The Influence of Time and Severity of Striga Infection on the Sorghum Bicolor - Striga Hermonthica Association

The influence of time and severity of Striga infection on the Sorghum bicolor - Striga hermonthica association

Promotor: Prof. dr. M.J. Kropff Hoogleraar in de Gewas- en Onkruidecologie

Co-promotor: Dr. ir. L. Bastiaans Universitair docent, leerstoelgroep Gewas- en Onkruidecologie

Promotiecommissie: Prof. dr. W.H.O. Ernst (Vrije Universiteit Amsterdam) Prof. dr. J.M. van Groenendael (Radboud Universiteit Nijmegen) Prof. dr. ir. H. van Keulen (Wageningen Universiteit) Prof. dr. Th.W. Kuyper (Wageningen Universiteit)

Dit onderzoek is uitgevoerd binnen de onderzoekschool: Production Ecology and Resource Conservation

The influence of time and severity of Striga infection on the Sorghum bicolor - Striga hermonthica association

Aad van Ast

Proefschrift ter verkrijging van de graad van doctor op gezag van de rector magnificus van Wageningen Universiteit prof. dr. M.J. Kropff in het openbaar te verdedigen op vrijdag 2 juni 2006 des namiddags te vier uur in de Aula.

Aad van Ast (2006)

The influence of time and severity of Striga infection on the Sorghum bicolor - Striga hermonthica association. van Ast, A. - [S.l.: s.n.]. Ill. PhD thesis Wageningen University. - With ref. - With summaries in English, French and Dutch ISBN: 90-8504-399-9

“Gelukkig hij, die de oorzaken van de dingen heeft leren inzien”. “Felix qui potuit rerum cognoscere causas” Vergilius

Aan Ans

Abstract

Ast, A. van, 2006. The influence of time and severity of Striga infection on the Sorghum bicolor - Striga hermonthica association. PhD thesis, Wageningen University, Wageningen, The Netherlands, with summaries in English, French and Dutch, 154 pp.

This thesis presents the results of a study on the interaction between the parasitic weed Striga hermonthica (Del.) Benth. and sorghum (Sorghum bicolor [L.] Moench). The main objective of the study was to investigate the effects of time and level of Striga infection on the interaction between host plant and parasite. Consequences for sorghum performance and the growth and development of the parasite were examined. A comparison between two sorghum cultivars differing in level of Striga tolerance, revealed that the absence of a negative effect of Striga infection on photosynthetic rate and a delayed time of first Striga infection both contributed to the lower extent of yield reduction of the tolerant cultivar. Likewise, in an experiment with a wide range of Striga seed infestation levels, it was observed that higher soil infestations levels did not only result in a higher Striga infection level, but also in an advanced time of first Striga infection. The importance of time of infection was further investigated in a pot experiment in which the time of infection was artificially delayed. Striga parasitism and reproduction, and the detrimental effects of Striga on crop performance could be strongly reduced by delaying the time of first infection. Prospects of reducing Striga parasitism by means of cultural control methods that are based on the principle of a delayed onset of Striga attachment were assessed. In a pot experiment, the combination of shallow soil tillage, deep planting and the use of transplants resulted in a four-week delay in first emergence of the parasite, a strongly reduced infection level of the sorghum host and highly improved sorghum yields. Evaluation of these methods under field conditions resulted in a 85% reduction in Striga-infection level, but as no delay in time of parasite infection was established, no beneficial effect on crop yield was obtained. Potential causes of the absence of a delay in Striga infection time under field conditions were discussed and alternative options for establishing a delayed infection in the field were proposed.

Keywords: Striga hermonthica, Sorghum bicolor, infection time, infection level, tolerance.

Preface

When writing this preface, I realize only too well how many people have contributed, directly or indirectly, in the foreground or in the background, to the completion of this thesis. Here, I like to take the opportunity to thank all of them for their contributions. I will thank you all personally, but here I can only thank a few people by name. First of all I thank my promotor, Prof. dr. Martin Kropff, for the confidence he showed in me when I asked him, shortly after his appointment as professor in Crop and Weed Science, if he was willing to supervise the PhD project I envisaged. Martin, I also like to thank you for your stimulating enthusiasm for the project, for your scientific input, and for the opportunity you gave me to carry out my research besides my regular work for the department. I am deeply indebted to my daily supervisor and co-promotor Dr. Lammert Bastiaans. His stimulating attitude, his clear ideas when discussing the experimental designs, his thorough and constructive comments on the manuscripts, and the lively discussions that resulted from them, strongly contributed to the realization and quality of this thesis. Lammert, I am very grateful for all the time you invested in me, and for your continuous confidence in a good final result. Besides, I recall with great pleasure our travels to Mali and Burkina Faso, where we combined business with pleasure. Also our common interest in sports, both active and passive, contributed much to our friendship. Lammert, I look forward to our continued collaboration in education and research for many years to come. My first steps in research on parasitic plants and weeds I took under the inspirational supervision of two persons who had since long won their spurs in this field : Dr. Siny ter Borg of the former department of Vegetation Science, Plant Ecology and Weed Science of Wageningen Agricultural University, and Dr. Arnold Pieterse of the Royal Tropical Institute in Amsterdam. I gained my first knowledge and experience in the field of parasitic plants as a research assistant in the projects on Orobanche (broomrape) conducted by Siny and Arnold. I am very grateful to the both of you for transferring your knowledge, and exciting my interest in this group of extraordinary plant species. Arnold, I also like to thank you for being my guide in Mali during my first visits to this fascinating country and its friendly population. I thank Prof. dr. Thom Kuyper, Dr. Tjeerd Jan Stomph and Dr. Harro Bouwmeester for their interest in my research and for the discussions that put my work in the broader perspective of food production and security. The research was enabled by many people around me. First, I like to thank Jet Drenth and Ans Hofman, not only for their assistance in the greenhouse experiments,

but also for taking over several of my duties in the department, which gave me time to carry out this PhD-research. I am grateful to Peter van de Putten and Peter van Leeuwen for their assistance in the photosynthesis measurements, and to Jacques Withagen for his advice on the lay-out of the experiments and the statistical analysis of the results. Gon van Laar edited the manuscript of the thesis in her well-known, accurate way, for which I am very thankful. I sincerely thank Gijsbertje Berkhout for her expertise in the financial settlement of my research activities. Une partie de la recherche a été effectuée à l’Institut d’Economie Rurale (IER) du Mali à la Station de Recherche Agronomique de Cinzana. Je voudrais remercier la direction de l’IER, et particulièrement dr. Bourema Dembélé, Directeur Scientifique, qui a bien voulu mettre à ma disposition les services et structures de son organisme. Monsieur Seriba Katilé, mon cher ami, merci beaucoup pour ton assistance dans l’organisation et l’installation des expérimentations au champ. En suite je voudrais remercier Madame Katilé et Madame Hawa Coulibaly pour leur hospitalité et les bons repas durant mes séjours à Cinzana. Les assistants de recherche, Siaka Dembélé, Amadou Tembely, Mamadou Ely Coulibaly, Mahamadou Ousmane Diarra et Badjan Bouaré sont sincèrement remerciés pour leurs contributions dans les expérimentations au champ. Je vous remercie tous pour m’avoir accueilli et m’avoir mis à l’aise à la station de recherche agricole à Cinzana. The research presented in this thesis has partly been conducted within the framework of the EU INCO-DEV program “Improved Striga Control in Maize and Sorghum” (ISCIMAS). The collaboration between the various research groups was very inspirational to me. For this I thank all the participants of the programme: Prof. dr. Mike Beale and Mrs. Vicky Child (IACR – Long Ashton Research Station, UK), Dr. Doulaye Traoré, Dr. Oumar Ouédraogo and Mrs Rokia Naomo Traoré (Institut de l’Environnement et des Recherches Agricoles; INERA, Burkina Faso), Mr. Sériba Katilé (Institut d’Économie Rurale; IER, Mali), Dr. Harro Bouwmeester (Plant Research International; PRI, The Netherlands) and Dr. Paula Westerman (Crop and Weed Ecology group, Wageningen University; CWE, The Netherlands). I also thank my fellow PhD students in Striga research. I am grateful to Jonne Rodenburg, Tom van Mourik, Venasius Wirnkar Lendzemo and Gideon Ayongwa for the interesting discussions on our mutual research projects, and for the friendly atmosphere in which the discussions took place. Venasius, I also like to thank you for your friendship and for being such an excellent host during my stay in Cameroon. I am looking forward to future collaboration. In the course of the years, many students contributed to the Striga research of the department. I am grateful, therefore, to the following students for their dedication and stimulating collaboration: Yvette Bakker, Paul Belder, Harm-Jan Bisschop, Anaïs

Bodé, Matthijn de Boer, Roy Gorkink, Petra Hogenboom, Slavica Ivanovic, Erik Jansen, Hans Keller, Diaz Kroeze, Marjolein Kruidhof, Marcius Kuster, Venasius Lendzemo, Cyril Maissan, Bianca van Mourik, Lester Narvaez, William Quaedvlieg, Barbara Remerij, Emiel Schreur, Emiel Snel, Esther Tempel, Tineke Troost, Ramon Welsing, Gerbrand Wilbrink, Erik Wijnker, Tieme Zeilmaker and Chiel Zwarts. When performing the experiments in the tropical greenhouses of Wageningen University, I could always call on the assistance of the staff of the Wageningen Plant Sciences Experimental Centre. I want to mention in particular Henk Meurs, Harry Scholten, Willem Scholten and Taede Stoker. Even though my asking for assistance was often in the nick of time, I was always assisted as soon as possible and in a friendly way. Many thanks. A special thanks to my paranymphs, Leo Vleeshouwers and Leo Oyen, for their help and moral support during the last phase of preparing this thesis. My gratitude also concerns Claire Seitz and Maja Slingerland for their help with the translation of the Summary into French and Rik Schuiling for his help with the design of the cover of this thesis. Furthermore I thank my parents for their unremitting interest and stimulating attitude towards my research. It is a profound disappointment to me that my father has not lived to witness the finalization of this thesis. He would have been very proud. I also want to thank my daughters, Mirjam, Sandra, and Irma, and their partners, Diego Heim and Peter, for their moral support, understanding, and patience. It must at times have been difficult to share the ups and downs inherent to my PhD research. The joy and relaxation I experience when being with my grandchildren, Maira and Lara, are invaluable to me. Ans, the last words of this preface are for you. I thank you in particular for your unremitting interest in my work, and for the freedom you gave me to spent so much of our spare time on the experiments and on writing this thesis. I appreciate very much that you could grant me this time and freedom, the more so as these activities were often at the expense of our common interests outside scientific research. Now the job is done, time has come for many other things that were postponed under the motto “after finishing my PhD-project”. Thanks for everything! To you I dedicate this thesis.

Aad van Ast Wageningen, June 2006

Contents

CHAPTER 1 General introduction 1

CHAPTER 2 A comparative study on Striga hermonthica interaction with a sensitive and a tolerant sorghum cultivar 11

CHAPTER 3 Time course of density effects of Striga hermonthica on the photosynthetic activity of Sorghum bicolor 31

CHAPTER 4 Effects of Striga seed infestation level on the performance of the host plant - parasite association Sorghum bicolor - Striga hermonthica 53

CHAPTER 5 The role of infection time in the differential response of sorghum cultivars to Striga hermonthica infection 71

CHAPTER 6 Cultural control measures to diminish sorghum yield loss and parasite success under Striga hermonthica infestation 91

CHAPTER 7 General discussion 109

References 117 Summary 131 Résumé 137 Samenvatting 143 Publications of the author 149 Curriculum vitae 153 Funding 154

CHAPTER 1

General introduction

Agriculture in the Sahelian zone in West Africa Agriculture is the most important economic sector in the Sahelian zone of West Africa, employing about 80% of the labour force (Kieft et al., 1994; Defoer, 2000). In traditional farming systems in sub-Saharan West Africa, shifting cultivation is the predominant cultural practice. The system is characterized by mixed cropping and the ability to provide moderate yields. In this system, farmers move to more fertile lands after short periods of cropping (2-7 years) as soil fertility and crop yields decline and weeds and pests become more abundant (Ker, 1995; Sauerborn et al., 2000). The abandoned land then reverts to natural vegetation for some years, the number of which varies greatly between and within agroecosystems, to replenish soil fertility and to reduce disease pressure and the weed seedbank. Rapid annual human population growth rate in sub-Saharan Africa (about 3%), concurrent with a general decline in yields per hectare due to drought and soil degradation, have led to an intensification of the traditional cropping systems, with shortened or even eliminated fallow periods, and cultivation on poor marginal and communal grazing lands (Jayne et al., 1989; Sauerborn et al., 2000; Vogt, 1993). The reduction of the length of fallow periods has continued to the point that the fallow system is losing its effectiveness in natural restoration of soil fertility. A downward spiral in productivity is observed, caused by depletion of nutrients, increased acidity, loss of organic matter and greater soil compaction (IITA, 1992). The predominant soils used for crop growth in this zone are classified as arenosols and luviosols. These soils are sandy, acidic, deficient in most nutrients (particularly in phosphorus and nitrogen) and low in organic matter content. The nutrient and water- holding capacities of these soils are low (Sivakumar, 1992). In these marginal areas, sorghum (Sorghum bicolor (L.) Moench) is one of the most important staple crops and grown as rainfed crop with little or no external inputs by subsistence farmers.

Sorghum Historians believe that sorghum (Sorghum bicolor (L.) Moench) originated in north- east Africa, where a large variability in wild and cultivated species is still found (Wayne Smith and Frederiksen, 2000). It probably was domesticated in Ethiopia, 5000 to 7000 years ago. From there, it was distributed along trade routes around the African continent and through the Middle East to India at least 3000 years ago.

Chapter 1

Sorghum, an important staple food crop in Africa, South Asia, and Central America, is the fifth major cereal crop in the world in terms of tonnage after maize, wheat, rice and barley (FAOSTAT, 2005). Ninety percent of the world’s area cultivated to sorghum is in developing countries, mainly in Africa and Asia. In terms of tonnage, sorghum is Africa’s second most important cereal after maize (FAOSTAT, 2005). In Africa, about 74% of sorghum produced is consumed in the home, primarily as thick or thin porridges, as a steam-cooked product such as couscous, or as a traditional beer (Africancrops, 2005). Sorghum stover is an important source of animal feed in mixed farming situations. Sorghum, a largely self-pollinating crop, primarily is a crop of hot, semi-arid tropical areas with 400-600 mm rainfall that are too dry for maize cultivation. Its deep- penetrating roots are mainly responsible for this drought tolerance, but the plant has other morphological and physiological characteristics that also contribute to its adaptation to dry conditions (Langer and Hill, 1982). For instance, waxy bloom on the leaves and stem reduce epidermal water loss. Sorghum also conserves moisture by reducing transpiration by rolling its leaves and closing the stomata under conditions of soil moisture deficit. Sorghum is among the most photosynthetically efficient plant species. It has one of the highest dry matter production rates and it is one of the fastest maturing food plants; certain cultivars can mature within 75 days (OIA, 1996). Over the past 40 years sorghum production in Africa has increased steadily from 10.9 million metric tonnes in 1965 to 20.8 million metric tonnes in 2005 (FAOSTAT, 2005). However, this is mostly due to an increase in the total area of sorghum harvested (70%) in this period and not to an improvement in sorghum yield per hectare (13%). Sorghum yields per hectare in Africa are very low (853 kg ha–1 in 2005), due to constraints of nutrient depletion, loss of organic matter and poor and erratic rainfall. Sorghum production also is negatively influenced by the incidence of pests and diseases such as bird damage, damage by storage pests, shoot fly (Atherigona soccata), sorghum midge (Contarinia sorgicola), anthracnose (Colletotrichum graminicola), leaf blight (Helminthosporium turcicum) and, increasingly, by the parasitic weed Striga.

Striga in sub-Saharan cereal production In low-input, cereal cropping systems in sub-Saharan Africa, the genus Striga (Orobanchaceae; formerly Scrophulariaceae)) is one of the most important biotic constraints affecting crop production. The genus is most widespread in western Africa where it covers 64% (17 million hectares) of the cereal production area (Gressel et al., 2004). Yield losses in staple cereal crops from damage by Striga range from a few percentages up to complete crop failure depending on factors such as crop species, the

General introduction inherent sensitivity of the crop cultivar, level of infestation, rainfall pattern and soil degradation (Weber et al., 1995). Sauerborn (1991) estimated that over 40% of the cereal production area in West Africa is heavily infested and estimated annual losses of more than one million metric tonnes of grain due to Striga in just five West African countries. Infestation may even reach critical levels, making continued cultivation impossible and leading to the abandonment of fields (Doggett, 1965; Parker, 1991). S. hermonthica (Del.) Benth. and S. asiatica (L.) Kuntze are the two most widespread and the most economically significant species that parasitize on sorghum (Sorghum bicolor [L.] Moench), pearl millet (Pennisetum glaucum [L.] R. Br.), maize (Zea mays [L.]) and rice (Oryza sativa [L.]), whereas S. gesnerioides (Willd.) Vatke attacks crops such as cowpea (Vigna unguiculata L. Walp.) and peanut (Arachis hypogaea L.) (Parker, 1991; Oswald, 2005). The various Striga species have coevolved with their respective hosts and have been present in these cropping systems for many years. Because these systems included prolonged fallow, crop rotations, and intercropping, populations of this parasitic weed seldom reached epidemic proportions. However, since cropping systems changed to more permanent monocropping, Striga poses an increasing problem especially in cereal-based cropping systems (Kroschel, 1999). Striga is considered an indicator of low soil fertility (Bebawi and Farah, 1981; Pieterse and Verkleij; 1991; Orr and Ritchie, 2004; Oswald, 2005). Consequently, Striga is most problematic on infertile or nutrient-depleted soils with low organic matter content and a weakened host crop. Consequently, Striga mostly affects resource-poor subsistence farmers. Under conditions of successive and sole cultivation of susceptible crops and crop varieties, a large Striga seedbank can rapidly build up. This results from increasing inputs of Striga seeds on the one side and, on the other hand, degrading soil conditions under continuous cultivation that diminish presence and activity of micro-organisms that could affect decomposition of Striga seeds in the soil. As a result, the extent and intensity of Striga spp. infestations have rapidly increased and become threats to food production in practically the entire semi-arid and sub-humid region’s farming systems of Africa. From all parasitic plant species, S. hermonthica (Striga) is considered as the most damaging and widespread parasitic weed species on a world scale.

The lifecycle of Striga Striga spp. are obligate hemi-parasitic plants that attach to the roots of their host to obtain water, nutrients and carbohydrates (Kuijt, 1969; Parker and Riches, 1993). They are native to the grasslands of the African tropics, reaching their greatest diversity in the region where they have co-evolved with the cereals (Gressel et al., 2004). Striga spp. have a very complex life cycle (Fig. 1), which is intimately tied to that

Chapter 1 of its host and that follows a series of developmental stages from seed to seed producing plants. After dispersal, the seeds are in a state of primary dormancy for up to six months (Vallance, 1950; Gbèhounou et al., 1996; Kuiper et al., 1996). Following after-ripening, a second prerequisite for germination is the preconditioning of the seeds, which requires an imbibition period of several weeks under humid and warm (25-35 °C) conditions (Okonkwo, 1991; Kebreab and Murdoch, 1999). After reaching maximum sensitivity, prolonged preconditioning induces secondary dormancy (Matusova et al., 2004). Preconditioned Striga seeds require various secondary metabolites (xenognosins) derived from the host roots and some non-host plants to induce germination and to develop (Estabrook and Yoder, 1998; Yoder 2001). These chemical compounds have been identified as sesquiterpene lactones, released in trace amounts in the root exudates (Bouwmeester et al., 2003). This germination stimulant mainly is exuded in a region 3 to 6 mm from the root apex (Hess et al., 1991; Riopel and Baird, 1987; Sunderland, 1960). The germinating seed produces a root-like structure, the radicle. For successful host attachment, germination must take place within 3 to 4 mm of the host root since Striga radicles have limited growth potential

Seed bank

Seed production Conditioning of seeds

Flowering

Germination

Attachment and penetration

Emergence Underground development

Figure 1. Life cycle of Striga hermonthica. Adapted from: Dembélé et al. (1994).

General introduction

(Ramaiah et al., 1991). Radicle growth is directed toward the host root under the influence of a gradient of chemical concentration of the root exudates (chemotropism) (Saunders, 1933; Williams, 1960, 1961). Within four days after germination the radicle needs to find a host root; if not, it will die. After contact with the host root the development of a haustorium begins which also is initiated and guided by host-derived secondary metabolites (Keyes et al., 2001; Yoder, 2001; Hirsch et al., 2003). The haustorium is a globular-shaped root structure that attaches the parasite to the host root, invades the host tissue, and establishes a vascular continuity through which the parasite translocates host resources (Kuijt, 1977; Press and Graves, 1995; Riopel and Timko, 1995; Dörr, 1996). The penetration occurs by development of intrusive cells at the tip which penetrate the cortex of the host root (Kuiper, 1997; Lane et al., 1991; Olivier et al., 1991a). Once the haustorium is inside the stele, direct links between parasite and host xylem systems develop. This can be established within a few days after attachment (Ramaiah et al., 1991; Riopel and Timko, 1995). Phloem connections are not formed between Striga and its host, thus the transfer of nutrients apparently depends on the xylem bridge and some limited diffusion through parenchyma tissues (Parker and Riches, 1993). The parasite seedlings connected with the host grow underground for approximately 3-6 weeks (Parker, 1965; Olivier et al., 1991a). During this time the parasite depends totally upon the host for all the substances it needs for growth and development. Subsequently, adventitious roots are formed at the base of the Striga shoot which, by means of secondary haustoria, attach to the same root to which the primary haustorium is attached or to nearby roots (Parker and Riches, 1993). After emergence the parasite forms stems and leaves with chlorophyll and becomes a hemi-parasite that produces assimilates, but remains partially dependent on the host for water, minerals and some assimilates. About one month after emergence, the parasite initiates flowers and, depending upon pollination, seed production begins shortly thereafter. Striga seeds are minute (0.20-0.50 mm long), weighing only approximately 3.7-12.4 µg each and are produced in very large numbers. Estimates of numbers of seeds produced per reproductive Striga plant can vary from several thousands to over 85,000 depending on species and growing conditions (Saunders, 1933; Kuijt, 1969; Pieterse and Pesch, 1983; Parker and Riches, 1993; Mohamed et al., 1998; Rodenburg et al., 2006a). Seeds are dispersed by cattle, wind, water and shared use of contaminated farm implements and contamination of sowing seed (Press and Gurney, 2000). Striga spp. are highly variable due to their obligate out-crossing character, requiring insect pollinators as bee-flies (Bombyliidae Diptera, West Africa) and butterflies (Lepidoptera, Sudan) for fertilization and seed production. Seeds reach maturity 2-4 weeks after pollination (Musselman, 1987; Webb and Smith, 1996). Large quantities of the newly produced

Chapter 1 seeds can survive the next dry season and a series of cropping seasons with Striga- sensitive host crops will lead to a quick build-up of the Striga seedbank (Gbèhounou et al., 2003; Weber et al., 1995).

Impact of S. hermonthica on the host S. hermonthica can affect its host in different ways. Only part of the reduction in the growth of the host results from competition for carbon assimilates, water, mineral nutrients and amino acids (Graves et al., 1989; Cechin and Press, 1994; Pageau et al., 1998; Taylor and Seel, 1998). However, Striga does not only act as an additional sink but the parasite also has a strong ‘toxic’ or ‘pathological’ effect on the host (Press and Stewart, 1987; Graves et al., 1990; Press and Gurney, 2000). Part of these effects are caused by the disturbed hormonal balance in Striga-infected host plants, characterized by increased levels of abscisic acid and decreased levels of cytokinins and gibberellins (Drennan and El Hiweris, 1979; Frost et al., 1997; Taylor et al., 1996). By altering the host’s hormonal balance, Striga affects host biomass allocation, resulting in the root systems of infected plants being greatly stimulated, while the shoot is stunted and reduced (Parker and Riches, 1993). The parasite also negatively affects host photosynthesis (Press and Stewart, 1987; Graves et al., 1989; Gurney et al., 1995; Smith et al., 1995; Watling and Press, 2001). Parasite induced reduction in host photosynthesis has been reported as the most important mechanism of growth reduction. Graves et al. (1989) estimated that 80% of the decrease in host growth rate can be attributed to the impact Striga has on host photosynthesis. Furthermore, Striga strongly affects the water economy of its host by its high transpiration rate and by reducing the stomatal conductance of the host plant (Gebremedhin et al., 2000; Gurney et al., 1995; Frost et al., 1997; Taylor et al., 1996).

Striga control and its constraints Striga primarily is a problem in small-scale subsistence farming systems with few options for external inputs as pesticides and fertilizers (Kroschel, 1999; Ransom, 2000). Under these conditions, there are major constraints to effective control of S. hermonthica. One of the main problems is the fact that these root-parasitic weeds cause most of their overall damage to the host crop during their subterranean stage (Parker and Riches, 1993; Eplee and Norris, 1995). This implies that controlling the emerged S. hermonthica plants gives limited immediate benefit. A large variety of agronomic, genetic, chemical and biological methods to combat Striga have been developed so far, ranging from measures that are directed on improvement of the fertility status of the soil and measures that have an immediate effect on emerged Striga plants (Oswald, 2005). Various studies have shown that

General introduction

Striga infestation frequently is associated with low soil fertility, reflected in low organic matter content, low biological activity and a weakened host crop (Parker, 1991; Debrah et al., 1998; Oswald, 2005). A substantial amount of work has been carried out to study the effects of soil fertility on Striga infestation (Abunyewa and Padi, 2003; Combari et al., 1990; Gacheru and Rao, 2005; Reda et al., 2005). Increasing soil fertility not only stimulates the growth of the host but also adversely affects longevity of the seeds in the soil, germination and attachment. Hence, improved soil fertility conditions, whether by crop rotation, fallowing or farmyard manure, are likely to lead to reduced infestation. However, these methods often require several years of intervention to effect soil fertility status and, consequently, Striga infestation level. Furthermore, reductions in Striga infestation level do not always immediately result in an improved host performance, because few Striga attachments can already seriously harm the host plant. This leads to lack of motivation by farmers to continue this long-term approach of Striga control. Other, more direct, Striga control methods are directed on prevention of reproduction of the parasite and the reduction of the viable seedbank (Pieterse and Pesch, 1983; Gbèhounou and Adango, 2003; Carsky et al., 1994; Eplee and Norris, 1987; Lendzemo et al., 2005; Haussmann et al., 2000; Oswald, 2005). These control methods have an immediate effect on emerged Striga (hand-weeding, herbicides, biological control), reduce the number of parasites attached to the roots of the host (transplanting, genetic resistance, seed dressing) or directly reduce the seedbank in the soil (trap cropping, ethylene). An important element of Striga control strategies is host plant defence. Two main groups of defence mechanisms against Striga can be distinguished: resistance and tolerance. Resistant varieties are defined as those that show reduced infection level of the host plant. Tolerance on the other hand enables a host plant to perform well, despite the parasitic infection. Yet, complete resistance, or immunity, against Striga has not been developed. Because only a small number of Striga attachments can already cause high levels of yield reduction, resistance alone may not be enough to prevent crop losses. It is therefore recommended to direct breeding efforts towards developing varieties that combine resistance with high levels of tolerance (Rodenburg, 2005; Shew and Shew, 1994). Despite the high potential of some of those control methods, none of these methods alone has proven to efficiently reduce Striga and to improve crop yield in a sustainable way. Many authors (Akobundu, 1991; Berner et al., 1995; Ejeta et al., 1992; Oswald, 2005) have emphasized the need for implementing several control methods as an integrated control approach that may improve efforts to maintain the Striga population at manageable levels and at the same time reduce crop losses. The development of

Chapter 1 high-yielding, Striga-resistant and -tolerant crop varieties, combined with cultural and mechanical control methods, may play a key-role in Striga-infested areas.

Objectives and approach of the research The main objective of this study was to gain more insight into the dynamics of the Striga hermonthica - Sorghum bicolor relationship, which is characterized by mutual interactions. Infection by Striga deregulates the physiology of sorghum, which results in a reduced growth of the plant, and, as a consequence, directly interferes with the quality of the sorghum plants as a host. This has implications for the further development of the parasite, both at the individual (growth of attached parasites) and the population level (increase in number of attachments). In turn, the development of the parasite influences the extent to which a host is affected. In this thesis, research was focussed on the level and timing of infection of the parasite. A better insight into the dynamics of the host plant-parasite association is expected to result in a better appreciation of the consequences of time and level of infection for host plant performance and parasite reproduction. Increased knowledge might help to further develop integrated Striga control strategies which are, on the one hand, directed to improve the yield of the crop and, at the same time, reducing Striga- infestation levels of farmers’ fields. To achieve the overall objectives of this thesis, several studies were conducted at different sites: pot experiments in the greenhouse at Wageningen University and field experiments at the Institut d’Economie Rurale, Research Station in Cinzana, Mali. The parasite-host plant interactions were studied with Striga hermonthica (Del.) Benth and sorghum (Sorghum bicolor [L.] Moench). In these studies two sorghum genotypes with different levels of tolerance to Striga were used: the sensitive cultivar CK60-B and the tolerant landrace Tiémarifing.

Outline of the thesis Following the General introduction, Chapter 2 gives a detailed description of the effects of Striga hermonthica (Del.) Benth. on the performance of a sensitive and a tolerant sorghum genotype. Together with the growth of the different genotypes of the host under Striga-infested conditions, development and biomass production of Striga was followed. In Chapter 3, the photosynthetic response of the two sorghum genotypes to S. hermonthica infection was compared. In an additional experiment, the photosynthetic response of the sensitive cultivar was studied at a wide range of Striga infestation levels. Measurements were taken throughout the growth period of sorghum plants to follow the response over time. In Chapter 4, the consequences of Striga infestation level for growth and development of host plant and parasite were studied

General introduction using the sensitive sorghum cultivar CK60-B. To establish the importance of time of Striga infection for the host plant-parasite interaction, the moment of first Striga- attachment was artificially delayed in a pot experiment (Chapter 5). Based on these results, possibilities to reduce Striga hermonthica parasitism by means of cultural control measures which are aimed at delaying the moment of first Striga-attachment were investigated. In Chapter 6, the results of these explorations are presented. Finally, the overall conclusions of this study and the implications for Striga management are discussed in Chapter 7.

