, G.L., Pedersen, J.F., Kofoid, K.D., & Tuinstra, M.R. (2001) Genetic analysis of kafirins and their phenotypic correlations with feed quality traits, in vitro digestibility, and seed weight in grain sorghum. Cereal Chemistry, 78, 4, pp 412-416 ://000169870100008

Hillmann, G. (1996) ‘Late Pleistocene changes in wild food plants available to huntergatherers of the northern Fertile Crescent:possible preludes to cereal cultivation. In The Origin and Spread of Agriculture and Pastoralism in Eurasia, (ed D.R. Harris), pp. 159–203, p 189. University College Press, London

Hirpara, D.S., Patel, J.C., Patel, B.S., & Khanpara, V.D. (1992) Response of Rainy-Season Hybrid Sorghum (Sorghum-Bicolor) to Nitrogen and Phosphorus. Indian Journal of Agronomy, 37, 3, pp 581-583 ://A1992KX20000054

Hoang-Tang & Liang, G.H. (1988) The genomic relationship between cultivated sorghum [Sorghum bicolor (L.) Moench] and Johnsongrass [S. halepense (L.) Pers.]: a re-evaluation. TAG Theoretical and Applied Genetics, 76, 2, pp 277-284 325

http://dx.doi.org/10.1007/BF00257856 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Hoang-Tang-S- halepense-Reevaluation-1988.pdf

Hoangtang, Dube, S.K., Liang, G.H., & Kung, S.D. (1991) Possible Repetitive DNA Markers for Eusorghum and Parasorghum and Their Potential Use in Examining Phylogenetic Hypotheses on the Origin of Sorghum Species. Genome, 34, 2, pp 241-250 ://A1991FJ24900010 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Hoang-Tang-DNA-Markers- 1991.pdf

Hoangtang & Liang, G.H. (1988) The Genomic Relationship between Cultivated Sorghum Sorghum-Bicolor (L) Moench and Johnsongrass Sorghum-Halepense (L) Pers - a Re-Evaluation. Theoretical and Applied Genetics, 76, 2, pp 277-284 ://A1988P756400017

Hodnett, G.L., Burson, B.L., Rooney, W.L., Dillon, S.L., & Price, H.J. (2005) Pollen-pistil interactions result in reproductive isolation between Sorghum bicolor and divergent Sorghum species. Crop Science, 45, 4, pp 1403-1409 ://000230713700027 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Hodnett-Sorghum-Pistil-Pollen- 2005.pdf

Holm, L.G., Plucknett, J.D., Pancho, L.V., & Herberger, J.P. (1977) The World's Worst Weeds, Distribution and Biology. In University Press of Hawaii, Vol. pp. 609, Honolulu

Hookstra, G.H. & Ross, W.M. (1982) Comparison of F1s and Inbreds as Female Parents for Hybrid Sorghum Seed Production. Crop Science, 22, 1, pp 147-150 ://A1982NL80000034

House, L.R. (1984) Sorghum Breeding - an International Program Outlined. Span, 27, 3, pp 132-134 ://A1984TY37500017

Howarth, F.G. (1991) Environmental Impacts of Classical Biological-Control. Annual Review of Entomology, 36, pp 485-509 ://A1991EQ80700021

Hughes, S. (2006) Opinion piece: Genomics and crop plant science in Europe. Plant Biotechnology Journal, 4, 1, pp 3-5 http://www.blackwell-synergy.com/doi/abs/10.1111/j.1467-7652.2005.00164.x

Hulbert, S.H., Richter, T.E., Axtell, J.D., & Bennetzen, J.L. (1990) Genetic-Mapping and Characterization of Sorghum and Related Crops by Means of Maize DNA Probes. Proceedings of the National Academy of Sciences of the United States of America, 87, 11, pp 4251-4255 ://A1990DG09200048 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Hulbert-Gen-Mapp-Sorghum- 1990.pdf

Hulse, J.M., Laingm, E.M., & Pearson, O.K. (1980) Sorghum and the Millets Their Composition and Nutritive Value Academic Press, London, New York, pp

Hultquist, S.J., Vogel, K.P., Lee, D.J., Arumuganathan, K., & Kaeppler, S. (1996) Chloroplast DNA and nuclear DNA content variations among cultivars of switchgrass, Panicum virgatum L. Crop Science, 36, 4, pp 1049-1052 ://A1996UV54900039

Hultquist, S.J., Vogel, K.P., Lee, D.J., Arumuganathan, K., & Kaeppler, S. (1997) DNA content and chloroplast DNA polymorphisms among switchgrasses from remnant midwestern prairies. Crop Science, 37, 2, pp 595-598 ://A1997WU90500047

ICRISAT (1993) Electronic Source: Descripteurs Du Sorgho (ed ICRISAT), published by: ICRISAT http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/ICRISAT-Descripteur-Sorghum-1993.pdf

ICRISAT & FAO (1996) Electronic Source: The World Sorghum and Millet Economies: Facts, Trends and Outlook, FAO Corporate Repository published by: ©FAO AND ICRISAT 1996

http://www.fao.org/docrep/W1808E/W1808E00.htm 326

Ilic, K., SanMiguel, P.J., & Bennetzen, J.L. (2003) A complex history of rearrangement in an orthologous region of the maize, sorghum, and rice genomes. Proceedings of the National Academy of Sciences of the United States of America, 100, 21, pp 12265-12270 ://000186024300061 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Ilic-Complexhistory-genome- 2003.pdf

Imik, H., Hayirli, A., Turgut, L., Lacin, E., Celebi, S., Koc, F., & Yildiz, L. (2006) Effects of additives on laying performance, metabolic profile, and egg quality of hens fed a high level of sorghum (Sorghum vulgare) during the peak laying period. Asian-Australasian Journal of Animal Sciences, 19, 4, pp 573-581 ://000234962400019

Iptas, S. & Brohi, A.R. (2003) Effect of nitrogen rate and stubble height on dry matter yield, crude protein content and crude protein yield of a sorghum- sudangrass hybrid Sorghum bicolor (L.) Moench x Sorghum sudanense (Piper) Stapf. in the three-cutting system. Journal of Agronomy and Crop Science, 189, 4, pp 227-232 ://000185415500002

Ito, Y., Arikawa, K., Antonio, B.A., Ohta, I., Naito, S., Mukai, Y., Shimano, A., Masukawa, M., Shibata, M., Yamamoto, M., Yokoyama, J., Sakai, Y., Sakata, K., Nagamura, Y., Namiki, N., Matsumoto, T., Higo, K., & Sasaki, T. (2005) Rice Annotation Database (RAD): a contig-oriented database for map-based rice genomics. Nucleic Acids Research, 33, pp D651- D655 ://000226524300134

Ives, A.R. & Carpenter, S.R. (2007) Stability and diversity of ecosystems. Science, 317, 5834, pp 58-62 ://WOS:000247776700038 AND http://www.botanischergarten.ch/Biotech-Biodiv/Ives-Stability-Diversity-Ecosystems- 2007.pdf

Jacomet, S. & Kreuz, A. (1999) Archäobotanik, Aufgaben, Methoden und Ergebnisse vegetations- und agrargeschichtlicher Forschung Eugen Ulmer, Stuttgart, IS: 3- 8252 8158 2, pp 368

Jacot, Y., Ammann, K., Rufener Al Mazyad, P., Chueca, C., Davin, J., Gressel, J., Loureiro, I., Wang, H., & Benavente, E. (2004) Hybridization between wheat and wild relatives, a European Union research programme. In Introgression from Genetically Modified Plants into Wild Relatives (ed D.B. H. den Nijs, and J. Sweet), pp. 63-74. CABI Publishing http://www.botanischergarten.ch/Geneflow/Jacot-et-al-Amsterdam-2003.pdf

Jambunathan, R., Kherdekar, M.S., & Bandyopadhyay, R. (1990) Flavan-4-Ols Concentration in Mold-Susceptible and Mold-Resistant Sorghum at Different Stages of Grain Development. Journal of Agricultural and Food Chemistry, 38, 2, pp 545-548 ://A1990CP91700046

Jambunathan, R., Kherdekar, M.S., & Vaidya, P. (1991) Ergosterol Concentration in Mold-Susceptible and Mold-Resistant Sorghum at Different Stages of Grain Development and Its Relationship to Flavan-4-Ols. Journal of Agricultural and Food Chemistry, 39, 10, pp 1866-1870 ://A1991GL32300038

Jambunathan, R., Mertz, E.T., & Axtell, J.D. (1975) Fractionation of Soluble-Proteins of High-Lysine and Normal Sorghum Grain. Cereal Chemistry, 52, 1, pp 119-121 ://A1975V659600014

Jambunathan, R., Rao, N.S., & Gurtu, S. (1983) Rapid Methods for Estimating Protein and Lysine in Sorghum (Sorghum-Bicolor (L) Moench). Cereal Chemistry, 60, 3, pp 192-194 ://A1983QR48800003

Jarvis, D.I. & Hodgkin, T. (1999) Wild relatives and crop cultivars: detecting natural introgression and farmer selection of new genetic combinations in agroecosystems. Molecular Ecology, 8, 12, pp S159-S173 ://000085737800015 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Jarvis-wildrelatives-Sorghum- 1999.pdf

Jeoung, J.M., Krishnaveni, S., Muthukrishnan, S., Trick, H.N., & Liang, G.H. (2002) Optimization of sorghum transformation parameters using genes for green fluorescent protein and beta-glucuronidase as visual markers. Hereditas, 137, 1, pp 20-28 ://000180573100004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Jeoung-Optimization-Transf- Sorgh-2002.pdf

327

Johari, R.P., Mehta, S.L., & Naik, M.S. (1977a) Changes in Protein-Fractions and Leucine- C-14 Incorporation During Sorghum-Grain Development. Phytochemistry, 16, 3, pp 311- 314 ://A1977DB80100003

Johari, R.P., Mehta, S.L., & Naik, M.S. (1977b) Protein-Synthesis and Changes in Nucleic-Acids During Grain Development of Sorghum. Phytochemistry, 16, 1, pp 19-24 ://A1977CW31100002

Johnson, B. (1958) The Staple Food Economics of Tropical West Africa, Stanford University pp Stanford University Press Stanford, California (Report)

Johnson, W.G., Bradley, P.R., Hart, S.E., Buesinger, M.L., & Massey, R.E. (2000) Efficacy and economics of weed management in glyphosate-resistant corn (Zea mays). Weed Technology, 14, 1, pp 57-65 ://000086220300009

Jones, M.D. & Sieglinger, J.B. (1951) Effect of Wind Direction on Natural Crossing in Sorghum, Sorghum-Vulgare Pers. Agronomy Journal, 43, 11, pp 571-572 ://A1951UC64400011 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Jones-Wind-Pollination- 1951.pdf

Jordan, D.R., Tao, Y.Z., Godwin, I.D., Henzell, R.G., Cooper, M., & McIntyre, C.L. (2004) Comparison of identity by descent and identity by state for detecting genetic regions under selection in a sorghum pedigree breeding program. Molecular Breeding, 14, 4, pp 441-454 ://000226417400009

Kabore, B.K., Couture, L., Dostaler, D., & Bernier, L. (2001) Phenetic variability of Colletotrichum graminicola from sorghum. Canadian Journal of Plant Pathology-Revue Canadienne De Phytopathologie, 23, 2, pp 138-145 ://000169247800004

Kandasamy, O.S. & Subramanian, S. (1979) Response of Hybrid Sorghum (Csh.5) to Irrigation Regimes and N-Rates. Indian Journal of Agronomy, 24, 1, pp 54-57 ://A1979HH32900014

Kane, E.J., Wilson, A.J., & Chourey, P.S. (1992) Mitochondrial Genome Variability in Sorghum Cell-Culture Protoclones. Theoretical and Applied Genetics, 83, 6-7, pp 799-806 ://A1992HP30700019

Kao, T.C., Ho, J.H., Liang, C.L., & Chiou, L.I. (1987) An Improvement and Cause of Low Germinability in Hybrid Sorghum-Taichung No-51. Journal of the Agricultural Association of China, 137, pp 55-71 ://A1987H136500005

Karp, A. & Isaac, P. (1998) Molecular tools for screening biodiversity Chapman and Hall, London, pp

Karp, A., Kresovich, S., Bhat, K., Ayad, W., & Hodgkin, T. (1997) Electronic Source: Molecular tools in plant genetic resources conservation: a guide to the technologies (ed IPGRI), IPGRI Technical Bulletin No. 2 published by: International Plant Genetic Resources Institute 1997 http://www.ipgri.cgiar.org/publications/pdf/138.pdf

Kato, A., Vega, J.M., Han, F.P., Lamb, J.C., & Birchler, J.A. (2005) Advances in plant chromosome identification and cytogenetic techniques. Current Opinion in Plant Biology, 8, 2, pp 148-154 ://000227932600006 http://www.botanischergarten.ch/Maize/Kato-Chromosome-Ident-2005.pdf

