568 Cytologia 33

Hybridization Range of Dichanthium annulatum (Forssk.) Stapf.

A. P. Singh1

Department of Botany and Plant Pathology, Oklahoma State University, Stillwater, Oklahoma, U. S. A.

Received June 24, 1968

The species D. annulatum (Forssk.) Stapf is widely distributed extending from the West coast of Northern Africa across India and Southeast Asia to Australia and the Fiji Islands. Along the range of its distribution this species is sympatric with a number of other species of Dichanthium Willemet as well as several species belonging to the two related genera Bothriochloa O. Kuntze and Capillipedium Stapf. Natural introgression among the members of the three genera is also known (Harlan et al. 1958, 1961; and Harlan and de Wet 1963). Thus, in order to understand the extent of genetic exchange between D. annulatum and rest of the taxa belonging to the generic group Bothriochloininae, tribe Andropogoneae, artificial crosses were made. Dichanthium annulatum is characterized by sexually reproducing diploids and facultatively apomictic tetraploid races. Representative of D. annulatum used specially as the female parent under the present study is a tetraploid (2n=40) of the Tropical ecotype which, however, is a sexual and self-sterile derivative, designated as X-98. For this reason, this plant has made possible a more extensive study of range of hybridization than has been possible with the conventional hand emasculation and pollination techniques used so effectively by Richardson (1958).

Materials and methods In all, 62 different pollen parents comprising 9 species of Dichanthium (22 collections), 10 species of Bothriochloa (35 collections), 3 species of Capillipedium (4 collections) and one species of Eremopogon (1 collection) were used. These represent plants with 2n=20 , 40, 50 and 60 chromosomes. Plants used in crossing programmes were all grown in a green house under short-day photoperiod condition. Panicles of the plant X-98 were bagged just prior to stigma exertion and the following morning pollen from desired male parents were dusted on the blooming panicles. These were immediately rebagged and harvested after about a month from the date of pollination . The seeds thus obtained in the crossing studies were germinated under controlled conditions of light, temperature and humidity according to the

1 Present address: Division of Plant Improvement , Indian Grassland and Fodder Re search Institute, Jansi, (U P.), India. 1968 Hybridization Range of Dichaethium annulatum (F orssk.) Stapf. 569

technique described by Ahring and Harlan (1961) . The sprouted seedlings were first transferred to soil in paper bands and then transplanted to a uniform nursery as outlined by Celarier and Harlan (1956) . Hybrids were identified essentially by morphological characteristics as listed by Harlan et al . (1961). These were also supported by cytological studies of the bud materials fixed in Carnoy's solution (6:3:1) and stained in acetocarmine .

Experimental results

The crossing data and the hybridization range of D . annulatum are presented in Fig. 1. On an average, 0.05 seed per panicle was obtained from 80 panicles of X-98 when selfed. Thus, theoretically , any increase in seed set over this when pollinated with pollen from other collection should represent hybrids. In general, when emasculation was not practiced , pollen of X-98 seems to interfere with foreign pollen. Most hybrids , however, were produced without emasculation.

A. Hybridization within the Dichanthium annulatum complex Artificial hybrids of X-98 with other tetraploid races of D. annulatum were easy to produce. These races are the Tropical and Mediterranean eco types and the Senegal type which may be regarded as a biotype of the Mediterranean ecotype. Crosses between the tetraploid parents resulted in hybrids with 2n=40 and 60 chromosomes, indicating that both the cytologically reduced as well as unreduced female gametes can function sexually. How ever, with the diploid race of the Tropical ecotype used as the male parent, the triploid hybrids obtained were all sterile. The hexaploid collections of D. papillosum, two (A-4080 and 4083) from South Africa and one (A-3716) from Southern Rhodesia, produced vigorous hybrids (2n=50). Such plants are primarily identified by more erect growth habit, and racemes longer and more pubescent than D. annulatum (X-98). Similarly, hybrids between D. annulatum and D. fecundum (2n=40) a collection (A-6525) from Australia, have 2n=40 chromosomes and are characterized by the dominant nature of the bisexual pedicellate spikelet (diagnostic character in D. fecundum).