CHAPTER 2

A comparative study on Striga hermonthica interaction with a sensitive and a tolerant sorghum cultivar1

A. van Ast, L. Bastiaans, M.J. Kropff

Crop and Weed Ecology Group, Department of Plant Sciences, Wageningen University, P.O. Box 430, 6700 AK Wageningen, The Netherlands

Abstract In this study, the effects of Striga hermonthica (Del.) Benth. on a sensitive and a tolerant sorghum genotype were investigated, and the underlying tolerance mechanisms were distinguished. The sensitive sorghum cv. CK60-B and the tolerant sorghum landrace Tiémarifing were grown in pots with and without seed infestation. Both sorghum genotypes responded to infection by the parasite, but it was evident that CK60-B was more strongly affected than Tiémarifing. Sorghum plant height, final leaf number, green leaf area, kernel yield, number of kernels and 1000-kernel weight were significantly reduced by infection, which also had a marked effect on the phenological development of CK60-B; the majority of the plants remained vegetative and, in the remaining plants, flowering was delayed by about 2 weeks. No effect on phenological development of Tiémarifing was observed. The tolerant landrace showed significantly lower and delayed emergence of S. hermonthica than the sensitive cultivar, and this could be explained by a delay in the onset of attachment. Significantly higher numbers of reproductive parasitic plants were found in the pots with the sensitive sorghum plants. It is concluded that differences in root architecture and the resulting early infection and higher S. hermonthica numbers are partly responsible for the stronger effects of the parasite on the sensitive cultivar.

Keywords: Striga hermonthica, sorghum, tolerance, biomass allocation, emergence, infection time.

1 Published in: Weed Research (2000) 40, 479-493. 11

Chapter 2

Introduction Parasitic weeds of the genus Striga (Orobanchaceae) strongly affect host crops such as sorghum (Sorghum bicolor (L.) Moench), pearl millet (Pennisetum americanum (L.) Beeke), maize (Zea mays L.), rice (Oryza sativa L.) and cowpea (Vigna unguiculata (L.) Walpers) and, as a consequence, they are important growth-reducing factors in crops in vast areas of the savannah zone in Africa (Parker and Riches, 1993). Although Striga spp. have long been pests in traditional cropping systems, they did not cause severe problems until 20-30 years ago, because these systems generally involved crop rotations, mixed cropping and, in particular, prolonged periods of fallow. As a result of the rapid human population growth in Africa and the increased demand for food supplies, cereal monocropping and intensive land use with little or no fallow have replaced the traditional cropping systems. These changes in cropping practices have resulted in soil depletion and degradation and a rapid increase in the extent and intensity of infestations of Striga spp. (Parker and Riches, 1993; Riches and Parker, 1995). Of all Striga species, Striga hermonthica (Del.) Benth. is the most economically important parasitic weed in Africa, causing losses in many cereal crops. The growth of its host is retarded, and crop yields are lowered or reduced to zero under conditions of severe infestation. Most traditional cultivars of cereal crops sensitive to S. hermonthica possess some kind of defence mechanism that limits yield reduction to at least some extent, most probably as a result of selection in infested sites. Tolerance, defined as the ability of host plants that are parasitized to the same extent to endure the presence of the parasite with reduced disease symptoms and/or damage compared with standard cultivars (Parker and Riches, 1993), is often mentioned as one of the mechanisms responsible for reduced yield loss as a result of S. hermonthica. So far, little is known about the mechanisms of host plant tolerance to S. hermonthica. Gurney et al. (1995) observed that, after S. hermonthica infestation, the photosynthetic capacity of sorghum cv. Ochuti, known to show some tolerance, remained unaffected, whereas the photosynthetic capacity of the very sensitive sorghum cv. CK60 was strongly reduced. The tolerant cv. Ochuti also showed some delay in the emergence of S. hermonthica. Both the maintenance of photosynthetic capacity of the host and the delayed emergence of the parasite could indicate the presence of a certain level of tolerance per se. However, Gurney et al. (1995) did not determine whether the observed delay in emergence resulted from a delayed attachment or other mechanisms. If attachment is delayed, avoidance might be the defence mechanism that, to at least some extent, is responsible for the relatively small effects of the parasite. The first objective of the present study was therefore to gain more insight into the defence mechanisms of sorghum that are responsible for a reduced effect of S. hermonthica on crop production. The main

Striga tolerance mechanisms in sorghum emphasis was to find out whether various defence mechanisms operate within the S. hermonthica-sorghum pathosystem, and whether specific morphological attributes of the host plant are related to them. All defence mechanisms, except complete resistance, still allow the parasite to complete its life cycle and produce seeds. Seed production determines the long-term consequences of host-parasite interactions, as it is an important component in determining the future level of infestation. Yet, parasite growth, development and seed production have been under-researched components of the host-parasite interaction. For this reason, the second objective of this research was to study the growth and development of the parasite on sorghum cultivars differing in sensitivity to the parasite.

Materials and methods Two pot experiments to study the interaction between sorghum and S. hermonthica were conducted in a greenhouse in Wageningen in the summers of 1996 and 1998. In both experiments the S. hermonthica-tolerant local Malian landrace Tiémarifing and the highly sensitive inbred cultivar CK60-B, both provided by ICRISAT, Mali, were compared. Striga hermonthica seeds for these experiments were collected in 1995 from sorghum hosts at experimental fields of ICRISAT, Samanko, Mali. In both experiments, daytime temperature in the glasshouse in Wageningen ranged from 30 °C to 35 °C, whereas night temperature was kept at about 20 °C. Day length was artificially kept at 11 h with black screens during the whole growing period, and relative humidity varied between 40% (day) and 85% (night). Plants were grown in 18L pots (l × w × h = 0.25 m × 0.25 m × 0.30 m) containing a mixture of sand and arable soil, collected from the top layer (0 - 0.25 m) of an arable field near Wageningen. The arable soil had a pH-KCl of 5.6, an organic matter content of 0.7 % and a total N content of 0.04 g N 100 g–1 soil. In the 1996 experiment, sand and arable soil were mixed in a ratio of 1:1, whereas in the 1998 experiment, a ratio of 3:1 was used to decrease the basic level of nutrients. Ten days after sorghum emergence, each pot received 0.38 g of N, 0.19 g of P and 0.31 g of K, equivalent to a field rate of 60 kg N ha–1, 30 kg P ha–1 and 50 kg K ha–1 respectively. Pots were watered every 2 days. Seeds of S. hermonthica (20 mg per pot, approximately 4900 seeds) were mixed through the upper 6 cm of the soil in half the pots. Ten days after adding the seeds to the soil to allow conditioning of the seeds, three pre-germinated sorghum seeds were planted per pot. After emergence, seedlings were thinned to one plant per pot.

Experiment 1 The 1996 experiment was laid out using a split-plot design in 24 blocks, in which S.

Chapter 2 hermonthica was used as main plot factor and cultivar as the sub-plot factor. Blocking was undertaken to take account of possible variation in different parts of the glasshouse. S. hermonthica was selected as main plot factor as its effect was expected to be much larger than that of the cultivars. The sensitive sorghum cultivar CK60-B and the tolerant sorghum landrace Tiémarifing were grown in pots with and without parasite seed infestation, resulting in four treatments. Within each block, all four treatments were represented by four pots, resulting in a total of 384 (24 × 4 × 4) pots. Each block was surrounded by border rows, consisting of additional pots with sorghum seeds planted in them, to create the same climatological conditions in all pots used for the experiment itself. Within a block, both S. hermonthica treatments were separated by additional pots to avoid interplot interference between infected and uninfected pots. Plant height of the sorghum was measured every 2 weeks in the period between emergence and the formation of the flag leaf. Plant height was defined as the distance from the base of the stem to the ligule of the youngest fully developed leaf. Both the number of sorghum leaves and the number of emerged S. hermonthica plants were counted every 2 days. The onset of flowering of sorghum and parasite was also recorded. Two harvests were conducted to determine biomass of host and parasite. The first harvest was taken at the onset of flowering of the infected CK60-B plants, at 67 days after sorghum emergence (DAE). For this harvest, all pots in eight randomly selected blocks were used. The 16 remaining blocks were used for the second harvest, after ripening of the kernels, at 94 DAE and 102 DAE for CK60-B and Tiémarifing, respectively. At both harvests, sorghum and S. hermonthica plants were dissected into leaf, stem, root and reproductive organs. Green leaf area was determined with a leaf area meter (Li-3100 area meter; Li-cor, Lincoln, NB, USA) and, subsequently, the plant material was put in an oven at 80 °C for 48 h for determination of plant organ dry weight. Treatment effects on plant weight, leaf area and number of leaves were analysed by ANOVA, using the statistical package Genstat. Data on the emergence of S. hermonthica plants were analysed in two steps. First, log-logistic curves (Campbell and Madden, 1990) were used to describe the emergence pattern of individual pots:

c f (t) = t −bx ln( ) τ 1+ e where f(t) is the number of emerged plants, c is the final number of emerged plants, b is a shape parameter, t is the time in days and τ is the mid-emergence time, at which 50% of the final number of plants was emerged. For each individual pot, the non-linear

Striga tolerance mechanisms in sorghum regression option of Genstat was used to estimate b and τ, whereas c was set equal to the observed final number of emerged S. hermonthica plants. Subsequently, the rate of emergence (dg/dt50) at the mid-emergence time (τ) was computed as:

dg/dt50 = b/(4 ×τ )

After fitting, ANOVA was conducted on c, τ and dg/dt50.

Experiment 2 In the 1998 experiment, a completely randomized block design was used with five blocks. Each block consisted of two factors (parasite and cultivar) at two levels. Within each block, each treatment was represented by 10 pots, resulting in a total of 200 (5 × 4 × 10) pots. Blocks were surrounded by border rows. Ten destructive harvests, in which a single pot per treatment was harvested per block, were carried out within the first 5 weeks after emergence of the sorghum plants at 3, 6, 10, 13, 17, 21, 25, 28, 32 and 35 DAE. For each harvest per treatment, one pot out of 10 was used from each block. Sorghum plants were dissected into leaf, stem and root, and green leaf area was determined. The number of S. hermonthica attachments on the roots were counted and root length was determined according to the line intersect method described by Tennant (1975). Plant organ dry weight was determined after drying at 80 °C for 24 h. Before drying, parasite attachments were removed and dry weight was determined separately. At the last harvest, at 35 DAE, the soil in the pot was dissected into two layers. The first layer consisted of the upper 6 cm, the layer through which S. hermonthica seeds were mixed, and the second layer of the remaining 24 cm of soil. Root length, root dry weight and number of S. hermonthica attachments (>0.5 mm) were determined separately for each layer. Treatment effects on plant weight, root length and number of attachments were analysed by ANOVA, using the statistical package Genstat.

Results

Experiment 1

Characteristics of sorghum cultivars Flowering of the uninfected plants of both sorghum cultivars was observed 54 d after sowing. At this development stage, Tiémarifing plants were about 2.7 times taller than CK60-B (Table 1). Plant dry weight of Tiémarifing was also significantly (P<0.001)

Chapter 2 higher than that of CK60-B. No difference in the allocation of dry weight over shoot and root was found between cultivars, although there was a marked difference in dry matter allocation to the various plant organs within the shoot. CK60-B had a relatively high fraction of leaf material and a relatively low fraction of stem compared with Tiémarifing. No differences in specific leaf area (SLA; cm2 g–1 leaf) was observed in either cultivar and, consequently, the leaf area ratio (LAR; cm2 g–1 plant) of CK60-B plants was significantly higher than that of Tiémarifing. No significant differences were found in leaf appearance rate, final number of leaves or total green leaf area. At maturity, significant differences in total biomass (P<0.001) and kernel yield (P<0.001) were observed, whereas there was no difference in harvest index (HI). Kernel yield of Tiémarifing was about 25% higher than that of CK60-B. This difference in kernel yield reflected the difference in the number of kernels per panicle, as the 1000-kernel weight of both cultivars was about the same.

Effects of Striga hermonthica on host performance Although both cultivars responded to infection by S. hermonthica, it was evident that CK60-B was more strongly affected than Tiémarifing (Table 1). Infection by the parasite had a marked effect on the phenological development of CK60-B. Approximately 60% of the plants of this cultivar remained vegetative and, in the remaining 40%, flowering was delayed by about 2 weeks. No effect on the phenological development of Tiémarifing was observed. Final plant height of Tiémarifing and CK60-B plants was reduced by 15% and 47%, respectively. Also, biomass accumulation at flowering was reduced in both cultivars, but to a much lower extent in Tiémarifing (19%) than in CK60-B (55%). A marked effect of infection by the parasite on dry matter partitioning over the various plant organs was observed. In both cultivars, the shoot fraction, i.e., shoot dry weight as a fraction of the total dry weight, decreased significantly (P<0.001), whereas infection caused a relatively greater allocation of dry matter to the root system. Again, this effect was more pronounced in CK60-B. Also in this cultivar, an altered allocation pattern within the shoot was observed. Dry matter partitioning to the leaves increased at the expense of stem and inflorescence. For Tiémarifing, no effect of infection on dry matter allocation within the shoot was found. A clear effect of S. hermonthica infection was observed on leaf development in CK60-B. Final leaf number and green leaf area were reduced by 10% and 34% respectively; SLA increased by 11%. As fewer leaves were produced over a longer period of time, leaf appearance rate was reduced by 21%. Despite the absolute reduction in green leaf area, the LAR of CK60-B was significantly increased as a

Striga tolerance mechanisms in sorghum

SED SED sensitive )

S. hermonthica

comparison for SED sorghum.

(tolerant) Tiémarifing

) and infected (+

S. hermonthica

and of the ering at maturity

(sensitive) CK60-B

uninfectedinfected uninfected infected

) 0.29 0.23 0.30 0.28 (0.003) –1

shoot) 0.16 0.01 0.16 0.16 (0.01) –1 shoot) shoot) 0.31 0.05 0.28 0.31 (0.02) total plant) total plant) 0.81 0.59 0.85 0.73 (0.02)

–1 shoot) shoot) 0.32 0.66 0.23 0.26 (0.02) shoot) shoot) 0.52 0.34 0.61 0.58 (0.02) –1 –1 257 (23) 321 18 249 –1

) 8.1 0.4 10.0 6.6 (0.7) –1 –1 ) 1156 759 1248 1095 (52) 1095 1248 759 ) 1156 2

plant) 66 98 46 50 (4) leaf) leaf) 235 (8) 238 261 271 –1 –1 g

g 2 2 Parameter values (based on dry weight) of uninfected (– .

SLA (cm SLA Stem fraction (gstem g 1000-kernel weight (g) 100% flowering. in (39%); Tiémarifing CK60-B of plants on the few flowering # Based 31.1 9.2 31.0 23.7 (1.4) Final numberleaves of Maturity Internode length (cm) Total dry weight (g) g kernels (g Index Harvest 12.3 4.8 11.0 30.4 2.9 12.6 8.8 12.2 12.9 44.4 (0.1) 11.4 28.4 (0.3) (1.4) LAR (cm Inflorescence fraction (g inflor. g No. of kernels panicle kernels No. of Kernel yield (g plant (g Kernel yield Shoot fraction (g shoot g shoot (g fraction Shoot of any parasite × cultivar combination (Expt cultivar combination 1, 1996). × parasite any of Flowering Table 1 Table (DAE)flowering Time of (cm) Height Total weight (g) Green leaf area (cm 54 67# 17.9 59.7 54 32.3 8.1 55 163.6 27.2 (0.1) 139.1 22.0 (3.7) (0.8) Leaf fractionleaves (g g Leaf appearance rate (no of leaves day at flow cultivars sorghum (Tiémarifing) and tolerant (CK60-B)

Chapter 2 result of parasite infection. For Tiémarifing, most of these effects were smaller. In this cultivar, total green leaf area was moderately (12%) reduced, whereas SLA increased by 14%. At maturity, the negative effect of infection on production in CK60-B was severe; total biomass and kernel yield were reduced by 71% and 95%, respectively. The strong reduction in kernel yield resulted, to a large extent, from the fact that only 40% of the plants produced seeds. For Tiémarifing, reduction in total biomass and kernel yield were both about 35%. In both cultivars, reduction in kernel yield was also the result of –1 a strong effect of infection on both number of kernels panicle and 1000-kernel weight.

Striga hermonthica characteristics With CK60-B, first emergence of the parasite was observed at 32 DAE (Fig. 1). Tiémarifing showed a significant delay, with first emergence at 50 DAE. In line with this difference, a significant difference (P<0.001) was observed in τ (mid-emergence time), which was 56 DAE and 63 DAE for CK60-B and Tiémarifing, respectively, indicating that the difference between cultivars in mid-emergence time was less pronounced than the difference in first emergence. Consequently, the slopes in the emergence curves at mid-emergence time, further

14 No. of emerged S. hermonthica plants

0 0 102030405060708090 DAE of sorghum Figure 1. Time course (d) for S. hermonthica emergence on the sensitive sorghum cv. CK60- B (solid line) and the tolerant cv. Tiémarifing (dotted line). Lines are based on averaged values of parameter estimates obtained after fitting observed emergence data of S. hermonthica plants in individual pots to a log-logistic model. The averaged parameter estimates of the log-logistic model are presented in Table 2 (Exp 1, 1996).

Striga tolerance mechanisms in sorghum

Table 2. Parameters of a log-logistic model fitted to data of Striga hermonthica emergence, with the sensitive cv. CK60-B and the tolerant cv. Tiémarifing as hosts (Exp 1, 1996). CK60-B Tiémarifing (sensitive) (tolerant) SED Maximum no. of emerged S. hermonthica plants 13.7 12.1 (0.4) Mid-emergence time (DAE) 56 63 (0.5) Rate of emergence (number d–1) 0.040 0.069 (0.012)

referred to as rate of emergence, also differed (Table 2). There was some evidence that rate of emergence at mid-emergence time was higher (P=0.084) in Tiémarifing (0.069 d–1) than in CK60-B (0.040 d–1), and also that the maximum number of emerged plants was higher (P=0.053) in the highly sensitive CK60-B, although the differences were not statistically significant. In Table 3, total number of S. hermonthica plants and their developmental stage are presented at flowering and at maturity of sorghum. At flowering of sorghum, development was more advanced in plants attached to CK60-B compared with those produced on Tiémarifing. This was also reflected in the dry weights of S. hermonthica plants produced on CK60-B and on Tiémarifing. The total dry weight and average dry weight of S. hermonthica biomass in the sensitive cultivar were about 2.5-3.0 times as high as in the more tolerant cultivar. The parasite also developed significantly (P<0.01) more leaf area (3.5 times) when attached to the sensitive CK60-B. At maturity of sorghum, differences in reproductive plants (P<0.05), in total dry weight (P<0.001) and average dry weight (P<0.001) of S. hermonthica biomass, were smaller, but still significant. S. hermonthica biomass with the sensitive cultivar was about 1.4 times as high as with the tolerant cultivar. Differences between hosts were also reflected in S. hermonthica leaf area development. The parasite developed 1.5 times more leaf area (P<0.001) when attached to the sensitive CK60-B. At maturity of the sorghum, the total number of emerged S. hermonthica plants was somewhat lower than the maximum number of emerged plants mentioned in Table 2, as at that stage, some parasite plants had died and disappeared.

Experiment 2 The clear difference in emergence of the parasite between both cultivars was examined further in the 1998 experiment, in which detailed observations on above- and below- ground biomass of sorghum and S. hermonthica root attachments were made during the first 35 d after emergence of sorghum.

Chapter 2

Maturity

hum cv. and theCK60-B tolerant cv.

Flowering

CK60-B Tiémarifing Tiémarifing CK60-B Tiémarifing SED CK60-B SED attachments from the sensitive sorg hermonthica

)/sorghum plant )/sorghum 106 31 (10) 236 154 (6) hum plant 0.04 0.05 (0.003) 0.25 0.26 (0.01) 2 Striga d plants 1 0 (0.2) 43.5 36.8 (1.5) plants (cm

) 85 84 (3) 48 43 (2) 43 48 (3) 84 ) 85 plants/sorghum plant plants/sorghum 0.1 0 (0.02) 5.3 4.1 (0.3) –1 plants/sorghum plant plants/sorghum 9.7 4.9 (0.6) 7.8 7.4 (0.4) g 2 (cm S. hermonthica S. hermonthica LAR Tiémarifing at floweringTiémarifing ofand at maturity sorghum (Exp 1, 1996). Table 3. Numbers and characteristics of No. of subterranean plant attachments/sorghum No. of emerged vegetative No. of emerged reproductive % reproductive plants of total emerge 15.2 17.2 (0.8) 11.6 14.3 (0.8) Total no. of attachments/sorghum plant weightDry of subterranean biomass (g)/sorg weightDry of emerged biomassplant (g)/sorghum weightDry plant total biomass (g)/sorghum perAverage attachment weight dry (g)/sorghum plant Total leaf area of 1.25 25.0 0.05 0.37 1.29 22.1 0.02 (0.12) (0.003) 0.42 (1.17) 4.96 0.26 (0.12) 24.7 3.58 5.21 0.19 25.8 (0.12) (0.01) 3.84 (1.0) (0.13)

Striga tolerance mechanisms in sorghum

Characteristics of sorghum cultivars For the control plants, total biomass accumulation of CK60-B at 35 DAE was again significantly lower than that of Tiémarifing (Table 4). Differences in root dry weight were even more pronounced, as root dry weight in Tiémarifing was three times as high as in CK60-B. As a result root weight as a fraction of the total in Tiémarifing was 0.36 compared with 0.22 in CK60-B. A similar difference was found in root length, indicating that significant differences in specific root length (SRL; m g–1) were present. In CK60-B, about 46% of total root weight was found in the upper 6 cm soil layer, which covered about 6% of total root length. In Tiémarifing, these percentages were 23% en 3%, respectively, indicating that, in CK60-B, a much higher fraction of total root biomass was concentrated in the upper soil layer. SRL in the upper layer was about 5 m g–1 for both cultivars, which was much lower than SRL in the lower soil layer, with values of 60 and 40 m g–1 for CK60-B and Tiémarifing respectively. Consequently, the large differences in overall root weight and overall root length between cultivars were mainly caused by roots in the lower soil layer, whereas differences in the upper layer were small.

Effects of Striga hermonthica on host performance As in the 1996 experiment, both cultivars responded to a different extent to infection. At 35 DAE, reductions in biomass accumulation of 40% and 20% were found for CK60-B and Tiémarifing, respectively (Table 4). For both cultivars, total root dry weight and total root length of infected plants did not differ significantly from that of the control. As a result, S. hermonthica infection resulted in a relative increase in allocation of biomass to the roots. Clear differences were found in the response observed in the upper and lower soil layers. For both cultivars, root dry weight in the upper layer tended to decrease, but these differences were not significant. Root length, however, increased significantly, indicating that SRL increased, meaning relatively thin roots, as a result of S. hermonthica infection. In CK60-B, this response was stronger than in Tiémarifing. In the lower soil layer, an effect of S. hermonthica infection was only observed for CK60-B. Root dry weight tended to increase, although this increase was not significant. Root length remained unaffected, indicating a reduction in SRL. Consequently, in CK60-B, infection resulted in an altered distribution of root dry weight over both layers, with a relatively large fraction of root dry weight in the lower soil layer. In Figs. 2a and b, the effects of infection on sorghum shoot and root biomass of sorghum are presented over time. At 32 DAE of sorghum, the first significant negative effects of the parasite on shoot biomass production were observed in both cultivars. In

Chapter 2

SED

) sensitive (CK60-B) and tolerant Tiémarifing (tolerant)

S. hermonthica

cm and > 6cm). Data from final harvest (35 DAE). SED SED for DAE). (35 harvest final from cm and > 6cm). Data

CK60-B (sensitive) ) and infected (+

infected uninfected uninfected infected

hermonthica S.

total plant) 0.22 0.43 0.36 0.35 (0.06) –1

) 5.08 3.04 9.43 7.53 (1.18) 7.53 9.43 3.04 ) 5.08 root - 4.5 - 2.5 (0.004) ) 1.20 1.61 3.61 3.21 (0.63) (0.16) (0.65) 3.21 0.59 3.61 2.62 0.64 2.97 1.61 0.32 ) 1.20 1.25 ) 0.50 ) 0.70 –1 –1 –1 –1 –1 total root) total root) total 0.46 0.26 0.23 0.54 0.28 0.74 (0.09) 0.77 0.73 (0.09) m

–1 –1 attachments/plant - 13.8 - 10.2 (0.8)

root) root) root) 38.3 31.7 34.9 5.0 41.1 10.0 62.9 5.3 38.4 7.3 41.4 48.9 –1 –1 –1 S. hermonthica S. hermonthica General SRL (m g (m SRL Root dry weight (g plant (g weight dry Root No. of g (m SRL (Tiémarifing) sorghum cultivars from different soil layers (0-6 soil layers different from cultivars sorghum (Tiémarifing) cultivar and parasite comparison (Exp 2, 1998).

Table 4. Parameter values of uninfected (– (– uninfected of values Parameter 4. Table Total dry weight plant (g Total root length (m) g root (g fraction weight Root Root(m) length Root fractionroot (g g 46 Root(m) length Root fractionroot (g g 51 126 2.5 132 3.2 (4) 44 3.4 4.3 48 (0.3) 123 128 (4) Root length fraction (root length/total root length) root length/total (root fraction Root length No. of 0.06 0.06 length) root length/total (root fraction Root length g (m SRL 0.03 0.94 0.03 0.94 (0.01) 0.97 0.97 (0.01) Root dry weight (g plant (g weight dry Root plant (g weight dry Root Upper soil layer (0 - 6 cm) 6 cm) - (0 soil layer Upper 6 cm) (> Lowersoil layer

Striga tolerance mechanisms in sorghum

CK60-B 8 shoot dry wt (g) a

6 shoot/root 3.2 4

2 0.9 DAE of sorghum 0 0 5 10 15 20 25 30 35 40

-22

-44

root dry wt (g) -66

8 Tiémarifing shoot dry wt (g) b shoot/root 6 1.6

1.3 4

DAE of sorghum 0 0 5 10 15 20 25 30 35 40

-22

-44

-66 root dry wt (g)

Figure 2. Influence of S. hermonthica infestation (dotted line, uninfected; solid line, infected) on the time course of per plant shoot and root dry weight of the sensitive sorghum cv. CK60- B (a) and the tolerant cv. Tiémarifing (b). Error bars indicate standard error of means for shoot (upper) and root (lower) biomass, respectively (Exp 2, 1998).

Chapter 2

16 No. of S. hermonthica first emergence of S. hermonthica attachments / pot above-ground in CK60-B 14

0 0 5 10 15 20 25 30 35 40 DAE of sorghum Figure 3. Time course (d) for S. hermonthica attachments on the sensitive sorghum cv. CK60- B (solid line) and the tolerant cv. Tiémarifing (dotted line) (Exp 2, 1998).

this phase, root biomass production was not affected. At the end of the growing period (35 DAE), shoot:root ratio of infected plants of CK60-B was 0.9 and of Tiémarifing 1.3, indicating a clear decrease compared with uninfected control plants, in which shoot:root ratios of 3.2 and 1.6 were observed for CK60-B and Tiémarifing, respectively. This shift was mainly the result of a strong reduction in absolute shoot growth, as absolute root growth in this initial infection phase was not yet significantly affected.

Striga hermonthica characteristics At 10 DAE of sorghum, the first S. hermonthica attachments were observed on the roots of CK60-B; for Tiémarifing, first attachments were seen at 17 DAE (Fig. 3). As in expt 1, at 32 DAE, the first emergence of the parasite was observed in CK60- B, whereas at the end of this experiment (35 DAE) no plants had emerged in pots with Tiémarifing. Throughout the experiment, the number of attachments increased gradually for both cultivars but, for CK60-B, a higher number of attachments was always observed than for Tiémarifing. At the end of the experiment, at 35 DAE, CK60-B hosted 13.8 attachments per plant, representing 4.5 attachments m–1 root length of the upper contaminated soil

Striga tolerance mechanisms in sorghum layer, whereas 10.2 attachments per plant, representing 2.5 attachments m–1 root length, were observed in Tiémarifing (Table 4).