Kato, T.A. & Sanchez, J.J. (2002) Introgression of chromosome knobs from Zea diploperennis into maize. Maydica, 47, 1, pp 33-50 ://000176851100004

Kaul, M.L.H. (1988) Male Sterilility in Higher Plants Springer, Berlin, pp

Kenga, R., Alabi, S.O., & Gupta, S.C. (2004) Combining ability studies in tropical sorghum (Sorghum bicolor (L.) Moench). Field Crops Research, 88, 2-3, pp 251-260 ://000221729800014

328

Kier, G., Mutke, J., Dinerstein, E., Ricketts, T.H., Kuper, W., Kreft, H., & Barthlott, W. (2005) Global patterns of plant diversity and floristic knowledge. Journal of Biogeography, 32, 7, pp 1107-1116 ://00022970590000

Kilian, A., Kudrna, D.A., Kleinhofs, A., Yano, M., Kurata, N., Steffenson, B., & Sasaki, T. (1995) Rice-Barley Synteny and Its Application to Saturation Mapping of the Barley Rpg1 Region. Nucleic Acids Research, 23, 14, pp 2729- 2733 ://A1995RN26800024

Kim, J.S., Childs, K.L., Islam-Faridi, M.N., Menz, M.A., Klein, R.R., Klein, P.E., Price, H.J., Mullet, J.E., & Stelly, D.M. (2002) Integrated karyotyping of sorghum by in situ hybridization of landed BACs. Genome, 45, 2, pp 402-412 ://000174333500021

Kim, J.S., Islam-Faridi, M.N., Klein, P.E., Stelly, D.M., Price, H.J., Klein, R.R., & Mullet, J.E. (2005a) Comprehensive molecular cytogenetic analysis of sorghum genome architecture: Distribution of euchromatin, heterochromatin, genes and recombination in comparison to rice. Genetics, 171, 4, pp 1963-1976 ://000234407100046

Kim, J.S., Klein, P.E., Klein, R.R., Price, H.J., Mullet, J.E., & Stelly, D.M. (2005b) Chromosome identification and nomenclature of Sorghum bicolor. Genetics, 169, 2, pp 1169-1173 ://000227697200055 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Kim-Chromosome- Nomenclature-2005.pdf

Kim, J.S., Klein, P.E., Klein, R.R., Price, H.J., Mullet, J.E., & Stelly, D.M. (2005c) Molecular cytogenetic maps of sorghum linkage groups 2 and 8. Genetics, 169, 2, pp 955-965 ://000227697200038 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Kim-Sequencing-Sorghum-2005.pdf

Kimber, C.T. (2000) Origins of Domesticated Sorghum and its Early Diffusion in India and China. In Sorghum, Origin, History, Technology and Production (eds C.W. Smith & R.A. Frederiksn). Wiley and Sons, New York http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Kimber-Origins-Sorghum-2000.PDF

King, S.R. & Hagood, E.S. (2006) Herbicide programs for the control of ALS-resistant shattercane (Sorghum bicolor) in corn (Zea mays). Weed Technology, 20, 2, pp 416-421 ://000238476300022

Kirby, L.K., Nelson, T.S., Johnson, Z.B., & York, J.O. (1983) The Effect of Seed Coat Color of Hybrid Sorghum Grain on the Ability of Chicks to Digest Dry-Matter and Amino-Acids and to Utilize Energy. Nutrition Reports International, 27, 4, pp 831-836 ://A1983QJ54400020

Klein, P.E., Klein, R.R., Cartinhour, S.W., Ulanch, P.E., Dong, J.M., Obert, J.A., Morishige, D.T., Schlueter, S.D., Childs, K.L., Ale, M., & Mullet, J.E. (2000) A high-throughput AFLP-based method for constructing integrated genetic and physical maps: Progress toward a sorghum genome map. Genome Research, 10, 6, pp 789-807 ://000087638200006 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Klein-High-Troughput- mapping-Sorghum-2000.pdf

Kresovich, S., Barbazuk, B., Bedell, J.A., Borrell, A., Buell, C.R., Burke, J., Clifton, S., Cordonnier-Pratt, M.M., Cox, S., Dahlberg, J., Erpelding, J., Fulton, T.M., Fulton, B., Fulton, L., Gingle, A.R., Hash, C.T., Huang, Y.H., Jordan, D., Klein, P.E., Klein, R.R., Magalhaes, J., McCombie, R., Moore, P., Mullet, J.E., Ozias-Akins, P., Paterson, A.H., Porter, K., Pratt, L., Roe, B., Rooney, W., Schnable, P.S., Stelly, D.M., Tuinstra, M., Ware, D., & Warek, U. (2005) Toward sequencing the sorghum genome. A US National Science Foundation-Sponsored Workshop Report. Plant Physiology, 138, 4, pp 1898-1902 ://000231206600009 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Kresovich-Sequencing- Workshop-2005.pdf

Kresovich, S., Drawe, H.J., & Molina, J.J. (1986) Sweet Sorghum Breeding Line Evaluations - 1985. Texas Agricultural Experiment Station Progress Report, 4358, pp 1-10 ://A1986A768300001

Krishnasamy, V. & Ramaswamy, K.R. (1985) Effect of Date of Harvest on the Yield, Size, Germination and Vigor of the Seeds of Csh-5 Hybrid Sorghum. Indian Journal of Agricultural Sciences, 55, 8, pp 526-530 ://A1985AQF7500008

Kristensen, C., Morant, M., Olsen, C.E., Ekstrom, C.T., Galbraith, D.W., Moller, B.L., & Bak, S. (2005) 329

Metabolic engineering of dhurrin in transgenic Arabidopsis plants with marginal inadvertent effects on the metabolome and transcriptome. Proceedings of the National Academy of Sciences of the United States of America, 102, 5, pp 1779-1784 ://000226877300093

Kruskal, J.B. (1964a) Multidimensional-Scaling by Optimizing Goodness of Fit to a Nonmetric Hypothesis. Psychometrika, 29, 1, pp 1-27 ://A1964CBT8100001

Kruskal, J.B. (1964b) Nonmetric Multidimensional-Scaling - a Numerical-Method. Psychometrika, 29, 2, pp 115-129 ://A1964CBT8200003

Kumar, V., Elangovan, A.V., & Mandal, A.B. (2005) Response of broiler chicks fed high tannin sorghum diets supplemented with DL-methionine. Indian Journal of Animal Sciences, 75, 12, pp 1417-1422 ://000234315800016

Kumaravadivel, N. & Rangasamy, S.R.S. (1994) Plant-Regeneration from Sorghum Anther Cultures and Field-Evaluation of Progeny. Plant Cell Reports, 13, 5, pp 286-290 ://A1994MY96200010

Kuper, W., Sommer, J.H., Lovett, J.C., Mutke, J., Linder, H.P., Beentje, H.J., Van Rompaey, R., Chatelain, C., Sosef, M., & Barthlott, W. (2004) Africa's hotspots of biodiversity redefined. Annals of the Missouri Botanical Garden, 91, 4, pp 525-535 ://000226362700001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Kueper-Hotspots-2004.pdf

Kush, G. (1997) Origin, dispersal, cultivation and variation of rice. Plant Molecular Biology, 35, 1-2, pp 25-37 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Kush-Origin-Rice-1997.pdf

Lansac, A.R., Sullivan, C.Y., Johnson, B.E., & Lee, K.W. (1994) Viability and Germination of the Pollen of Sorghum Sorghum-Bicolor (L) Moench. Annals of Botany, 74, 1, pp 27-33 ://A1994NX07600004

Larsson, H. (1996) Relationships between rainfall and sorghum, millet and sesame in the Kassala province, eastern Sudan. Journal of Arid Environments, 32, 2, pp 211-223 ://A1996UF94000010

Laurie, D.A. & Bennett, M.D. (1985) Nuclear-DNA Content in the Genera Zea and Sorghum - Intergeneric, Interspecific and Intraspecific Variation. Heredity, 55, pp 307- 313 ://A1985AWY7900003

LeConte, J. (1965) Enquête sur les sorghos et mils du Tchad Institut de Recherches Agronomiques Tropicales et des Cultures Vivribres, Paris, pp

Lee, S.H., Muthukrishnan, S., Sorensen, E.L., & Liang, G.H. (1989) Restriction Endonuclease Analysis of Mitochondrial-DNA from Sorghum with Fertile and Male-Sterile Cytoplasms. Theoretical and Applied Genetics, 77, 3, pp 379-382 ://A1989T895900012

Lekule, F.P. & Kyvsgaard, N.C. (2003) Improving pig husbandry in tropical resource-poor communities and its potential to reduce risk of porcine cysticercosis. Acta Tropica, 87, 1, pp 111-117 ://000183576700015

Leslie, J.F. (2003) Sorghum and Millets Diseases Blackwell Publishing, IS: 0813803896, pp 528 pages, 48 illustrations. http://www.blackwellpublishing.com/book.asp?ref=0813803896&site=1

Levin, D.A., Francisco-Ortega, J., & Jansen, R.K. (1996) Hybridization and the Extinction of Rare Plant Species. 10, pp 10-16 http://links.jstor.org/sici?sici=0888-8892%28199602%2910%3A1%3C10%3AHATEOR%3E2.0.CO%3B2-X

Li, G., Gu, W., Hicks, A., & Chapman, K. (1998) Electronic Source: A training manual for sweet sorghum. (ed FAO), This manual is an output of the FAO Project TCP/CPR/0066 "Development of Sweet Sorghum for Grain, Sugar, Feed, Fiber, and Value-Added By-products, in the Arid, Saline-Alkaline Regions". published by: EcoPort version (revised) by Peter Griffee. 330

http://ecoport.org/ep?SearchType=earticleView&earticleId=172&page=-2

Li, X.Q., Dodson, J., Zhou, X.Y., Zhang, H., & Masutomoto, R. (2007) Early cultivated wheat and broadening of agriculture in Neolithic China. Holocene, 17, 5, pp 555-560 ://WOS:000248755500001 AND http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Li-Early-Wheat-China- 2007.pdf

Li, Y. & Li, C.Z. (1998) Genetic contribution of Chinese landraces to the development of sorghum hybrids. Euphytica, 102, 1, pp 47-55 ://000074500500006

Lin, Y.R., Zhu, L.H., Ren, S.X., Yang, J.S., Schertz, K.F., & Paterson, A.H. (1999) A Sorghum propinquum BAC library, suitable for cloning genes associated with loss-of-function mutations during crop domestication. Molecular Breeding, 5, 6, pp 511-520 ://000084983100003 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Lin-propinquum-BAClibrary- 1999.pdf

Lo, S.C.C., De Verdier, K., & Nicholson, R.L. (1999) Accumulation of 3-deoxyanthocyanidin phytoalexins and resistance to Colletotrichum sublineolum in sorghum. Physiological and Molecular Plant Pathology, 55, 5, pp 263-273 ://000083648700002

Long, B. (1930) Spikelets of Johnson Grass and Sudan Grass. Botanical Gazette, 89, 2, pp 154-168 Stable URL: http://links.jstor.org/sici?sici=0006-8071%28193004%2989%3A2%3C154%3ASOJGAS%3E2.0.CO%3B2-D AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Long-Johnson-Sudan-1930.pdf

Lughadha, E.N., Baillie, J., Barthlott, W., Brummitt, N.A., Cheek, M.R., Farjon, A., Govaerts, R., Hardwick, K.A., Hilton-Taylor, C., Meagher, T.R., Moat, J., Mutke, J., Paton, A.J., Pleasants, L.J., Savolainen, V., Schatz, G.E., Smith, P., Turner, I., Wyse-Jackson, P., & Crane, P.R. (2005) Measuring the fate of plant diversity: towards a foundation for future monitoring and opportunities for urgent action. Philosophical Transactions of the Royal Society B-Biological Sciences, 360, 1454, pp 359-372 ://000228214600011

Luo, Y.W., Yen, X.C., Zhang, G.Y., & Liang, G.H. (1992) Agronomic Traits and Chromosome Behavior of Autotetraploid Sorghums. Plant Breeding, 109, 1, pp 46-53 ://A1992JJ53500006

Ma, Z., Wood, C.W., & Bransby, D.I. (2000) Impacts of soil management on root characteristics of switchgrass. Biomass & Bioenergy, 18, 2, pp 105-112 ://000085496800002

Macdonald, K.C., Nesbitt, M., & Samuel, D.J. (1995) Domestic Plants and Animals - the Egyptian Origins - Brewer,Dj, Redford,Db, Redford,S. Antiquity, 69, 264, pp 637-639 ://A1995RW15100033

Madakadze, I.C., Coulman, B.E., Peterson, P., Stewart, K.A., Samson, R., & Smith, D.L. (1998) Leaf area development, light interception, and yield among switchgrass populations in a short-season area. Crop Science, 38, 3, pp 827-834 ://000074754600034