B. Hybridization within the genus Dichanthium Triploid, tetraploid and hexaploid hybrids were obtained when the tetraploid D. annulatum was crossed with diploid (A-8452 from Poona, India) and tetraploid (A-5599 from the Fiji, Islands) races of D. caricosuin. Hybrids are recognised by their comparatively glabrous and somewhat obovate lower glume of the sessile spikelets. Morphological data indicated that the hexaploid hybrid (2n=60) was obtained through fertilization of a cytologically unreduced egg of D. annulatum (2n=40) by the normal male gamete of D. caricosum (2n=40). Artificial crosses of D. annulatum with the tetraploid D. aristatum 570 A. P. Singh Cytologia 33

were unsuccessful. However, when the latter was made female parent, tetraploid hybrids were obtained. Hybrids between the two species are charac terized by less pilose peduncle as compared to D. aristatum where it is almost wooly. Contrary to the tetraploid races, the diploids of both D. caricosum and D. aristatum (A-7199 from Sangli, India) crossed more easily with D. annulatum giving rise to the triploid hybrids which were reproductively sterile. Similarly, the diploid species, D. panchganiense (A-8278 from Panchgani, India) crossed readily with tetraploid D. annulatum to produce sterile triploids. Crosses with other diploid representatives of Dichanthium were not successful. With D. sericeum, a few poorly developed seeds were obtained but the seedling died in the juvenile stage.

C. Intergeneric hybridization Intergeneric crosses were attempted between D. annulatum and various species of the genus Bothriochloa, e. g. B. ambigua (2n=60), B. ewartiana (2n=40, 60), B. glabra (2n=40), B. grahamii (2n=40), B. insculpta (2n=40), B. intermedia (2n=40), B. ischaemum (2n=40, 50 and 60), B. longifolia (2n=20), B. pertusa (2n=20, 40) and B. radicans (2n=40). The only species of this genus which has so far been able to cross readily with D. annulatum is B. grahmii (2n=40). Among the collections of this species used as the pollen parents are those from British Guiana (A-2655), India (A-5450) and X-750 a synthetic hybrid (Harlan et al. 1961). In these crosses, plants with 2n=40 and 60 chromosomes were obtained. They were of normal vigor and produced a large number of seeds. None of the attempted crosses of D. annulatum with the three species of Capillipedium (e.g. C. huegelii 2n=20; C. spicigerum 2n=40 and C. parviflorum 2n=40) and one species of Eremopogon (e.g. E. feveolatus 2n=40) were successful.

Discussion Hybrids between the biotypes and ecotypes of D. annulatum as discussed by Celarier et al. (1958), and Mehra (1962), can be produced with relative ease. These, as indicated by Celarier and Harlan (1957) and Harlan et al . (1958, 1961), are facultative apomict tetraploids. Morphological variation is closely correlated with adaptation to either a tropical or desert habitat (Singh et al. 1962). For this reason the Tropical and Mediterranean types could be referred to as ecotypes, as described by Turesson (1922). Furthermore, the ease with which artificial hydrids were produced within the D. annulatum complex suggests very little or no genetic barriers between the various ecotypes of this species complex except for apomixis . In reality , the ecotypes of D. annulatum overlap in their geographic distribution and also, natural hybrids between the different ecotypes are known (Mehra 1966) . 1968 Hybridization Range of Dichanthium annulatum (Forssk .) Stapf. 571

This suggests that these ecotypes represent merely ecological variation coupled with intraspecific polyploidy. Typical representative of the genus Dichanthium are morphologically subdivided by de Wet and Richardson (1963) into two tetraploid agamospecies and two diploid sexual species groups. Turesson (1929) defined agamospecies as an apomict-population, the constituents of which for morphological, cytological or other reasons have a common origin. Within the first agamospecies con sisting of D. annulatum, D. andringitrense, D. fecundum, D. mucronulatum and D. papillosum artificial hybrids, at least among the species available for study, are possible with relative ease. Harlan et al. (1961), Borgaonkar and Singh (1962) and Mehra (1964) suggested that D. papillosum originated from