Discussion Striga hermonthica infection clearly had a stronger effect on the sensitive cultivar, although the parasite affected growth and dry matter allocation in both cultivars (Table 1). The reduction in biomass production was accompanied by a relatively increased allocation of dry matter to the roots. These observations are well-known symptoms of infected host plants (Parker and Riches, 1993; Press and Graves, 1995). Although the absolute green leaf area was reduced, mainly as a result of a reduction in individual leaf size and a faster senescence of older leaves, the final number of leaves was hardly affected. The reduction in kernel yield resulted from a reduced number of kernels per panicle and a reduced 1000-kernel weight. These observations are in agreement with those of Wylde (1991) and of Clark et al. (1994) in sorghum and millet. Grain fill is sustained largely by photosynthesis of the flag leaves, and this is impaired in cereals affected by S. hermonthica. The consequences of these changes in grain fill may have far-reaching effects on the nutritive value of the food crop, as it is known that the quality of seed protein is strongly affected in S. hermonthica infected millet (Wylde, 1991). In general, the effects of S. hermonthica on plant production are attributed to two damage mechanisms: directly through the role of the parasite as an additional sink for carbon, inorganic solutes and water on the one hand and, indirectly, through the so- called phytotoxic or pathological effects of the parasite on the host. A clear demon- stration of the latter can be found in Figs. 2a and b, in which 25-45% reduction in total biomass was observed at about 1 month after emergence of the sorghum, when parasite attachments were still so small that they were hardly visible to the naked eye. Changes in growth regulators in the host are thought to be responsible for this reduced shoot growth (Drennan and El Hiweris, 1979; Taylor et al., 1996; Frost et al., 1997). Kropff and Schippers (1986), Hibberd et al. (1996) and Manschadi (1999) demonstrated that yield reductions in faba bean, tomato and tobacco as a result of Orobanche spp. (broomrape) infection could be completely explained by translocation of assimilates from host to parasite. For hosts infected by S. hermonthica, it was shown that reallocation of carbon to the parasite was unlikely to account for the entire biomass reduction of infected plants (Press and Stewart, 1987; Graves et al., 1989; Cechin and Press, 1993a; Frost et al., 1997). The present results confirm that, for both CK60-B and Tiémarifing, S. hermonthica dry matter was equal to only 24% of the overall reduction in sorghum biomass. It is most likely that growth reduction due to withdrawal by the parasite is even smaller, as this is a hemi-parasitic species and, therefore,

Chapter 2 part of the S. hermonthica biomass produced results from the photosynthetic activity of the parasite itself. Press and Graves (1991) found that, for S. hermonthica grown on sorghum, 48% of carbon of the parasite was derived from parasite photosynthesis. It is known that the effect of S. hermonthica infection on photosynthesis causes the most prominent effect of the parasite on host production (Cechin and Press, 1994; Clark et al., 1994; Gurney et al., 1995). Reductions in photosynthetic rate at light saturation of up to 60% have been reported. The assimilate production of an infected plant will also be reduced as a result of a reduction in leaf area, as observed, for instance, in the current experiment. This reduction in leaf area might simply result from the positive feedback that exists between assimilate production, leaf area formation and light interception, especially during the exponential growth phase. Apart from a reduced production of assimilates, an increased respiration of infected root tissue has been reported (Sallé et al., 1987). Furthermore, below-ground parts of S. hermonthica appeared to have relatively high rates of respiration (Graves et al., 1989; Press and Graves, 1991). The relatively mild effects of S. hermonthica on Tiémarifing, when compared with CK60-B, were accompanied by a delayed emergence of the parasite plants in pots with Tiémarifing. Detailed observations during the initial growth period of sorghum in the second experiment showed that this delay in emergence can be explained, at least to some extent, by a delay in the onset of attachment. Attachment of the parasite to roots of Tiémarifing was delayed by approximately 1 week. These observations leave us with various questions relating to the causes of the observed difference between the onset of attachment and the potential relationship between the timing of attachment and tolerance. It is well documented that differences in production of germination stimulants exist between cultivars (Hess et al., 1992; Reda et al., 1994). The cultivars used in the present study were both qualified as high stimulant-producing ones (D.E. Hess, pers. comm.). Further research is needed to find out whether differences exist in the onset of stimulant excretion by the roots of these cultivars. Two obvious differences in root development of CK60-B and Tiémarifing were observed. Overall root weight and root length of Tiémarifing exceeded that of CK60-B by a factor of 2.5-3.0. Apart from that, a clear difference was present in the distribution of root material over the two soil layers. The absolute amount of root material in the top 6 cm of the soil hardly differed between cultivars. For Tiémarifing, this amount represented nearly 20% of overall root dry weight, indicating that over 80% of total root biomass was present in the lower soil layer. For CK60-B, over 40% of total root biomass was present in the upper layer. Assuming a more or less similar root activity (i.e. uptake of water and nutrients) for both cultivars and assuming that within a root system, root activity is more or less proportional to root biomass, this observation

Striga tolerance mechanisms in sorghum would indicate that root activity of CK60-B in the upper layer is substantially higher than that of Tiémarifing, despite an almost identical amount of root biomass in the upper layer. If root activity and the production of root exudates are positively correlated, this might explain the higher number of parasite attachments on CK60-B at this early growth stage. According to this reasoning, differences in root architecture might play a pronounced role in determining differences in the onset of attachment between cultivars. A few studies (Cherif-Ari et al, 1990; Baltus et al., 1994; van Delft, 1997) also indicate that differences in host root characteristics may confer avoidance or delay of attachment. A second question that arises is to what extent a delayed onset of S. hermonthica attachment is functionally related to tolerance. In particular, if plant growth is disturbed during the initial phases, slight differences in the onset of these disturbances might have a large influence on the final level of yield reduction. This is mainly because of the positive feedback that exists between biomass production, leaf area formation and light interception. Studies by Cechin and Press (1993b) and van Delft (1997) clearly demonstrated that early attachment of S. hermonthica had a greater effect on the growth of its host compared with later attachment. A systems analytical approach in which the effects of the parasite on physiological plant processes are quantified experimentally and introduced in a mechanistic model of crop growth seems an appropriate way to estimate the importance of a delayed onset and explain the underlying mechanisms. Earlier studies on the effects of biotic stress factors on crop production demonstrated the usefulness of this approach in unravelling the mechanisms of yield reduction through biotic stress factors (Kropff, 1989; van Oijen, 1991a; Bastiaans, 1993a; Schans, 1993). It is unlikely that the large difference in biomass production and final yield between infected and uninfected plants of both cultivars can be explained entirely by the observed delay in the onset of attachment of about 1 week. Apart from avoidance, tolerance per se, as observed for example in the effects of the parasite on photosynthesis of the host (Gurney et al., 1995; unpubl. obs., this experiment) might also explain in part the difference in response to S. hermonthica infection. Combining data on sorghum and S. hermonthica, it becomes clear that, as expected, a stronger negative effect on growth of the host plant coincided with a stronger increase in development of the parasite. At flowering, total number of parasite attachments was almost similar, but average dry weight per attachment on CK60-B was 2.5 times that on Tiémarifing. Probably, this difference will be caused, at least partly, by the delayed attachment of the parasite on Tiémarifing plants. However, it is also possible that a relatively poor host-parasite compatibility, as expressed in a relatively high level of tolerance of the host, is partly responsible. In the case of both

Chapter 2 number of emerged S. hermonthica plants and their individual dry weight, the large differences between host plant cultivars observed during the first part of the growing period decreased substantially towards the end of the growing period. This implies that emergence and growth of parasite plants on Tiémarifing during the last part of the growing period was faster than for the plants attached to CK60-B. This effect is clearly demonstrated when comparing the rate of emergence of S. hermonthica plants at 50% of final emergence, when values of 0.04 and 0.07 were found for CK60-B and Tiémarifing, respectively (Table 2). At crop flowering, overall S. hermonthica biomass attached to Tiémarifing was only 30% of the biomass attached to CK60-B, whereas at maturity, S. hermonthica biomass on the tolerant cultivar represented 70% of the biomass attached to the sensitive cultivar (Table 3). It is hypothesized that two factors determine the growth of the parasite. The first factor is the host-parasite compatibility, where a good compatibility is reflected in a strong growth-reducing effect of the parasite on the host, and a strong sink activity of the parasite resulting in a fast initial growth. The second factor is the vigour of the host, determining the amount of resources potentially available for the parasite. Because in the early growth stages, vigour of the hosts was comparable, a stronger initial growth of the parasite on CK60-B must result from its better compatibility. At the same time, this higher compatibility resulted in a stronger growth reduction of the CK60-B plants. As a result, the reduced availability of resources during the later part of the growing season limited the rapid growth of the parasite plants. For Tiémarifing, the initial growth and development of S. hermonthica plants was relatively slow, indicating a less compatible host; this host plant was only marginally affected by the parasite. Resource availability was maintained throughout the growing season and, consequently, growth and development of the parasite plants during the later phases of sorghum growth were relatively high. This feedback mechanism might explain why only marginal differences in final dry weight of S. hermonthica plants were observed between cultivars that differ in the level of tolerance. Assuming a more or less linear relationship between S. hermonthica biomass and seed production, and the observation that the number of reproductive parasite plants in the tolerant cultivar was only slightly smaller, this would suggest that tolerance has just a small effect on S. hermonthica seed production. Observed differences between cultivars in timing of the first parasite attachments and their expected influence on the final yield reduction are a strong indication that control practices based on a delay in first attachments could contribute to a reduction in the S. hermonthica problem (van Delft, 1997). The final degree of yield reduction can be further affected by breeding activities concentrated on (physiological) tolerance of the host to the parasite. As tolerant cultivars have the disadvantage of encouraging

Striga tolerance mechanisms in sorghum the build-up of a S. hermonthica seedbank, breeding for tolerant cultivars with lower S. hermonthica emergence than a sensitive cultivar might be important for reducing problems with this weed in the long term.

CHAPTER 3

Time course of density effects of Striga hermonthica on the photosynthetic activity of Sorghum bicolor

A. van Ast, L. Bastiaans, M.J. Kropff

Crop and Weed Ecology Group, Department of Plant Sciences, Wageningen University, P. O. Box 430, 6700 AK Wageningen, The Netherlands

Abstract Reduction in leaf photosynthetic rate of Sorghum bicolor (L.) Moench following infection by the root hemi-parasitic weed Striga hermonthica (Del.) Benth. has often been reported as the most important mechanism of growth reduction of sensitive genotypes. In this study, the relationship between Striga infestation level and the magnitude of the reduction in photosynthetic rate was analyzed. In a pot experiment, sorghum plants were exposed to a wide range of Striga infestation levels. Photosynthesis was regularly measured over time and related to Striga infestation level, Striga infection level, represented by the number of parasite attachments, and the number of emerged Striga plants. The maximum reduction in leaf photosynthetic rate was 40-50% and this reduction was already obtained with few Striga attachments, often well before the first Striga plant emerged. Consequently, Striga infestation level will mainly enlarge the extent of photosynthesis-related damage through advancing the onset of the negative effect on photosynthesis, following an earlier Striga infection time. Reductions in leaf photosynthetic rate were always accompanied with lower rates of transpiration and stomatal conductance, but only during the reproductive phase of sorghum the reduced stomatal conductance was found to act as one of the causes of a reduced photosynthetic rate.

Keywords: Infection level, infestation level, photosynthesis, sorghum, Striga hermonthica, transpiration.

Chapter 3

Introduction Obligate hemi-parasitic plants of the genus Striga (Orobanchaceae) form an increasing problem in crop production systems in the savannah regions of Africa. These root parasites constrain the growth of crops like sorghum, millet, maize, rice, teff and cowpea. Striga hermonthica (Del.) Benth. is the most noxious of the Striga species in Africa, causing losses in cereal food crops which are estimated to range from circa 10% upwards (Doggett, 1988). Severe infestations can result in total crop failure (Riches and Parker, 1995). The effects of S. hermonthica on its host are diverse. The parasite competes with the host for resources, it changes host plant architecture and it reduces the photosynthetic rate and the water use efficiency of the host (Cechin and Press, 1994; Gebremedhin et al., 2000; van Ast et al., 2000, see also Chapter 2; Watling and Press, 2001). Striga induced reduction in host photosynthesis has been reported as the most important mechanism of growth reduction of the host (Press et al., 1987; Graves et al., 1989; Gurney et al., 1995; Smith et al., 1995; Watling and Press, 2001). This reduction develops rapidly after infection, as significant reductions are often observed before emergence of the parasite (Gurney et al., 1995; Smith et al., 1995; Watling and Press, 2000). Most of the studies on the negative effect of Striga on leaf photosynthetic rate of the host have been conducted at high levels of photon flux density (Cechin and Press, 1994; Smith et al., 1995; Watling and Press, 1997, 2000; Gurney et al., 2002). Consequently, the parasitic effect on initial light use efficiency was not established, and no comparison could be made between effects on the photosynthetic responses at low and high photon flux densities. Studies on the effect of Striga on the magnitude of the reduction in leaf photosynthetic rate of sorghum have revealed large differences in response to parasitic infection (Cechin and Press, 1993b; Gurney et al., 1995, 1999; Frost et al., 1997). This may partly be attributed to genetic differences of the host. Various studies demonstrated that the ability to maintain high rates of photosynthesis in Striga-infected plants may be an important correlate of tolerance to the parasite (e.g., Gurney et al., 1999, 2002; Rodenburg et al., 2006c). Differences in host plant resistance against Striga, resulting in differences in Striga infection level, might also have their effect on the size of the reduction in leaf photosynthetic rate. Such differences in number of Striga infections may also be brought about by differences in Striga seed infestation level in the soil. Smith et al. (1995) reported that with maize the number of emerged Striga shoots appeared to have little effect on the reduction in photosynthetic rate of the host plant. This kind of observations is however obscured by the lag-time that exists between Striga infection and emergence of the parasite. As a consequence, the number of observed Striga plants does not reflect the actual infection level, which also

Density effects of Striga on photosynthesis of sorghum includes the number of not yet emerged Striga attachments. Finally, it has also been found that differences in photosynthetic rate between Striga-infected and control plants changed over time. In this respect, Graves et al. (1990), Ramlan and Graves (1996) and Frost et al. (1997) reported that differences were largest when host plants were young and steadily decreased as host plants became older. In our study a comparison was made between the photosynthetic response of a sensitive (CK60-B) and a tolerant (Tiémarifing) sorghum cultivar to Striga infection. Photosynthesis was measured at a range of photon flux densities in order to compare the response at low and high light intensities. In an additional experiment, the photosynthetic response of the sensitive cultivar was studied at a wide range (8-2000 seeds dm–3 soil) of Striga infestation levels. Measurements were taken throughout the growing period of sorghum plants to follow the response over time.

Materials and methods

1996 experiment In 1996, a pot experiment was conducted in which the interaction between Sorghum bicolor (L.) Moench and the parasitic weed Striga hermonthica (Del.) Benth. was studied for the Striga-tolerant local Malian sorghum landrace Tiémarifing and the highly Striga-sensitive inbred sorghum cultivar CK60-B. The experiment was conducted in a greenhouse in Wageningen, The Netherlands, and was laid out using a split-plot design with S. hermonthica infestation as main plot factor (infested; non- infested) and sorghum cultivar as sub-plot factor (CK60-B and Tiémarifing). A detailed description of this experiment can be found in van Ast et al. (2000, see also Chapter 2). From 52 - 60 days after emergence (DAE) of sorghum, gas exchange measurements were conducted on flag leaves of infected and non-infected plants of both sorghum cultivars. For each combination of genotype × infestation level photosynthetic rate was measured on nine plants. Leaves were mounted horizontally in a flat chamber (25 × 15 –1 cm). Average conditions within the leaf chamber were 360 µmol mol CO2, a temperature of 23 °C, and a relative humidity between 45-60%. Light was provided by HPI/T lamps of 400 W each and the light flux could be reduced by inserting metal screens of different light transmission below the lamps. Photosynthetic rate was measured at increasing photosynthetic photon flux densities of approximately 0 (dark), 120, 350, 600, 900 and 1200 µmol m–2 s–1. Before measurement, plants were given at least 15 minutes to adapt to the next light level. The part of the plant outside the leaf chamber was subjected to almost the same conditions of light and temperature as the part in the chamber. CO2 uptake was determined in an open system modified with an

Chapter 3 automatic regulatory and recording system for gas flow, comparable to the type described by Louwerse and van Oorschot (1969). Rates of photosynthesis and transpiration, stomatal conductance and internal CO2 concentration (Ci) were calculated following the procedure of Goudriaan and van Laar (1994). For each leaf, the measured net photosynthetic rates at different radiant fluxes were fit to a negative exponential function

A = (Amax + Rd) (1 – exp (– I ε / (Amax + Rd))) – Rd (equation 1)

in which A and Amax represent the actual and the maximum net CO2 assimilation rate –2 –1 (µmol CO2 m s ), respectively, ε represents the initial light use efficiency (µmol –2 CO2 fixed per µmol photon absorbed), Rd represents dark respiration (µmol CO2 m s–1) and I represents the photosynthetic photon flux density (µmol m–2 s–1). After photosynthesis measurement, SPAD was determined by using the Minolta Chlorophyll Meter SPAD-502. Four readings were taken half way along the length of the flag leaf. Additionally, leaf thickness was determined by using an electronic micrometer (Tesa Digico 11, Brown and Sharpe, Renens, CH). In the Striga-infested pots, the number of emerged Striga plants was counted. At 63 DAE, discs of 9 mm diameter were taken from the measured flag leaves and placed in vials containing 5 ml of N,N-dimethylformamide (DMF) 99%, for extraction of chlorophyll. The absorbance of chlorophyll in DMF was measured at wavelengths of 647 (chlorophyll a) and of 664.5 nm (chlorophyll b), with the UV visible recording spectrophotometer (model Shimadzu UV-160A). Total chlorophyll was determined, using the calculations of Inskeep and Bloom (1985) and equations of Moran (1982).

2003 experiment

Plant material and growth conditions In a pot experiment conducted in 2003, in a naturally lit greenhouse in Wageningen, the photosynthetic response of the sensitive sorghum cultivar CK60-B to S. hermonthica infection was further examined. In a randomized complete block design in 16 replicates sorghum plants were exposed to Striga seed densities of 0 (control), 8, 16, 32, 64, 125, 250, 500, 1000 and 2000 germinable Striga seeds per dm3 of soil. Striga seeds were collected in 1998 from sorghum host plants in an experimental field of the Institute of Rural Economy (IER), Cinzana, Mali. A mixture of quartz sand and arable soil (3:1 v/v) was used to fill 17-L pots. Striga seeds were mixed through the upper 15 cm of soil. Ten days after adding the S. hermonthica seeds to the soil, three pre-germinated sorghum seeds were planted per pot. Thinning to one plant per pot was

Density effects of Striga on photosynthesis of sorghum done at 4 DAE, while at 10 DAE fertilizer was applied in a single dose, equivalent to 50 kg N, 42 kg P and 75 kg K per hectare. Pots were watered every 2-3 days. Daytime temperature in the greenhouse ranged from 28 to 37 °C and mean night temperature was 24 °C. Supplemental light was provided by high-pressure sodium lamps (400 W SON-T, Agro Philips lamps), which automatically switched on during day time when photosynthetically active radiation (PAR) outside the greenhouse dropped below 910 µmol photons m–2 s–1. Black screens were used to create a day length of 12 h throughout the entire growing period. Relative humidity varied between 50 and 70%.

Gas exchange and additional measurements To follow the photosynthetic response of sorghum to Striga-infection over time, leaf photosynthetic rate near light saturation was measured for each Striga-density at 19- 20, 29-30, 36-37, 43-44, 57-58, and 71-72 DAE, using four sorghum plants per density. Measurements were done with an infrared gas analyser (Model LCA-2 and PLC-N leaf chamber, Analytical Development Company (ADC), Ltd. Hoddesdon, UK) and taken halfway along the length of the youngest fully expanded leaf. An external, heat filtered light source was used to maintain the irradiance at a photosynthetic photon flux density of around 1800 µmol m–2 s–1. Environmental conditions at the time of measurement were comparable to the growing conditions. A reading was taken when the rate of CO2 exchange had been steady for 5 minutes. Leaves were always measured between 10.30 h and 14.30 h. Leaf chamber temperature ranged between 28 and 34 °C and the inlet CO2 concentration was around 350 µmol mol–1. At 64 DAE, the same equipment was used to determine the photosynthetic rate of S. hermonthica leaves. From eight Striga plants, several leaves were measured, starting from the oldest (lowest) leaf and including all younger leaves that were sufficiently large to fit into the leaf chamber. At 29 DAE, detailed gas exchange measurements were made on the youngest fully expanded leaves of control plants and sorghum plants exposed to the highest Striga seed density (2000 seeds dm–3 of soil), using a LICOR-6400-40 (LI-COR Bioscience, Lincoln, Nebraska, USA). Measurements were conducted on plants from eight replicates. Simultaneously with gas exchange measurements, chlorophyll fluorescence was measured using the same equipment. After a dark adaptation period of 10 minutes, photosynthesis and fluorescence responses were measured at photosynthetic photon flux densities of 0, 100, 350, 700, 1200 en 1500 µmol m–2 s–1. The negative exponential photosynthesis-light response curve (equation 1) was fitted to the measured net photosynthetic rates of each plant, to obtain estimates of Amax, ε and Rd. Leaf temperature ranged between 27 and 31 °C (mean: 30 °C) and the inlet CO2

Chapter 3 concentration was 400 µmol mol–1. Chlorophyll fluorescence measurements were used to calculate relative electron transport rate (ETR; µmol m–2 s–1) through photosystem II (PSII), as well as the level of photochemical (Pq) and non-photochemical quenching

(NPq). For the derivation of ETR, the electron transport efficiency of PSII (Φ2) was calculated as

Φ2 = 1 – Ft / F'm where Ft is the steady-state fluorescence emission, and F'm is the maximum fluorescence emission induced by a saturating light pulse in the light (Genty et al., 1989). Relative ETR was then calculated as

ETR = Φ2 ρƒabs I where ρ is the factor to account for the partitioning of energy between the two photosystems (PSI and PSII), ƒabs is the absorbtivity of the leaf and I is the photosynthetic photon flux density (Genty et al., 1989). Commonly, ρ is set to 0.5, assuming that at any light level the excitation energy is equally distributed between PSI and PSII (Maxwell and Johnson, 2000; Rascher et al., 2000). The absorbtivity was set to 0.8, which indicates that of the incoming PAR 80% is absorbed by the leaf (Goudriaan and van Laar, 1994). Photochemical quenching of fluorescence (Pq) was calculated as

Pq = (F'm – Ft)/(F'm – F'o) where F'o is the basic fluorescence in the light when all PSII centres are oxidized by a period of far-red light (Schreiber, 1986). Finally, non-photochemical quenching (NPq) was calculated as

NPq = (Fm – F'm)/F'm where Fm is the maximum fluorescence emission induced by a saturating light pulse in the dark (Genty et al., 1989). Throughout the duration of the experiment emergence of S. hermonthica was frequently determined, using a time interval of 2-3 days. Destructive harvests on four replicates were conducted at 31 and 45 DAE to determine the number of underground S. hermonthica attachments. Roots were separated from the soil by carefully washing over a 2-mm sieve after which Striga attachments were counted with the help of a magnifying (8×) glass.

Density effects of Striga on photosynthesis of sorghum

Statistical analysis Data were subjected to analysis of variance (ANOVA) followed by a comparison of means with the least significant difference (LSD.) using the statistical package Genstat (release 8.1). Prior to analysis, data on number of Striga attachments and number of emerged Striga plants were subjected to log(x+1) transformations, in which x is the original observation, to meet the assumptions of the analysis of variance. Characteristics of S. hermonthica leaves, including photosynthetic rate, were subjected to a two sided t-test.

Results

1996 experiment The negative exponential function (equation 1) gave an accurate description of the net

CO2 assimilation light response of flag leaves of both healthy and Striga-infected plants (Fig. 1). Clear differences were present between the non-infected plants of CK60-B and

Tiémarifing. Net photosynthetic rate at light saturation (Amax) of CK60-B was

a b

CO -assimilation CO2-assimilation 50 2 50 –2 –1 (µmol CO m –2 s –1) (µmol CO2 m s ) 2 40 40

30 30

20 20

10 10

0 0 0 250 500 750 1000 1250 0 250 500 750 1000 1250 -10 -10 –2 –1 –2 –1 photon flux density (µmol m s ) photon flux density (µmol m s ) –2 –1 Figure 1. CO2 assimilation rate (A; µmol CO2 m s ) of sorghum leaves measured at different photosynthetic photon flux densities (I; µmol m–2 s–1) for a sensitive (CK60-B; (a) and a tolerant (Tiémarifing; (b) cultivar at 52-60 days after emergence. Curves represent control plants (solid line) or plants infected with Striga hermonthica (dotted line) and are constructed by fitting a negative exponential function (equation 1) to the measured data. Bars represent standard error of mean. Parameter estimates of the negative exponential function are presented in Table 1.

Chapter 3

) were all were ) s

g ) were derived

d R

28.4 b 28.1 b 2.010

; Dark respiration, ; Dark

ε ) and Stomatal conductance ( conductance Stomatal and ) E

<0.05) different. P

CK60-B Tiémarifing Tiémarifing CK60-B

-infected sorghum plants of the sensitive cv CK60-B and 52-60 DAE. Parameters ofthe photosynthesis light response

1.37 a a 1.37 0.93 b 0.94 b 1.01 b 0.077 12.11a 11.98 a 12.97 a 12.89 a 0.784 45.7a 23.6 c 30.5 b 30.0 b 2.61 Control Infected Control Infected SED Control Infected Striga

( rate Transpiration ), A . SED for cultivar x infection comparison. comparison. x infection cultivar for . SED –1 efficiency, use light ; Initial s ) 12.23a 5.67 c 8.41 b 7.53 b 0.522 max –2 ) ) 39.4 a 22.4 c –1

A –1 s

) 90.9 a 58.0 b 64.6 ab 71.6 ab 13.45 –2 –1 ) a 3.30 1.92 b 2.21 b 2.20 b 0.213 m 2

–1 ) 0.27a 0.13 c 0.17 b 0.17 b 0.016 s –1

–2 )

s –1 –2 O µmol mol O m

2 ; 2 i ; µmol CO ; µmol assimilation rate, rate, assimilation A (C

2 ) 0.082 a 0.060b 0.074 a 0.081 a 0.0049 ; mol m ; mol s –1

) g s mmol H

–1 ) –2 ; mmolH –2 –2 s –1 2 E –2

s –2 m 2 m 2 µmol m µmol concentration 2 2 assimilation rate ( assimilation rate 2 (µmol CO ratio (µmol CO (µmol ratio

(µmol CO max d (µmol CO derived at a photon 1200flux of µmoldensity m through curve fitting. Values of Net photosynthesis Net ( Values of curve fitting. through A/E Stomatal conductance ( thicknessLeaf (cm) SPAD-value chlorophyllTotal (a+b)leaf (mg fresh g weight 0.064 a 0.058b 47.8 a 0.057 b 35.8 c 0.058 b 0.0026 42.6 b 41.6 b 2.27

Transpiration rate ( rate Transpiration values in the same row followed by a different letter are significantly ( significantly are letter a different by row followed same in the values

Net CO Internal CO the tolerant cv Tiémarifing. Measurements were made from curve (Maximum netCO and uninfected of leaves flag of Characteristics 1. Table A ε R

Density effects of Striga on photosynthesis of sorghum

significantly higher and this higher Amax was accompanied with significantly higher values for leaf dark respiration (Rd), SPAD, leaf thickness and total chlorophyll content (Table 1). Initial light use efficiency (ε) did not differ between control plants of CK60-B and Tiémarifing. In the Striga-infested pots, first emergence of the parasite on CK60-B (36 DAE) occurred significantly earlier than on Tiémarifing (43 DAE). During photosynthesis measurements, conducted from 52 to 60 DAE, the number of emerged Striga plants on CK60-B was 7.6 (± 1.14) plants per pot, which was significantly higher than on Tiémarifing with 3.8 (± 0.77) plants per pot. Both cultivars differed considerable in their response to Striga-infection. With CK60-B all three parameters characterising the leaf photosynthesis light-response curve (Amax, ε, Rd) were significantly reduced (48, 27 and 32%, respectively) following infection with Striga. These reductions were accompanied with significant reductions in SPAD, leaf thickness and total chlorophyll content of the flag leaves. With Tiémarifing, Striga-infection did not affect the parameters of the leaf photosynthesis light-response curve, nor did it affect the other measured leaf characteristics. A closer look at the measurements at the highest photosynthetic photon flux density (1200 µmol m–2 s–1) provided further information on the differences in photosynthetic response between both cultivars. Apart from net CO2-assimilation rate (A) also the transpiration rate (E) and the stomatal conductance (gs) of control plants of CK60-B were significantly higher than those of control plants of Tiémarifing. The ratio between photosynthesis and transpiration (A/E-ratio) was identical for both cultivars, whereas also no significant difference in internal CO2-concentration (Ci) was observed. Striga infection reduced A and E of CK60-B to a similar extent, and consequently A/E-ratio of infected plants was not affected. Stomatal conductance of Striga infected plants of CK60-B was only half of that of uninfected plants and this resulted in a significantly reduced Ci. With Tiémarifing, similar to what was observed for photosynthesis, Striga infection did not affect E, A/E-ratio, gs or Ci.

2003 experiment In the 2003 experiment, plants of CK60-B were exposed to a logarithmically increased range of Striga seed densities ranging from 8 - 2000 viable seeds per dm3 of soil. At 19 DAE, when photosynthesis was measured for the first time, no Striga plants had emerged. However, the first symptoms of Striga infection were already visible on host plants exposed to the highest densities. All measured leaves of plants infested with densities of 256 seeds dm–3 of soil and higher contained at least some purple specks of pinpoint size. At lower infestation levels these leaf symptoms were rare and only present at infestation levels of 125 seeds dm–3 (two out of four plants) and 32 seeds

Chapter 3

45 number of emerged Striga plants a 45 number of emerged Striga plants b 40 40 35 35 30 30 25 25 20 20

15 15 10 10 5 5 0 0 8 16 32 64 125 250 500 1000 2000 8 16 32 64 125 250 500 1000 2000 –3 number of Striga seeds dm–3 of soil number of Striga seeds dm of soil

45 number of emerged Striga plants c 45 number of emerged Striga plants d 40 40 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 8 16 32 64 125 250 500 1000 2000 8 16 32 64 125 250 500 1000 2000 –3 number of Striga seeds dm–3 of soil number of Striga seeds dm of soil

45 number of emerged Striga plants e 40 35 30 25 20 15 10 5 0 8 16 32 64 125 250 500 1000 2000

number of Striga seeds dm–3 of soil

Figure 2. Number of emerged Striga hermonthica plants per host plant as a function of Striga- infestation level (seeds dm–3). Observations were made on the sensitive sorghum cultivar CK60-B at 31 days after sorghum emergence (DAE) (a), 38 DAE (b), 45 DAE (c), 58 DAE (d) and 72 DAE (e). Bars represent standard error of mean and are based on log(x + 1) transformed data.

Density effects of Striga on photosynthesis of sorghum

35 number of attached Striga plants

30 25

10 5

0 8 16 32 64 125 250 500 1000 2000 number of Striga seeds dm–3 of soil

Figure 3. Total number of Striga hermonthica attachments per host plant as a function of Striga-infestation level (seeds dm–3). Observations were made on the sensitive cultivar CK60- B at 31 (open bars) and 45 days (shaded bars) after sorghum emergence (DAE). Bars represent standard error of mean and are based on log(x + 1) transformed data.

per dm3 (one out of four plants). From then on, Striga plants started to appear, and emergence was more rapidly at higher infestation levels (Fig. 2). At 31 DAE, emerged Striga plants were present in at least one pot of the treatments with an infestation level of 125 seeds dm–3 and higher. In contrast, the first pot with emerged Striga in the lowest density treatment (8 seeds dm–3) was observed at 58 DAE. Only at 72 DAE all infested pots contained at least one emerged Striga plant (data not shown). Striga seed density was found to have a significant effect on the average number of emerged Striga plants (31 DAS: P=0.049; later dates: P<0.001). The number of emerged Striga plants increased with time and at each observation date a gradual increase of Striga plants with infestation level was observed. At 31 and 45 DAE, sorghum roots were rinsed with water to count the total number of Striga attachments. The number of attachments showed a similar pattern as the number of emerged Striga plants; on both dates a gradual increase with Striga infestation level was observed and at 45 DAE a higher number of attachments was observed than at 31 DAE (Fig. 3). Already at 31 DAE, all treatments contained at least one pot with Striga attached to the roots of its host.