Madsen, K.H., Holm, P.B., Lassen, J., & Sandøe, P. (2002) Ranking Genetically Modified Plants According to Familiarity. Journal of Agricultural and Environmental Ethics, 15, 3, pp 267-278 http://www.springerlink.com/openurl.asp?genre=article&id=doi:10.1023/A:1015729011895 and http://www.botanischergarten.ch/Familiarity/Madsen-Familiarity-2002.pdf

Magoon, M.L. & Shambuli.Kg (1961) Karyomorphology of Sorghum Propinquum and Its Bearing on Origin of 40-Chromosome Sorghum. Chromosoma, 12, 4, pp 460-& ://A19618737A00005

Magoon, M.L. & Shambuli.Kg (1962) Karyomorphology of Sorghum Propinquum and Its Bearing on Origin of 40-Chromosome Sorghum. Chromosoma, 12, 4, pp 460-& ://A19628743A00005 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Magoon-propinquum- 1961.pdf

Majisu, B.N. (1971) Effects of autoploidy on grain sorghum. E. Afr. agric. For. J., 36, pp 235

Mann, J.A., Gbur, E.E., & Miller, F.R. (1985) 331

A Screening Index for Adaptation in Sorghum Cultivars. Crop Science, 25, 4, pp 593-598 ://A1985AKX5400004

Manzelli, M., Benedettelli, S., & Vecchio, V. (2005) Agricultural biodiversity in Northwest Somalia - an assessment among selected Somali sorghum (Sorghum bicolor (L.) Moench) germplasm. Biodiversity and Conservation, 14, 14, pp 3381-3392 ://000234053800005 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Manzelli-Sorghum-biodiv- Somalia-2005.pdf

Marier, D. (1985) Cogeneration-Based Ethanol Plant Could Generate over 2 Million Gallons of Anhydrous Ethanol Annually - Hybrid Sorghum Is Key to Success. Alternative Sources of Energy, 73, pp 70-70 ://A1985AGQ8700014

Marley, P.S., Aba, D.A., Shebayan, J.A.Y., Musa, R., & Sanni, A. (2004) Integrated management of Striga hermonthica in sorghum using a mycoherbicide and host plant resistance in the Nigerian Sudano- Sahelian savanna. Weed Research, 44, 3, pp 157-162 ://000221355400002

Marshall, D.R. & Downes, R.W. (1977) Test for Obligate Apomixis in Grain-Sorghum R473. Euphytica, 26, 3, pp 661-664 ://A1977EQ02900015

Mathews, S., Spangler, R.E., Mason-Gamer, R.J., & Kellogg, E.A. (2002) Phylogeny of Andropogoneae inferred from phytochrome B, GBSSI, and NDHF. International Journal of Plant Sciences, 163, 3, pp 441-450 ://000174783500010 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Mathews-Phylog- Andropgoneae-2002.pdf

Mattei, M.R. (1980) The Breeding System of Populations of Sorghum, Sorghum-Bicolor (L) Moench .3. Correlated Effects of Mass Selection for Several Quantitative Characters on Outcrossing. Acta Cientifica Venezolana, 31, 6, pp 604-613 ://A1980MH64100019

Maunder, A.B. & Sharp, G.I. (1963) Localisation of outcrosses within the panicle of fertile sorghum. Crop Science, 3, pp 149-158 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Maunder-Localization-1963.pdf

McGuire, S. (2004) The Human Ecology of the Seed System: Farmer and formal management of sorghum in Ethiopia. PhD Thesis, unpublished, University of Wageningen, Wageningen Thesis, pp

McVetty, P. (1997) Cytoplasmatic Male Sterility. In Pollen biotechnology for crop production and improvement. (eds K. Shivanna & V. Sawheney). Cambridge Univ. Press, Cambridge

McWhorter, C.G. & Anderson, J.M. (1981) The Technical and Economic-Effects of Johnsongrass (Sorghum-Halepense) - Control in Soybeans (Glycine-Max). Weed Science, 29, 3, pp 245-253 ://A1981LM71700001

Menkir, A., Goldsbrough, P., & Ejeta, G. (1997) RAPD based assessment of genetic diversity in cultivated races of sorghum. Crop Science, 37, 2, pp 564-569 ://A1997WU90500042

Menz, M.A., Klein, R.R., Unruh, N.C., Rooney, W.L., Klein, P.E., & Mullet, J.E. (2004) Genetic diversity of public inbreds of sorghum determined by mapped AFLP and SSR markers. Crop Science, 44, 4, pp 1236-1244 ://000222582400016

Merwine, N.C., Gourley, L.M., & Blackwell, K.H. (1981) Inheritance of Papery Glume and Cleistogamy in Sorghum. Crop Science, 21, 6, pp 953-956 ://A1981MY38600035

Merx, R.J.H.M., Lofti, M., Naber-Van Den Heuvel, P., & Venkatesh Mannar, M. (1996) Micronutrient Fortification of Foods. Current Practices, Research and Opportunities., International Agricultural Centre (IAC) pp 107 Wageningen NL (Report)

Messerli, B. & Winiger, M. (1992) 332

Climate, Environmental-Change, and Resources of the African Mountains from the Mediterranean to the Equator. Mountain Research and Development, 12, 4, pp 315-336 ://A1992KE27600003 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Messerli-Environmental- Change-1992.pdf

Mgonja, M.A., Chandra, S., Monyo, E.S., Obilana, A.B., Chisi, M., Saadan, H.M., Kudita, S., & Chinhema, E. (2006) Stratification of SADC regional sorghum testing sites based on grain yield of varieties. Field Crops Research, 96, 1, pp 25-30 ://000235502900004 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Mgonja-SADC-Sorghum-2006.pdf

Mickelbart, M.V., Peel, G., Joly, R.J., Rhodes, D., Ejeta, G., & Goldsbrough, P.B. (2003) Development and characterization of near-isogenic lines of sorghum segregating for glycinebetaine accumulation. Physiologia Plantarum, 118, 2, pp 253-261 ://000182797600011

Miller, F.R. (1980) The Breeding of Sorghum. Texas Agricultural Experiment Station Miscellaneous Publication, 1451, pp 128-136 ://A1980KS86700012

Millhollon, R. (2000) Loose kernel smut for biocontrol of Sorghum halepense in Saccharum sp hybrids. Weed Science, 48, 5, pp 645-652 ://000089783100016

Ming, R., Liu, S.C., Bowers, J.E., Moore, P.H., Irvine, J.E., & Paterson, A.H. (2002a) Construction of a Saccharum consensus genetic map from two interspecific crosses. Crop Science, 42, 2, pp 570-583 ://000174093500036 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Ming-Sach-Consens-2002.pdf

Ming, R., Liu, S.C., Lin, Y.R., da Silva, J., Wilson, W., Braga, D., van Deynze, A., Wenslaff, T.F., Wu, K.K., Moore, P.H., Burnquist, W., Sorrells, M.E., Irvine, J.E., & Paterson, A.H. (1998) Detailed alignment of Saccharum and Sorghum chromosomes: Comparative organization of closely related diploid and polyploid genomes. Genetics, 150, 4, pp 1663-1682 ://000077401300029 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Ming-Alignment-Sach-Sorgh- 1998.pdf

Ming, R., Wang, Y.W., Draye, X., Moore, P.H., Irvine, J.E., & Paterson, A.H. (2002b) Molecular dissection of complex traits in autopolyploids: mapping QTLs affecting sugar yield and related traits in sugarcane. Theoretical and Applied Genetics, 105, 2-3, pp 332-345 ://000177739800021 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Ming-Dissection-QTLS-Sach- Sorgh-2002.pdf

Mohammad, D., Cox, P.B., Posler, G.L., Kirkham, M.B., Hussain, A., & Khan, S. (1993) Genotype-X Environment Interaction and Its Implications in Sorghum (Sorghum-Bicolor) Breeding Program. Indian Journal of Agricultural Sciences, 63, 3, pp 153-156 ://A1993KV01300003

Monaghan, N. (1979) Biology of Johnson Grass (Sorghum-Halepense). Weed Research, 19, 4, pp 261-267 ://A1979HQ82500006

Monyo, E.S., Ejeta, G., Nyquist, W.E., & Axtell, J.D. (1988) Combining Ability of High Lysine Sorghum Lines Derived from P-721 Opaque. Crop Science, 28, 1, pp 70-75 ://A1988L709500017

Moore, G., Devos, K.M., Wang, Z., & Gale, M.D. (1995) Cereal Genome Evolution - Grasses, line up and form a circle. Current Biology, 5, 7, pp 737-739 ://A1995RL95300014 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Moore-Cereal-Genomes- 1995.pdf

Moran, J.L. & Rooney, W.L. (2003) Effect of cytoplasm on the agronomic performance of grain sorghum hybrids. Crop Science, 43, 3, pp 777-781 ://000182540300004

Morden, C.W., Doebley, J., & Schertz, K.F. (1990) Allozyme variation among the spontaneous species of Sorghum section Sorghum (Poaceae). TAG Theoretical and Applied Genetics, 80, 3, pp 296-304 http://www.springerlink.com/openurl.asp?genre=article&id=doi:10.1007/BF00210063 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Morden-Allyzyme--Variation-Sorghum-1990-better.doc.pdf

Morden, C.W., Doebley, J.F., & Schertz, K.F. (1989) 333

Allozyme Variation in Old-World Races of Sorghum Bicolor (Poaceae). American Journal of Botany, 76, 2, pp 247-255 ://A1989T473300011

Morgan, P.W., Finlayson, S.A., Childs, K.L., Mullet, J.E., & Rooney, W.L. (2002) Opportunities to improve adaptability and yield in grasses: Lessons from sorghum. Crop Science, 42, 6, pp 1791-1799 ://000181430200003 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Morgan-Lessons-Sorghum- 2002.pdf

Morrell, P.L., Lundy, K.E., & Clegg, M.T. (2003) Distinct geographic patterns of genetic diversity are maintained in wild barley (Hordeum vulgare ssp. spontaneum) despite migration. Proceedings of the National Academy of Sciences of the United States of America, 100, 19, pp 10812-10817 http://www.pnas.org/content/100/19/10812.abstract AND http://www.ask-force.org/web/Africa-Harvest-Sorghum-Lit-1/Morrell- Distinct-Patterns-Hordeum-ö2003.pdf

Morrell, P.L., Williams-Coplin, T.D., Lattu, A.L., Bowers, J.E., Chandler, J.M., & Patterson, A.H. (2005) Crop-to-weed introgression has impacted allelic composition of johnsongrass populations with and without recent exposure to cultivated sorghum. Molecular Ecology, 14, 7, pp 2143-2154 ://000229190300024 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Morrell-Introgression-2005.pdf

Morris, W.F., Kareiva, P.M., & Raymer, P.L. (1994) Do Barren Zones and Pollen Traps Reduce Gene Escape from Transgenic Crops. Ecological Applications, 4, 1, pp 157-165 ://A1994MU60500017 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Morris-Barrenzones- 1994.pdf

Moschini, G., Bulut, H., & Cembalo, L. (2005) On the segregation of genetically modified, conventional and organic products in European agriculture: A multi-market equilibrium analysis. Journal of Agricultural Economics, 56, 3, pp 347-372 ://000234524100001

Mulatu, E. & Belete, K. (2001) Participatory varietal selection in lowland sorghum in eastern Ethiopia: Impact on adoption and genetic diversity. Experimental Agriculture, 37, 2, pp 211-229 ://000169626400004

Mulcahy, C., Hedges, D.A., Rapp, G.G., & Wheeler, J.L. (1992) Correlations among Potential Selection Criteria for Improving the Feeding Value of Forage Sorghums. Tropical Grasslands, 26, 1, pp 7-11 ://A1992JT84300002

Munjal, S.V. & Narayan, R.K.J. (1995) Restriction Analysis of the Mitochondrial-DNA of Male-Sterile and Maintainer Lines of Pearl-Millet, Pennisetum-Americanum L. Plant Breeding, 114, 3, pp 256-258 ://A1995RT33800015

Murdock, G.P. (1960) Staple Subsistence Crops of Africa. Geographical Review, 50, 4, pp 523-540 ://A1960CBE6500004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Murdock-StapleCrops- 1960.pdf

Murty, B.R., Arunacha.V, & Saxena, M.B.L. (1967) Classification Anc Catalogue of a World Collection of Sorghum. Indian Journal of Genetics and Plant Breeding, 27, pp 1-& ://A1967A434000004

Murty, U.R. (1993) Appraisal on the Present Status of Research on Apomixis in Sorghum. Current Science, 64, 5, pp 315-318 ://A1993KW86600015 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Murty-Apopmixis-1993.pdf

Murty, U.R., Kirti, P.B., Sridhar, P., & Bharati, M. (1984) Genetic-Markers to Detect Apomixis in Sorghum-Bicolor L Moench. Current Science, 53, 1, pp 49-51 ://A1984RZ28200022