Fig. 1. Hybridization range of Dichanthium annulatum (Forssk.) Stapf. a cross between tetraploids of the Tropical and Mediterranean ecotypes of D. annulatum. The ability of the latter species to cross with D. papillosum could thus be expected. Natural hybridization apparently takes place in Australia between D. annulatum and D. fecundum (Borgaonkar and de Wet 1964) and can also be produced artificially. The relative ease with which hybrids within this agamospecies could be produced indicates the close rela tionships between the species recognised taxonomically. The second agamospecies includes D. aristatum, D. caricosum, D. tenue, D. pallidum and D. theinlwinii. Hybrids between members of this group of facultative apomicts and D. annulatumm are produced with more difficulty. However, crosses between D. annulatum (4n) and the diploid and tetraploid 572 A. P. Singh Cytologia 33 races of D. caricosum and D. aristatum are possible. Celarier et al. (1961) indicated the existence of natural hybrids between D. annulatum and D. aristatum. Crossing data suggest that D. caricosum is more closely related to D. annulatum than is D. aristatum. Furthermore, as the attempted crosses of D. annulatum with other species or biotypes of this agamospecies were not successful it appears members of this groups are more distantly related to D. annulatum than are those of the agamospecies to which the latter species belongs. As mentioned above, two diploid species groups are recognised in Di chanthium. Those endemic to Australia are D. sericeum, D. setosum, D. superciliatum and D. humilius while D. armatum, D. maccanni and D. panchganiense are confined to India (de Wet and Richardson 1963). The fact that except for D. panchganiense, no hybrid plant could be obtained between any of these species and D. annulatum, indicates that they are only distantly related to the D. annulatum species complex. The difficulties encountered in obtaining crosses at the intergeneric level as a result of failure in seed setting, frequency of lethals in some crosses and the sterility of the hybrids indicate that many of the species and genera involved in this study are valid and no gene flow is possible. However, hybrids with 2n=70 chromosomes can be produced from crosses between D. annulatum of the tropical ecotype (2n=40) and B. ischaemum (2n=50) only when cyto logically unreduced egg of the latter species is fertilized by normal male gamete of D. annulatum (Singh 1965). Crosses with D. annulatum as the female parent were all unsuccessful. This indicates some degree of genetic barrier and that the two species are phylogenetically widely separated. Weak and sterile hybrids could also suggest that in such wide crosses even apomixis could not salvage fertility. In one of their earlier papers, Harlan et al. (1958) indicated that there are some morphological reasons to suspect that the Mediterranean type of D. annulatum involves genetical material from B. ischaemum, but no experimental evidence has yet been obtained on this point. In the genus Bothriochloa, the only species with which successful artificial hybrids of D. annulatum were produced was B. grahamii which according to Bor (1960) is intermediate between D. annulatum and the gangetica race of B. intermedia. Hybrids between D. annulatum and B. grahamii would thus be expected. Crossing data presented by Harlan et al. (1962) indicated that barriers to gene flow are least between B. grahamii and D. annulatum. Indeed, on genetic grounds , B. grahamii could more naturally be included in Dichanthium. Roberty (1960) had thus rightly sug gested that Dichanthium and the genus Bothriochloa are so closely related that they should be combined into one taxonomic unit. Capillipedium is also able to cross with B. grahamii (Harlan et al. 1961) . Thus the latter may be considered, systematically as a connecting link between D. annulatum and species of Capillipedium. No direct relationship can, however , be established 1968 Hybridization Range of Dichanthium annulatum (Forssk .) Stapf. 573 between the latter two taxa. Thus, as indicated by Harlan et al . (1962), on the basis of available evidence from experimental taxonomy , lumping of the three genera Dichanthium, Bothriochloa and Capillipedium is the most nearly fitting solution at the generic level. The genus Eremopogon is completely isolated genetically from D. annulatum.