Chapter 3

a Stomatal conductance 50 CO2 assimilation rate 6 Transpiration rate 0.5 40 5 0.4 4 30 0.3 3 20 0.2 2 10 0.1 1 0 0 0 1 10 100 1000 10000 1 10 100 1000 10000 1 10 100 1000 10000

b 50 6 0.5

5 40 0.4 4 30 0.3 3 20 0.2 2

10 1 0.1 0 0 0 1 10 100 1000 10000 1 10 100 1000 10000 1 10 100 1000 10000

c 50 6 0.5 40 5 0.4 4 30 0.3 3 20 0.2 2 10 1 0.1

0 0 0 1 10 100 1000 10000 1 10 100 1000 10000 1 10 100 1000 10000 d 50 6 0.5

5 40 0.4 4 30 0.3 3 20 0.2 2 10 1 0.1 0 0 0 1 10 100 1000 10000 1 10 100 1000 10000 1 10 100 1000 10000

e 50 6 0.5

40 5 0.4 4 30 0.3 3 20 0.2 2 10 1 0.1 0 0 0 1 10 100 1000 10000 1 10 100 1000 10000 1 10 100 1000 10000

f 0.5 50 6 40 5 0.4 4 30 0.3 3 20 0.2 2 0.1 10 1 0 0 0 1 10 100 1000 10000 1 10 100 1000 10000 1 10 100 1000 10000 number of Striga seeds dm–3 of soil number of Striga seeds dm–3 of soil number of Striga seeds dm–3 of soil

Density effects of Striga on photosynthesis of sorghum

–2 The net leaf CO2 assimilation rate (A) of control plants was about 40 µmol CO2 m –1 –2 –1 s , started to decline at 57 DAE (35.4 µmol CO2 m s ) and was considerably lower –2 –1 at 72 DAE (25.7 µmol CO2 m s ), when flag leaves were measured (Fig. 4). Already at 19 DAE, A was significantly affected by Striga density (P=0.014). Photosynthesis of plants exposed to a Striga density of 250 seeds dm–3 and above was significantly reduced and on average 30% lower than that of the control plants. At 29 (P<0.001), 36 (P=0.07) and 43 DAE (P<0.001), photosynthesis was again significantly affected by Striga density. A of all treatments was significantly reduced compared to the control treatments, except for the lowest two densities at 36 DAE. At all dates a gradual decrease in A with Striga density was observed with maximum reductions of 45-50% at the highest Striga density. At 57 and 72 DAE significant reductions in A were only found when all Striga densities were combined and a comparison was made between control and Striga-infested treatments. At 57 DAE the average reduction in A due to Striga infestation was just 17% (P=0.004), whereas at 72 DAE, when flag leaves were measured, the reduction was much higher again with 42% (P=0.011). For ordinary leaves of control plants transpiration rate (E) varied between 5.9 to 4.4 –2 –1 mmol H2O m s . Flag leaves showed a considerably lower transpiration rate of 3.3 –2 –1 mmol H2O m s . Transpiration rate of infected plants largely followed the same pattern as the assimilation rate, though at 19 DAE Striga did not yet have a significant effect on transpiration rate. It often appeared that photosynthesis was slightly more reduced than transpiration, but only at 29 DAE Striga density had a significant effect on A/E-ratio. On that date A/E-ratio of control plants was 6.6, and this ratio gradually reduced with increasing Striga density to a value of 5.0 at the highest Striga density. The A/E-ratio of treatments with a Striga density of 64 seeds dm–3 and higher was significantly reduced compared to that of the control treatment.

Stomatal conductance (gs) was closely linked to photosynthetic rate. For control plants an average value of 0.361 mol m–2 s–1 was found during the first four observations dates. This value was reduced to 0.239 and 0.191 mol m–2 s–1 at 57 and

–2 –1 –2 Figure 4. CO2 assimilation rate (A; µmol CO2 m s ), transpiration rate (E; mmol H2O m –1 –2 –1 s ) and stomatal conductance (gs; mol m s ) as a function of Striga hermonthica-infestation level (seeds dm–3). Measurements were made on the sensitive sorghum cultivar CK60-B at 19 days after sorghum emergence (DAE) (a), 29 DAE (b), 36 DAE (c), 43 DAE (d), 57 DAE (e) and 72 DAE (f). The dotted lines represent values of the uninfected control plants. Bars represent standard error of mean, whereas S.E.D. is presented in the upper right corner of each graph.

Chapter 3

–1 Table 2. Intercellular CO2 concentrations (Ci; µmol mol ) of Striga-infected and uninfected (control) plants of the sorghum cultivar CK60-B at different days after sorghum emergence (DAE). Data of Striga-infected plants are averages of all Striga-infested treatments. Control Infected SED P-value 19 DAE 97.4 95.9 19.54 0.940 29 DAE 83.2 77.2 7.31 0.412 36 DAE 74.6 79.9 14.29 0.714 43 DAE 61.0 63.8 11.6 0.813 57 DAE 57.6 53.3 10.71 0.689 72 DAE 81.4 58.7 11.03 0.048

72 DAE, respectively. Stomatal conductance was significantly affected by Striga density at 29, 36 and 43 DAE (P<0.001). On these dates conductance decreased gradually with Striga infestation level. At 57 and 72 DAE a significant reduction in gs (P=0.01 and P=0.001, respectively) was only obtained if all Striga-infested treatments were combined and compared to gs of the control plants.

The internal CO2-concentration (Ci) of control plants gradually dropped from 97.4 –1 –1 µmol mol to 57.6 µmol mol at 57 DAE, whereas Ci of flag leaves was again 81.4 –1 µmol mol (Table 2). Ci was not affected by Striga density, except at 72 DAE when

Ci of Striga-infected plants was significantly lower (P=0.048) than that of control plants. Results of the photosynthesis measurements conducted at 31 and 45 DAE were combined and grouped according to number of attachments and number of emerged Striga plants. If less than two observations were available, the observations were combined with that of the neighbouring group. Finally, A was plotted against total number of attachments (Fig. 5a) and number of emerged Striga plants (Fig. 5b). –2 –1 Fig. 5a shows that A of control plants was on average 40 µmol CO2 m s . Photosynthetic rate of plants that were exposed to Striga infested soil, but on which roots no Striga attachments could yet be detected (n = 21), already showed a marked reduction in A (19%), whereas on plants with just one Striga attachment (n = 7) the reduction in A was 29%. On plants with two Striga attachments or more A stabilised –2 –1 around 23 µmol CO2 m s , corresponding to a reduction of 42%. From Fig. 5b it can be learned that this maximum reduction was in general already obtained before the first Striga plant emerged, which is not surprising as on sorghum plants with just one emerged Striga plant on average eight Striga attachments (range 1-22) were observed. At 29 DAE additional photosynthesis measurements were taken on uninfected control plants and plants exposed to the highest Striga seed density (2000 seeds dm–3

Density effects of Striga on photosynthesis of sorghum

45 –2 –1 45 CO assimilation rate (µmol CO m–2 s–1) CO2 assimilation rate (µmol CO2 m s ) a 2 2 b 40 40

35 35

30 30

25 25 20 20 15 15 10 10 5 5

0 0 0 5 10 15 20 25 30 35 0 5 10 15 total Striga attachments total emerged Striga plants

–2 –1 Figure 5. CO2 assimilation rate (A; µmol CO2 m s ) for the sensitive sorghum cultivar CK60-B as a function of total number of Striga hermonthica attachments (a) and as a function of total number of emerged Striga hermonthica plants (b). Results of the photosynthesis measurements conducted at 31 and 45 days after sorghum emergence (DAE) were combined and grouped according to total number of Striga attachments (a) or to total number of emerged Striga plants (b). Open circle represents plants of the Striga-free (control) treatment; open triangle represents plants from infested treatments on which no Striga attachments were observed; open square represents plants from treatments on which Striga attachments, but no emerged Striga plants were observed Bars represent standard error of mean.

of soil), using a LICOR-6400-40. In contrast to the ADC-measurements, CO2 assimilation rate was determined at a range of light intensities and next to gas- exchange, fluorescence was determined. The net CO2 assimilation light response for both control and infected plants could be accurately described by the negative exponential function (Fig. 6). Striga infection resulted in significant reductions in Amax and ε, whereas Rd was not affected. At the highest photosynthetic photon flux density (1500 µmol m–2 s–1) the reduction in A was 44% (Table 3). This reduction was accompanied with a significant reduction in E and gs. Photosynthesis was more strongly reduced than transpiration, resulting in a small (8%), but significant, reduction in A/E-ratio. Ci was not affected. Fluorescence measurements indicated that Striga infection reduced the electron transport rate through PSII (ETR) (Fig. 7) and this reduction was accompanied by a reduction in photochemical quenching of fluorescence (Pq), whereas the level of non- photochemical quenching (NPq) was increased.

Chapter 3

CO assimilation 40 2 –2 –1 (µmol CO2 m s )

0 0 400 800 1200 1600 -10 photon flux density (µmol m–2 s –1)

–2 –1 Figure 6. CO2 assimilation rate (µmol CO2 m s ) measured at different photosynthetic photon flux densities (I; µmol m–2 s–1) for the sensitive sorghum cultivar CK60-B at 29 days after emergence in the 2003 experiment. Curves represent control plants (solid line) or plants from the highest Striga hermonthica-infestation level (2000 seeds dm–3) (dotted line) and are constructed by fitting a negative exponential function (equation 1) to the measured data. Bars represent standard error of mean. The parameter estimates of the negative exponential function are presented in Table 3.

Table 3. Characteristics of flag leaves of uninfected and Striga-infected sorghum plants of the sensitive cv CK60-B. Measurements were made at 29 DAE. Values of infected plants refer to sorghum plants exposed to the highest Striga-infestation level. Parameters of the

photosynthesis light response curve (Maximum net CO2 assimilation rate, Amax; Initial light

use efficiency, ε; Dark respiration, Rd,) were derived through curve fitting. Values of Net

photosynthesis (A), Transpiration rate (E) and Stomatal conductance (gs) were all derived at a photon flux density of 1500 µmol m–2 s–1. Control Infected SED –2 –1 Amax (µmol CO2 m s ) 41.9 a 24.2 b 3.08 –2 –1 ε (µmol CO2 µmol m s ) 0.045 a 0.026 b 0.0048 –2 –1 Rd (µmol CO2 m s ) 1.4 a 1.4 a 0.354 –2 –1 Net CO2 assimilation rate (A; µmol CO2 m s ) 32.9 a 18.4 b 0.610 –2 –1 Transpiration rate (E; mmol H2O m s ) 3.56 a 2.19 b 0.120 –2 –1 A/E ratio (µmol CO2 mmol H2O ) 9.27 a 8.49 b 0.254 –2 –1 Stomatal conductance (gs; mol m s ) 0.17 a 0.10 b 0.004 –1 Internal CO2 concentration (Ci; µmol mol ) 63.9 a 67.8 a 7.24 Specific Leaf Weight (g cm–1) 0.0029 0.0027 0.0002 values in the same row followed by a different letter are significantly (P<0.05) different.

Density effects of Striga on photosynthesis of sorghum

Photosynthesis of Striga plants On 64 DAE, the photosynthetic rate of leaves of S. hermonthica plants was determined. Based on leaf position, leaves were classified in three categories: young (upper leaves), middle and old (lower leaves). Compared to the youngest leaves, older leaves were thicker, had a higher SPAD-value and a higher A (Table 4). E, gs and Ci were not affected by leaf age and consequently the A/E-ratio of older leaves was significantly higher than that of younger leaves. Compared to sorghum, Striga was characterized by a relatively low A and high values for E, gs and Ci.

Discussion The 1996 experiment demonstrated that Striga infection did not affect the photosynthetic response of flag leaves of the tolerant sorghum cultivar Tiémarifing. Additional observations on leaf thickness, total leaf chlorophyll content and SPAD of the flag leaves of Tiémarifing demonstrated that also these characteristics, which have a clear relation with leaf photosynthetic rate, were not affected by Striga infection. The absence of a reduction in leaf photosynthetic rate of tolerant sorghum cultivars after Striga infection was earlier reported by Gurney et al. (1995) and Frost et al. (1997). Our observation thus confirms that tolerant cultivars often possess the ability to

Table 4. Parameter values of S. hermonthica leaves of plants parasitizing on the sensitive sorghum cv CK60-B. Leaves were classified according to leaf position. Measurements were conducted at 64 days after sorghum emergence. P- P- P- young middle old value value value (upper; y) (m) (lower; o) y ↔ m m ↔ o y ↔ o Number of leaves measured 10 9 7

CO2 assimilation rate –2 –1 (A; µmol CO2 m s ) 13.5 18.6 19.5 0.081 0.912 0.033 Transpiration rate –2 –1 (E; mmol H2O m s ) 6.64 7.54 7.61 0.036 0.804 0.0391 Stomatal conductance –2 –1 (gs; mol m s ) 0.74 0.72 0.78 0.924 0.673 0.880

Internal CO2 concentration –1 (Ci; µmol mol ) 338 338 326 0.960 0.147 0.345 SPAD-value 23.6 31.9 38.2 0.013 0.042 <0.001 Specific Leaf Weight (SLW; g cm–1) 0.0031 0.0035 0.0040 0.070 0.032 0.002

Chapter 3

150 ETR (µmol m–2 s–1) a

120

0 0 500 1000 1500 2000 – – photon flux density (µmol m 2 s 1)

1.2 Pq b

1.0

0.8

0.6

0.4

0.2

0.0 0 500 1000 1500 2000 photon flux density (µmol m–2 s–1)

1.8 NPq c 1.5

1.2

0.9

0.6 0.3

0 0 500 1000 1500 2000 -0.3 photon flux density (µmol m–2 s–1) Figure 7. Electron transport rate (ETR: µmol m–2 s–1) (a), photochemical quenching (Pq) (b) and non-photochemical quenching (NPq) (c) for the sensitive sorghum cultivar CK60-B for control (open symbols) and Striga hermonthica-infected (closed symbols) plants at 29 DAE. Striga-infected plants originate from the highest Striga-infestation level (2000 Striga seeds dm–3 of soil). Bars represent standard error of mean.

Density effects of Striga on photosynthesis of sorghum maintain high rates of photosynthesis when infected with Striga, supporting the hypothesis that this capacity is one of the mechanisms behind tolerance. Reduced rates of photosynthesis after Striga-infection were observed for flag leaves of the sensitive cultivar CK60-B. Lower rates of photosynthesis after Striga infection are commonly observed and, just as in this experiment, often accompanied with lower rates of transpiration and stomatal conductance (e.g., Taylor and Seel, 1998; Gurney et al., 2002). Whether the latter is a cause or a consequence of other changes in photosynthetic metabolism has not been resolved. Press et al. (1987) for instance concluded that stomatal conductance was responding to, rather than causing a decrease in photosynthesis. Frost et al. (1997) on the other hand concluded that lower values of stomatal conductance are a primary cause of lower rates of photosynthesis in leaves of Striga-infected sorghum. Based on the two years of experimentation in our study it can be concluded that the role of stomatal conductance in the reduction of leaf photosynthetic rate following Striga infection changes with time. Initially, the reduction in stomatal conductance is just a response to a decrease in photosynthesis, whereas later on stomatal conductance is at least partially responsible for the reduction in leaf photosynthetic rate at higher levels of photon flux density. In 1996, the observations during the last part of the growing season were conducted on flag leaves. These observations showed a reduction of 27% in initial light use efficiency (ε) combined with a reduction of 48% in maximum leaf photosynthetic rate (Amax). Effects on CO2 assimilation can be divided into effects on photosynthetic events per se and effects on photosynthesis related processes. The first category of effects will result in an increased carboxylation resistance, whereas an increased stomatal resistance is an example of the second category. A reduced ε as a result of a direct effect on stomatal behaviour is unlikely. Even if diffusion of CO2 into the leaf is impaired, it is doubtful whether CO2 will become a limiting factor at low light intensities. Therefore, the reduction in ε is more likely to result from a hindered photosynthetic event per se. The larger reduction at high light intensities suggests that in this case another mechanism is contributing to the overall effect. At high levels of photon flux density the reduction in photosynthetic rate was accompanied by significant reductions in transpiration rate and stomatal conductance, but more important also with a significant reduction in internal

CO2 concentration. This implies that at high radiation levels, CO2 supply was insufficient to meet demand, and consequently that the lower stomatal conductance contributed to the reduction in leaf photosynthetic rate. The measurements in 2003, when leaf photosynthetic rate was only measured at relatively high photosynthetic photon flux densities (1800 µmol m–2 s–1), support this last finding. At 72 DAE, measurements were conducted on flag leaves, and also in this case the reduction in leaf photosynthetic rate of infected sorghum plants was accompanied with a reduced

Chapter 3

internal CO2 concentration. The observations in 2003, conducted during earlier stages of the growing period of sorghum plants on ordinary leaves, show a different pattern. With the exception of the first observation date, reductions in leaf photosynthetic rate following Striga infection were also in this case accompanied by a reduced stomatal conductance. Internal CO2 concentration however was not affected, indicating that during these stages the lower stomatal conductance was not responsible for the reduction in leaf photosynthetic rate. The photosynthesis light response characteristics, determined at 29 DAE, strengthened this conclusion as reductions in ε and Amax (42%) were identical. Simultaneous chlorophyll fluorescence measurements confirmed the presence of changes in photochemistry of the host during these early stages of Striga infection. Photochemical quenching and the electron transport rate through PSII were both reduced, whereas non-photochemical quenching was increased. These responses have often been observed in plants exposed to abiotic stress conditions, like drought, that reduce the capacity for photochemistry of leaves (e.g., Havaux and Lannoye, 1985; Schapendonk et al., 1992) and have also often been used as a measure for stress severity (Schreiber, 1986) or stress tolerance (Harbinson, 1995). Earlier, Cechin and Press (1994) conducted chlorophyll fluorescence measurements on Striga-infected sorghum plants. They also found a lower maximum efficiency of excitation energy capture by open photosystem II reaction centres and a lower quantum efficiency of photosystem II electron transport. All of these responses in themselves are still not conclusive on the primary cause behind the reduction in photosynthetic rate, as it can not be excluded that the responses are brought about by the feedback effect of down regulated photosynthesis following reduced stomatal aperture in infected plants. Yet again, the observations at low light intensities in our study make it very unlikely that stomatal closure is the primary cause, as significant reductions in Pq and ETR were already observed at photosynthetic photon flux densities as low as 350 µmol m–2 s–1 and a significant increase in NPq was even present at 100 µmol m–2 s–1. The observation that a reduced stomatal conductance contributed to a reduction in leaf photosynthetic rate during the later part of the growing period of sorghum, might be related to the increased water loss of the host-parasite system after emergence of the parasite. In the aerial phase of the life cycle of Striga, transpiration is generally regarded as a means of facilitating the flow of resources from the host to the parasite. Earlier observations demonstrated that the transpiration rate of the parasite expressed per unit leaf area is relatively high (Shah et al., 1987; Ackroyd and Graves, 1997), and –2 –1 our observations, with values between 6.6-7.6 mmol H2O m s , confirm this. Taylor and Seel (1998) studying Striga hermonthica-induced changes in soil matric potential found no indication of an increase in the rate of water depletion in the rooting zone of

Density effects of Striga on photosynthesis of sorghum infected maize plants until approximately 63 DAE. During the final growth stages, when the total leaf area of the parasite was rapidly increasing, the Striga-infected plants used nearly 50% more water than uninfected maize plants during the same period. The withdrawal of water from the host might induce water stress, resulting in stomatal closure. As in our experiment plants were regularly watered, this water stress can not have resulted from a lack of water supply, but rather it might have resulted from an increased transpiration rate that exceeded the water uptake capacity of the host. Rather than the identification of the mechanism behind the reduction in photosynthesis, the 2003 experiment was designed to resolve the relation between Striga infection level and the size of the reduction in photosynthetic rate. Observations of Graves et al. (1989) on Striga-infected sorghum and Smith et al. (1995) on Striga- infected maize indicated that photosynthesis of the host plant was already significantly affected before emergence of the parasite. The current research confirmed these findings and demonstrated that just a few Striga attachments were sufficient to bring about the maximum reduction in photosynthetic rate, which was shown to be 40-50% of the photosynthetic rate of uninfected control plants. In most instances, this maximum reduction was already obtained before any of the attachments had developed into an emerged Striga plant. Consequently, the relation between number of emerged Striga plants and the reduction in leaf photosynthetic rate could simply be depicted by a horizontal line representing the maximum reduction level. This maximum reduction in leaf photosynthetic rate due to Striga infection was initially around 45%, a value similar to the maximum reduction reported by Gurney et al. (1995) for the same cultivar. At 57 DAE the maximum reduction in leaf photosynthetic rate of infected sorghum plants was considerably smaller (circa 25%) than earlier observed. Also Graves et al. (1990), Ramlan and Graves (1996) and Frost et al. (1997) found that differences in rates of photosynthesis between infected and control plants were largest with young host plants and decreased over time. In our research the reduction in leaf photosynthetic rate increased again during the final stages of the growing period of sorghum plants when flag leaves were measured. Such measurements were not included in the other studies. As discussed earlier, this increased reduction was at least partially caused by a reduced stomatal conductance, which in turn was hypothesized to result from an increased transpiration rate of the host-parasite complex in this period. As CK60-B is known to be an extremely sensitive cultivar (van Ast et al., 2000, see also Chapter 2; Rodenburg et al., 2005), it can not be excluded that for less sensitive cultivars the reduction in leaf photosynthetic rate increases more gradually with an increase in number of emerged Striga plants. In our research, the relation between

Chapter 3 photosynthetic rate and infestation level (Fig. 4), observed at various times throughout the growing period of sorghum plants, gradually evolved, and perhaps falsely suggesting that after infection leaf photosynthetic rate drops slowly. Analysis of the average photosynthetic response at a specific infestation level and a specific moment in time learns however that intermediate average responses are mainly the result of a number of non-infected plants showing no reduction combined with a number of infected plants showing the maximum reduction. Only rarely is this response the result of intermediate responses of individual plants. Infestation level thus merely has an effect on the average photosynthetic rate by influencing the fraction of infected and not-yet infected plants. If just a few attachments are capable of causing a maximum reduction in photosynthetic rate, one might argue whether control strategies aiming at a reduction in Striga soil seedbank are worthwhile to pursue. However, even though the number of emerged Striga plants was found not to have a direct influence on the magnitude of the reduction in photosynthesis, Striga infestation level still has a clear indirect effect on the extent of photosynthesis-related-damage caused by the parasite; Striga infestation level determines the time of Striga infection and through that the onset of the negative effect on photosynthesis. Due to this strong relation between infestation level and infection time it is likely that lower infestation levels will have a clear yield benefit, though the extent of such yield improvements is difficult to foresee.

CHAPTER 4

Effects of Striga seed infestation level on the performance of the host plant - parasite association Sorghum bicolor - Striga hermonthica

A. van Ast, L. Bastiaans

Crop and Weed Ecology Group, Department of Plant Sciences, Wageningen University, P.O. Box 430, 6700 AK Wageningen, The Netherlands

Abstract Striga hermonthica (Del.) Benth. is an obligate root-parasitic weed and has been identified as one of the major biological threats to sorghum production in the savannah zones of sub-Saharan Africa. A pot experiment with the sensitive sorghum cultivar CK60-B was conducted to establish the responses of host and S. hermonthica at a wide range of Striga-seed infestation levels in the soil. Reductions in kernel weight of sorghum due to parasite infection could be attributed to a reduced dry matter production of the host, the withdrawal of host resources by the parasite and an altered allocation of assimilates within the host plant. Yield reduction of sorghum increased with Striga infestation level, and this increase was associated with an earlier time of first Striga emergence and a higher maximum number of emerged Striga plants (NSmax). The relation between Striga infestation level and the reduction in sorghum kernel dry weight developed according to a rectangular hyperbola, mainly because of a non-proportional increase in NSmax with infestation level. A proportional 4.3% increase in yield reduction per emerged Striga plant was found. Despite the reduction in host plant vigor at higher infestation levels, the parasite was able to at least maintain its dry matter production. This was mainly because the fraction Striga of the total dry matter production of the host-parasite association significantly increased with infestation level, reaching a value of nearly 70% at the highest infestation level. Despite differences in emergence time of Striga plants, following differences in Striga soil infestation level, first flowering of the parasite was always obtained around 71 days after sorghum emergence, indicating a clear synchronization of flowering time of Striga plants. It was concluded that in highly infested fields, large reductions in infestation level are required before significant decreases in crop yield loss and parasite pressure will be obtained. Delaying the time of infection was hypothesized to offer greater potential for reducing Striga-induced damage to cereal production.

Keywords: Striga hermonthica, Sorghum bicolor, infection level, infection time, attachment.

Chapter 4

Introduction Striga hermonthica (Del.) Benth. is an obligate root-parasitic weed and has been identified as one of the major biological threats to cereal production in the savannah zones of sub-Saharan Africa (Sauerborn, 1991; Parker and Riches, 1993). The intensification of land use accompanied by a reduction in fallow periods and an expansion of cereal monocropping resulted in agro-ecological conditions that gave good reproduction opportunities for Striga plants (Weber et al., 1995; Berner et al., 1996). As a result, incidence and severity of S. hermonthica strongly increased in the last 20-30 years and caused severe losses in important food crops as sorghum (Sorghum bicolor [L.] Moench), pearl millet (Pennisetum glaucum [L.] R. BR.) and maize (Zea mays [L.]) (Sauerborn, 1991; Ejeta et al., 1992). Striga produces numerous very small seeds per plant. Estimates of seed production vary from 6,700 to 85,000 per reproductive Striga plant (Rodenburg et al., 2006a; Stewart, 1990; Webb and Smith, 1996). The seeds can stay viable for many years in the soil (Eplee and Westbrooks, 1990; Weber et al., 1995). Infection by Striga can cause yield losses of a few percentages up to complete crop failure, depending on crop variety and Striga seed infestation level of the soil (Adetimirin and Aken’Ova, 2000; Riches and Parker, 1995; Rodenburg et al., 2005; Sauerborn, 1991; van Ast et al., 2000, see also Chapter 2). There are, however, only few data relating the severity of Striga seed infestation to Striga infection level, and the subsequent dry matter production of host and parasite. As Striga is a root parasitic plant, determination of the actual infection level requires uprooting of the host plant, which is a laborious and destructive measurement. Consequently, not much is known on the relation between soil infestation and infection level. For practical reasons, the number of above-ground Striga plants is often used as an indication of parasitic pressure. This number is however a reflection of two successive processes: attachment and emergence. Doggett (1965) suggested that at low infestation levels the majority of the Striga seedlings on the host root are able to emerge, indicating that under such conditions the number of emerged Striga plants provides a good reflection of the total number of attachments. At the same time he estimated that at high infestation levels, due to intraspecific competition, only 10-30% of the attached parasites is able to emerge. Density dependence during the subterranean stages of the life cycle of the parasite was also reported by Smith and Webb (1996), van Delft et al. (1997) and Rodenburg et al. (2006a). From these observations it becomes evident that there is no linear relationship between Striga infestation level and the number of emerged Striga plants. In another pot experiment, with infestation levels ranging from 62 to 16,000 seeds dm–3 of soil, Rodenburg et al. (2006b) even observed a decline in the number of emerged Striga plants at extremely high infestation levels (≥ 8,000 seeds dm–3). The authors suggested that apart from intraspecific

Effects of Striga seed infestation level on the sorghum-Striga association competition this response might result from a reduced carrying capacity of the host plant at higher infestation levels, following reduced host vigour. In the same experiment, Rodenburg et al. (2006b) studied the relationship between Striga infection level and sorghum yield loss, and observed that this response was genotype specific. For the susceptible, sensitive genotypes no grain yield was obtained at the highest infestation levels and even at the lowest infestation level yield reductions of 55% (E36-1) and 80% (CK60-B) were already observed. For the poorly resistant but tolerant cultivar Tiémarifing yield reduction did not vary much with infestation level and was always around 57%, indicating that in this range of infestation and infection levels (9-27 Striga plants per pot) yield reduction was independent of number of parasites. Earlier, Press and Stewart (1987) with sorghum and Carsky et al. (1998) with maize, reported that the response of the host plant was not significantly related to Striga plant numbers. For the resistant genotype Framida, Rodenburg et al. (2006b) observed a linear increase in yield reduction with infection level. These observations are in line with those of Sinebo and Drennan (2001), who found strong evidence that sorghum growth was dependent on the extent of Striga infection. Apart from sorghum yield, Striga reproduction is another important attribute of the sorghum-Striga relationship, with particular relevance for the medium and the long term. Weber et al. (1995), analysing farmers’ fields, reported that the number of newly produced S. hermonthica seeds on infected sorghum and maize plants showed to be largely independent of the initial Striga seed infestation level. In contrast, Webb and Smith (1996) found no evidence of a relationship between the number of emerged Striga plants and seed production per plant, indicating that seed production was largely proportional to the number of emerged Striga plants. Contrary to this, Rodenburg et al. (2006a) observed a hyperbolic relation between seedbank density and Striga reproduction. The density dependent reduction in seed production resulted mainly from intraspecific competition between aboveground Striga plants. For the highly resistant genotype N13, however, an increase in seed production per Striga plant was observed with increasing Striga infestation level. The sorghum-Striga relationship is characterized by a mutual interaction. Infection by Striga affects the host plant through several effects of which withdrawal of carbon and nutrients and a reduction in photosynthetic rate are considered the most important (e.g., Parker and Riches, 1993). These effects are known to have severe effects on sorghum yield. At the same time, the Striga-related effects also have direct implications for the quality of the sorghum plant to act as a host and through that on the growth and development of the attached parasites. To obtain a better understanding of the dynamic interaction between host plant and parasite a density experiment was conducted to establish the performance of the host plant-parasite association at a broad

Chapter 4 range of Striga seed-infestation levels.