Murty, U.R., Schertz, K.F., & Bashaw, E.C. (1979) Apomictic and Sexual Reproduction in Sorghum. Indian Journal of Genetics and Plant Breeding, 39, 2, pp 271-278 ://A1979JE65800023 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Murty-Apomixis-Sorghum- 1979.pdf

Musabyimana, T., Sehene, C., & Bandyopadhyay, R. (1995) 334

Ergot Resistance in Sorghum in Relation to Flowering, Inoculation Technique and Disease Development. Plant Pathology, 44, 1, pp 109-115 ://A1995QN58700013

Muthusubramanian, V., Bandyopadhyay, R., Tooley, P.W., & Reddy, D.R. (2005) Inoculated host range and effect of host on morphology and size of macroconidia produced by Claviceps africana and Claviceps sorghi. Journal of Phytopathology, 153, 1, pp 1-4 ://000226471200001

Myers, N., Russell A. Mittermeier, Cristina G. Mittermeier, Gustavo A. B. da Fonseca & Jennifer Kent (2000) Biodiversity hotspots for conservation priorities. Nature, 403, pp 853 - 858 http://www.botanischergarten.ch/BiodivVorles/MyersHotSpotsNature.pdf

Naeem, S. (2002) Ecosystem consequences of biodiversity loss: The evolution of a paradigm. Ecology, 83, 6, pp 1537-1552 ://000176400800006

Naik, M.S. (1968) Lysine and Tryptophan in Protein Fractions of Sorghum. Indian Journal of Genetics and Plant Breeding, 28, 2, pp 142-& ://A1968C120600005

Nair, N.V., Selvi, A., Sreenivasan, T.V., Pushpalatha, K.N., & Mary, S. (2006) Characterization of intergeneric hybrids of Saccharum using molecular markers. Genetic Resources and Crop Evolution, 53, 1, pp 163-169 ://000234874500017 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Nair-Intergeneric-2006.pdf

Nature Genomics Gateway (2003 ff) Electronic Source: Genomics Gateway, Nature http://www.nature.com/genomics/papers/

Navi, S.S., Bandyopadhyay, R., Reddy, R.K., Thakur, R.P., & Yang, X.B. (2005) Effects of wetness duration and grain development stages on sorghum grain mold infection. Plant Disease, 89, 8, pp 872-878 ://000230773100014

Needham, J. & Bray, F. (1984) Science and Civilisation in China Volume 6, Biology and Biological Technology, Part 2 Agriculture Cambridge University Press, Cambridge UK, IS: (ISBN-13: 9780521250764 | ISBN-10: 0521250765), pp 768 http://www.cambridge.org/catalogue/catalogue.asp?isbn=9780521250764

Neto, M.M.G., Obeid, J.A., Pereira, O.G., Cecon, P.R., de Queiroz, A.C., Zago, C.P., Candido, M.J.D., & Miranda, L.F. (2004) Sorghum (Sorghum bicolor (L.) Moench) hybrids cultivated under increasing fertilization levels. Agronomic characteristics, soluble and structural carbohydrates of the plant. Revista Brasileira De Zootecnia-Brazilian Journal of Animal Science, 33, 6, pp 1975-1984 ://000228133900008

Nevo, E. (1998) Genetic diversity in wild cereals: Regional and local studies and their bearing on conservation ex situ and in situ. Genetic Resources and Crop Evolution, 45, 4, pp 355-370 ://000075583700010

Nicoll, K. (2004) Recent environmental change and prehistoric human activity in Egypt and Northern Sudan. Quaternary Science Reviews, 23, 5-6, pp 561-580 ://000220163100003

Nordborg, M. (2000) Linkage disequilibrium, gene trees and selfing: An ancestral recombination graph with partial self-fertilization. Genetics, 154, 2, pp 923-929 ://000085178700037

Nordborg, M., Borevitz, J.O., Bergelson, J., Berry, C.C., Chory, J., Hagenblad, J., Kreitman, M., Maloof, J.N., Noyes, T., Oefner, P.J., Stahl, E.A., & Weigel, D. (2002) The extent of linkage disequilibrium in Arabidopsis thaliana. Nature Genetics, 30, 2, pp 190-193 ://000173708700020

O'Brien, L. (1999) Genotype and environment effects on feed grain quality. Australian Journal of Agricultural Research, 50, 5, pp 703-719 ://000080634300005

335

O'Kennedy, M.M., Grootboom, A., & Shewry, P.R. (2006) Harnessing sorghum and millet biotechnology for food and health. Journal of Cereal Science, 44, 3, pp 224-235 ://000242699500002 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/OKennedy-Harnessing- 2006.pdf

Ogunbayo, S.A., Ojo, D.K., Guei, R.G., Oyelakin, O.O., & Sanni, K.A. (2005) Phylogenetic diversity and relationships among 40 rice accessions using morphological and RAPDs techniques. African Journal of Biotechnology, 4, 11, pp 1234-1244 ://000234300800006

Ogunlela, V.B. (1983) Striga, Fertilizer Application and Hybrid Sorghum Performance in the Nigerian Savanna. Zeitschrift Fur Acker Und Pflanzenbau-Journal of Agronomy and Crop Science, 152, 3, pp 208-218 ://A1983RA04700005

Ogunlela, V.B. & Eastin, J.D. (1985) Aftereffect of Elevated Night Temperature and Heat-Preconditioning on Net Carbon-Dioxide Exchange and Grain Development in Sorghum-Bicolor L. Zeitschrift Fur Acker Und Pflanzenbau-Journal of Agronomy and Crop Science, 154, 3, pp 182-192 ://A1985AHN9100005

Okonkwo, C.A.C. & Onoenyi, F.I. (1998) Morphological characters as indicators of drought resistance in diverse sorghum varieties in the scrub savannah, Nigeria. Tropical Agriculture, 75, 4, pp 440-444 ://000082263300006

Olah, B. (1981) Analysis of the Morphological-Characteristics of the Panicle in 4 Populations of Sorghum Halepense (L) Pers. Acta Agronomica Academiae Scientiarum Hungaricae, 30, 3-4, pp 383-406 ://A1981MA16200011

Oliver, A.L., Pedersen, J.F., Grant, R.J., Klopfenstein, T.J., & Jose, H.D. (2005) Comparative effects of the sorghum bmr-6 and bmr-12 genes: II. Grain yield, stover yield, and stover quality in grain sorghum. Crop Science, 45, 6, pp 2240-2245 ://000233147100011

Ollitrault, P., Arnaud, M., & Chantereau, J. (1989) Enzyme Polymorphism in Sorghum .2. Genetic and Evolutionary Constitution of Cultivated Sorghum. Agronomie Tropicale, 44, 3, pp 211-222 ://A1989DL10100008

Orak, A. (2001) The effect of different sowing dates on yield and yield components of hybrid sorghum (Sorghum bicolor Moench.) in dryland conditions. Cereal Research Communications, 29, 1-2, pp 183-187 ://000169442000024

Ortiz-Garcia, S., Ezcurra, E., Schoel, B., Acevedo, F., Soberon, J., & Snow, A.A. (2005) Absence of detectable transgenes in local landraces of maize in Oaxaca, Mexico (2003-2004). PNAS, 102, 35, pp 12338-12343 http://www.pnas.org/cgi/content/abstract/102/35/12338 and http://www.botanischergarten.ch/Mexico/Ortiz-Snow-PNAS-2005.pdf

Oswalt, D.L. (1973) Nutritional Quality of Sorghum Bicolor (L.) Moench Grain in Inheritance and Improvement of Protein Quality and Content in Sorghum., Dept. Agron.Agric. Exp. Sta. Purdue, Lafayette, Ind pp 144 Res. Prog. Pep.. Lafayette, Ind. (Report)

OTA (1993) Harmful Non-Indigenous Species in the United States. In Office of Technology Assessment (ed U.S. Congress). U.S. Government Printing Office, Washington DC http://www.wws.princeton.edu/~ota/ns20/alpha_f.html

Pagnol, V., Padhukasahasram, B., Wall, J.D., Marjoram, P., & Nordborg, M. (2006) Relative influences of crossing over and gene conversion on the pattern of linkage disequilibrium in Arabidopsis thaliana. Genetics, 172, 4, pp 2441-2448 ://000237225800035

Parodi, L. (1943) Una nueve especie de Sorghum cultivada en la Argentina. Rev. Argent. Agron., 10, pp 361-372

Pasquet, R.S. (2000) Allozyme diversity of cultivated cowpea Vigna unguiculata (L.) Walp. Theoretical and Applied Genetics, 101, 1-2, pp 211-219 336

://000088403800031

Paterson, A.H., Bowers, J.E., & Chapman, B.A. (2004) Ancient polyploidization predating divergence of the cereals, and its consequences for comparative genomics. Proceedings of the National Academy of Sciences of the United States of America, 101, 26, pp 9903-9908 ://000222405600069 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Paterson-Polyploidization-2004.pdf

Paterson, A.H., Lin, Y.R., Li, Z.K., Schertz, K.F., Doebley, J.F., Pinson, S.R.M., Liu, S.C., Stansel, J.W., & Irvine, J.E. (1995a) Convergent Domestication of Cereal Crops by Independent Mutations at Corresponding Genetic-Loci. Science, 269, 5231, pp 1714- 1718 ://A1995RV94800036

Paterson, A.H., Schertz, K.F., Lin, Y.R., & Li, Z. (1998) Case history in plant domestication: sorghum, an example of cereal evolution. In Molecular dissection of complex traits (ed A.H. Paterson), pp. 187-195. CRC Press Boca Raton, Florida, New York

Paterson, A.H., Schertz, K.F., Lin, Y.R., Liu, S.C., & Chang, Y.L. (1995b) The Weediness of Wild Plants - Molecular Analysis of Genes Influencing Dispersal and Persistence of Johnsongrass, Sorghum- Halepense (L) Pers. Proceedings of the National Academy of Sciences of the United States of America, 92, 13, pp 6127-6131 ://A1995RF05000076 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Paterson-PNAS-1995.pdf

Pazoutova, S., Bandyopadhyay, R., Frederickson, D.E., Mantle, P.G., & Frederiksen, R.A. (2000) Relations among sorghum ergot isolates from the Americas, Africa, India, and Australia. Plant Disease, 84, 4, pp 437-442 ://000085994700009

Pazoutova, S. & Frederickson, D.E. (2005) Genetic diversity of Claviceps africana on sorghum and Hyparrhenia. Plant Pathology, 54, 6, pp 749-763 ://000233515500005

Pecetti, L. & Damania, A.B. (1996) Geographic variation in tetraploid wheat (Triticum turgidum ssp turgidum convar durum) landraces from two provinces in Ethiopia. Genetic Resources and Crop Evolution, 43, 5, pp 395-407 ://A1996VF75500002

Pedersen, J.F., Marx, D.B., & Funnell, D.L. (2003) Use of A(3) cytoplasm to reduce risk of gene flow through sorghum pollen. Crop Science, 43, 4, pp 1506-1509 ://000183762800035 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Pedersen-Reduce-2003.pdf

Pedersen, J.F. & Toy, J.J. (2001) Germination, emergence, and yield of 20 plant-color, seed-color near-isogenic lines of grain sorghum. Crop Science, 41, 1, pp 107- 110 ://000169469600019 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Pedersen-Isogenic-Sorghum- 2001.pdf

Pedersen, J.F., Toy, J.J., & Johnson, B. (1998) Natural outcrossing of sorghum and sudangrass in the central great plains. Crop Science, 38, 4, pp 937-939 ://000075054600008

Pendleton, B.B., Teetes, G.L., & Peterson, G.C. (1994) Phenology of Sorghum Flowering. Crop Science, 34, 5, pp 1263-1266 ://A1994PF01600023

Peng, Y., Schertz, K.F., Cartinhour, S., & Hart, G.E. (1999) Comparative genome mapping of Sorghum bicolor (L.) Moench using an RFLP map constructed in a population of recombinant inbred lines. Plant Breeding, 118, 3, pp 225-235 ://000081518500005

Piper, J.K. & Kulakow, P.A. (1994) Seed Yield and Biomass Allocation in Sorghum-Bicolor and F1 and Backcross Generations of Sorghum-Bicolor X Sorghum-Halepense Hybrids. Canadian Journal of Botany-Revue Canadienne De Botanique, 72, 4, pp 468-474 ://A1994NN38200008 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Piper-hybrid-yield-1994.pdf

Plumley, J. (1970) Qasr Ibrim. J. Egypt Archaeol., 56, pp 12-18

Poletti, S. & Sautter, C. (2005) Biofortification of the crops with micronutrients using plant breeding and/or transgenic strategies. Minerva Biotecnologica, 17, 1, pp 1-11 337

://000231035900001

Porteres, R. (1962) History of Primary Food Crops in Africa. Journal of African History, 3, 2, pp 195-210 ://A1962CGF5900002

Posnansky, M. (1969) 'Bigo bya Mugenyi'. Uganda Journal, 33, pp 125-150

Potrykus, I., Saul, M.W., Petruska, J., Paszkowski, J., & Shillito, R.D. (1985) Direct gene transfer to cells of a graminaceous monocot. Molecular and General Genetics MGG (now Molecular Genetics and Genomics), 199, 2, pp 183-188 http://dx.doi.org/10.1007/BF00330257 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Potrykus-Transfer- 1985.pdf

Prain, D. (1903) ‘Flora of the Sundribuns, pp 231–370, p 357.