Acknowledgments The author wishes to express his sincere gratitude to late Prof. R. P. Celarier under whose kind guidance the study was carried out in the Department of Botany, Oklahoma State University, Stillwater, U. S. A. Thanks are extended also to Drs. J. M. J. de Wet and J. R. Harlan presently at the University of Illinois, U. S. A., for their valuable interest during the course of this investi gation.

References

Ahring, R. M. and Harlan, J. R. 1961. Germination studies on the Dichanthium annulatum complex. Okla. Agr. Exp. Sta. Tech. Bull. T-90: 1-15. Bor, N. L. 1960. The grasses of Burma, Cylon, India and Pakistan. Pergamon Press, Oxford. Borgoankar, D. S. and Singh, A. P. 1962. Species relationship in Dichanthium. IV. D. annulatum and D. papillosum. Phyton (Argentina) 19: 101-107. - and de Wet, J. M. J. 1964. Cytogenetical study of hybrids between Dichanthium an nulatum and D. fecundum III. Phyton (Argentina) 21: 55-60. Celarier, R. P. and Harlan, J. R. 1956. An Andropogoneae garden in Oklahoma. Taxon 5: 138-186. - and - 1957. Apomixis in Bothriochloa, Dichanthium and Capillipedium. Phytomorph. 7: 93-102. -, Mehra, K. L. and Wulf, M. L. 1958. Cytogeography of the Dichanthium annulatum complex. Brittonia 10: 59-72. - , de Wet, J. M. J. and Richardson, W. L. 1961. Species relationships in Dichanthium. I. Hybrids betweed D. caricosum, D. aristatum and D. annulatum. Phyton (Argentina) 20: 19-28. de Wet, J. M. J. and Richardson, W. L. 1963. Morphological variation and species rela tionships in Dichanthinm. Phyton (Argentina) 20: 19-28. Harlan, J. R., Celarier, R. P., Richardson, W. L., Brooks, W. H. and Mehra, K. L. 1958. Studies on Old World Blue Stems. II. Okla. Agri. Expt. Sta. Tech. Bull. T-72: 1-23. - , de Wet, J. M. J., Richardson, W. L. and Chheda, H. R. 1961. Studies on Old World Blue Stems. III. Okla. Agro. Exp. Sta. Tech. Bull. T-29: 1-30. -, Chheda, H. R. and Richardson, W. L. 1962. Range of hybridization with Bothriochloa intermedia (R. Br.) A. Camus. Crop. Sci. 2: 480-483.- and de Wet, J. M. J. 1963. The compilo-species concept. Evolution 17: 497-501. Mehra, K. L. 1962. The Dichanthium annulatum complex. I. Morphology. Phyton (Argentina) 18: 87-93. - 1964. The Dichanthium annulatum complex. III. Origin and artificial synthesis of Dichanthium papillosum Stapf. Caryologia 17: 545-556.- 1966. The Dichanthium annulatum complex. IV. Study of natural hybridization from herbarium specimens. Revista De Biologia 5: 295-302. Richardson, W. L. 1958. A technique of emasculating small grass florets. Indian Jour. Genet. and Plant Breed. 18: 69-73. 574 A. P. Singh Cytologia 33

Roberty, G. 1960. Monographic systematique des Andropogonees du globe. Boissiera 9, 1960. Singh, A. P., Harlan, J. R. and de Wet, J. M. J. 1962. Relationships within the Dichanthium annulatum complex. Proc. Okla. Acad. Sci. 42: 50-54. - 1965. Intergeneric cross of Dichanthium annulatum with Bothriochloa ischaemum. Cytologia 30: 54-57. Turesson, G. 1922. The species and the variety as ecological units. Hereditas 3: 100-113.- 1929. Zur Natur und Begrenzung der Arteinheiten. Hereditas 12: 323-333.

Cytologia Vol. 33, no. 3-4 (pp . 337-574) Issued December 25, 1968 Ausgegeben am 25 . Dezember 1968 Palu le 25 decembre 1968