Materials and methods

Plant material and growth conditions In a pot experiment conducted in 2003 in a greenhouse in Wageningen, The Netherlands, the response of the sensitive sorghum cultivar CK60-B and the parasitic weed S. hermonthica to different Striga seed infestation levels in the soil was examined. In a randomized complete block design with 16 replicates sorghum plants were exposed to different Striga seed densities in the soil. A mixture of coarse sand and arable soil (3:1 v/v) was used to fill the 17-L pots. Different infestation levels were created by thoroughly mixing 0 (control), 8, 16, 32, 64, 125, 250, 500, 1000 and 2000 germinable Striga seeds dm–3 of soil, through the upper 15 cm of soil. Striga seeds were collected in 1998 from sorghum host plants in an experimental field of the Institute of Rural Economy (I.E.R.), Cinzana, Mali. Seeds were stored under dark and dry conditions, in a plastic container at room temperature. Before the experiment, seed viability was determined after ten days of pre-conditioning using a synthetic germination stimulant (GR-24; 2 mg L–1; provided by the Department of Organic Chemistry, Radboud University, Nijmegen, The Netherlands). At June 12, ten days after adding the S. hermonthica seeds to the soil to allow conditioning of the seeds, three pre-germinated sorghum seeds were planted per pot. The seedlings were thinned to one per pot 4 days after emergence (DAE), while at 10 DAE each pot received a single dose N-P-K-fertilizer (12-10-18), equivalent to 50 kg N, 42 kg P and 75 kg K per hectare. Pots were watered with tap water every 2-3 days. Daytime temperature in the greenhouse ranged from 28 °C to 37 °C and mean night temperature was 24 °C. Relative humidity varied between 50 and 70%. Supplemental light was provided by high-pressure sodium lamps (400W SON-T, Agro-Philips lamps), which automatically switched on during day time when photosynthetically active radiation (PAR) outside the greenhouse dropped below 910 µmol photons m–2 s–1. Black screens were used to create a day length of 12 h throughout the entire growing period. Throughout the duration of the experiment, emergence of S. hermonthica was frequently (time interval 2-3 days) determined in those replicates that were assigned for final harvest (8 out of 16). From these observations time of first emergence, maximum number of emerged Striga plants (NSmax) and the Area under the Striga Number Progress Curve (ASNPC; Haussmann et al., 2000) was calculated. The observations were also used to determine the moment of first flower appearance of the parasite in each pot. Three destructive harvests, at 31 DAE (using 4 replicates), 45

Effects of Striga seed infestation level on the sorghum-Striga association

DAE (4 replicates), and 98-102 DAE (8 replicates) were conducted. To determine the number of underground S. hermonthica attachments, roots were separated from the soil by carefully washing over a 2-mm sieve after which the number of Striga attachments was counted and separated from the roots. During the final harvest the number of Striga flowers per pot was counted. Sorghum plants were dissected into leaf, stem, root and, if present, panicle. Striga attachments and sorghum plant organs were dried in an oven at 70 °C for at least 48 h after which their dry weight was determined.

Data analysis Data were subjected to analysis of variance (ANOVA) and, if appropriate, followed by a comparison of means using the least significant difference (L.S.D.). To meet the assumptions of the analysis of variance, data on number of Striga attachments and number of emerged Striga plants were subjected to log(x+1) transformation, where x is the original observation, whereas data on the fraction Striga biomass of the total sorghum-Striga association were subjected to arcsin-transformation. All statistical analyses were carried out by using the Genstat statistical software package (release 8.1).

Results

Sorghum yield There was a significant effect of Striga infestation level on sorghum plant and kernel dry weight (Table 1; P < 0.001). Already at the lowest infestation level (8 seeds dm–3) sorghum plant and kernel dry weight were reduced by 31 and 35%, respectively. Higher infestation levels resulted in increased reductions in sorghum plant dry weight, leading to a maximum reduction of 83% at the highest infestation level (2000 seeds dm–3). Kernel dry weight was even stronger affected, as Striga infestation level also had a significant negative effect on harvest index (P<0.001), resulting from a reduced number of kernels per plant (P<0.001) and a reduced 1000-kernel weight (P<0.001). Harvest index started to decline from an infestation level of 32 seeds dm–3 onwards, whereas significant reductions for the individual components were observed from 125 seeds dm–3 onwards. At the highest infestation levels only small panicles were formed or panicles were even completely absent, resulting in yield reductions of, or near to, 100%. In Fig. 1, a three-quadrant representation of the relationship between Striga seed infestation level, the maximum number of emerged Striga plants (NSmax) and relative kernel yield loss of the host is presented. The upper-left quadrant (I) represents the

Chapter 4 relation between Striga infestation level of the soil and the relative yield loss of the host. This relationship developed according to a rectangular hyperbola with an initial slope of 3.8% yield loss per Striga seed present per dm3 of soil and a maximum yield loss of 100%. From this quadrant it is evident that with CK60-B complete or nearly complete yield losses were already attained at relatively low Striga seed infestation levels. The two other quadrants provide further information on how the relation between Striga infestation level and relative yield loss was achieved. The lower-right quadrant

(III) represents the relation between Striga infestation level and NSmax. Also this relationship developed according to a rectangular hyperbola, with an initial slope of 0.55 emerged Striga plants per Striga seed present per dm3 of soil. Since a Striga density of 1 seed dm–3 of soil corresponded to a total of 7.6 Striga seeds per pot, it can be calculated that at the lowest densities approximately every 14 Striga seeds gave rise to one emerged Striga plant. This value gives an indication of the level of host resistance. The maximum number of emerged Striga plants was estimated to be 27.6 ± 2.2, a value obtained at the two highest infestation levels. In the upper-right quadrant (II)

Table 1. Effects of various Striga seed infestation levels on plant dry weight, kernel yield. Harvest index, number of kernels and 1000-kernel weight of sorghum cultivar CK60-B plants. Striga seed Plant dry Kernel Number of 1000- infestation level weight yield Harvest kernels kernel (seeds dm–3 of soil) (g plant–1) (g plant–1) index per plant weight (g) 0 69.3 a 22.1 a 0.68 a 916 a 24.71 a 8 47.5 b 14.2 bc 0.61 ab 569 a 24.52 a 16 49.2 b 15.5 ab 0.65 a 736 a 21.19 ab 32 42.8 bc 9.6 cd 0.43 bc 395 ab 21.02 ab 64 36.7 c 7.2 de 0.35 cd 313 ab 22.27 ab 125 26.4 d 3.7 ef 0.26 de 161 bc 16.94 bc 250 28.5 cd 3.7 fg 0.22 def 174 bc 15.46 bc 500 22.3 de 1.4 gh 0.1 ef 67 cd 9.77 cd 1000 15.6 ef 0.1 h 0.01 fg 7 de 7.17 d 2000 11.5 f 0 h 0 g 0 e -

SED 4.54 0.142 1 0.028 1 0.354 1 3.612 1 ANOVA was based on log(x+1) transformations; SED values of transformed data are given. Values in the same column followed by a different letter are significantly (P<0.05) different.

Effects of Striga seed infestation level on the sorghum-Striga association

Kernel yield loss (%) 100

I II

2000 35

Infestation level NS max

III

Infestation level Infestation level 2000 Figure 1. Three-quadrant representation of the relations between Striga infestation level,

Striga infection level (NSmax) and relative kernel yield loss for sorghum genotype CK60-B. Dotted lines indicate fitted models using non-linear regression on a rectangular hyperbolic function (quadrant I and III) and linear regression (quadrant II).

the relation between NSmax and relative yield loss is presented. This relationship developed according to a straight line until at higher infection levels a nearly complete yield loss was attained. The estimated slope of the increase in yield reduction, which can be interpreted as a measure of host tolerance, was about 4.3% yield reduction per emerged Striga plant.

Striga infection The destructive harvests conducted at 31 and 45 DAE provided information on Striga infection. At both dates, the number of Striga attachments increased linearly with Striga infestation level, until a density of 1000 seeds dm–3 (Fig. 2). No further increase –3 in number of attachments was observed between densities of 1000 and 2000 seeds dm .

Chapter 4

Number of Striga attachments 35

5 0 1 10 100 1000 10000

Number of Striga seeds dm–3 of soil

Figure 2. Number of Striga attachments per sorghum plant as a function of the number of Striga seeds dm–3 of soil. Observations were made at 31 days after sorghum emergence (DAE) (open circles), at 45 DAE (closed diamonds) and at 102 DAE (open triangles). Sorghum cultivar CK60-B was used as a host.

Average number of attachments per sorghum plant was 0.014 and 0.028 attachments per seed present dm3 at 31 and 45 DAE, respectively. Meaning that at 31 DAE about 0.2% of the Striga seeds had realized a successful attachment, whereas 14 days later this number had increased to 0.4%. At 102 DAE, considerable increases in number of attachments were observed for Striga infestation levels up to 64 seeds dm–3. For densities from 125 to 500 seeds dm–3 small increases were observed, whereas for the highest two infestation levels the number attachments had reduced. At this final harvest date, the number of attachments and the number of emerged Striga plants were nearly identical, meaning that all attachments had emerged, or that those attachments that had not managed to emerge had died and disappeared. In Table 2, characteristics of emerged Striga plants are presented. Striga infestation level had a significant effect on both the time of first Striga emergence (P<0.001) and the maximum number of emerged Striga plants (P<0.001). The average time of first emergence ranged from about 31 DAE for the highest infestation levels to around 49 DAE for the lowest infestation levels. Apart from this difference in average time of first emergence there was also a considerable difference in the variation in this characteristic among individual pots (Fig. 3a). At high densities the difference in first emergence between the first and the last pot was about two weeks, whereas at the lower densities this time span was about five weeks. For reasons of clarity only five out of nine infestation levels are presented in Fig. 3, but the response of the other

Effects of Striga seed infestation level on the sorghum-Striga association

Striga of Total number Total number flowers per pot

seed infestation levels ) –1

Striga Striga

(g pot dry weightdry Total 3.50 0.228

2 1

0.081

2 Striga <0.05) different. P

Final number number Final pot plants per ASNPC of emerged

0.088 Striga

D values of transformed data are given.

cultivar following CK60-B sorghum various for infection

pot plants per of emerged Maximum number

Striga (DAE) Striga hermonthica emergence First

Number Progress Curve.

Striga

of soil)

–3

seed

Area under the was based on log(x) transformations;ANOVA SE of the soil. of Striga infestation level (seeds dm Table 2. Characteristics of 8 16 32 64 125 250 500 46 a 1000 49 a 2000 53 a 42 b 34 bc 36 bc 6.9 f 8.6 ef 33 c 11.5 de 30 c 18.1 bc 31 c 13.4 cd 21.0 ab 22.1 ab 6.8 e 29.0 a 8.4 de 10.8 d 30.3 a 16.1 bc 12.0 cd 203 f 287 ef 16.7 abc 306 e 585 cd 17.6 abc 529 d 816 bc 10.5 b 24.5 a 11.0 b 11.2 b 19.4 a 20.0 ab 862 ab 24.4 a 20.6 a 1223 a 276 bc 292 bc 1135 ab 23.6 a 127 c 569 ab 1178 a 977 a 20.7 a 24.3 a 692 ab 869 a 973 a SED 4.2 0.090 ( significantly are letter different a by followed column same the in Values 1 2

Chapter 4

Number of pots with emerged Striga plants a 9

8 7 6 5 4 3 2 1 0 20 30 40 50 60 70 80 90 100

Number of emerged Striga plants per pot b 35

30 25 20 15

0 20 30 40 50 60 70 80 90 100 Days after sorghum emergence

Figure 3. Time course (days after sorghum emergence; DAE) of the number of pots (maximum 8) with emerged Striga plants (a) and of the number of emerged Striga plants per pot (b). Responses of five Striga infestation levels are presented: 8 (open diamonds), 32 (closed triangles), 125 (open squares), 500 (closed circles) and 2000 (open triangles) Striga seeds dm–3 of soil. Sorghum cultivar CK60-B was used as a host.

infestation levels confirmed the observed trends. The maximum number of emerged Striga plants increased with higher infestation levels and ranged from 6.9 at the lowest seed density to 30.3 at the highest seed density. From Table 2, it is evident that at the end of the growing period, particularly at higher infestation levels, the number of emerged Striga plants decreased. Still a significant effect of Striga density on the final number of emerged Striga plants was observed (P<0.001), with the highest number

Effects of Striga seed infestation level on the sorghum-Striga association obtained at the highest seed densities. In Fig. 3b, the time course of the number of emerged Striga plants is presented. This figure illustrates that at higher infestation levels Striga plants emerged earlier, their number increased faster and mortality of some of the Striga plants occurred during the last part of the growing period. The area under these curves is referred to as the ASNPC, a parameter that integrates time and level of Striga infection. Striga infestation level had a significant effect on ASNPC (P<0.001), with values ranging from 200 Striga-days at low infestation levels to around 1200 Striga-days for the highest infestation levels (Table 2).

Striga growth and reproduction Striga infestation level had a significant effect on total Striga dry weight (P<0.001) and the number of Striga flowers at final harvest (P<0.001). For the lowest three infestation levels total Striga dry weight was around 11 g per host plant and significantly lower than that of the other infestation levels (Table 2). For infestation levels of 64 seeds dm–3 and higher total Striga dry weight did not differ significantly and was around 22 g per host plant. For these highest infestation levels also more or less identical numbers of flowers (around 875 Striga flowers per host plant) were observed. Also for the lowest three infestation levels the number of flowers (around 230 flowers per host plant) did not differ significantly.

23.3 5.1

2000

soil of 1000 3 – 500 250

dm seeds 125 64 Striga 32 16 8 Number of of Number

0 20406080

Days after sorghum emergence Figure 4. Time until first Striga emergence (shaded bars) and first Striga flower initiation (open bars) for the various Striga seed infestation levels (8-2000 Striga seeds dm–3 of soil). Sorghum cultivar CK60-B was used as a host.

Chapter 4

A clear synchronization of flowering time of Striga plants was observed. Even though the average time of first Striga emergence differed markedly between Striga infestation levels (a range of 23 days; Table 2), onset of first Striga flowering occurred in a time span of only 5 days, between 69 and 74 DAE (Fig. 4). Consequently, no significant effect of Striga infestation level on first Striga flowering was observed.

Sorghum - Striga interaction In Fig. 5, the relationship between relative loss in sorghum plant dry weight and Striga dry weight is presented. Both characteristics largely follow a similar pattern in relation to infestation level, meaning that a fixed ratio exist between the relative loss in sorghum plant dry weight and the produced amount of Striga dry weight. On average, each gram of parasite dry weight corresponded to a 3% loss in host plant dry weight. A good correspondence was also observed between the relative loss in sorghum kernel dry weight and Striga dry weight, characterized by a 4% loss in kernel dry weight for each gram of Striga dry weight. In Fig. 6, the absolute amounts of sorghum and Striga plant dry weight are presented. Infestation level had a significant effect on the total dry weight of the host- parasite association (P<0.001). Total biomass decreased with an increase in Striga

90 30

) 75 25 ) 60 20

45 15 dry weight (g dry weight

30 10 Striga

Dry weight loss sorghum (% sorghum loss Dry weight 15 5

0 0 0 8 16 32 64 125 250 500 1000 2000

Number of Striga seeds dm–3 of soil

Figure 5. Relative loss in sorghum plant dry weight (shaded bars) and Striga dry weight per sorghum host plant (dotted line) as a function of Striga infestation level (Striga seeds dm–3 of soil). Sorghum cultivar CK60-B was used as a host. Error bars represent the standard error of the mean.

Effects of Striga seed infestation level on the sorghum-Striga association

fraction Striga 0.20 0.18 0.21 0.35 0.48 0.39 0.52 0.54 0.69 ef f e d bcd cd bc b a 80 a b 60 b bcd bc cd c d

40 e e

Dry weight (g) 20

0 0 8 16 32 64 125 250 500 1000 2000

Number of Striga seeds dm–3 of soil

Figure 6. Sorghum dry weight (shaded bars) and Striga dry weight (open bars) as a function of the Striga infestation level. Bars indicated by a different letter are significantly (P<0.05) different. The dotted line represents the average dry weight of non-infected sorghum plants. On top of the figure the fractions Striga dry weight of the total dry weight of the sorghum - Striga association are presented. Values followed by different letters are significantly (P<0.05) different. ANOVA of Striga fractions is based on arcsin-transformations. Sorghum cultivar CK60-B was used as a host.

infestation level, indicating that Striga affected its host more than just through a reallocation of biomass from host to parasite. The fraction Striga of the total biomass produced by the host-parasite association was significantly affected by Striga infestation level (P<0.001). Whereas at low infestation levels 20% of the total biomass was captured by the parasite, this percentage increased considerably at higher infestation levels and exceeded 50% at the highest infestation levels.

Discussion

Striga infestation level In the current experiment a broad range of Striga infestation levels, ranging from 8- 2000 seeds per dm3 of soil, was used. As seeds were mixed through the upper 15 cm of the soil, these densities correspond to densities of around 1200-300,000 seeds per m2.

Chapter 4

Average Striga seed densities of infested fields reported by Weber et al. (1995) and Smith and Webb (1996) ranged from 20,000-40,000 seeds m–2 at 0-15 cm depth, though for individual fields a much wider range was reported (0-200,000 seeds m–2 at 0-15 cm depth). Our range thus covers a realistic range of seedbank densities, though the highest level seems relatively high for natural circumstances.

Sorghum yield Striga infection was found to have a severe effect on dry matter production of its host. Reductions in kernel dry weight of sorghum following Striga infection could be attributed to three processes. First, as is obvious from Fig. 6, the total biomass production of the host plant-parasite association was considerably reduced compared to the dry matter production of a non-infected sorghum plant. This might be attributed to the so-called pathological effect of Striga, of which a reduced photosynthetic rate is often mentioned as the most important mechanism (e.g., Press et al., 1987; Gurney et al., 1995; Watling and Press, 2001). Photosynthesis measurements conducted in this experiment and described in Chapter 3 demonstrated significant reductions in leaf photosynthetic rate of up to 40-50%. Reductions in total biomass production are commonly observed for host plant-parasite associations where S. hermonthica is involved, even though the parasite is known to be a hemi-parasite with photosynthetic activity during the aerial part of its life-cycle. In contrast, host plant-parasite associations involving holo-parasitic Orobanche species often result in total biomass productions identical or nearly identical to that of non-infected host-plants, indicating that in these cases biomass losses of the host plant merely result from allocation of host plant resources to the parasite (ter Borg, 1986; Hibberd et al., 1999; Labrousse et al., 2001). Also with Striga this mechanism is operative. Particularly during its subterranean phase, the parasite completely relies on the withdrawal of resources from its host. An accurate estimate of the contribution of this second mechanism to the sorghum kernel yield reduction is difficult to make, as the Striga dry weight presented in Fig. 6 also partly results from the photosynthetic activity of the parasite. Graves et al. (1989) for instance estimated that at the end of the growing season host-derived carbon accounted for around 40% of the total parasite carbon. Next to the pathological effect and the withdrawal of assimilates and nutrients by the parasite, a reduced harvest index contributed to the reduction in kernel dry weight of Striga infected plants. The lowered harvest index was associated with an altered allocation of assimilates within the host plant, leading to a fewer number of kernels per panicle and a reduced 1000-kernel weight and at the highest infestation levels even to the complete absence of emerged panicles. Yield reduction of sorghum increased with Striga infestation level, resulting in

Effects of Striga seed infestation level on the sorghum-Striga association complete yield loss at the highest infestation levels. This increase in yield reduction with increasing infestation levels was associated with two phenomena: an earlier moment of first Striga emergence and a higher maximum number of emerged Striga plants. The importance of infection time for the consequences of biotic disturbances was previously demonstrated for various fungal leaf pathogens (e.g., Lim and Gaunt, 1986; van Oijen, 1991b; Bastiaans, 1993b). Cechin and Press (1993b) and Gurney et al. (1999) demonstrated, with experiments in which time of attachment was artificially delayed, that also for the S. hermonthica-host plant system, time of infection was an important determining factor of the level of yield reduction. The photosynthesis measurements described in Chapter 3 revealed that Striga infection level was not an important determinant of the size of the reduction in leaf photosynthetic rate. Already with few attachments the maximum reduction in leaf photosynthetic rate was obtained. This implies that, at least for the yield reduction resulting from this damage mechanism, time of infection is more important than infection level. This observation might also explain why kernel yield was more closely correlated to time of infection

(r = 0.753; P<0.01) than to infection level (NSmax; r = –0.540; P<0.01). However, as both factors were confounded (r = –0.363; P<0.01), it remains difficult to assess the relative importance of both factors. The relation between soil infestation level and the reduction in sorghum kernel dry weight developed according to a rectangular hyperbola. Already at the lowest Striga infestation level sorghum plant and kernel dry weight were severely reduced, confirming the lack of resistance and tolerance of cultivar CK60-B (e.g., Olivier et al., 1991b; van Ast et al., 2000; Rodenburg et al., 2005). From soil infestation levels of 250 seeds dm–3 onwards, yield reduction exceeded 90% and at the highest infestation levels kernel yield loss was complete. In Fig. 1, the relation between infestation level and kernel yield loss was dissected into two separate components: the relation between infestation level and NSmax and the relation between NSmax and kernel yield loss. The first component relates to host plant resistance, whereas the second component reflects host tolerance. Ideally, NSmax should be replaced by the total number of attachments, but for practical reasons the aboveground number of Striga plants is often used to express the infection level.

The relation between soil infestation level and NSmax also developed according to a rectangular hyperbola. The absence of a proportional relation indicates that density- dependence occurred during the subterranean phase. Density dependence during the subterranean phase was earlier reported by Smith and Webb (1996) and Rodenburg et al. (2006a, b). The latter even found a reduction in NSmax when seed densities over 2000 seeds dm–3 soil were used (Rodenburg et al., 2006b). Generally, the less than proportional increase in number of emerged Striga plants is ascribed to intra-specific

Chapter 4 competition between attached Striga plants. At the highest infestation level in this experiment, density dependence was however already evident during an earlier developmental stage, expressing itself in a less than proportional increase in number of attachments. Both at 31 and 45 DAE, the number of Striga attachments obtained at a density of 2000 seeds dm–3 was identical to the number of attachments obtained at a density of 1000 seeds dm–3, whereas at lower densities a nearly linear increase was observed.

A proportional relation was found between NSmax and kernel yield loss, until at the highest infection levels a nearly complete yield loss was attained. The estimated 4.3% increase in yield reduction per emerged Striga plant was very close to the 4% increase reported for the more resistant cultivar Framida (Rodenburg et al., 2006b). These results suggest that both cultivars mainly differ in level of resistance, whereas the consequences of Striga infection for grain production, expressing their level of tolerance, are largely identical.

Striga growth and reproduction Apart from the consequences of infestation level for the host plant, the consequences for the parasite were investigated. To some extent these consequences have already been discussed. A higher infestation level resulted in an advanced first Striga emergence, and a higher maximum number of emerged Striga plants. Because of intra- specific competition during the subterranean phase there was no proportional increase in maximum number of emerged Striga plants with Striga infestation level. At the higher infestation levels, intra-specific competition was also observed during the aerial phase of the parasite and reflected in a decrease in the number of emerged Striga plants during the final part of the growing period of the host plant. This confirms that density dependence is able to manifest itself in more than one stage of the Striga lifecycle (Rodenburg et al., 2006a). The increase in Striga dry weight with infestation level largely followed the same pattern as the loss in plant and kernel dry weight of the sorghum host plant (Fig. 5). On average, the production of each gram of Striga dry weight corresponded to losses of 3 and 4% for sorghum plant and kernel dry weight, respectively. This indicates that despite the reduction in host plant vigour obtained at higher Striga infestation levels, the parasite did not suffer and was able to increase (moderate infestation levels) or at least maintain (highest infestation levels) its dry matter production. Consequently, the fraction Striga of the total dry matter production of the host plant-parasite association steadily increased with infestation level, reaching a value of nearly 70% at the highest infestation level (Fig. 6). Again it is likely that both an increased withdrawal of host- produced assimilates, as well as an increased autotrophic production of assimilates of

Effects of Striga seed infestation level on the sorghum-Striga association the parasite, are responsible for this. Both increases are facilitated by the earlier moment of first Striga emergence and the higher number of emerged Striga plants that were obtained at higher infestation levels. Striga dry weight gave a good indication of the number of Striga flowers. This is in line with observations of Rodenburg et al. (2006a), who observed a good correlation between Striga dry weight and number of seed capsules. To some extent this will be due to the synchronization in flowering time of the parasite. For a cross-pollinating species such as S. hermonthica, synchronization of flowering seems an essential trait, to guarantee a successful reproduction.

Outlook The current research showed that each emerged Striga plant represented a yield reduction of about 4%, indicating that Striga control strategies that are able to reduce the number of Striga infections will have a significant effect on sorghum yield. However, the saturation type of response between infestation and infection level that was observed indicates that, particularly at higher infestation levels, extremely large reductions in infestation level are required before significant reductions in number of infections will be obtained. A similar saturation type of response was found between infestation level and Striga seed production. Both observations imply that every effort taken to gradually reduce the Striga infestation level will only become effective and paying when pursued for a longer time period. Such a long time lag before any return on investment is however unacceptable for small-scale subsistence farmers in Africa, particularly in view of the crucial role of cereal food crops in the food security of the farm household and the limited financial means of these farmers. Studying the response of the host plant-parasite association in dependence of Striga infestation level indicated that the influence of seed density on sorghum yield and Striga seed production was largely governed through two factors: the time of first Striga emergence and the number of Striga attachments. Though in this density study time of infection and infection level were confounded, some indications, as earlier discussed, suggest that time is more important than number. The importance of infection time was earlier identified in studies in which infection time was artificially delayed (e.g., Cechin and Press, 1993b; Gurney et al., 1999), though, unfortunately, the response on sorghum kernel yield and the effect on the growth of the parasite were not determined. These findings imply that a delayed Striga infection time might offer great potential as a starting point for strategies to reduce the negative impact of Striga on cereal production. Establishing the consequences of delaying the time of infection for sorghum yield and Striga seed production seems therefore a logical next step.

CHAPTER 5

The role of infection time in the differential response of sorghum cultivars to Striga hermonthica infection1

A. van Ast, L. Bastiaans

Crop and Weed Ecology Group, Department of Plant Sciences, Wageningen University, P.O. Box 430, 6700 AK Wageningen, The Netherlands

Abstract Striga hermonthica is an important parasitic weed that severely reduces yields of sorghum in sub-Saharan Africa. Pot experiments with the sensitive sorghum cultivar CK60-B and the tolerant Tiémarifing were conducted in 1999 and 2000 to investigate the role of infection time on the interaction between sorghum and Striga hermonthica. Timing of Striga-inoculation was used to establish delays of one and two weeks in first attachment of the parasite. In 1999, early Striga inoculation resulted in a relatively early first Striga-attachment on CK60-B. Though first infection of Tiémarifing occurred one week later, an identical final number of emerged Striga plants was observed. Plants of CK60-B were more severely affected and supported a higher total Striga biomass. Only with this cultivar the interaction between host and parasite was significantly affected by delayed infection. Parasite biomass was most sensitive and already significantly reduced following a one-week delay in infection time. With a further week delay, an additional reduction in parasite biomass was accompanied by a strong and significant increase in total and panicle dry weight of the host plant. In 2000, first infection of CK60-B was relatively late and occurred simultaneously with first infection of Tiémarifing and no significant effect of delayed infection on Striga biomass or host plant performance was observed. The results indicate that the influence of delayed infection strongly depends on actual infection time and confirm that earlier observed differences in time of first infection between the two cultivars do contribute to the more tolerant response of Tiémarifing to Striga-infection.

Keywords: Striga hermonthica, Sorghum bicolor, delayed infection, infection time, tolerance.

1 Published in Weed Research (2006), 46, (in press) 71

Chapter 5

Introduction The parasitic weed Striga hermonthica (Del.) Benth. is a major constraint to cereal production in semi-arid, sub-Saharan Africa (Parker and Riches, 1993). While S. hermonthica is a hemi-parasite, it is obligate in its relationship as it cannot establish and develop independently of its host. Initiation of the host-parasite relationship takes place belowground where root exudates released by the host plant stimulate the germination of S. hermonthica seeds. The seeds then produce a radicle and ultimately a haustorium to penetrate the root of their host. S. hermonthica withdraws mineral nutrients and water and competes directly for carbon fixed by the host. It, thus, represents an additional sink for host photosynthates. Following attachment, the parasite also causes non-sink-related effects on its host, such as reductions in leaf photosynthetic rate and changes in plant architecture (Watling and Press, 2001). Host plants parasitized by S. hermonthica also display a range of leaf symptoms including chlorosis, necrotic lesions and wilting (Parker and Riches, 1993; Gurney et al., 1995). The extent of yield reduction caused by the root parasite is highly variable, depending on factors such as host plant genotype, parasite infestation level and environmental conditions. Most effects occur almost immediately following infection (Press et al., 1996; van Ast et al., 2000, see also Chapter 2), indicating that shortly after attachment the parasite interferes with host metabolism, ultimately resulting in yield reduction. Small differences in the onset of such physiological disturbances might have a considerable impact on the ultimate level of yield reduction, particularly if infection occurs during the initial growth stages of the host. Consequently, it is not surprising that for the S. hermonthica-host plant system, time of infection was identified as an important factor in determining levels of yield reduction (Cechin and Press, 1993b; Gurney et al., 1999). Tolerance refers to the innate capacity of host genotypes to suffer less damage than standard varieties, even though they are parasitized to the same extent (Parker and Riches, 1993). Differences in time of infection have been suggested as one of the possible causes behind these genotypic differences in response to parasite infection. Gurney et al. (1999) made a comparison between two cultivars of the cereal Sorghum bicolor (L.) Moench cultivar CSH-1 and the tolerant landrace Ochuti. They concluded that the time of parasite attachment might explain much of the variation in host tolerance. In the study of van Ast et al. (2000, see also Chapter 2), a comparison was made between two other sorghum cultivars; the tolerant landrace Tiémarifing and the sensitive cultivar CK60-B. Infection of Tiémarifing with S. hermonthica occurred one week later than infection of CK60-B, and this difference in time of infection was suggested to contribute to the more tolerant response of this cultivar to Striga infection.