Price, H.J., Dillon, S.L., Hodnett, G., Rooney, W.L., Ross, L., & Johnston, J.S. (2005a) Genome evolution in the genus Sorghum (Poaceae). Annals of Botany, 95, 1, pp 219-227 ://000226370900015 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Price-Genome-Evolution-Sorghum- 2005.pdf

Price, H.J., Hodnett, G.L., Burson, B.L., Dillon, S.L., & Rooney, W.L. (2005b) A Sorghum bicolor x S-macrospermum hybrid recovered by embryo rescue and culture. Australian Journal of Botany, 53, 6, pp 579- 582 ://000232225600010 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Price-hybrid-embryorescue- 2005.pdf

Prom, L.K., Isakeit, T., Odvody, G.N., Rush, C.M., & Kaufman, H.W. (2005a) Survival of Claviceps africana within sorghum panicles at several Texas locations. Plant Disease, 89, 1, pp 39-43 ://000226014800007

Prom, L.K., Lopez, J.D., & Latheef, M.A. (2003) Transmission of Claviceps africana spores from diseased to non-infected sorghum by corn earworm moths, Helicoverpa zea. Journal of Sustainable Agriculture, 21, 4, pp 49-58 ://000182683300004

Prom, L.K., Lopez, J.D., & Mayalagu, G.P. (2005b) Passive transmission of sorghum ergot (Claviceps africana) by four species of adult stink bugs. Southwestern Entomologist, 30, 1, pp 29-34 ://000232393300005

Quist, D. & Chapela, I. (2001) Transgenic DNA introgressed into trditional maize landraces in Oaxaca, Mexico. Nature, 414, pp 541-543 http://www.botanischergarten.ch/Bt/Quist-Chapela-Oaxaca-2002.pdf

Rabinowicz, P.D. & Bennetzen, J.L. (2006) The maize genome as a model for efficient sequence analysis of large plant genomes. Current Opinion in Plant Biology, 9, 2, pp 149- 156 ://000236167100010 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Rabinowicz-Maize- Modelgenome-2006.pdf

Radchenko, E.E. (2000) Identification of genes for resistance to greenbug in sorghum. Russian Journal of Genetics, 36, 4, pp 408-417 ://000086934500010

Raghuwanshi, R.K.S., Nema, M.L., Dubey, D.D., Umat, R., & Thakur, H.S. (1988) Manurial Requirement of a Long-Term Hybrid Sorghum-Wheat Cropping Sequence. Indian Journal of Agronomy, 33, 3, pp 295-299 ://A1988AK41100017

Rami, J.F., Dufour, P., Trouche, G., Fliedel, G., Mestres, C., Davrieux, F., Blanchard, P., & Hamon, P. (1998) Quantitative trait loci for grain quality, productivity, morphological and agronomical traits in sorghum (Sorghum bicolor L. Moench). Theoretical and Applied Genetics, 97, 4, pp 605-616 ://000076269000013

Rao, N.G.P., Narayana, L.L., & Reddy, R.N. (1978) Apomixis and Its Utilization in Grain-Sorghum .1. Embryology of 2 Apomictic Parents. Caryologia, 31, 4, pp 427-433 338

://A1978GX66000002

Rao, N.P.G. & Murty, U.R. (1973) Further Studies on Obligate Apomixis in Grain-Sorghum. Indian Journal of Genetics and Plant Breeding, 32, 3, pp 379-383 ://A1973Q784900010

Rao Prasada, K.E., Mengesha, M.H., & Reddy, V.G. (1989) International Use of Germ Plasm Collection. In The Use of Plant Genetic Resources (ed A.H.D. Brown), pp. 49-67. Cambridge University Press, Cambridge

Rao Prasada, K.E.P. & Rao, N.K. (1990) Sorghum-Nitidum (Vah1) Pers, Occurrence, Morphology and Cytology. Proceedings of the Indian Academy of Sciences-Plant Sciences, 100, 5, pp 333-336 ://A1990FD10600006

Rao, S.A., Rao Prasada, K.E.P., Mengesha, M.H., & Reddy, V.G. (1996) Morphological diversity in sorghum germplasm from India. Genetic Resources and Crop Evolution, 43, 6, pp 559-567 ://A1996VQ35800008 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Rao-Diversity-1996.pdf

Ravi, S.B. (1993) Apomixis in Sorghum Line R473 - Truth or Myth - a Critical Analysis of Published Work. Current Science, 64, 5, pp 306-314 ://A1993KW86600014 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Ravi-Apomixis-1993.pdf

Reddy, N.P.E. & Jacobs, M. (2002) Sorghum lysine-rich cultivar verification by SDS-PAGE and Southern blot. Acta Physiologiae Plantarum, 24, 3, pp 285-290 ://000178657100009

Reddy, R.N., Narayana, L.L., & Rao, N.G.P. (1979) Apomixis and Its Utilization in Grain-Sorghum .2. Embryology of F1 Progeny of Reciprocal Crosses between R473 and 302. Proceedings of the Indian Academy of Sciences-Plant Sciences, 88, 6, pp 455-461 ://A1979HW88000006

Redfearn, D.D., Burton, D.R., & Devine, T.E. (1999) Sorghum intercropping effects on yield, morphology, and quality of forage soybean. Crop Science, 39, 5, pp 1380-1384 ://000082813800017

Repellin, A., Baga, M., Jauhar, P.P., & Chibbar, R.N. (2001) Genetic enrichment of cereal crops via alien gene transfer: New challenges. Plant Cell Tissue and Organ Culture, 64, 2-3, pp 159-183 ://000168821300007

Retta, A., Vanderlip, R.L., Higgins, R.A., & Moshier, L.J. (1996) Application of SORKAM to simulate shattercane growth using forage sorghum. Agronomy Journal, 88, 4, pp 596-601 ://A1996UZ53900016

Retta, A., Vanderlip, R.L., Higgins, R.A., Moshier, L.J., & Feyerherm, A.M. (1991) Suitability of Corn Growth-Models for Incorporation of Weed and Insect Stresses. Agronomy Journal, 83, 4, pp 757-765 ://A1991GA44300021

Rich, P.J., Grenier, U., & Ejeta, G. (2004) Striga resistance in the wild relatives of sorghum. Crop Science, 44, 6, pp 2221-2229 ://000225153000052

Riley, K.W. (1980) Inheritance of Lysine Content, and Environmental Responses of High and Normal Lysine Lines of Sorghum bicolor (L) Moench in the Semi-arid Tropics of India., University of Manitoba, Canada., Manitoba Thesis, pp

Ritland, K. (2000) Marker-inferred relatedness as a tool for detesting heritability in nature. Molecular Ecology, 9, 9, pp 1195-1204 ://000089424200001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Ritland-Marker-Relatedness- 2000.pdf

Roberts, E.H. (1975) Problems of long-term storage of seed and pollen for genetic resources conservation. In Crop Genetics Resources for Today and Tomorrow. (eds O.H. Frankel & J.G. Hawkes). Cambridge University Press, Cambridge, England

Robinson, K. (1966) The Leopard's Kopje culture, it., position in the Iron Age of Southern Rodesia. South African Archaeol. Bull., 21, pp 5-51

339

Rodriguez-Herrera, R., Waniska, R.D., Rooney, W.L., Aguilar, C.N., & Contreras-Esquivel, J.C. (2006) Antifungal proteins during sorghum grain development and grain mould resistance. Journal of Phytopathology, 154, 9, pp 565-571 ://000239860600009

Rooney, W.L. (2004) Sorghum improvement - Integrating traditional and newtechnology to produce improved genotypes. In Advances in Agronomy, Vol. 83, pp. 37-109. ELSEVIER ACADEMIC PRESS INC, San Diego Advances in Agronomy ://000223638100002 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Rooney-Sorghum-Improvement- 2004-A.pdf

Rosalesrobles, E. (1993) Postemergence Shattercane (Sorghum-Bicolor) Control in Corn (Zea-Mays) in Northern Tamaulipas, Mexico. Weed Technology, 7, 4, pp 830-834 ://A1993MW67000006

Ross, W.M. & Wilson, J.A. (1969) Polyembryony in Sorghum. Crop Science, 9, 6, pp 842-& ://A1969F147000058 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Ross-Polyembryony-Sorghum- 1969.pdf

Ruskin, F.R., Axtell, J., Burton, G.W., Harlan, J.R., & Rachie, K. (1996) Sorghum. In Lost Crop of Africa, Vol.1, Grains (eds N. Borlaugh, G.W. Burton, J.R. Harlan & K. Rachie), Vol. 1, pp. 127 ff. Natiional Academy Press, Washington.NAS Lost Crops http://www.nap.edu/catalog/2305.html

Sai, N.S., Visarada, K., Lakshmi, Y.A., Pashupatinath, E., Rao, S.V., & Seetharama, N. (2006) In vitro culture methods in sorghum with shoot tip as the explant material. Plant Cell Reports, 25, 3, pp 174-182 ://000235903300003 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Sai-invitro-Sorghum-2006.pdf

Sairam, R., Chennareddy, S., Parani, M., Zhang, S.L., Al-Abed, D., Abou-Alaiw, W., & Goldman, S. (2005) OBPC Symposium: Maize 2004 & Beyond - Plant regeneration, gene discovery, and genetic engineering of plants for crop improvement. In Vitro Cellular & Developmental Biology-Plant, 41, 4, pp 411-423 ://000231708100009

Salinas, I., Pro, A., Salinas, Y., Sosa, E., Becerril, C.M., Cuca, M., Cervantes, M., & Gallegos, J. (2006) Compositional variation amongst sorghum hybrids: Effect of kafirin concentration on metabolizable energy. Journal of Cereal Science, 44, 3, pp 342-346 ://000242699500011 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Salinas-Compositional- 2006.pdf

Sane, A.P., Nath, P., & Sane, P.V. (1996) Cytoplasmic male sterility in sorghum: Organization and expression of mitochondrial genes in Indian CMS cytoplasms. Journal of Genetics, 75, 2, pp 151-159 ://A1996WJ56500002

Sangduen, N. & Hanna, W.W. (1984) Chromosome and Fertility Studies on Reciprocal Crosses between 2 Species of Autotetraploid Sorghum - Sorghum-Bicolor (L) Moench and Sorghum-Halepense (L) Pers. Journal of Heredity, 75, 4, pp 293-296 ://A1984TB43800012 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Sangduen-Crosses-1984.pdf

SAS Instiute (1992) SAS/STAT Users’ Guide.Volume 1 SAS Institute Inc., Cary, NC, pp

Savolainen, V., Cuenoud, P., Spichiger, R., Martinez, M.D.P., Crevecoeur, M., & Manen, J.F. (1995) The Use of Herbarium Specimens in DNA Phylogenetics - Evaluation and Improvement. Plant Systematics and Evolution, 197, 1-4, pp 87-98 ://A1995TE07200007

Schechter, Y. & De Wet, J.M.J. (1975) Comparativa electrophoresis and isozyme analysis of seed proteins from cultivated races of sorghum. Amer. J. Bot., 62, 3, pp 254- 261

Schertz, K.F. & Johnson, J.W. (1984) A Method for Selecting Female Parents of Grain-Sorghum Hybrids. Crop Science, 24, 3, pp 492-494 ://A1984SS50100015

Schertz, K.F. & Ritchey, J.M. (1978) Cytoplasmic-Genic Male-Sterility Systems in Sorghum. Crop Science, 18, 5, pp 890-893 340

://A1978FW50200055

Schiemann, J. (2003) Co-existence of genetically modified crops with conventional and organic farming. Environmental Biosafety Research, 2, 4, pp 213- 214 http://www.botanischergarten.ch/Coexistence/Schiemann-Editorial-EBR-2003.pdf

Schmidt, M. & Bothma, G. (2006) Risk Assessment for Transgenic Sorghum in Africa: Crop-to-Crop Gene Flow in Sorghum bicolor (L.) Moench. Crop Sci, 46, 2, pp 790- 798 http://crop.scijournals.org/cgi/content/abstract/46/2/790 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum- Lit/Schmidt-Geneflow-CropSci-2006.pdf

Scopel, A.L., Ballare, C.L., & Ghersa, C.M. (1988) Role of Seed Reproduction in the Population Ecology of Sorghum-Halepense in Maize Crops. Journal of Applied Ecology, 25, 3, pp 951-962 ://A1988R565500014

Sharma, D.C. (1977) Some Considerations in Hybrid Sorghum Seed Production in Nicaragua. Turrialba, 27, 2, pp 202-203 ://A1977DW52300016