The role of infection time in the response of sorghum cultivars

The aim of the current research was to further resolve the role of time of infection on the subsequent interaction between host and parasite. Because of the mutual interaction between host and parasite and the involvement of feedback mechanisms, time of infection is likely to have a major effect on the host-parasite relationship. A better insight into the dynamics of this interaction is expected to result in a better appreciation of the consequences of time of infection for host plant performance and parasite reproduction. For that reason, regular observations on parasite infection level and the production and allocation of host plant biomass were conducted in experiments in which the time of infection of S. hermonthica was artificially delayed.

Material and methods

Plant material and growth conditions Three experiments were conducted in a glasshouse at Wageningen University, The Netherlands, to examine the influence of time of infection on the sorghum-S. hermonthica interaction. In each experiment, the highly-sensitive, inbred sorghum cultivar CK60-B and the tolerant, local Malian sorghum landrace Tiémarifing were used. Seeds of both cultivars were provided by ICRISAT, Samanko, Mali. S. hermonthica seeds were collected in 1998 from sorghum host plants in experimental fields at the Institute of Rural Economy, Cinzana, Mali. Seeds were stored under dark and dry conditions, in a plastic container at room temperature. Before each experiment, seed viability was determined after ten days of preconditioning using a synthetic germination stimulant (GR-24; 2 mg L–1; provided by the Department of Organic Chemistry, Radboud University, Nijmegen, The Netherlands). Experiments were conducted between May and September in 1999, 2000 and 2002. Daytime temperature in the glasshouse ranged from 25 °C to 35 °C. Night temperature was kept at 20 °C in both the 1999 and the 2000 experiments and at 16 °C in the 2002 experiment. Black screens were used to create a day length of 12 h throughout the entire growing period. Relative humidity varied between 40% (day) and 85% (night). Plants were grown in 12 L pots, containing a mixture (3:1) of sand and arable sandy soil, collected from the top layer (0-0.25 m) of an arable field near Wageningen. Ten days after sorghum emergence, each pot received 3.13 g N-P-K-fertilizer (12-10-18), equivalent to field rates of 60 kg N ha–1, 50 kg P ha–1 and 90 kg K ha–1, respectively. Pots were watered every 2 days or as needed. After a pre-germination period of 24 h, three sorghum seeds were planted in the centre of each pot. After emergence, seedlings were thinned to one plant per pot. Before sowing, four plastic coffee cups (height 8.5 cm; volume 200 cm3; perforated bottom) containing the 3:1 mixture of sand and arable soil were buried at fixed

Chapter 5 positions around the centre of each pot with the top of the cups flush with the surrounding soil. Approximately 1250 germinable Striga seeds were thoroughly mixed with the soil in the cups. These seeds were given a conditioning period of ten days, after which the cups were removed and the mixture of soil and conditioned Striga seed was used to refill the holes. From then on the roots of the sorghum plants were able to exploit the Striga-infested soil. To create differences in infection time, Striga seeds were mixed with the soil in the cups at four different moments in time. The four inoculation times (time of removal of the cups) were set at 0, 7, 14 and 21 days after sorghum sowing (DAS). To study the eventual effects of the presence of coffee cups as such, control pots for each inoculation time, containing four cups with Striga-free soil, were included in the 1999-experiment.

1999 experiment A split-plot design with 4 replications was used, in which sorghum cultivar was the main plot factor (2 levels; CK60-B, Tiémarifing). Within each main plot, eight treatments consisting of combinations of infestation level (with and without S. hermonthica) and inoculation time (0, 7, 14 and 21 DAS) were present. Treatments were randomly assigned over the pots. Blocking was undertaken to account for possible variation in the glasshouse. Within each block, each treatment was represented by seven pots to allow periodic sampling. Each block was surrounded by border rows, consisting of additional pots with non-infected sorghum plants. For each pot, the number of emerged S. hermonthica individuals was counted every 2-3 days. Seven destructive harvests, at 23, 30, 37, 44, 51, 58 and 84 DAS were conducted to determine the number of Striga attachments. For this purpose roots were separated from the soil by careful washing over a 2 mm sieve after which S. hermonthica attachments were counted. At 30, 44, 58 and 84 DAS, sorghum plants were dissected into leaf, stem, root and inflorescence. Plant material was dried in an oven at 70 °C for at least 48 h for the determination of plant organ dry weight. Dry weight of emerged Striga plants was determined at final harvest (84 DAS).

2000 experiment The experimental design was basically identical to the experiment conducted in 1999. However, as the results of the first experiment revealed that the presence of coffee cups throughout the first three weeks after sowing did not have an effect on growth and development of sorghum, just one Striga-free control treatment (without cups) was included. This resulted in a split-plot design with 4 replications, with sorghum cultivar as the main plot factor (2 levels; CK60-B and Tiémarifing) and inoculation time as sub-plot factor. The latter consisted of five treatments; a Striga-free control and four

The role of infection time in the response of sorghum cultivars inoculation times at 0, 7, 14 and 21 DAS. In each block, each treatment was represented by seven pots, to allow destructive harvests for the determination of the number of Striga attachments at 23, 30, 37, 44, 51, 58 and 71 DAS. At 23, 37, 51 and 71 DAS, dry weight of sorghum plants was determined after dissection of the plants into leaf, stem, root and inflorescence. Dry weight of emerged Striga plants was determined at final harvest (71 DAS).

2002 experiments In 2002, two additional experiments were conducted to address questions that arose during analysis of the first two experiments. In the first experiment (2002-A), the influence of an aggregated infestation pattern was evaluated, as use of the coffee-cup method brought about an infestation of just over 15% of the upper soil layer. Treatments consisted of (a) no Striga infestation (control), (b) an aggregated infestation by using the “coffee-cup” method and (c) a homogeneous infestation whereby seeds were mixed through the entire upper soil layer (8.5 cm). In the infested treatments (b and c) a total of 5000 germinable S. hermonthica seeds per pot were used. These seeds were preconditioned for 10 days, after which three pre-germinated sorghum seeds were sown in the centre of each pot. Just prior to sowing, the coffee cups in treatment b were removed. In the second experiment (2002-B), the influence of the length of the preconditioning period was evaluated. In the earlier experiments this period was fixed at 10 days, but under field conditions, a successful postponement of the contact between Striga seeds and sorghum roots would automatically result in a prolonged preconditioning period of Striga seeds. In this experiment all three treatments contained four coffee cups, which were removed three weeks after sowing of sorghum. Treatments consisted of (a) a Striga-free control and two Striga infested treatments (5000 germinable seeds per pot), of which the seeds were either (b) preconditioned for 10 days, similar to the earlier experiments, or (c) preconditioned for 31 days, to represent prolonged preconditioning. Both experiments were laid out as a split-plot design with 5 replications, in which sorghum cultivar (2 levels; CK60-B, Tiémarifing) was the main plot factor and the sub-plot factor consisted of the described treatments. Treatments were randomly assigned over the pots within a block. In both experiments, the number of emerged S. hermonthica plants was counted every 2-3 days. At maturity, total plant dry weight and kernel yield of sorghum was determined.

Data analysis Data were subjected to analysis of variance (ANOVA) and, if appropriate, followed by

Chapter 5 a comparison of means using the least significant difference (L.S.D.). To meet the assumptions of the analysis of variance, data on number of Striga attachments and number of emerged Striga plants were subjected to log(x+1) transformations, where x is the original observation. For the same reason, the data on allocation of dry matter expressed as leaf, stem, root and panicle fraction of total plant dry weight were subjected to square root transformations. All statistical analyses were carried out by using the Genstat statistical software package, release 7.1.

Results Both in 1999 and 2000, Striga inoculations at 0 and 7 DAS yielded similar results. The first Striga attachments were detected at the same observation date and also at the following observation dates no significant differences in number of Striga attachments between both treatments were observed. Not surprisingly, the response of the sorghum host plant to Striga infection did not differ between the two inoculation times. Since sorghum seedlings only emerged at four days after sowing, it is not unlikely that with inoculation at 0 DAS root growth of sorghum through the Striga-infested soil columns during the first week after sowing was either marginal or completely absent. As a consequence, there was hardly a difference between both treatments in the moment that roots of the host plant started to exploit the Striga-infested soil. For this reason, the results of the first two inoculation treatments were combined and are further referred to as inoculation at 7 DAS.

Striga attachment Between 23 and 58 DAS, sorghum plants were uprooted weekly to determine the number of Striga attachments on the roots of the sorghum host. In 1999, Striga attachments on CK60-B were first observed at 23, 30 and 37 DAS for inoculations at 7, 14 and 21 DAS, respectively (Fig. 1). This means that for all treatments the first Striga attachments were observed at 16 days after inoculation. Also for Tiémarifing no effect of inoculation time on the time period between Striga inoculation and first attachment was found, but with this cultivar first Striga attachments were observed at 23 days after inoculation, one week later than on CK60-B. In 2000, first Striga attachments on both cultivars were observed at 23 days after inoculation, except on plants of Tiémarifing inoculated at 14 DAS where the first Striga attachments were detected at 30 days after inoculation (44 DAS). These results indicate that in nearly all situations a delay in inoculation did result in an identical delay in first Striga attachment. At final harvest, no significant effect of inoculation time on final number of attachments was observed, indicating that in the later inoculation treatments a faster

The role of infection time in the response of sorghum cultivars

b d 21 DAS 21

21 DAS

84 58 51 44 3037 23 23 30 37 44 51 58 71 58 51 44 30 37 23 4 Days after sorghum sowing sorghum after Days

sorghumDays after sowing achments per host plant) as

14 DAS 14 DAS

h observation time below-ground 23 23 30 37 44 51 58 8

71 58 51 44 37 30 23 attachments

attachments

Striga Striga 7 DAS

7 DAS

Number of Number of 23 3023 37 44 51 58 84 attachments (number of att Number of of Number 23 30 37 44 51 58 71 0 0 80 60 40 20 80 60 40 20 100

100

resent the standard error of mean. p 2000). Arrows indicate at whic S. hermonthica

a c 21 DAS21 cultivar CK60-B (a. 1999; c. sorghum ). were made on the sensitive Observations

21 DAS 23 30 37 44 51 58 84 23 30 37 44 51 58 71

Days after sorghum sowing sorghum after Days Days after sorghum sowing

14 DAS re Error bars observed. were first 14 DAS 14 23 30 37 44 51 58 84 71 44 51 37 58 30 23

attachments attachments

Striga Striga

S. hermonthica 7 DAS 7

DAS 7 Number of 23 30 37 44 51 58 84 58 51 44 37 23 30 Number of of Number 71 58 51 44 37 30 23 0 0 80 60 40 20 80 60 40 20 100

100 Figure 1. Time course (days after sorghum sowing; DAS) of 2000) and on the tolerant cultivar Tiémarifing (b. 1999; d. 2000) and on the tolerant cultivar Tiémarifing affected by inoculation time (7, 14 and 21 DAS of attachments

Chapter 5 increase in number of attachments compensated the shorter time span available for establishing infections. Significant differences in number of Striga attachments between the two cultivars were also absent at final harvest. In 1999, this was partly due to a decrease in number of attachments on CK60-B between the last and the second-to-the last observation date.

Striga emergence The earliest emergence of Striga was obtained with the first inoculated plants (7 DAS). In 1999, the earliest Striga-emergence on CK60-B was observed at 34 DAS, which was 19 days earlier than on Tiémarifing (53 DAS). In 2000, this difference in first Striga emergence between the two cultivars was only three days, as first emergence on CK60-B did not occur until 45 DAS, and the earliest emerged Striga plants on Tiémarifing were observed at 48 DAS. Inoculations that occurred one (at 14 DAS) or two (at 21 DAS) weeks later resulted in delays in first Striga emergence of on average 6 and 13 days, respectively. This indicates that the time span between inoculation and first emergence was not influenced by the time of inoculation. Both in 1999 and 2000, cultivar and time of inoculation did not have a significant effect on maximum number of emerged Striga plants per pot. Average maximum number of emerged Striga plants was 8.2 in 1999 and 9.6 in 2000.

Sorghum biomass accumulation and allocation Biomass increased with time for the uninfected control plants of CK60-B and Tiémarifing (Fig. 2). In 1999, this observation was based on four different control treatments, consisting of pots in which coffee cups with Striga-free soil were removed at 0, 7, 14 or 21 DAS. Between these control treatments no significant differences in biomass accumulation or biomass allocation of sorghum were observed and for that reason all data were combined. In 1999, total final plant dry weight of control plants of cultivar CK60-B was significantly higher than that of Tiémarifing (P<0.01), whereas in 2000 no significant difference between both cultivars was observed. A significant reduction in total plant dry weight of CK60-B due to Striga infection was first observed at 44 DAS (P<0.001). At that time, early infected plants were significantly smaller than non-infected and later infected plants (Fig. 2). Also at 58 and 84 DAS, total plant dry weight of infected plants was significantly reduced (P<0.001). At 58 DAS, plants inoculated at 14 and 21 DAS were larger than those inoculated at 7 DAS. At final harvest, the difference between inoculation at 7 and 14 DAS had disappeared, but plants inoculated at 21 DAS were to a lesser extent affected by Striga infection. In 2000, effects of Striga infection on plants of CK60-B were much smaller.

The role of infection time in the response of sorghum cultivars

d b

) for control plants plants ) for control

–1 Days after sorghum sowing after Days sorghum after Days sowing

cultivar CK60-B ((a) 1999; (c) of inoculation times. imes (7, (open triangles) 14

ght of sorghum (g plant

Sorghum (g) wt dry Sorghum (g) wt dry 0 20406080100 0 20406080 0 5 0 Line bars indicate SED

50 40 30 20 10 30 25 20 15 10

made on the sensitive sorghum seeds at different inoculation t ) of total plant dry wei

a c

S. hermonthica

arifing ((b) 1999; (d) 2000).

Days after sorghum Days after sowing sorghum Days after sowing

Sorghum wt (g) dry

Sorghum dry wt(g) 020406080 0 20406080100 5 0 Figure 2. Time course (days after sorghum sowing; DAS Figure 2. Time course (days after sorghum sowing; (closed circles) and plants inoculated with inoculated plants and circles) (closed circles) and 21 DAS (closed triangles)). Observations were 2000) and on the tolerant cultivar Tiém 0 30 25 20 15 10 50 40 30 20 10

Chapter 5

A significant effect on total plant dry weight was only observed (P<0.05) at final harvest (71 DAS). Inoculation at 7 and 14 DAS resulted in plants that were significantly smaller than the non-inoculated plants. Inoculation at 21 DAS did not result in a significant reduction in total plant dry weight compared with the control plants. For Tiémarifing in 1999, significant reductions in host plant dry weight following inoculation with S. hermonthica could only be demonstrated at final harvest (84 DAS) and only when data of all inoculation dates were combined (P<0.01). In 2000, the effect of inoculation time on total plant dry weight only became apparent at final harvest (71 DAS). Non-inoculated plants were larger than inoculated plants (P<0.01), but no effect of time of inoculation was observed.

Fraction 30 44 58 84 DAS a Fraction 30 44 58 84 DAS b 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 Fraction 23 37 51 71 DAS c Fraction 23 37 51 71 DAS d 1.0 1.0

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 c 7 14 21 Inoculation time (DAS) Figure 3. Distribution of accumulated dry matter over root (grey shading), stem (cross- hatching), leaf (stippled) and panicle (white) at different moments in time (30, 44, 58 and 84 days after sorghum sowing (DAS) in 1999; 23, 37, 51 and 71 DAS in 2000) for control plants (c), and plants inoculated with S. hermonthica seeds at different inoculation times (7, 14 and 21 DAS). Data represent observations on the sensitive sorghum cultivar CK60-B (a. 1999; c. 2000) and on the tolerant cultivar Tiémarifing (b.1999; d. 2000).

The role of infection time in the response of sorghum cultivars

Reductions in dry matter production due to Striga infection were accompanied by changes in dry matter allocation (Fig. 3). Striga infection caused reductions in stem and panicle fraction, whereas root and leaf fraction were increased. With CK60-B in 1999, the first changes were already observed at 30 DAS, when both the root fraction (P<0.01) and the stem fraction (P<0.001) were altered. Late inoculations caused milder alterations in dry matter allocation than early inoculation. The smaller effect of Striga infection on dry matter production of CK60-B in 2000, was accompanied with smaller effects on dry matter allocation, which also appeared later in the season. A significantly reduced stem fraction (P<0.01) in early inoculated sorghum plants

Dry weight (g) Dry weight (g) 40 40 a b total plant total plant

30 panicle 30 panicle

20 20

10 10

0 0 control 7 14 21 control 7 14 21 Days after sorghum sowing Days after sorghum sowing

Dry weight (g) Dry weight (g) 25 25 c d total plant total plant 20 20 panicle panicle 15 15

10 10

5 5

0 0 control 7 14 21 control 7 14 21 Days after sorghum sowing Days after sorghum sowing Figure 4. Total sorghum (shaded bars) and panicle (open bars) dry weight (g plant–1) of control and Striga-infected sorghum plants at final harvest as influenced by inoculation time (7, 14 and 21 days after sorghum sowing ; DAS) for the sensitive sorghum cultivar CK60-B ((a) 1999; (c) 2000) and the tolerant cultivar Tiémarifing ((b) 1999; (d) 2000). Error bars represent standard error of the mean. Line bars indicate SED of total plant dry weight and panicle dry weight, respectively.

Chapter 5 observed at 37 DAS was the first indication that Striga influenced the dry matter allocation of its host. At 71 DAS, stem fraction was significantly reduced (P<0.01), whereas root (P<0.001) and leaf fraction (P<0.05) were significantly increased. Panicle fraction was not significantly affected. Again the observed changes in allocation pattern were, in general, stronger in early-inoculated plants. For Tiémarifing, differences in allocation pattern following Striga inoculation were only observed in 2000. At 51 DAS, a significant increase (P<0.05) in root fraction of inoculated plants was observed. At 71 DAS, root fraction was significantly (P<0.01) increased, whereas stem fraction was significantly (P<0.05) reduced. The alterations in dry matter allocation pattern were stronger when plants were inoculated at an earlier stage. Panicle dry weights of sorghum at final harvest are presented for both cultivars (Fig. 4). A significant effect (P<0.001) of time of inoculation on panicle dry weight was observed with CK60-B in 1999. Inoculations at 7 and 14 DAS resulted in reductions in panicle dry weight that were far stronger than the reductions in total plant dry weight. However, panicle dry weight of plants inoculated at 21 DAS did not differ from that of the control. Also in 2000, inoculations at 7 and 14 DAS resulted in significantly lower panicle dry weights (P<0.05), whereas plants inoculated at 21 DAS did not differ significantly from the control. For Tiémarifing, no significant reductions in panicle dry weight were observed, irrespective of inoculation time and year of experimentation.

Biomass accumulation of Striga In 1999, total dry weight of emerged S. hermonthica plants parasitizing CK60-B decreased significantly with inoculation time (P<0.001) (Fig. 5). Inoculation at 21 DAS resulted in 26% of the S. hermonthica biomass produced with early inoculation (7 DAS). Plants of CK60-B supported greater biomass than Tiémarifing (P<0.001), for which no significant effect of inoculation time was observed. In 2000, differences between CK60-B and Tiémarifing were not significant and the effect of inoculation time on biomass accumulation of S. hermonthica was not significant for either cultivar.

2002 experiments No interactions between the specific treatment (2002-A: seed-distribution pattern; 2002-B: duration of seed conditioning) and cultivar (CK60-B and Tiémarifing) were observed in either experiment (Table 1). Consequently, results presented are a combination of both cultivars. The presence of Striga seeds in the entire upper soil layer did not cause any significant difference in Striga characteristics compared with the aggregated infestation pattern characteristic for the coffee-cup inoculation method. In all situations, Striga inoculation resulted in significantly lower total plant dry weight

The role of infection time in the response of sorghum cultivars

6 Dry weight emerged Striga shoots (g)

1999 2000 5 CK60-B Tiémarifing CK60-B Tiémarifing 4

0 7 14 21 7 14 21 7 14 21 7 14 21

Days after sorghum sowing

Figure 5. Dry weight of emerged S. hermonthica shoots (g pot–1) at final harvest for the sensitive sorghum cultivar CK60-B and the tolerant cultivar Tiémarifing in 1999 and 2000 as influenced by inoculation time (7, 14 and 21 days after sorghum sowing). Error bars represent standard error of the mean. Line bars indicate SED for comparison of any cultivar × inoculation time combination.

and panicle dry weight of sorghum (control: 56.3 and 11.8 g plant–1, respectively). Total plant dry weight of sorghum was not affected by Striga distribution pattern, but a homogeneous infestation pattern caused a significantly smaller panicle dry weight compared to the aggregated infestation pattern (P<0.05). The length of the preconditioning period of S. hermonthica seeds (either 10 or 31 days) did not have a significant effect on any of the Striga characteristics nor on dry weights (total and panicle) of sorghum.

Discussion

Methodology In this study, the effect of delayed parasitism by S. hermonthica on the host plant- parasite interaction was studied. Previously, Cechin and Press (1993b) studied the effect of delayed infection of S. hermonthica on sorghum growth. In their study, the roots of the sorghum plants were grown on moist glass fibre paper, to facilitate the manipulation of infection time. Non-destructive observations, particularly sorghum plant height and photosynthetic activity, were made. Gurney et al. (1999) described a

Chapter 5 method of delaying Striga attachment in pot-grown plants. At specific moments after planting of sorghum, the sand was washed from the base of the stem to expose the root system. Preconditioned seeds of S. hermonthica were then placed on the surface of the roots, after which the root-system was recovered with sand. This procedure might easily result in root injuries. Furthermore, Striga seeds are only applied in the vicinity of roots that are growing in the upper few centimetres of the soil, whereas under natural conditions seeds are found throughout the ploughing layer (Van Delft et al., 1997). In the current research, an alternative methodology for delaying first Striga- attachment on the roots of pot-grown sorghum plants was applied. Weekly observations on carefully washed root systems of sorghum plants demonstrated that application of this alternative methodology resulted in the establishment of a progressive delay in moment of first Striga attachment. Compared with inoculation at

Table 1. Effect of distribution of S. hermonthica seeds (inoculation at 0 DAS) and of duration of Striga seed conditioning (inoculation at 21 DAS) on first emergence time of Striga, total number of Striga attachments, final number of emerged Striga plants and on total plant dry weight panicle weight of sorghum. Presented data are averages of data obtained with sorghum cultivars CK60-B and Tiémarifing. Data on Striga attachments and emerged Striga plants were analysed after log(x+1) transformation. SED-values of transformed data are given. S. hermonthica Sorghum Total Final Number number number of of days of attach- emerged Total dry Panicle until first ments plants per weight dry weight Treatment emergence per pot pot (g plant–1) (g plant–1) Distribution seeds of Striga aggregated 56 9.8 9.4 30.7 5.15 seeds seeds randomized 51 11.5 10.0 22.8 2.82

SED 7.9 0.20 0.19 5.11 0.739 Duration of 10 days 73 9.3 8.5 38.5 6.43 Striga seed conditioning 31 days 74 12.1 10.0 37.7 7.28

SED 9.4 0.09 0.09 5.83 1.800

The role of infection time in the response of sorghum cultivars

7 DAS, the removal of the cups at 14 and 21 DAS resulted in an average postponement of first attachment with one and two weeks, respectively. Similar delays were observed for first Striga emergence. At the same time, the use of the cups facilitated the inoculation of S. hermonthica seeds to a depth of 8.5 cm, without damaging the roots. One of the consequences of the application of the described methodology is an aggregated horizontal distribution pattern of Striga seeds, as after inoculation the seeds are present in four soil columns, which together only represent 15% of the upper soil layer. The 2002-experiment demonstrated that, compared to a homogeneous distribution pattern, this aggregated distribution of Striga seeds did result in a lower reduction of sorghum kernel yield. Though it is likely that this lower reduction resulted from a delayed first contact between host roots and seeds of the parasite, no significant difference in first emergence of S. hermonthica between both seed distribution patterns was observed.

Comparison of cultivars In 1999, early Striga-inoculation resulted in clear differences between the responses of both sorghum cultivars. On the sensitive cultivar CK60-B, first Striga-infections were observed one week earlier than on the tolerant cultivar Tiémarifing (23 DAS and 30 DAS, respectively), whereas first Striga emergence occurred 19 days earlier. Total plant dry weight and panicle dry weight of CK60-B were both strongly reduced, whereas the effects on Tiémarifing were significant, but less severe. No difference in final number of emerged Striga plants was found, but Striga plant dry weight on CK60-B was significantly higher than on Tiémarifing. These differences correspond with earlier reported differences between both cultivars (van Ast et al., 2000; Rodenburg et al. (2006b). Differences in infection pattern between CK60-B and Tiémarifing following first Striga-inoculation were absent in 2000. On both cultivars, first Striga attachment was observed at 30 DAS and first Striga-emergence was observed about two weeks later. Surprisingly, reductions in dry weight accumulation for both cultivars also were nearly identical. The observation that dry matter yields of CK60-B and Tiémarifing are similar when infection occurred at the same time is consistent with the hypothesis that differences in infection time between CK60-B and Tiémarifing, as earlier reported and also observed in 1999, do contribute to differences in response to Striga infection between both cultivars. However, a clear difference in reduction in panicle dry weight was observed. For Tiémarifing, the reduction in panicle weight was proportional to the overall reduction in total plant dry weight, whereas for CK60-B panicle weight was much more strongly reduced than total biomass. The ability of Tiémarifing to maintain its harvest index under Striga-infested conditions might well be another aspect of the tolerant character of this cultivar.

Chapter 5

In most situations, reductions in dry matter production due to Striga infection were preceded by alterations in dry matter partitioning of the host plant. The most important of these changes was a reduced allocation to the stem, combined with an increased allocation to the roots. This change in allocation pattern, which was more evident with CK60-B, remained throughout the growing season, a response consistent with effects of S. hermonthica reported by others (Graves et al., 1990; Gurney et al., 1995; Frost et al., 1997).

Effects of delayed inoculation Observations on the sorghum root system demonstrated that a later time of inoculation did not result in a lower maximum number of attachments (Fig. 1). This indicates that with later inoculated plants a similar number of attachments was realized in a shorter time span. For this observation, two explanations can be given. First, with a later time of inoculation the host plants are taller at the time of inoculation and consequently the opportunities for a rapid and more complete exploitation of the infested soil are larger. Additionally, the root biomass of late inoculated sorghum plants is larger than that of early inoculated plants. This particularly holds for sensitive cultivars, such as CK60-B. The final root biomass of late inoculated plants in 1999 was for instance twice as large as that of early inoculated plants. In the 1999 and 2000 experiments, Striga seeds were preconditioned for 10 days, irrespective of inoculation time. Under real field conditions however, the establishment of a delayed contact between Striga seeds and host root will result in a proportional increase in the preconditioning period of Striga seeds. Preconditioned Striga seeds that are not exposed to germination stimulants are reported to enter a state of secondary dormancy (Vallance, 1950). Based on this observation it can be hypothesised that an additional benefit of a delayed contact will be a reduction in the number of Striga attachments, because of the development of secondary or enforced dormancy. Results of the 2002-experiment showed, however, that an extension of the preconditioning period by three weeks did not result in any significant reduction in Striga infection level. A reduction in number of attached Striga under real field conditions associated with the appearance of secondary dormancy is, therefore, highly unlikely. Also, the observation in the 1999 and 2000 experiments that the number of attachments still increased till at least 58 DAS, demonstrates that the capability of Striga seeds to germinate was maintained for a considerable time span. The aim of the current research was to delay the timing of first Striga-infection in order to investigate its effect on yield reduction in sorghum. For the tolerant cultivar Tiémarifing, delayed infection did not affect total plant dry weight or panicle yield. Also total biomass of S. hermonthica was not significantly affected by inoculation

The role of infection time in the response of sorghum cultivars time. This observation corresponds to the results of Gurney et al. (1999) with the tolerant cultivar Ochuti. Infection time did have a significant effect on total host plant dry weight for the sensitive cultivar CK60-B, but only in 1999. Delayed infection resulted in smaller reductions in panicle dry weight. Also, a significant reduction in Striga dry weight with inoculation time was observed in 1999, whereas Striga dry weight was not affected in 2000. Since first Striga infection on CK60-B in 1999 was already observed at 23 DAS, it is suggested that host plants can particularly benefit from delayed infection when first infection occurs early. Previously, such a strong sensitivity to infection time during early growth stages was demonstrated for various fungal leaf pathogens like powdery mildew (Erysiphe graminis) and stripe rust (Puccinia hordei) on spring barley (Lim and Gaunt, 1986), late blight (Phytophthora infestans) on potato (van Oijen, 1991b) and leaf blast (Magnaporthe grisea) on rice (Bastiaans, 1993b). It is obvious that the consequences of a physiological disturbance following infection will be stronger with smaller host plants. During the early stages of growth, the plants are in their exponential growth phase, characterized by a strong positive feedback between biomass production, leaf area formation and light interception (Blackman, 1919). Consequently, the effects of any disturbance in this period will be magnified. The hypothesis that the effects of delayed infection depend very much on the actual moment of first infection is visualized in Fig. 6, where the relation between first infection time and (a) relative total plant dry weight and (b) relative panicle dry weight are presented. With a very early first infection, the effect of a small delay in infection time on total plant dry weight is still insignificant. In this situation a later initial infection is compensated by a faster infection rate during later stages of growth, all resulting in nearly identical reductions in biomass and a nearly complete absence of panicle yield. A further delay in first infection is followed by a rapid improvement in the performance of the host plant. This typically represents the period in which the host is very sensitive to infection time. Any further delay in first infection only results in relatively minor increases in host performance, as the opportunities for further improvement are small. This results in an S-shaped curve for the relation between infection time and host plant performance. The effect on panicle weight is represented by a much steeper curve than that on total plant dry weight. Apart from biomass production, Striga infection also reduced the harvest index, and this effect was strongest with early infection. The results obtained with Tiémarifing fit very well in the upper right part of the S- shaped curves derived from the results obtained with CK60-B (Fig. 6). The results suggest that differences in initial infection time are at least partly responsible for the

Chapter 5

1.0 Relative plant dry weight a

0.8

0.6

0.4

0.2

0.0 0 1530456075 Day of first Striga emergence

b 1.2 Relative panicle dry weight

1.0

0.8

0.6

0.4

0.2

0.0 0 1530456075 Day of first Striga emergence

Figure 6. Relation between first emergence of S. hermonthica and relative sorghum yield (based on a. total plant dry weight and b. panicle dry weight) as observed for CK60-B (closed triangles) and Tiémarifing (open triangles) in the 1999 experiment and CK60-B (closed circles) and Tiémarifing (open circles) in the 2000 experiment. The curve illustrates the hypothesis on the sensitivity of the response of the host plant to delayed infection in dependence of infection time.

genotypic differences of sorghum cultivars in their response to infection by S. hermonthica (Gurney et al., 1999). It demonstrates that with this host plant-parasite system, it is very difficult to make an accurate distinction between the contribution of various defence mechanisms, like avoidance, tolerance and resistance to host plant performance under Striga-infested conditions (Rodenburg et al., 2005). The frequent observations on attachment level and host response improved the understanding of the

The role of infection time in the response of sorghum cultivars system, and also endorsed the dynamic characteristic of this host-parasite relationship, which is characterised by mutual interactions. Infection by S. hermonthica deregulates the physiology of sorghum, which results in a reduced growth of the plant, and through that directly interferes with the quality of the plant as a host. This has implications for the further development of the parasite, both at the individual (growth of attached parasites) and the population level (increase in number of attachments). In turn, the development of the parasite influences the extent to which the host is affected. The feedbacks that are operative in this complex relationship suggest that dynamic simulation models, which have been successfully applied with other host-pest (e.g., Bastiaans, 1993b) and crop-weed relationships (e.g., Kropff et al., 1992b) might be an appropriate tool for obtaining a better quantitative understanding of the host-parasite association. The results further indicate that in situations with a severe Striga infestation, the parasite is more sensitive to infection time than the host plant. With CK60-B, a one week delay in infection time did not affect sorghum plant dry weight or panicle yield; however, Striga dry weight was significantly reduced. This observation corresponds to the relation between grain dry weight of sorghum and total S. hermonthica dry weight presented by Gurney et al. (1999). They also observed that, when taking situations with a high parasitic load and a low grain dry weight as a reference point, any gradual decrease in parasitic pressure, through, for instance, a reduction in infection level or a later infection time, first resulted in reductions in dry weight of the parasite, whereas only more severe alleviations of parasite pressure resulted in reductions in S. hermonthica dry weight that were accompanied with improvements in sorghum grain yield.