Sharma, H.C. & Hariprasad, K.V. (2002) Flowering events in sorghum in relation to expression of resistance to sorghum midge, Stenodiplosis sorghicola. Euphytica, 127, 3, pp 411-419 ://000178239200012

Shepard, R.N. & Kruskal, J.B. (1964) Nonmetric Methods for Scaling and for Factor-Analysis. American Psychologist, 19, 7, pp 557-558 ://A1964CCK5600459

Shertz, K.F. & Dalton, L.G. (1980) Sorghum. In Hybridization of crop plants. (eds W. Fehr & H. Hadley), pp. 577–588. ASA,, Madison, WI

Singh, M., Singh, T., & Singh, H. (1986) Studies on Plant Geometry and Nitrogen-Fertilization of Hybrid Sorghum. Indian Journal of Agronomy, 31, 1, pp 33-36 ://A1986E556300008

Singh, R. & Axtell, J.D. (1973) High Lysine Mutant Gene (Hl) That Improves Protein Quality and Biological Value of Grain-Sorghum. Crop Science, 13, 5, pp 535-539 ://A1973R255700012 AND NEBIS

Singh, S., Srivastava, U.S.L., & Singh, B. (1983) Studies on Planting System and Intercropping in Hybrid Sorghum under Rainfed Conditions. Indian Journal of Agronomy, 28, 1, pp 92- 94 ://A1983RY63300024

Singh, S.P. (1976) Modified Vitreous Endosperm Recombinants from Crosses of Normal and High Lysine Sorghum. Crop Science, 16, 2, pp 296-297 ://A1976BP29200037

Singh, T., Singh, H., & Singh, M. (1987) Effect of Plant Geometry and Nitrogen on Yield and N Uptake in Hybrid Sorghum. Indian Journal of Agronomy, 32, 4, pp 447-448 ://A1987T775500035

Slingerland, M.A., Klijn, J.A.E., Jongman, R.H.G., & van der Schans, J.W. (2003) The unifying power of sustainable development. Towards balanced choices between People, Planet and Profit in agricultural production chains and rural land use: the role of science., Wageningen University pp 1-94 WUR-report Sustainable Development Wageningen (Report) http://www.botanischergarten.ch/Discourse/Slingerland-Unifying-Power-2003.pdf

Slingerland, M.A., Traore, K., Kayode, A.P.P., & Mitchikpe, C.E.S. (2006) Fighting Fe deficiency malnutrition in West Africa: an interdisciplinary programme on a food chain approach. Njas-Wageningen Journal of Life Sciences, 53, 3-4, pp 253-279 ://000238275500002 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Slingerland-Deficiency-Sorghum- 2006.pdf AND http://www.botanischergarten.ch/Facultyof1000/Slingerland-F1000-2006.pdf

Smeda, R.J., Currie, R.S., & Rippee, J.H. (2000) 341

Fluazifop-P resistance expressed as a dominant trait in sorghum (Sorghum bicolor). Weed Technology, 14, 2, pp 397-401 ://000166693500022 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Smeda-Fluazifop-2000.pdf

Smith, C.W. & Frederickson, R.A., eds. (2000) Sorghum: Origin, history, technology, and production., pp 840, Wiley Series in Crop Science, Wiley & Sons, New York., IS: ISBN-10: 0471242373 ISBN-13: 978-0471242376

Smith, J.S.C., Kresovich, S., Hopkins, M.S., Mitchell, S.E., Dean, R.E., Woodman, W.L., Lee, M., & Porter, K. (2000) Genetic diversity among elite sorghum inbred lines assessed with simple sequence repeats. Crop Science, 40, 1, pp 226-232 ://000085505300036

Snowden, J. (1936) The cultivated races of Sorghum. Allard and Son, London, pp

Snowden, J. (1955) The wild fodder sorghums of the section Eu-sorghum. J. Linnean Soc. Bot., 55, pp 191-260 cannot be found in CH

Snowden, J. (1961) The classification and nomenclature of Sorghum Section Sorghum. Sorghum Newsletter, Sorghum Research Committee U.S.A., 4, pp 67

Sobral, B.W.S., Braga, D.P.V., Lahood, E.S., & Keim, P. (1994) Phylogenetic Analysis of Chloroplast Restriction Enzyme Site Mutations in the Saccharinae Griseb Subtribe of the Andropogoneae Dumort Tribe. Theoretical and Applied Genetics, 87, 7, pp 843-853 ://A1994MY86300012 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Sobral-Phylogenetic- Saccharinae-1994.pdf

Sokal, R. & Mitchener, C. (1958) A statistical method for evluating systematic relationships. Univ. of Kansas Sci. Bull., 38, pp 1409-1438

Sokal, R. & Sneath, P. (1963) Principles of Numerical Taxonomy. W.H. Freeman, San Francisco, pp

Soltis, D.E. & Soltis, P.S. (1993) Molecular-Data and the Dynamic Nature of Polyploidy. Critical Reviews in Plant Sciences, 12, 3, pp 243-273 ://A1993LJ92200004

Soltis, D.E. & Soltis, P.S. (1995) The Dynamic Nature of Polyploid Genomes. Proceedings of the National Academy of Sciences of the United States of America, 92, 18, pp 8089-8091 ://A1995RR84400001 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Soltis-Dynamic-Polyploids- 1995.pdf

Song, K.M., Lu, P., Tang, K.L., & Osborn, T.C. (1995) Rapid Genome Change in Synthetic Polyploids of Brassica and Its Implications for Polyploid Evolution. Proceedings of the National Academy of Sciences of the United States of America, 92, 17, pp 7719-7723 ://A1995RP74800025 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Song-Rapid-Polyploids- 1995.pdf

Song, Z.P., Lu, B.R., & Chen, J.K. (2004a) Pollen flow of cultivated rice measured under experimental conditions. Biodiversity and Conservation, 13, 3, pp 579-590 ://000187427000007

Song, Z.P., Lu, B.R., Wang, B., & Chen, J.K. (2004b) Fitness estimation through performance comparison of F-1 hybrids with their parental species Oryza rufipogon and O-sativa. Annals of Botany, 93, 3, pp 311-316 ://000220199500009

Songa, J.M., Overholt, W.A., Mueke, J.M., & Okello, R.O. (2002) Farmers' perceptions of aspects of maize production systems and pests in semi-arid eastern Kenya: factors influencing occurrence and control of stemborers. International Journal of Pest Management, 48, 1, pp 1-11 ://000173233900001

Sonune, S.P. & Kalra, G.S. (1984) Irrigation and Fertilization Studies on Hybrid Sorghum (Sorghum Bicolor L Moench) Variety Csh-8 During Rabi-Cum-Hot Weather. Indian Journal of Agronomy, 29, 2, pp 222-224 ://A1984AFP4600016 342

Spangler, R., Zaitchik, B., Russo, E., & Kellogg, E. (1999) Andropogoneae evolution and generic limits in Sorghum (Poaceae) using ndhF sequences. Systematic Botany, 24, 2, pp 267-281 ://000081954100010 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Spangler-Andropogoneae- 1999.pdf

Spangler, R.E. (2003) Taxonomy of Sarga, Sorghum and Vacoparis (Poaceae : Andropogoneae). Australian Systematic Botany, 16, 3, pp 279-299 ://000183802200001

Srivasta.Sp & Singh, A. (1970) Maturity of Hybrid Sorghum as Influenced by Fertilizer Application and Intra-Row Spacings. Indian Journal of Agricultural Sciences, 40, 12, pp 1056-& ://A1970I821000007

Stebbins, G.L. (1956) Cytogenetics and Evolution of the Grass Family. American Journal of Botany, 43, 10, pp 890-905 ://A1956WP72300018 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Stebbins-Evolution-Grass- Scheme-1956.pdf

Stemler, A.B.L., Harlan, J.R., & De Wet, J.M.J. (1975a) Caudatum Sorghums and Speakers of Chari-Nile Languages in Africa. Journal of African History, 16, 2, pp 161-& ://A1975AF93900001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Stemler-Caudatum-1975.pdf

Stemler, A.B.L., Harlan, J.R., & De Wet, J.M.J. (1975b) Symposium on Biochemical Systematics, Genetics and Origins of Cultivated Plants .4. Evolutionary History of Cultivated Sorghums (Sorghum-Bicolor- Linn Moench) of Ethiopia. Bulletin of the Torrey Botanical Club, 102, 6, pp 325-333 ://A1975BP87100005 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Stemler-Evolution-Sorghum- Ethopia-1975.pdf

Stenhouse, J., Prasada Rao, K., Gopal Reddy, V., & S., A.R. (1997) Sorghum. In Biodiversity in Trust (eds D. Fucillo, L. Sears & P. Stapleton), pp. 292-308. Cambridge University Press, Cambridge

Stephens, J.C. & Holland, R.F. (1954) Cytoplasmic Male-Sterility for Hybrid Sorghum Seed Production. Agronomy Journal, 46, 1, pp 20-23 ://A1954UC64600006

Stewart, C.N. (2005) Monitoring the presence and expression of transgenes in living plants. Trends in Plant Science, 10, 8, pp 390-396 ://000231484900007

Stewart, C.N., Halfhill, M.D., & Warwick, S.I. (2003) Transgene introgression from genetically modified crops to their wild relatives. Nature Reviews Genetics, 4, 10, pp 806-817 http://www.botanischergarten.ch/Monitoring/Stewart-Introgression-Review-2003.pdf

Sticklen, M.B. & Oraby, H.F. (2005) Invited review: Shoot apical meristem: A sustainable explant for genetic transformation of cereal crops. In Vitro Cellular & Developmental Biology-Plant, 41, 3, pp 187-200 ://000230506000001

Stokstad, E. (2001) ECOLOGY: Parasitic Wasps Invade Hawaiian Ecosystem. Science, 293, 5533, pp 1241- http://www.sciencemag.org

Sturgis, P., Cooper, H., & Fife-Schaw, C. (2005) Attitudes to biotechnology: estimating the opinions of a better-informed public. New Genetics and Society, 24, 1, pp 31-56 ://000228159800003 AND http://www.botanischergarten.ch/Discourse/Sturgis-Deficitmodel-2005.pdf

Subudhi, P.K. & Nguyen, H.T. (2000) Linkage group alignment of sorghum RFLP maps using a RIL mapping population. Genome, 43, 2, pp 240-249 ://000086581200004

Sullins, R.D., Rooney, L.W., & Rosenow, D.T. (1975) Endosperm Structure of High Lysine Sorghum. Crop Science, 15, 4, pp 599-600 ://A1975AN59700045

Sun, Y., Skinner, D.Z., Liang, G.H., & Hulbert, S.H. (1994) 343

Phylogenetic Analysis of Sorghum and Related Taxa Using Internal Transcribed Spacers of Nuclear Ribosomal DNA. Theoretical and Applied Genetics, 89, 1, pp 26-32 ://A1994PL44300005 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Sun-Phylog.Sorgh-taxa- 1994.pdf

Swanson, T. (1995) The Economics and Ecology of Biodiversity Decline Cambridge Univ. Press,, Cambridge, England, pp

Tadesse, Y., Sagi, L., Swennen, R., & Jacobs, M. (2003) Optimisation of transformation conditions and production of transgenic sorghum (Sorghum bicolor) via microparticle bombardment. Plant Cell Tissue and Organ Culture, 75, 1, pp 1-18 ://000184030600001AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Tadesse-Transform-Cond- Sorghum-2003.pdf

Tamado, T., Ohlander, L., & Milberg, P. (2002) Interference by the weed Parthenium hysterophorus L. with grain sorghum: influence of weed density and duration of competition. International Journal of Pest Management, 48, 3, pp 183-188 ://000176559700002

Tang, C.Y., Schertz, K.F., & Bashaw, E.C. (1980) Apomixis in Sorghum Lines and Their F1 Progenies. Botanical Gazette, 141, 3, pp 294-299 ://A1980KQ21800010

Tang, H.V. & Pring, D.R. (2003) Conversion of fertility restoration of the sorghum IS1112C (A3) male-sterile cytoplasm from two genes to one gene. Crop Science, 43, 5, pp 1747-1753 ://000185174300023 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Tang-Male-Sterile-2003.pdf

Tang, W., Luo, X., & Samuels, V. (2004) Regulated gene expression with promoters responding to inducers. Plant Science, 166, 4, pp 827-834 http://www.sciencedirect.com/science/article/B6TBH-4BBH614-3/2/6851a1739ee9b2e47f74742cc7c7c470 AND http://www.botanischergarten.ch/Terminator/Tang-Regul-geneexpression-2004.pdf

Tanksley, S.D. & McCouch, S.R. (1997) Seed banks and molecular maps: Unlocking genetic potential from the wild. Science, 277, 5329, pp 1063-1066 ://A1997XT03300024 AND http://www.botanischergarten.ch/Centers-Biodiv/Tanksley-seedbanks-maps-1997.pdf

Tanksley, T.D. & Escobosa, A. (1971) Protein Levels and Lysine Supplementation of Sorghum-Soy Diets for G-F Swine. Journal of Animal Science, 33, 1, pp 239-& ://A1971J782700216