Practical application The implication of our study is that if sorghum cultivars can be protected from parasitism through the critical initial period, significant yield increases can be expected accompanied by reductions in S. hermonthica reproduction. The results demonstrate that infection time is partly genetically determined and for that reason breeding focussed on attachment time represents a first option. Van Delft et al. (2000) demonstrated that deep planting and no-tillage resulted in an improved sorghum yield combined with a reduced number of emerged Striga plants. They hypothesized that the probable mechanism causing these results is a delay in the onset of Striga attachment. Transplanting also has been used as a means to avoid Striga infection during early stages. Oswald et al. (2001) found that transplanted maize was able to resist Striga- induced stress better than direct seeded maize. Also, Gbèhounou et al. (2004) found Striga-infection level was often reduced with transplanting of millet and sorghum

Chapter 5 under Striga-infested conditions, though not in all situations significant improvements of grain yield were found. Based on the current results, a further exploration of the possibilities for prevention of early infection through management operations seems a promising option.

CHAPTER 6

Cultural control measures to diminish sorghum yield loss and parasite success under Striga hermonthica infestation1

A. van Ast1, L. Bastiaans1, S. Katile2

1 Crop and Weed Ecology Group, Department of Plant Sciences, Wageningen University, P.O. Box 430, 6700 AK, Wageningen, The Netherlands 2 Institut d’Economie Rurale, Agricultural Research Station, Cinzana, P.O. Box 214, Ségou, Mali

Abstract Prospects of reducing Striga hermonthica (Del.) Benth. parasitism by means of cultural control measures were assessed. In a pot experiment, deep planting, the use of transplants and shallow soil tillage, strongly delayed and reduced Striga infection of a sensitive and a tolerant sorghum cultivar. Application of all three measures combined resulted in a four week delay in first emergence of the parasite, a reduced number of emerged Striga plants and a decimation of Striga dry weight. With the sensitive cultivar CK60-B a yield loss of 92% was reduced to 26%, whereas with the tolerant cultivar Tiémarifing application of the cultural control measures completely cancelled out a yield loss of 28%. Evaluation of these cultural measures under field conditions in Mali resulted in a strong reduction in Striga-infection level (85%). At the same time, the measures did not result in delayed first emergence of the parasite and had no beneficial effect on crop yield. It was argued that the presence of a natural Striga soil seedbank, with seeds throughout the tilth, might be an important reason for the reduced effectiveness of the proposed cultural control measures under field conditions.

Keywords: Striga hermonthica, Sorghum bicolor, shallow tillage, deep planting, transplanting, avoidance.

1 Published in: Crop Protection (2005) 24, 1023-1034. 91

Chapter 6

Introduction The semi-parasitic weeds of the genus Striga (Witchweed) are well-known constraints for grain production in the savannah regions of West Africa. Striga hermonthica (Del.) Benth. which parasitizes maize, sorghum, millet, teff and rice, is considered the economically most important species in the semi-arid tropics. This applies in particular to agro-ecosystems where a high human population density imposes a strong pressure on arable land. This factor has forced farmers to abandon the practice of shifting cultivation, characterized by long fallows, and to adopt a permanent cropping system with short fallows and little or no fertilizer or manure application. These changes in cropping practices have resulted in heavily Striga-infected crops and, consequently, in serious losses of crop yield (Parker and Riches, 1993; Riches and Parker, 1995). The extent to which the parasitic angiosperm S. hermonthica reduces the growth of its sorghum host is highly variable and depends on factors such as host plant genotype, parasite infestation level and environmental conditions. In a number of studies it was demonstrated that the age of the host plant when first attachment of the parasite occurs is another important factor determining the level of yield reduction (Cechin and Press, 1993; Gurney et al., 1999; van Ast and Bastiaans, 2006, see also Chapter 5). Based on the results of the latter study, it can be concluded that if a sensitive crop can be protected from parasitism throughout the first 40 days after sowing, significant yield increase can be expected, accompanied by considerable reductions in S. hermonthica reproduction. They also suggested that new studies have to be directed at developing or improving control methods that are based on the principle of a delayed onset of Striga attachment. One of the keys to prevent early attachment is the separation of Striga seeds from the root system of the host during early crop development. Information on the spatial distribution of Striga seeds in the soil showed that after seed shedding at the end of the cropping season, the majority of seeds were found on top and in the upper 5 cm of the soil profile (Eplee and Westbrooks, 1990; van Delft et al., 1997). However, subsequent ploughing and harrowing led to a more or less homogeneous distribution of the Striga seeds throughout the upper 10 - 15 cm of the soil. Through application of shallow- tillage (0-5 cm) Striga infestation might be restricted to the upper soil layer, reducing the likelihood of developing roots to become infected. Another possible way of separating Striga seeds and host roots is by planting the host below the soil layer where most of the Striga seeds are present. In a field study of van Delft et al. (2000), inverted cone-shaped plant holes to realize deep planting were used, which at the same time prevented an unwanted delay in emergence of crop seedlings. Application of this technique in tilled soil gave 74% more sorghum grain yield relative to standard sowing. Striga seed production was not reduced but when

Cultural control measures to diminish sorghum yield loss deep sowing was applied in non-tilled fields a complete suppression of Striga seed production was found. The authors hypothesized that these positive effects resulted from a delay in the onset of Striga attachment. Replacing seeds by transplants that are raised in Striga-free nurseries is yet another option for avoiding infection during the early stages of crop development. Oswald et al. (2001) found that transplanted maize was able to resist Striga-induced stress better because older plants possess a certain physiological tolerance to the phytotoxic effects exerted by the parasite. Results of Gbèhounou et al. (2004) were less consistent. In 2000, transplanting of millet compared to direct sowing showed no significant effect on S. hermonthica incidence, but transplanted sorghum yielded significantly more grain than direct seeded sorghum. In 2001, the opposite was observed. Transplanting caused a significant reduction in S. hermonthica infestation relative to direct sowing, but no significant difference in millet grain yield was found. In some areas in Africa, transplanting of cereal seedlings from irrigated nurseries is a traditional method of improving food security to compensate for a growing season that is too short for a complete crop cycle (Oswald et al., 2001). In the north of Cameroon for instance, at farmers’ fields situated on vertisols with high clay content, transplanting of sorghum has proven effective in off-season production. These soils are characterized by stickiness and high plasticity when wet. Nursery beds are established and the seedlings transplanted as the soils dehydrate at the end of the rainy season. In this way a crop stand is quickly established, allowing the crop to mature with the water that is still available (Chantereau and Nicou, 1994). Transplanting is also more widely used for gap filling after crop emergence, using seedlings from plant holes with several seedlings (Dawoud et al., 1996). The objective of the research herein reported was to find out whether shallow soil- tillage, deep planting, the use of transplants or a combination of them would lead to a delay in Striga attachment of sorghum. Moreover, the effects of these measures on sorghum yield and the reproductive success of the parasite were investigated.

Material and methods In 2002, a greenhouse and a field experiment were conducted. Both experiments focused on the effect of shallow tillage, type of planting material (seeds or transplants) and depth of planting on the ultimate infection level of S. hermonthica on sorghum (Sorghum bicolor (L.) Moench) and the time interval between planting and first parasite attachment. In the greenhouse experiment, the S. hermonthica-tolerant local Malian landrace Tiémarifing and the highly sensitive inbred cultivar CK60-B were used, whereas in the field experiment only CK60-B was grown. Seeds of both cultivars were provided by ICRISAT, Samanko, Mali. S. hermonthica seeds for both

Chapter 6 experiments were collected in 1998 from sorghum host plants in experimental fields at the Institut d’Economie Rurale (IER), Cinzana, Mali.

Greenhouse experiment The experiment was conducted in Wageningen, the Netherlands, between May and September 2002. Daytime temperature in the greenhouse ranged from 25 to 35 °C, night temperature was kept at 18 °C. To simulate short-day conditions, day length throughout the whole growing period was artificially kept at 12 h, using black screens. Relative humidity varied between 40% (day) and 85% (night). Plants were grown in 12 L pots, containing a mixture of sand and arable soil. The arable soil had a pH-KCl of 6.7, an organic matter content of 1.3% and a total N content of 0.05%. Sand and arable soil were mixed in a 3:1 ratio in order to reduce the basal level of nutrients. Ten days after sorghum emergence, each pot received 3.13 g N-P-K-fertilizer, equivalent to field rates of 60 kg N ha–1, 50 kg P ha–1 and 90 kg K ha–1, respectively. Pots were watered every 2 days. The experimental design was a split-split-split-plot in 4 blocks, with S. hermonthica infestation (+, –) as the main plot factor, cultivar as the sub-plot factor (CK60-B, Tiémarifing) and soil-tillage (standard, shallow) as the sub-sub-plot factor. The four combinations of planting material (seeds, transplants) and planting depth (shallow, deep) were randomly assigned over the pots in the sub-sub-plots. Blocking was undertaken to account for possible variation in different parts of the greenhouse. In each block, all treatments were represented by three pots. Destructive sampling was conducted twice: the first time at the onset of flowering of sorghum, using one pot and the second time at crop maturity, using two pots. This design resulted in a total of 384 pots. Each block was surrounded by border rows, consisting of additional pots with sorghum seeds planted in them. In Striga-infested pots, approximately 5000 germinable seeds of S. hermonthica (27 mg) were added on May 14. To allow conditioning of the Striga seeds, sorghum was introduced to the pots 10 days later. The distribution of the seeds of the parasite in the pots was determined by the factor ‘soil tillage’. In case of ‘tillage’, S. hermonthica seeds were mixed through the upper 10 cm of the soil, while in the ‘shallow-tillage’ treatment the seeds were mixed through the upper 2 cm of the soil (Fig. 1). Two different types of planting material were used. In the case of ‘sowing’, a pre- germinated sorghum seed was sown, while in case of ‘transplanting’ a 16 day-old sorghum plant was used. These transplants were raised under Striga-free conditions (Fig. 1). Two planting depths were used: ‘shallow’, at a depth of 2-3 cm, or ‘deep’, at a depth of 15 cm. For deep planting the holes were made with a metal cone-shaped tool after which seeds or plants were put in the bottom of each plant hole. PVC-rings (1.5

Cultural control measures to diminish sorghum yield loss cm in height and 10 cm in diameter) placed around the plant hole helped the holes to remain intact for several weeks (Fig. 1). The number of emerged S. hermonthica plants was counted with a time interval of 2-3 days. These observations were used to determine mid-emergence time, the moment at which 50% of the final number of S. hermonthica plants had emerged, and to calculate the Area under the Striga Number Progress Curve (ASNPC) according to Haussmann et al. (2000). A first sampling was taken at the onset of flowering of the infected CK60-B plants, 54 days after sowing or planting (DAS). A final sampling was taken after ripening of the kernels, at 95 DAS for CK60-B and 102 DAS for Tiémarifing. At both sampling

current practice proposed cultural control measures

soil tillage 10 - 15 cm 2 - 3 cm

x planting depth 2 - 3 cm 15 cm

planting material seed transplant

all measures combined

Striga-infested soil

Figure 1. Schematic representation of the proposed cultural control measures, particularly shallow soil-tillage, deep planting and the use of transplants, that were evaluated in the current experiments.

Chapter 6 times, number of subterranean and emerged S. hermonthica plants was counted and their dry weight recorded. If present, number of flowers and seed capsules was counted to get an estimate of the reproductive capacity of the parasite. Sorghum plants were dissected into leaf, stem, root and reproductive organs and put in an oven at 70 °C for 48 h for the determination of plant organ dry weight. At final harvest, additional observations were taken to determine the number of kernels per panicle and 1000- kernel weight.

Field experiment During the rainy season of 2002, a field experiment was conducted at the Institut d’Economie Rurale, Research Station of Cinzana (latitude 13°15’ N; longitude 5°57’ W; altitude 285 m) in Mali on a S. hermonthica-infested experimental field. Soils were ferruginous sandy loams with small, temporarily leached, depressions where concretions had been formed. The field had been planted with cowpea in the previous season to reduce Striga infestation level. On July 31, NPK fertilizer (15-15-15) was applied at a rate of 100 kg ha–1, followed by an application of urea around first flowering of sorghum (at August 26) at a rate of 50 kg ha–1. Plots were kept free of all weeds, except for S. hermonthica through hand-weeding. The experimental design was a split-split plot in 4 blocks, in which S. hermonthica infestation level (low and high) was used as the main plot factor and soil-tillage (standard, shallow) as the sub-plot factor. The four combinations of planting material (seeds, transplants) and planting depth (shallow, deep) were randomly assigned over four plots within each sub-plot. Individual plot size was 4.80 m × 4.00 m (19.2 m2). Each plot consisted of six rows, with 0.80 m spacing and within-row spacing of 0.40 m. Only the centre four rows were used for observations and harvesting. The sorghum cultivar CK60-B was grown under low (natural infestation level) and high S. hermonthica infestation. To create the latter, 0.45 g of S. hermonthica seeds, representing a seed number of around 100,000 were applied per m2. Before application, S. hermonthica seeds were mixed with sand to facilitate a regular spread of seeds over the plot. After S. hermonthica application, soil-tillage treatments were carried out using a hoe. With the standard soil-tillage a depth of about 10 cm was cultivated, whereas in the shallow soil-tillage treatment cultivation was restricted to the soil surface (approximately 2 cm). On July 5, sorghum seeds were sown in the seeding treatment at either 2.5 cm (shallow) or 15 cm (deep). For deep sowing, conically shaped plant holes were created using a hoe. Sowing density was 5 seeds per plant hole. After emergence, plants were thinned to one plant per plant hole. For the transplanting treatment three batches of seedlings were raised in 10 m2 Striga-free nurseries, with a sowing interval of 1 week. At July 11, when soil moisture conditions

Cultural control measures to diminish sorghum yield loss were suitable for transplanting, the size of the seedlings of the second batch was closest to the size of the transplants used in the greenhouse. These 28-days-old sorghum plants were transplanted into the experimental field in plant holes created as described above. Throughout the experiment the number of emerged Striga plants was counted at weekly intervals at two areas of 2.60 m2 per plot. At maturity (106 DAS), 16 plants per plot were harvested. Panicles were dried for 7 days in the open air before being threshed and weighed. Number of kernels per panicle and 1000-kernel weight was determined. At harvest time, total number of emerged S. hermonthica plants was counted and their dry weight recorded after drying in the open air. The response variables measured on S. hermonthica and the sorghum crop were subjected to ANOVA, using the statistical package Genstat 6.1 for Windows.

Results

Greenhouse experiment

Striga hermonthica characteristics In Table 1 results of the greenhouse experiment on Striga attachment, Striga emergence and Striga dry matter production are presented for cultivars CK60-B and Tiémarifing. In pots treated according to standard practice (standard tillage, combined with shallow sowing of seeds), the main difference between both cultivars was a slightly delayed first emergence of Striga with Tiémarifing (at 45 ± 1.6 DAS compared to 41 ± 1.3 DAS for CK60-B), and related to that a slightly reduced ASNPC value (402 ± 16.9 compared to 456 ± 29.2). Clear differences between CK60-B and Tiémarifing appeared when comparing the average results (grand mean) of all treatments. For Tiémarifing a lower number of attached (6.1 ± 1.39 compared to 10.2 ± 1.23) and maximum emerged (7.1 ± 0.68 compared to 12.5 ± 1.08) Striga plants was observed and Striga biomass production was on average only about 50% of that on CK60-B. This indicates that the investigated measures for delaying and reducing Striga infection were on average more effective with Tiémarifing than with CK60-B. With CK60-B, shallow-tillage and deep sowing/planting caused a nearly similar reduction in Striga biomass and number of reproductive organs. Both measures also caused a significant delay in attachment and emergence of Striga. The main difference between the two measures was that with deep sowing/planting the maximum number of emerged Striga plants and ASNPC were significantly reduced from 15.8 to 9.3 and from 438 to 194, respectively, whereas with shallow-tillage these characteristics were not significantly affected. The third measure, using transplants rather than seeds, gave

Chapter 6

pot

Striga flowers / flowers Total no of per capsules

dry Total pot (g) (g) pot Striga per weight

plants

Tiémarifing following variousculturalcontro per pot ASNPC of emerged

Striga Max. numbe Max.

time Mid- (DAS) emergence

>0.1). P

Striga (DAS) (DAS) days until first emergence emergence of Number

attach- 6.1 60 68 7.1 148 3.44 75 0 72 74 2.1 35 1.16 13

sorghum 10.2 55 65 12.5 316 6.65 89 infection for sorghum for sorghum infection cultivars CK60-B and flowering of of flowering <0.01; ns, non-significant ( non-significant ns, <0.01; Striga P Total of number of onset at ments

>0.01; *** *** >0.01; P Grand mean Grand mean Deep Transplants 9.1 4.5 58 61 Deep Transplants 66 70 2.4 4.3 10.2 9.3 249 194 64 62 5.36 5.01 72 70 70 50 5.0 6.5 88 128 1.95 3.06 21 51

hermonthica Striga

Number Progress Curve. Curve. Progress Number

** 0.05> >0.05; P

Striga All cultural control measures measures control cultural All combined 0.5 measures control cultural All combined 73 77 6.1 77 1.40 6 Tillage Planting depth material Planting Shallow 4.7 Tillage Planting depth material Planting 59 Shallow 68 0.9 12.7 309 65 4.81 69 44 3.9 72 2.11 42 Area under the Area under Standard practice 23.2 41 58 15.4 456 11.56 236 11.56 456 practice 15.4 58 23.2 41 Standard 272 Tiémarifing 8.96 402 practice 16.5 64 24.8 45 Standard Significance levels refer to the comparison of the two levels within each experimental factor. factor. experimental each within thelevels two of the to comparison refer levels Significance Significant at * 0.1> at Significant CK60-B Standard 15.8 *** *** 50 # ** 63 Standard 12.3 ns 11.2 * 323 ns 8.48 ** ** 133 ** 55 ns 67 10.3 * 223 ** 4.77 * ** 109 Shallow Seeds 15.9 *** 11.4 ns *** 48 *** 61 ** 51 Shallow 15.8 *** Seeds ns 64 438 *** 7.8 ns 8.28 *** 14.8 ** *** 128 9.7 ** 384 ** * 57 7.93 *** *** 55 ns 108 * 67 *** 65 7.7 ns 9.2 *** 167 ns 207 *** 3.83 ns 4.93 *** *** 130 *99 Table 1. Characteristics of 1. Characteristics Table measures, as observed in a greenhouse experiment (Wageningen, 2002).

Cultural control measures to diminish sorghum yield loss a slightly smaller, but still significant, reduction in Striga dry weight. For the number of Striga capsules/flowers no significant effect was observed. First emergence of Striga was 7 days delayed, but no significant reductions in number of Striga attachments at flowering and in mid-emergence time were found. Maximum number of emerged Striga plants as well as ASNPC was significantly reduced from 14.8 to 10.2 and from 384 to 249, respectively. For Tiémarifing, shallow-tillage and the use of transplants were the most effective measures. Both measures caused a significant reduction in Striga biomass production as well as in number of Striga capsules/flowers. A significant reduction in number of Striga attachments at flowering, a delayed first emergence and a reduced maximum number of emerged Striga plants accompanied these reductions. The third measure, deep sowing/planting, did not cause a significant reduction in total Striga dry weight, though the number of Striga capsules/flowers was significantly reduced from 99 to 51. Furthermore, a significant delay in first emergence of Striga (57 to 62 DAS) and mid- emergence time (67-70 DAS) were observed. The absence of any significant interaction between the three measures indicates that the effect of the various measures was additive. The effect that resulted from a combined application of all three measures is also presented in Table 1. Compared to standard practice, the effect of shallow-tillage and deep planting of transplants caused reductions of more than 85% in Striga dry weight for both cultivars. Reductions in number of capsules/flowers exceeded 95%. Hardly any attachments were found at flowering, compared to around 24 attachments per plant with standard practice. In line with this, the moment of first emergence was delayed by about 30 days in both cultivars. With CK60-B this resulted in a delay of 19 days in mid-emergence time, and a reduction in maximum number of Striga plants of 60%. With Tiémarifing, maximum number of Striga plants was reduced with 87%, whereas the delay in mid-emergence time was 10 days.

Effects of the cultural measures on sorghum In the course of the experiment, no S. hermonthica plants emerged in any of the control pots and also inspection of the roots of plants grown under Striga-free conditions, confirmed that these plants were free from S. hermonthica. For the uninfected sorghum plants raised according to standard practice, total biomass accumulation of CK60-B was lower than that of Tiémarifing, whereas differences in grain yield and yield components between cultivars were very small (Table 2). Whereas growth of CK60-B was not affected by any of the cultural measures, cultivar Tiémarifing turned out to be more sensitive, as all three measures significantly affected dry matter production. Shallow tillage had a positive effect, whereas deep

Chapter 6 planting and the use of transplants reduced final plant dry weight. Only the last two measures had a significant effect on grain yield. Both cultivars responded to a different extent to Striga infection (Table 3). Comparison of parameters of uninfected with infected sorghum plants under standard practice showed that the sensitive cultivar CK60-B was more strongly affected by the parasite than the tolerant cultivar Tiémarifing. This was most obvious for grain yield where a reduction of 92% was found for CK60-B, compared to a reduction of 28% for Tiémarifing.

Table 2. Effects of various cultural control measures on plant weight, grain yield, number of kernels and 1000-kernel weight of sorghum plants grown under Striga-free conditions, as observed in a greenhouse experiment (Wageningen, 2002). Plant Grain Number 1000- dry weight yield of kernels kernel (g plant–1) (g plant–1) per plant weight (g) CK60-B Grand mean 44.2 13.25 508 25.86 Tillage Standard 44.3 ns 13.33 ns 507 ns 25.98 ns Shallow 44.0 13.16 509 25.73 Planting depth Shallow 46.0 ns 13.48 ns 520 ns 25.66 ns Deep 42.3 13.01 496 26.05 Planting material Seeds 44.4 ns 13.18 ns 504 ns 25.83 ns Transplants 43.9 13.31 512 25.88

Standard practice 42.3 12.81 494 25.44 All cultural control measures combined 42.4 13.71 534 25.56 Tiémarifing Grand mean 68.2 11.84 459 25.73 Tillage Standard 64.7 * 11.83 ns 471 ns 25.35 ns Shallow 71.6 11.85 447 26.10 Planting depth Shallow 70.8 ** 13.29 ** 514 ** 26.23 ns Deep 65.5 10.38 404 25.22 Planting material Seeds 70.6 ** 10.57 * 417 ns 25.02 ns Transplants 65.7 13.10 502 26.44

Standard practice 72.4 12.08 464 26.44 All cultural control measures combined 67.3 11.90 442 27.33 Significant at * 0.1>P>0.05; ** 0.05>P>0.01; *** P<0.01; ns, non-significant (P>0.1). Significance levels refer to the comparison of the two levels within each experimental factor.

100

Cultural control measures to diminish sorghum yield loss

Focussing on the effects of individual measures clearly showed that with the sensitive cultivar CK60-B treatment effects were highly significant whereas with Tiémarifing individual treatments did not result in significant effects. With CK60-B, all treatments resulted in an increase in total sorghum dry weight, grain yield and kernel number whereas shallow-tillage also resulted in a significant increase of 1000- kernel weight.

Table 3. Effects of various cultural control measures on plant weight, grain yield, number of kernels and 1000-kernel weight of sorghum plants grown under Striga-infested conditions, as observed in a greenhouse experiment (Wageningen, 2002). Plant Grain Number of 1000- dry weight yield kernels kernel (g plant–1) (g plant–1) per plant weight (g) CK60-B Grand mean 19.2 3.42 144 23.17 Tillage Standard 15.1 ** 1.53 *** 68 *** 22.01 *** Shallow 23.3 5.31 219 24.33 Planting depth Shallow 16.9 ** 2.22 *** 95 *** 23.11 ns Deep 21.5 4.62 192 23.22 Planting material Seeds 16.7 *** 2.31 *** 96 *** 22.75 ns Transplants 21.6 4.53 191 23.58

Standard practice 13.5 1.01 48 20.92 All cultural control measures combined 30.4 9.49 394 24.06 Tiémarifing Grand mean 51.2 11.93 493 24.25 Tillage Standard 44.6 ns 11.67 ns 468 ns 24.48 ns Shallow 57.9 12.19 519 24.01 Planting depth Shallow 50.4 ns 10.90 ns 446 ns 24.46 ns Deep 52.0 12.97 541 24.03 Planting material Seeds 49.6 ns 11.81 ns 490 ns 23.87 ns Transplants 52.8 12.06 496 24.63

Standard practice 37.2 8.70 358 23.70 All cultural control measures combined 57.1 13.21 573 23.81 Significant at * 0.1>P>0.05; ** 0.05>P>0.01; *** P<0.01; ns, non-significant (P>0.1). Significance levels refer to the comparison of the two levels within each experimental factor.

101

Chapter 6

The effects that resulted from a combination of measures are also presented in Table 3. Compared to standard practice, deep planting of transplants after shallow- tillage of the soil more than doubled the biomass accumulation of CK60-B. The improvement in kernel yield and number of kernels was even stronger; both increased about 9 times. The combination of all measures showed a clear increase in 1000-kernel weight (15%). Compared to uninfected plants (Table 2), reduction of kernel yield due to Striga infection in CK60-B reduced from 92% to 26% when all three cultural measures were combined. Although for Tiémarifing the effects of individual treatments were not significant the combination of measures resulted in significant increases in biomass production, kernel yield and number of kernels. All of these traits were about 1.5 times higher than obtained with standard practice. No significant effect on 1000-kernel weight was found. The observed 28% reduction in kernel yield due to Striga infestation under standard practice completely disappeared when sorghum was transplanted at greater depth in shallow-tilled soil.

Field experiment

Striga hermonthica characteristics In Table 4, results of the field experiment on Striga emergence and Striga biomass production are presented at natural (low) and at artificially raised (high) Striga infestation levels. For CK60-B plants cultivated at standard practice the maximum number of emerged Striga plants, and concomitantly Striga dry weight, the number of reproductive Striga plants and ASNPC, were significantly higher in the high infestation plots. No difference in mid-emergence time was observed between natural and high infestation level, whereas first Striga emergence at the high infestation level was advanced by 6 days. Shallow-tillage, as well as the use of transplants, caused significant reductions in Striga dry weight and number of reproductive plants, both at low and high infestation level. This effect was to a large extent caused by a reduction in maximum number of emerged Striga plants. No clear pattern for timing of first emergence and mid- emergence time was discovered. Deep sowing/planting did not cause any significant response in Striga parameters both at low and high infestation level.