Tao, Y., Manners, J.M., Ludlow, M.M., & Henzell, R.G. (1993) DNA Polymorphisms in Grain-Sorghum (Sorghum-Bicolor (L) Moench). Theoretical and Applied Genetics, 86, 6, pp 679-688 ://A1993LR53800004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Tao-DNA-Polymorph- Sorghum-1996.pdf

Taylor, J. & Shewry, P. (2006) Preface to sorghum and millets reviews. Journal of Cereal Science, 44, 3, pp 223-223 ://000242699500001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Taylor-Preface-2006.pdf

Taylor, J.R.N., Schober, T.J., & Bean, S.R. (2006) Novel food and non-food uses for sorghum and millets. Journal of Cereal Science, 44, 3, pp 252-271 ://000242699500004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Taylor-Novel-Food-2006.pdf

Tefera, T. (2004) Farmers' perceptions of sorghum stem-borer and farm management practices in eastern Ethiopia. International Journal of Pest Management, 50, 1, pp 35-40 ://000187027300006

Tenaillon, M.I., Sawkins, M.C., Anderson, L.K., Stack, S.M., Doebley, J., & Gaut, B.S. (2002) Patterns of diversity and recombination along chromosome 1 of maize (Zea mays ssp mays L.). Genetics, 162, 3, pp 1401-1413 ://000179739900034

Tenaillon, M.I., Sawkins, M.C., Long, A.D., Gaut, R.L., Doebley, J.F., & Gaut, B.S. (2001) Patterns of DNA sequence polymorphism along chromosome 1 of maize (Zea mays ssp mays L.). Proceedings of the National Academy of Sciences of the United States of America, 98, 16, pp 9161-9166 ://000170216900045

344

Teshome, A., Baum, B.R., Fahrig, L., Torrance, J.K., Arnason, T.J., & Lambert, J.D. (1997) Sorghum Sorghum bicolor (L.) Moench landrace variation and classification in north Shewa and south Welo, Ethiopia. Euphytica, 97, 3, pp 255-263 ://A1997YL65500002 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Teshome-Landraces-Ethiopia- 1997.pdf

Teshome, A., Fahrig, L., Torrance, J.K., Lambert, J.D., Arnason, T.J., & Baum, B.R. (1999a) Maintenance of Sorghum (Sorghum bicolor, Poaceae) landrace diversity by farmers' selection in Ethiopia. Economic Botany, 53, 1, pp 79-88 ://000079800200007

Teshome, A., Torrance, J.K., Baum, B., Fahrig, L., Lambert, J.D.H., & Arnason, J.T. (1999b) Traditional farmers' knowledge of Sorghum (Sorghum bicolor Poaceae ) landrace storability in Ethiopia. Economic Botany, 53, 1, pp 69-78 ://000079800200006

Tesso, T., Ejeta, G., Chandrashekar, A., Huang, C.P., Tandjung, A., Lewamy, M., Axtell, J.D., & Hamaker, B.R. (2006) A novel modified endosperm texture in a mutant high-protein digestibility/high-lysine grain sorghum (Sorghum bicolor (L.) Moench). Cereal Chemistry, 83, 2, pp 194-201 ://000236606000013 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Tesso-Endosperm-Lysine- 2006.pdf

Thebault, E. & Loreau, M. (2005) Trophic interactions and the relationship between species diversity and ecosystem stability. American Naturalist, 166, 4, pp E95- E114 ://000232270600002

Tilman, D., Polasky, S., & Lehman, C. (2005) Diversity, productivity and temporal stability in the economies of humans and nature. Journal of Environmental Economics and Management, 49, 3, pp 405-426 ://000229345100001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Tilman-Diversity-Stability- 2005.pdf

Tinker, N.A., Fortin, M.G., & Mather, D.E. (1993) Random Amplified Polymorphic DNA and Pedigree Relationships in Spring Barley. Theoretical and Applied Genetics, 85, 8, pp 976- 984 ://A1993KQ85400009

Toler, J.E., Guice, J.B., & Murdock, E.C. (1996) Interference between johnsongrass (Sorghum halepense), smooth pigweed (Amaranthus hybridus), and soybean (Glycine max). Weed Science, 44, 2, pp 331-338 ://A1996UL53900018

Tsvetova, M.I. & Elkonin, L.A. (2002) Instability of the ploidy level in autotetraploid sorghum plants from a line with variable male fertility. Russian Journal of Genetics, 38, 5, pp 526-530 ://000175638600008

Tuinstra, M.R., Ejeta, G., & Goldsbrough, P. (1998) Evaluation of near-isogenic sorghum lines contrasting for QTL markers associated with drought tolerance. Crop Science, 38, 3, pp 835-842 ://000074754600035

Tuinstra, M.R., Ejeta, G., & Goldsbrough, P.B. (1997a) Heterogeneous inbred family (HIF) analysis: a method for developing near-isogenic lines that differ at quantitative trait loci. Theoretical and Applied Genetics, 95, 5-6, pp 1005-1011 ://A1997YF71500037

Tuinstra, M.R., Grote, E.M., Goldsbrough, P.B., & Ejeta, G. (1996) Identification of quantitative trait loci associated with pre-flowering drought tolerance in sorghum. Crop Science, 36, 5, pp 1337- 1344 ://A1996VK30900043

Tuinstra, M.R., Grote, E.M., Goldsbrough, P.B., & Ejeta, G. (1997b) Genetic analysis of post-flowering drought tolerance and components of grain development in Sorghum bicolor (L.) Moench. Molecular Breeding, 3, 6, pp 439-448 ://000072253900004 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Tuinstra-Drought-Grain-1997.pdf

345

Tuinstra, M.R. & Wedel, J. (2000) Estimation of pollen viability in grain sorghum. Crop Science, 40, 4, pp 968-970 ://000089410600007

Tunstall, V., Teshome, A., & Torrance, J.K. (2001) Distribution, abundance and risk of loss of sorghum landraces in four communities in North Shewa and South Welo, Ethiopia. Genetic Resources and Crop Evolution, 48, 2, pp 131-142 ://000168335700003

Turkhede, B.B. & Prasad, R. (1978) Effect of Rates and Timings of Nitrogen Application on Hybrid Sorghum (Csh-1). Indian Journal of Agronomy, 23, 2, pp 113-126 ://A1978GB17200006

Tyagi, A.K., Mohanty, A., Bajaj, S., Chaudhury, A., & Maheshwari, S.C. (1999) Transgenic rice: A valuable monocot system for crop improvement and gene research. Critical Reviews in Biotechnology, 19, 1, pp 41-79 ://000079465100002

Tzvelev, N.N. (1989) The System of Grasses (Poaceae) and Their Evolution. Botanical Review, 55, 3, pp 141-204 ://A1989AU61200001

Uptmoor, R., Wenzel, W., Friedt, W., Donaldson, G., Ayisi, K., & Ordon, F. (2003) Comparative analysis on the genetic relatedness of Sorghum bicolor accessions from Southern Africa by RAPDs, AFLPs and SSRs. Theoretical and Applied Genetics, 106, 7, pp 1316-1325 ://000183398100022 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Uptmoor-bicolor-genetic- 2003.pdf van Jaarsveld, P.J., Marais, D.W., Harmse, E., Nestel, P., & Rodriguez-Amaya, D.B. (2006) Retention of beta-carotene in boiled, mashed orange-fleshed sweet potato. Journal of Food Composition and Analysis, 19, 4, pp 321- 329 ://000237141100008 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Jaarsveld-Betacarotene- Sweetpotato-2006.pdf AND http://www.harvestplus.org/pubs.html

Vanscoyoc, S.W., Ejeta, G., & Axtell, J.D. (1988) Kernel Characteristics and Protein-Fraction Changes During Seed Development of High-Lysine and Normal Sorghums. Cereal Chemistry, 65, 2, pp 75-80 ://A1988M582000001

Vavilov, N.I. (1926) Studies on the origin of cultivated plants State Press, Leningrad, pp 248

Vavilov, N.I. (1951) The origin, variation, immunity, and breeding of cultivated plants. In Chronica Botanica (ed C. Botanica) Selected Writings of N.I. Vavilov

Vavilov, N.I. (1987) Vavilov,N.I. - on the Centenary of His Birth. Tsitologiya I Genetika, 21, 5, pp 323-& ://A1987K743500001

Vavilov, N.L. (1940) The new systematics of cultivated plants. In The new systematics (ed J. Huxley), pp. 549-566. University Press, Oxford

Vernaillen, S., Laureys, F., & Jacobs, M. (1993) A Potential Screening System for Identifying Sorghum Ecotypes with Increased Lysine in the Seeds. Plant Breeding, 111, 4, pp 295- 305 ://A1993MU91300005

Viguier, P. (1945) Les sorghos A grains et leur culture au Soudan Française. Revue de Botanique Appliquie et Agriculture Tropicale, 25, pp 163-221 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Vignier-Sorghos-Soudan-1945.pdf

Vila, M., Williamson, M., & Lonsdale, M. (2004) Competition experiments on alien weeds with crops: lessons for measuring plant invasion impact? Biological Invasions, 6, 1, pp 59-69 ://000220965900005

Virupaka.Tk & Sastry, L.V.S. (1967) Relation of Lysine Content to Protein Levels in Sorghum. Indian Journal of Biochemistry, 4, 2, pp 32-& 346

://A19679781500103

Vogel, K.P., Pedersen, J.F., Masterson, S.D., & Toy, J.J. (1999) Evaluation of a filter bag system for NDF, ADF, and IVDMD forage analysis. Crop Science, 39, 1, pp 276-279 ://000078747300042

Von Holle, B., Delcourt, H.R., & Simberloff, D. (2003) The importance of biological inertia in plant community resistance to invasion. Journal of Vegetation Science, 14, 3, pp 425-432 ://000184495400013 AND NEBIS

Von Holle, B. & Simberloff, D. (2004) Testing Fox's assembly rule: does plant invasion depend on recipient community structure? Oikos, 105, 3, pp 551-563 ://000221438100010

Von Holle, B. & Simberloff, D. (2005) Ecological resistance to biological invasion overwhelmed by propagule pressure. Ecology, 86, 12, pp 3212-3218 ://000234066700009 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/VonHolle-Invasion-2005.pdf

Waines, J.G. & Hegde, S.G. (2003) Intraspecific gene flow in bread wheat as affected by reproductive biology and pollination ecology of wheat flowers. Crop Science, 43, 2, pp 451-463 ://000181479700001

Waniska, R.D., Venkatesha, R.T., Chandrashekar, A., Krishnaveni, S., Bejosano, F.P., Jeoung, J., Jayaraj, J., Muthukrishnan, S., & Liang, G.H. (2001) Antifungal proteins and other mechanisms in the control of sorghum stalk rot and grain mold. Journal of Agricultural and Food Chemistry, 49, 10, pp 4732-4742 ://000171615100034 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Waniska-transg-Chitinase- Sorghum-2001.pdf

Waquil, J.M., Teetes, G.L., & Peterson, G.C. (1986) Adult Sorghum Midge (Diptera, Cecidomyiidae) Nonpreference for a Resistant Hybrid Sorghum. Journal of Economic Entomology, 79, 2, pp 455-458 ://A1986A926100034

Warkentin, T.D., Marshall, G., McKenzie, R.I.H., & I.N., M. (1988) Diclofop-methyl tolerance in cultivated oats (Avena sativa L.). Weed Research, 28, pp 27-35

Warmington, E.H. (1974) The commerce between the Roman empire and India in: (*first publ., 2 vols., Cambridge University Press, 1928) 2nd ed., rev. and enl., 488 p., Curzon Press, London/Octagon Books, New York; *repr., New Delhi 1995., Cambridge UK, pp

Warwick, S.I. (1990) Allozyme and life history variation in five northwardly colonizing North American weed species. Plant Systematics and Evolution, 169, 1 - 2, pp 41-54 http://dx.doi.org/10.1007/BF00935983 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Warwick-Allozymes- 1990.pdf

Warwick, S.I. & Black, L.D. (1983) The Biology of Canadian Weeds .61. Sorghum-Halepense (L) Pers. Canadian Journal of Plant Science, 63, 4, pp 997-1014 ://A1983RX35000025 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Warwick-halepense-1983.pdf

Warwick, S.I. & Stewart Jr., C.N. (2005) Crops come from wild plants: How domestication, transgenes, and linkage together shape ferality. In Crop Ferality and Volunteerism (ed J. Gressel). CRC Press, Boca Raton http://www.botanischergarten.ch/Feral-def/Warwick-Weeds-2005.pdf

Wasylikowa, K., Mitka, J., Wendorf, F., & Schild, R. (1997) Exploitation of wild plants by the early Neolithic hunter-gatherers of the Western Desert, Egypt: Nabta Playa as a case-study. Antiquity, 71, 274, pp 932-941 ://000071072600012

Weaver, C.A., Hamaker, B.R., & Axtell, J.D. (1998) Discovery of grain sorghum germ plasm with high uncooked and cooked in vitro protein digestibilities. Cereal Chemistry, 75, 5, pp 665-670 ://000076150200018

Weltzin, J.F., Muth, N.Z., Von Holle, B., & Cole, P.G. (2003) 347

Genetic diversity and invasibility: a test using a model system with a novel experimental design. Oikos, 103, 3, pp 505-518 ://000186985900007

Wendorf, F., Close, A.E., Schild, R., Wasylikowa, K., Housley, R.A., Harlan, J.R., & Krolik, H. (1992) Saharan Exploitation of Plants 8,000 Years Bp. Nature, 359, 6397, pp 721-724 http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Wendorf-Sahara-Sorghum-8000BP-1982.pdf

Wenzel, W.G., Ayisi, K.K., Mogashoa, A., Donaldson, G., Mohammed, J., Uptmoor, R., Ordon, F., & Friedt, W. (2001a) Improved sorghum varieties for smallholder farmers. Journal of Applied Botany-Angewandte Botanik, 75, 5-6, pp 207-209 ://000173454300006

Wenzel, W.G., Van Loggerenberg, M., & Ordon, F. (2001b) Quick screening methods for sorghum quality traits. Journal of Applied Botany-Angewandte Botanik, 75, 1-2, pp 43-45 ://000169043200007

White, P.S., Gilbert, M.L., Nguyen, H.T., & Vierling, R.A. (1995) Maximum Parsimony Accurately Reconstructs Relationships of Elite Sorghum Lines. Crop Science, 35, 6, pp 1560-1565 ://A1995TD27900007

Whitham, T.G., Martinsen, G.D., Floate, K.D., Dungey, H.S., Potts, B.M., & Keim, P. (1999) Plant hybrid zones affect biodiversity: Tools for a genetic-based understanding of community structure. Ecology, 80, 2, pp 416-428 ://000079036500007 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Witham-biodiversiy-hybrids- 1999.pdf

Whitham, T.G., Morrow, P.A., & Potts, B.M. (1994) Plant Hybrid Zones as Centers of Biodiversity - the Herbivore Community of 2 Endemic Tasmanian Eucalypts. Oecologia, 97, 4, pp 481-490 ://A1994NL97900008 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Wihtham-Oecologia- Hybridzone-1994.pdf

Wiersema, J.H. & Dahlberg, J. (2007) The nomenclature of Sorghum bicolor (L.) Moench (Gramineae). Taxon, 56, 3, pp 941-946 ://000249247400030 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Wiersema-Nomenclature- Sorghum-bicolor-2007.pdf AND http://ddr.nal.usda.gov/dspace/bitstream/10113/4229/1/IND43957178.pdf

Williams, J.T. (1990) Vavilov Centers of Diversity and the Conservation of Genetic-Resources. Biological Journal of the Linnean Society, 39, 1, pp 89-93 ://A1990CN43400008 AND http://www.botanischergarten.ch/biodiversity/Williams-Vavilov-1990.pdf

Wood, D. & Lenne, J. (2001) Nature's fields: a neglected model for increasing food production. Outlook on Agriculture, 30, 3, pp 161-170 ://000171396200003 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Wood-Natures-Fields.pdf

Worstell, J.V., Kidd, H.J., & Schertz, K.F. (1984) Relationships among Male-Sterility Inducing Cytoplasms of Sorghum. Crop Science, 24, 1, pp 186-189 ://A1984SF69200044

Wright, S. (1931) Evolution in Mendelian populations. Genetics, 16, 2, pp 0097-0159 ://000201597700001

Wright, S. (1965) Factor Interaction and Linkage in Evolution. Proceedings of the Royal Society of London Series B-Biological Sciences, 162, 986, pp 80- & ://A19656315800004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Wright-Factor-Interaction- 1965.pdf

Wright, S. (1978) Variability within and among natural populations. 4 University of Chicago Press, Chicago, pp

Wright, S. (1990) Evolution in Mendelian Populations (Reprinted from Genetics, Vol 16, Pg 97-159, 1931). Bulletin of Mathematical Biology, 52, 1-2, pp 241-295 ://A1990DC78200010

Wu, T.P. (1979) Cytological and morphological studies on Sorghum roxburghii, S. propinquum and their F1 hybrid. Proc. Natl. Sci. Counc. ROC., 3, pp 291-298 348

Wu, T.P. (1993) Cytological and Morphological Relationships between Sorghum-Laxiflorum and S-Bicolor. Journal of Heredity, 84, 6, pp 484-489 ://A1993MM56500011

Wu, X.R., Chen, Z.H., & Folk, M.R. (2003) Enrichment of cereal protein lysine content by altered tRNA(lys) coding during protein synthesis. Plant Biotechnology Journal, 1, 3, pp 187-194 ://000188439800004 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Wu-et-al-Plant-Biotech-J- 2003.pdf

Xu, G.W., Cui, Y.X., Schertz, K.F., & Hart, G.E. (1995) Isolation of Mitochondrial-DNA Sequences That Distinguish Male-Sterility-Inducing Cytoplasms in Sorghum-Bicolor (L) Moench. Theoretical and Applied Genetics, 90, 7-8, pp 1180-1187 ://A1995RE88200040

Xu, G.W., Magill, C.W., Schertz, K.F., & Hart, G.E. (1994) A Rflp Linkage Map of Sorghum-Bicolor (L) Moench. Theoretical and Applied Genetics, 89, 2-3, pp 139-145 ://A1994PN76800001

Xu, W.W., Subudhi, P.K., Crasta, O.R., Rosenow, D.T., Mullet, J.E., & Nguyen, H.T. (2000) Molecular mapping of QTLs conferring stay-green in grain sorghum (Sorghum bicolor L. Moench). Genome, 43, 3, pp 461-469 ://000087279800007 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Xu-QTL-Staygreen-2000.pdf

Yamasaki, M., Tenaillon, M.I., Bi, I.V., Schroeder, S.G., Sanchez-Villeda, H., Doebley, J.F., Gaut, B.S., & McMullen, M.D. (2005) A large-scale screen for artificial selection in maize identifies candidate agronomic loci for domestication and crop improvement. Plant Cell, 17, 11, pp 2859-2872 ://000232991700004

Yang, W.P., deOliveira, A.C., Godwin, I., Schertz, K., & Bennetzen, J.L. (1996) Comparison of DNA marker technologies in characterizing plant genome diversity: Variability in Chinese sorghums. Crop Science, 36, 6, pp 1669-1676 ://A1996VZ26100042

Yim, K.O. & Bayer, D.E. (1997) Rhizome expression in a selected cross in the Sorghum genus. Euphytica, 94, 2, pp 253-256 ://A1997WY84100016 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Yim-Rhizome-Expression- Sorghum-1997.pdf

Yoshida, T. (2002) Production and breeding of sorghum and pearl millet. Japanese Journal of Crop Science, 71, 2, pp 147-153 ://000176239800001

Young, N.D. (1999) A cautiously optimistic vision for marker-assisted breeding. Molecular Breeding, 5, pp 505 - 515 http://www4.ncsu.edu/~jholland/CS746/YoungMolBreed5.pdf

Zeller, F.J. (2000) Sorghum (Sorghum bicolor L. Moench): utilization, genetics, breeding. Bodenkultur, 51, 1, pp 71-85 ://000086842100008

Zeven, A.C. (1991) 400 Years of Cultivation of Dutch White Clover Landraces. Euphytica, 54, 1, pp 93-99 ://A1991FR27500012

Zeven, A.C. (1998) Landraces: A review of definitions and classifications. Euphytica, 104, 2, pp 127-139 ://000077301500008 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Zeven-Landraces-Review- 1998.pdf

Zeven, A.C. (1999) The traditional inexplicable replacement of seed and seed ware of landraces and cultivars: A review. Euphytica, 110, 3, pp 181-191 ://000083275200005 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Zeven-inexplicable- replacement-1999.pd

Zeven, A.C. (2000) Traditional maintenance breeding of landraces: 1. Data by crop. Euphytica, 116, 1, pp 65-85 349

://000089501700008 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Zeven-Landraces-bycrop- 2000.pdf

Zeven, A.C. (2002) Traditional maintenance breeding of landraces: 2. Practical and theoretical considerations on maintenance of variation of landraces by farmers and gardeners. Euphytica, 123, 2, pp 147-158 ://000175410800001 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Zeven-Landraces-Consid- 2002.pdf

Zeven, A.C. & Zhukovsky, P.M. (1975) Dictionary of cultivated plants and their centres of diversity Pudoc, Wageningen, pp 9-26

Zhan, X., Wang, D., Bean, S.R., Mo, X., Sun, X.S., & Boyle, D. (2006) Ethanol production from supercritical-fluid-extrusion cooked sorghum. Industrial Crops and Products, 23, 3, pp 304-310 ://000237152700010 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Zhan-Ethanol-2006.pdf

Zhao, Z.Y., Cai, T.S., Tagliani, L., Miller, M., Wang, N., Pang, H., Rudert, M., Schroeder, S., Hondred, D., Seltzer, J., & Pierce, D. (2000) Agrobacterium-mediated sorghum transformation. Plant Molecular Biology, 44, 6, pp 789-798 ://000165792000009 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Zhao-Agrobacterium- mediated-2000.pdf

Zhu, H., Krishnaveni, S., Liang, G.H., & Muthukrishnan, S. (1998) Biolistic transformation of sorghum using a rice chitinase gene. J. Genetic Breeding, 52, pp 243-252

Zongo, J.D., Gouyon, P.H., Sarr, A., & Sandmeier, M. (2005) Genetic diversity and phylogenic relations among Sahelian sorghum accessions. Genetic Resources and Crop Evolution, 52, 7, pp 869- 878 ://000234591100010 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit/Zongo-Sorghumraces-BurkinaF- 2005.pdf

Zuo, J. & Chua, N.-H. (2000) Chemical-inducible systems for regulated expression of plant genes. Current Opinion in Biotechnology, 11, 2, pp 146-151 http://www.sciencedirect.com/science/article/B6VRV-40199MT-8/2/2b6525e82db738dafe6ac8e3b60a339e AND http://www.botanischergarten.ch/Terminator/Zuo-Chem-Inducible-2000.pdf

Zwick, M.S., Islam-Faridi, M.N., Zhang, H.B., Hodnett, G.L., Gomez, M.I., Kim, J.S., Price, H.J., & Stelly, D.M. (2000) Distribution and sequence analysis of the centromere-associated repetitive element CEN38 of Sorghum bicolor (Poaceae). American Journal of Botany, 87, 12, pp 1757-1764 ://000165958000002 AND http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Lit-1/Zwick-Centromere-2000.pdf

350

11. Bibliographies on publications on the biology of the genus Sorghum

Over many months the author has collected relevant papers on the topic, but it is clear that there are still many lacunes, especially if one compares with the printed bibliography of (Doggett, 1988; Smith & Frederickson, 2000), it will be important to build all those publications in the system, but this needs to be done not by merely typing the data into a retrieving system, it will be necessary to track as many publications as possible down in electronic databases over the Web of Science. The report concentrated on the biology of the genus Sorghum and touched the agronomic aspects only marginally.

Most of the references can be traced back to the database system of the Web of Science, which is only available in some consortia as the Universities of Central Europe, in a few big libraries and international organizations.

This is why the author offers here three types of lists of the more than 4625 references (status December 10, 2006), all are given in downloadable files with a hyperlink to the server of the University of Bern, Switzerland.

A shorter list in pdf format for easy reading and retrieving keywords over the Adobe search system 6.46 MB, 296 with direct links to the original text, 310 pages http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Bibliography-Sorghum-20061205.pdf

A long list of citations including the abstracts, also in pdf format, 15.4 MB, 640 pages http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Bibliography-Sorghum-Annotated-20061205.pdf

A list of over 300 selected references with a direct link to the original paper in pdf format http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Biology-links-20061205.enl

A file in xx.enl format of the database retrieving system Endnote, Version 10, this will allow easy citation in all important journal formats and a more sophisticated search system will deliver lists created with multiple search questions of complex logics. All references are downloaded with the complete set of fields including authors addresses etc. in RIS format, compatible with the other usual citation manager systems, ready to be exported with 4625 references. Direct download time of the Endnote database in enl format from the server will be usually several minutes, since it is a 10.1 MB file, within the Endnote Database system it takes only a few seconds and is fully compatible to Word within Windows XP prof. http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-20061205.enl

643 references are cited in the Sorghum Biology Document and are exported into a Endnote database (Version 10), see below about more instructions how to use this database http://www.botanischergarten.ch/Africa-Harvest-Sorghum-Bibl/Sorghum-Biology-Citations-20061205.enl

You can only read directly and with a minimal download time of several seconds this file with the 351 downloaded reference management program Endnote, from Version 7 upwards. You can download the program (Version 10X or lower) for an affordable price. See http://www.endnote.com/enpurchase.asp