Effects of cultural measures on sorghum The average yield under high Striga infestation was about 23% lower than under conditions of natural Striga infestation (Table 5). Except for a significantly negative effect of shallow tillage under low Striga infestation no significant effects of the other

102

Cultural control measures to diminish sorghum yield loss

) –2

Striga Striga Number ofNumber (plants m

reproductive

) –2 Total Striga (g m dry weight weight dry

ASNPC

)

–2

Striga Max. no of (plants m

>0.1). P time time Mid-

(DAS)

emergence

Striga

(DAS)

emergence emergence ns, non-significant<0.01; ( P until first until first Number of days

infection for sorghum cultivar CK60-B following various cultural control measures at two the two levels within each experimental factor. factor. experimental within each the levels two

*** >0.01; P

Standard Shallow Shallow 7.6 64 Deep 1.5 76 Seeds 57 * 40 0.8 9.2 Transplants 4.8 62 61 5.5 60 ns 2.5 72 1.3 72 407 79 2.5 1.8 37 1.3 0.7 259 72 ** 8.2 61 ns 13.2 Standard Shallow 53 75 ns 70 4.4 ** Shallow 71 Deep 52 3.4 ns 76 * Seeds 143 *** 51 ** 2.1 53Transplants 2.5 4.3 * 4.3 134 68 104 ns 4.6 *** 53 ns 72 * 2.7 ns 3.0 ** 146 *** 52 ns 3.8 *** 71 ns 2.1 ns 19.6 * 3.1 *** 14.6 ns 73 *** 586 *** 23.4 *** 11.5 *** 438 ns 10.0 ** 711 *** 7.8 ns 14.5 *** 7.2 ns 12.7 *** Grand mean 61 3.0 74 92 2.6 1.9 Grand Grand mean 52 71 13.9 423 8.5 7.4 Striga hermonthica Number Progress Curve. Progress Number

>0.05; ** 0.05> P

Striga infestation levels, as observed in a field experimentas observedin a field (Cinzana, 2002). infestation levels,

Tillage Tillage depth Planting material Planting Standard practice Tillage depth Planting material Planting 57 Standard practice 75 8.0 51 260 74 6.0 41.9 5.5 1222 19.9 17.7 All cultural control measures combined 59 74 All cultural control measures combined 0.9 56 27 70 0.9 6.3 0.6 186 1.7 2.2

natural Striga infestation infestation Striga natural infestation Striga high Area under the under Area S. hermonthica

Table 4. Characteristics of # Significant at * 0.1> at Significant Significance levels refer of refer to levels comparison the Significance

103

Chapter 6 cultural control measures on sorghum yield were detected. Even the combination of cultural control measures did not result in an increased grain yield. This is surprising, as, both at low and high infestation level, a clearly lower Striga infection level was observed in plots that received a combination of control measures. This result suggests that the positive effect on crop production that is likely to follow the reduction in Striga infection level must have been counterbalanced by other unintentional effects of the measures as such.

Table 5. Effects of various cultural control measures on grain yield, number of kernels and 1000-kernel weight of sorghum cultivar CK60-B plants grown at two Striga-infestation levels, as observed in a field experiment (Cinzana, 2002). Number of Grain yield kernels 1000-kernel (g m–2) per plant weight (g) Grand mean 111.4 2276 17.78 Tillage Standard 124.7 ** 2389 ns 18.18 ns Shallow 98.0 2163 17.37 Planting depth Shallow 106.0 ns 2309 ns 17.38 ns Deep 116.7 2243 18.17 Planting material Seeds 106.2 ns 2067 ** 17.37 ns Transplants 116.5 2485 18.18

natural Striga natural Striga infestation Standard practice 122.8 2246 17.97 All cultural control measures combined 122.8 2351 19.04 Grand mean 86.0 1732 17.21 Tillage Standard 82.6 ns 1648 ns 17.62 ns Shallow 89.4 1816 16.80 Planting depth Shallow 82.8 ns 1674 ns 17.04 ns Deep 89.2 1789 17.38 Planting material Seeds 87.3 ns 1687 ns 16.88 ns Transplants 84.7 1777 17.54

high Striga infestation high Striga infestation Standard practice 100.6 1761 17.86 All cultural control measures combined 106.1 2067 18.21 Significant at * 0.1>P>0.05; ** 0.05>P>0.01; *** P<0.01; ns, non-significant (P>0.1). Significance levels refer to the comparison of the two levels within each experimental factor.

104

Cultural control measures to diminish sorghum yield loss

Discussion

Striga infection The objective of this study was to evaluate if management options aiming at a separation of host roots and Striga seeds during the early stages of crop development are able to delay, or avoid, infection of the host by the parasite. Shallow-tillage and deep sowing/planting are expected to operate through an initial spatial separation of host and parasite. The results from the greenhouse experiment showed that both measures resulted in a delayed emergence of S. hermonthica. The strong reduction in number of attachments at flowering of sorghum suggests that this delay in first emergence resulted from a delay in first Striga attachment. For Tiémarifing deep planting was less effective than for CK60-B. This might be related to differences in rooting pattern between both cultivars. Van Ast et al. (2000, see also Chapter 2) observed that with the sensitive cultivar CK60-B root development in the upper soil layer was much faster and more extensive than with Tiémarifing. At 35 DAE, about 46% of the total root weight of CK60-B was found in the upper 6 cm soil layer, whereas only 23% of the total root biomass of Tiémarifing was found in this layer. This suggests that deep planting of sorghum is in particular effective for shallow- rooting cultivars. The use of sorghum transplants raised in a Striga-free nursery causes a more direct separation of host roots and parasites. As introduction of sorghum seeds and transplants to the pots occurred at the same moment, the main difference between both treatments was the age at which sorghum roots were exposed to Striga seeds. For both cultivars the use of transplants resulted in a delayed first emergence of Striga. Oswald and Ransom (2002), studying the effect of transplanting on a range of maize varieties also observed a delayed emergence of S. hermonthica, irrespective of genotype. Dawoud et al. (1996), studying the effect of transplanting on Striga incidence and sorghum yield, explained the reduced number of attachments in transplanted sorghum by a reduced production of germination stimulants and an increased penetration resistance of older roots. Rolando et al. (1992) demonstrated that the resistance of older sorghum roots to penetration of S. hermonthica was because of the development of substances such as lipids, phenols, suberin and lignin in the root cells. Also for maize an age-related resistance to S. hermonthica was reported (Oswald et al., 2001). In the greenhouse experiment, all measures combined resulted in a delay of first emergence of ca 4 weeks, and this delay was followed by a reduced maximum number of emerged Striga plants. This reduced number might be a direct consequence of the shorter time span that remained for realizing successful host-parasite connections. At the same time, ageing of the host roots might have strengthened this effect.

105

Chapter 6

Under field conditions, the application of all cultural control measures combined resulted in a significant reduction in maximum number of emerged Striga plants. Surprisingly, the measures did not result in a substantial delay in first Striga emergence. The delay in first emergence was only 2-5 days, compared to 27-32 days in the greenhouse experiment. This overall result is in good agreement with results reported by Oswald et al. (2001), who studied the effect of transplanting on the host- parasite relationship for maize and S. hermonthica under field conditions. Transplanting resulted in significantly lower Striga densities than the direct-seeded control, but also in this case early emergence of Striga could not be prevented. The presence of a natural Striga soil seedbank, with Striga seeds evenly distributed across the cultivated layer of the soil, might be an important factor in this respect. Such a distribution is to be expected on Striga-infested fields that have regularly grown with susceptible crops (van Delft et al., 1997). At the same time such a distribution might undermine the effectiveness of measures that aim at a spatial separation of Striga seeds and host roots. Most noteworthy in this context was the absence of any effect of deep planting. In the pot experiment holes were created by using a metal cone-shaped tool and erosion of the plant hole was prevented by putting 1.5 cm high PVC rings at the top of the hole. In the field experiment a daba was used to dig the plant holes, as the soil was too hard to use the cone-shaped tool. Besides, PVC rings were not used and consequently holes gradually filled up with Striga-infested soil from the topsoil layer that was transported by run off water after heavy rain showers.

Sorghum yield The results from the greenhouse experiment indicate that the consequences of the proposed cultural control measures in terms of reduced yield loss are especially beneficial with sensitive sorghum cultivars. With cultivar CK60-B, all three measures substantially reduced yield loss due to Striga, whereas the strongest effect was obtained by combining the various measures. A nearly complete yield loss (92%) under standard practice was reduced to 26% when all three measures were combined. With Tiémarifing, individual measures did not result in significant yield improvements, though the combination of all three measures caused a significant increase in kernel yield. It should be realized that yield reduction of the tolerant Tiémarifing under standard practices was relatively small. Consequently, the potential for substantial improvements in kernel yield under Striga-infested conditions will also be comparatively small. These results correspond to observations reported by Gurney et al. (1999) and van Ast and Bastiaans (2006; see also Chapter 5), who compared the effect of an artificially delayed moment of Striga infection on a tolerant and a sensitive sorghum cultivar.

106

Cultural control measures to diminish sorghum yield loss

Results of the field experiment show that, despite the strong reduction in Striga infection level, the cultural control measures had no direct beneficial effect on crop yield. An important difference between the results obtained under greenhouse and field conditions was that in the latter no significant delay in first Striga emergence was established. For yield reduction, the moment of first Striga infection might be at least as important as the number of infections. With yield reduction due to resource competition from weeds, it is well established that relative emergence time is a critical factor next to weed density (Kropff and Spitters, 1991; Kropff et al., 1992a). A delay of one week already has tremendous effects, as competition is often asymmetric, meaning that relatively small differences in starting position are enlarged towards the end. Also with parasitic weeds it has been reported that early disturbances of the host plant have far larger consequences than late disturbances (Cechin and Press, 1993b; Gurney et al., 1999; van Ast and Bastiaans, 2006, see also Chapter 5). That a reduced Striga infection level as such does not necessarily result in an improved sorghum yield has more often been observed. Reda (2002), who studied the effect of relay cropping of different food legumes and oil crop species on Striga infection of sorghum, observed that after three cropping cycles sorghum grain yield in the relay cropped treatments was still identical to that of the control. At the same time, the number of emerged Striga plants in these treatments started to go down relative to the numbers in the control. Again the same phenomenon was found in a long-term experiment on hand weeding (Ransom and Odhiambo, 1994). After the first year, a decline in Striga infection level could be observed in the hand-weeded plots, whereas no consistent correlation was observed between Striga infection level and maize yield. These observations might be explained by the non-linear relationship between Striga biomass and grain production reported by Gurney et al. (1999). This relationship indicates that above a specific infection level host grain production is independent of parasite number. Particularly if the actual infection level is way above this “saturation level” any measure that reduces Striga infection will not automatically result in an increased sorghum production. Another factor that cannot be excluded is a direct negative effect of the cultural control measures on crop performance as such. As a result, the positive effects of the cultural control measures on Striga incidence might be levelled out by these negative effects. Van Delft et al. (2000) for instance reported that deep planting might have a detrimental effect on seedling growth and establishment, particularly if plant holes gradually refill with soil. Unfortunately, it was not possible to evaluate the direct consequences of the cultural control measures on crop performance as it was not feasible to include Striga-free control plots in the field experiment.

107

Chapter 6

Perspectives The results of this study confirm that the principle of separating Striga seeds and sorghum root system in space and time during early development stages of the crop bears great potential (e.g., van Delft et al., 2000). Under controlled conditions, such a separation could be successfully established, resulting in a delayed moment of first Striga infection, a reduced Striga infection level, and a strong reduction in Striga- related sorghum yield loss. On top, the reproductive success of the parasite, expressed in numbers of seed capsules and flowers, was markedly reduced. The study also demonstrated that in the field a complete separation between Striga seeds and sorghum roots was much more difficult to achieve, most likely because of the existence of a natural soil seedbank. As a result, the proposed cultural control measures failed to establish a significant delay in moment of first Striga emergence under field conditions, though a clear reduction in Striga infection level was attained. However, this lower infection level did not result in an increased sorghum yield. The absence of such an immediate positive effect on sorghum yield is an important constraint for adaptation of the proposed measures by small-scale subsistence farmers (Ransom, 2000). The main advantage of the proposed measures under field conditions is a reduced reproductive level of the parasite. As controlling the production of new Striga seeds is an important component of any long-term Striga control programme the proposed cultural control measures might still reveal their usefulness in the long run.

108

CHAPTER 7

General discussion

Origin of the Striga problem In sub-Saharan Africa, maize, sorghum and millet are the main cereal food crops. Maize is the most prominent cereal crop in the sub-humid region, whereas this crop is gradually replaced by sorghum in areas with lower water availability and by millet, the most drought resistant species, if even less water is available. Traditional cereal-based cropping systems in sub-Saharan Africa consisted of shifting cultivation practices, characterized by long fallow periods, crop rotation and mixed cropping. As short periods of cultivation were alternated with relatively long fallow periods, the abandoned land was able to restore its soil fertility and to reduce disease pressure and the size of the weed seedbank (Ker, 1995; Sauerborn et al., 2000). A rapid annual human population growth rate in this part of Africa led to an intensification of the traditional cropping systems. The length of the fallow periods was shortened and consequently there was often insufficient time for the natural restoration of soil fertility. This led to a downward spiral in average productivity per unit area, which was further strengthened when marginal lands were taken into production to cover the demand for basic food. For the same reason, crop rotations and intercropping practices were gradually replaced by permanent monocropping of cereals. The result of all of this was that production fields gradually developed into nutrient poor, acidic fields with low organic matter content and great soil compaction (IITA, 1992). Simultaneously, the parasitic weed Striga hermonthica (Del.) Benth. developed into one of the most serious biological constraints to cereal food production in this region (Gressel et al., 2004; Kroschel, 1999). To some extent the increased S. hermonthica pressure was a consequence of the above described developments. The nearly permanent presence of host plants combined with low soil fertility conditions created ideal conditions for a rapid built up of an extensive Striga soil seedbank. The continuous presence of host plants led to increasing inputs of S. hermonthica seeds, whereas the low microbial activity following the low organic matter content of the soil resulted in a reduced decomposition of S. hermonthica seeds. The rapid increase in the extent and intensity of S. hermonthica infestations was accompanied with increasing yield reductions due to Striga infection ranging from just a few percentages to complete crop failure (Sauerborn, 1991; Weber et al., 1995). This devastating effect of the parasite on the productivity of the cereal food crops made that S. hermonthica also

109

Chapter 7 contributed to the earlier described downward spiral in productivity per unit area and thus to the growing demand for cereal production fields. Striga species are obligate hemi-parasitic plants that attach to the roots of their hosts to obtain water, nutrients and carbohydrates (Kuijt, 1969; Parker and Riches, 1993). The species are native to the grasslands of the semi-arid and sub-humid tropics of Africa. In these grasslands they co-exist with their host plants at relatively low infestation levels. Only after the intensification of the traditional cereal-based cropping systems the parasite developed into a widespread parasitic weed species, threatening food production in practically the entire semi-arid and sub-humid region’s farming systems of Africa. This makes this weed species a typical example of a pre-adapted weed species. Such species are resident within a plant community within dispersal distance of the crop and come to dominate within the cropping system through a change in management practices.

Bewitching the host; detrimental or beneficial? Striga hermonthica affects its host plant not only through withdrawal of assimilates and nutrients, but also poses a strong pathological effect upon its host. The most important element of this pathological effect is a reduction in leaf photosynthetic rate (Press and Stewart, 1987; Graves et al., 1989; Smith et al., 1995). The current research demonstrated that with the sensitive cultivar CK60-B the maximum reduction in leaf photosynthetic rate was about 40-50%. This reduction was already obtained with few Striga attachments (Chapter 3). From an evolutionary point of view this response seems surprising, as it might be argued that a well-performing host plant is the best guarantee for a successful parasite. The Striga seed density experiment showed however that in situations where host plant vigour was most severely reduced (at the highest infestation level; 2000 seeds dm–3 of soil), this did not result in a reduced dry matter production of the parasite. At this infestation level the fraction parasite of the total host plant-parasite association was close to 70% and significantly higher than at lower infestation levels (Chapter 4). Whether this increased fraction was because of an increased autotrophic production in carbon by the parasite or resulted from a larger withdrawal of host produced carbon could not be established. For both mechanisms it might be argued that a weakened host plant might be favourable for the performance of the parasite. In case of parasitism, it cannot be excluded that the parasite is better able to withdraw assimilates and nutrients from a weakened host. At the same time a weakened, and often smaller host plant, might create better light conditions for the hemi-parasite to produce its own assimilates. Another hypothesis is that the observed reduction in leaf photosynthetic rate is simply an undesired side-effect, resulting from the physiological disturbances initiated by the parasite to withdraw resources from the

110

General discussion host plant. However, observations on the tolerant cultivar Tiémarifing, in which the photosynthetic rate remained unaffected after Striga infection (Chapter 3), clearly prove that a reduced photosynthetic rate is not a necessary prerequisite for the withdrawal of resources from the host.

Eradication of Striga weeds? In the USA, the occurrence of an epidemic of S. asiatica in the 1950s in maize in North and South Carolina led to the initiation of a drastic quarantine and eradication programme. At that time an area of about 150,000 hectares was infested. Main objectives of this programme were prevention of the spread of Striga seeds to non- infested areas and depletion of the Striga seed population in the soil of the already infested areas. To this purpose infested fields were treated with herbicides, methyl bromide and ethylene. Herbicides were also applied on road sides and idle land, to kill the parasites on alternative host plant species (Sand, 1987). By strictly enforcing quarantine of the area and through an extensive education and eradication effort (carried out for nearly 50 years), a further spread of the parasite was prevented, though even today the parasite is still present on a couple of thousand hectares (Chrispeels and Sadava, 2003). Implementation of such a drastic eradication program in Africa is unrealistic and seems to be out of place. Major achievement of the described programme was the prevention of a further spread of the invasive parasitic weed species to potential threatened areas with maize, sorghum and sugar cane production. In the sub-Saharan zone of Africa, Striga species are native plants and the total infested area is estimated to cover around 17 million hectares (Gressel et al., 2004). As a complete eradication of Striga is not realistic, the presence of these parasites is inevitable. The crop species that serve as host plants to Striga are staple food crop species in West Africa. The lack of alternative food crops makes it practically impossible to shift, on a large scale, to other crop species. Consequently, it is unrealistic to assume that in the near future the major cereal crops are completely replaced by other food crops. This implies that in the forthcoming decades both host plant and parasite will most likely continue to form important elements of cereal cropping systems in the semi-arid and sub-humid regions in tropical Africa.

The sorghum – Striga interrelationship Given the presence of sorghum and S. hermonthica it is vital to clarify the relationship between the two organisms and their mutual interactions. The Striga-infestation level in the soil is an important determinant of the Striga-infection level of the host. This infection level determines the level of kernel yield reduction of the host, as well as the seed production of the parasite. In turn, the reproductive effort of the parasite largely

111

Chapter 7 determines the seed input into the soil seedbank, and as such forms an important determinant of future infestation levels. One obvious way of influencing these relationships is through the selection of resistant and/or tolerant sorghum cultivars. Resistance against Striga reduces the infection level of a host plant, while tolerance enables a host plant to perform well, despite the parasitic infection. Host resistance is thought to be the most economical and potentially the most effective control option against root diseases and soil born pathogens (Shew and Shew, 1994) and therefore a potentially acceptable Striga control option to resource-poor farmers (Hess and Ejeta, 1992; Debrah, 1994). So far, complete resistance, or immunity, against Striga has not been found. Because few Striga infections already seriously harm the host plant, resistance alone may not be enough to prevent crop losses. It is therefore recommended to direct breeding efforts towards finding cultivars that combine resistance with high levels of tolerance (Haussmann et al., 2001; Rodenburg, 2005). Over the years, breeding, or the use of resistant and tolerant cultivars, has developed into one of the main elements of nearly all integrated Striga control programmes. The aim of the current research was not to focus specifically on these defence mechanisms, but rather to obtain a better understanding of the mutual interactions between host plant and parasite, as it is this interaction that has implications for both the short term (sorghum kernel yield) as well as the long term (Striga population development). For this reason a number of experiments were conducted using the highly susceptible and sensitive cultivar CK60-B. Based on the results of a comparative study between CK60-B and the tolerant cultivar Tiémarifing it was hypothesized that, apart from a reduced pathological effect of the parasite on the host, the reduced consequences of parasite infection for Tiémarifing might partly result from the observed delay in Striga attachment (Chapter 2). Whereas the first attachments on CK60-B were observed at 10 DAE, first attachments on Tiémarifing were only found 7 days later. Time of first emergence differed 18 days. Also in an experiment with a wide range of Striga infestation levels it was shown that a good correlation existed between the time of first Striga emergence and the observed kernel yield reduction of the sorghum host (Chapter 4). Despite strong indications of the role of time of infection, no final conclusions could be drawn, as in this last experiment infection level, expressed as the maximum number of emerged Striga plants, and time of first Striga emergence were confounded. An experiment in which infection time was artificially delayed clearly resolved that Striga infection time is an important factor in determining the yield reduction following Striga infection (Chapter 5). The influence of delayed infection strongly depended on actual infection time. With the tolerant Tiémarifing, on which infection under standard conditions already occurred relatively late, no significant increase in total plant dry

112

General discussion weight and kernel yield was obtained when infection time was even further delayed. With CK60-B, the effects were more pronounced in 1999, when infection under standard conditions happened relatively early, compared to 2000, when infection under standard conditions happened relatively late. These differential responses between cultivars and years further illustrate the relevance of infection time. Both in the density experiment as well as in the experiment in which time of attachment was artificially delayed a reasonable correspondence was observed between sorghum kernel yield loss and Striga dry matter production. Striga dry matter production was found to be a reasonable estimate of the reproductive effort (Rodenburg et al., 2006a), even if clear differences in time of first Striga emergence exist. To a large extent this will be due to the observed synchronization in flowering time of the parasite (Chapter 4). A later time of infection is thus clearly beneficial to the host, whereas it is disadvantageous to the parasite. This implies that in the host plant-parasite relationship the parasite is hindered more by a reduced growing period than that it is able to profit from a more vigorously growing host plant. Again this response is clearly related to the dual position of host vigour. Early infection results in a severely reduced host plant and a reduced dry matter production of the total host plant - parasite association. However, Striga showed to perform relatively well under these circumstances, as it was able to increase its share in the total association, and through that was able to increase or maintain its production in absolute terms. For practical applications it is highly relevant that a delay in time of infection showed to have positive implications for both the short (reduced kernel yield loss) and the long term (reduced seed input). The lack of financial means and the crucial role of cereal food crops in the food security of the farm household, forces small scale farmers to rely on Striga control measures that are at least beneficial on the short term.

Delayed infection time and Striga control Pot experiments demonstrated that the separation of Striga seeds from the root system of the host during early crop development can be established in different ways. Deep planting, the use of transplants and keeping the Striga seeds in the top soil layer all showed effective in delaying the moment of first infection. Even more interesting, the measures showed to have an additive effect. Application of all measures combined resulted in a four week delay in first emergence of the parasite, and a considerable reduction (around 75%) in the maximum number of emerged parasites (Chapter 6). The consequences of this delay for host and parasite were above expectation. With the sensitive cultivar CK60-B a yield loss of 92% was reduced to 26%, whereas with the tolerant cultivar Tiémarifing application of the cultural control measures completely cancelled out a yield loss of 28%. For both cultivars, Striga dry weight production

113

Chapter 7 was decimated. Application of the investigated cultural control measures under field conditions at a Research Station in Cinzana, Mali were however disappointing. Under both low and high Striga infestation levels the desired delay in Striga infection time, as derived from the time of first Striga emergence, was not achieved, though significant reductions in number of emerged Striga plants were obtained. This resulted in severe reductions in Striga dry matter production, but did not result in significant improvements in sorghum grain yield. The absence of such a short-term beneficial effect of the proposed measures makes farmers’ acceptance far less likely. One explanation for the large discrepancy between the results obtained under controlled conditions in the pot experiment and the results obtained under African field conditions might be the presence of a natural Striga soil seedbank throughout the tilth. Consequently, the separation of host roots and Striga seeds was far more difficult to achieve. Also the creation of planting holes in the compacted soil in the field was far more difficult to realize than in the sandy soil mixture in the pot experiment and it is not unlikely that Striga seeds ended up at the bottom of the planting holes during construction of the holes or shortly after, as a result of transportation by run-off water following heavy rain showers. As already a few Striga attachments are able to initiate severe physiological disturbances of the host, just a small number of seeds is able to completely undermine this control strategy. An additional observation in this experiment was that, as a result of the applied measures, and contrary to what was observed in the density experiment, Striga infection time and Striga infection level were not confounded anymore. The result suggests that sorghum grain yield is mainly determined by Striga infection time, whereas Striga dry weight is more closely related to Striga infection level. One explanation for a weak dependency of sorghum yield to infection level was found in Chapter 3. Here it was observed that with the sensitive CK60-B just a few Striga attachments were sufficient to realize the maximum reduction in leaf photosynthetic rate of around 45%. Consequently, the yield reduction following from this damage mechanism will be nearly independent on the number of attachments.

Outlook The potential value of a delayed contact between host and parasite, combined with the observed difficulties to establish a delay in infection under natural field conditions through the proposed cultural control measures, leads to the question whether a delayed infection can be achieved by alternative means. In this respect, herbicide seed dressings may provide a promising alternative. In maize, research on this topic has been conducted in recent years. Application of ALS-inhibiting herbicides such as

114

General discussion imazapyr on seeds of resistant crops, either as drench at planting or through seed priming or as seed coating has been shown to provide early season control to Striga hermonthica in maize (Ransom et al., 1995; Berner et al., 1997). Striga emergence was nearly completely prevented for 3 months, whereas Striga emerged from 6 to 12 weeks after planting on untreated maize (Kanampiu et al., 2001). More recently, investigations in collaboration between Kansas State University (Dr. M. Tuinstra; Center for Sorghum Improvement, Department of Agronomy) and the Department of Crop and Weed Ecology at Wageningen University were undertaken to apply the same methodology in sorghum. Tolerance to ALS-inhibiting herbicides was introgressed from shattercane (Sorghum bicolor), a common weed found in the USA. Preliminary results were promising in greenhouse and field trials. Significant delays in Striga attachment were achieved with concomitant increases in crop yield.

115

References

Abunyewa, A.A., Padi, F.K., 2003. Changes in soil fertility and Striga hermonthica prevalence associated with legume and cereal cultivation in the Sudan savannah zone of Ghana. Land Degradation & Development 14, 335-343. Ackroyd, R.D., Graves, J.D., 1997. The regulation of the water potential gradient in the host and parasite relationship between Sorghum bicolor and Striga hermonthica. Annals of Botany 80, 649-656. Adetimirin, V.O., Aken’Ova, M.E., 2000. Effects of Striga hermonthica on yield components in maize. Journal of Agricultural Sciences135, 185-191. Akobundu, I.O., 1991. Integrated weed management for Striga control in cropping systems in Africa. In: Combating Striga in Africa. Ed S.K. Kim, Proceedings International Workshop organised by IITA, ICRISAT and IDRC, 22-24 August 1988, IITA, Ibadan, Nigeria, pp. 122-125. Africancrops, 2005. http://www.africancrops.net Ast, A. van, Bastiaans, L., Kropff, M.J., 2000. A comparative study on Striga hermonthica interaction with a sensitive and a tolerant sorghum cultivar. Weed Research 40, 479-493. Ast, A. van, Bastiaans, L., 2006. The role of infection time in the differential response of sorghum cultivars to Striga hermonthica infection. Weed Research (in press). Baltus, P.C.W., Ransom, J.K., Odhiambo, G.D., Egbers, W.S., ter Borg, S.J., Verkleij. J.A.C., Pieterse, A.H., 1994. A comparative study on the defence mechanism(s) of the maize variety “Katumani” against S. hermonthica under field conditions in Kenya. In: Biology and Management of Orobanche. Eds A.H. Pieterse, J.A.C. Verkleij, S.J. ter Borg. Proceedings Third International Workshop on Orobanche and related Striga research, Amsterdam, The Netherlands. The Alder Press, Oxford, UK, pp. 373-381. Bastiaans, L., 1993a. Understanding yield reduction in rice due to leaf blast. PhD Thesis, Wageningen Agricultural University, The Netherlands. Bastiaans, L., 1993b. Effects of leaf blast on growth and production of a rice crop. 2. Analysis of the reduction in dry matter production of rice due to leaf blast, using two models with different complexity. Netherlands Journal of Plant Pathology 99, 19-28. Bebawi, F.F., Farah, A.F., 1981. Effects of parasitic and non-parasitic weeds on sorghum. Experimental Agriculture 17, 415-418. Berner, D.K., Kling, J.G., Singh, B.B., 1995. Striga research and control. A perspective from Africa. Plant Disease 79, 652-660.

117

References

Berner, D.K., Alabi, M.O., Di-umba, U., Ikie, F.O., 1996. Proposed Integrated Control Program for Striga hermonthica in Africa. In: Advances in parasitic plant research. Eds M.T. Moreno, J.I. Cubero, D. Joel, L.J. Musselman, C. Parker. Proceedings Sixth International Parasitic Weed Symposium. Cordoba, Spain, pp. 817- 825. Berner, D.K., Ikie, F.O., Green, J.M., 1997. ALS-inhibiting herbicide seed treatments control Striga hermonthica in ALS-modified corn (Zea mays). Weed Technology 11, 704-707. Blackman, V.H., 1919. The compound interest law and plant growth. Annals of Botany 33, 353-360. Borg, S.J. ter, 1986. Effects of environmental factors on Orobanche-host relationships; a review and some recent results. In: Biology and control of Orobanche. Ed. S.J. Borg. Proceedings of a workshop on biology and control of Orobanche. L.H./VPO, Wageningen, The Netherlands, pp. 57-69. Bouwmeester, H.J., Matusova, R., Zhongkui, S., Beale, M.H., 2003. Secondary metabolite signalling in host-parasitic plant interactions. Current Opinion in Plant Biology 6, 358-364. Campbell C.L., Madden, L.V., 1990. Introduction to Plant Disease Epidemiology. Wiley, New York, USA. Carsky, R.J., Singh, L., Ndikawa, R., 1994. Effect of herbicide and hand weeding on current and subsequent season Striga hermonthica density on sorghum. International Journal of Pest Management 40, 111-116. Carsky, R.J., Nokoe, S., Lagoke, S.T.O., Kim, S.K., 1998. Maize yield determinants in farmer-managed trials in the Nigerian northern guinea savanna. Experimental Agriculture 34, 407-422. Cechin, I., Press, M.C., 1993a. Nitrogen relations of the sorghum-Striga hermonthica host-parasite association: growth and photosynthesis. Plant, Cell and Environment 16, 237-247. Cechin, I., Press, M.C., 1993b. The influence of nitrogen on growth and photosynthesis of sorghum infected with Striga hermonthica from different provenances. Weed Research 33, 289-298. Cechin, I., Press, M.C., 1994. Influence of nitrogen on growth and photosynthesis of a

C3 cereal, Oryza sativa, infected with the root hemiparasite Striga hermonthica. Journal of Experimental Botany 45, 925-930. Chantereau, J., Nicou, R., 1994. The Tropical Agriculturalist. Sorghum. The Macmillan Press Ltd., London. Cherif-Ari, O., Housley, T.L., Ejeta, G., 1990. Sorghum root length density and the potential for avoiding Striga parasitism. Plant and Soil 121, 67-72. Chrispeels, M.J., Sadava, D.E., 2003. Plants, Genes, and Crop Biotechnology. Jones

118

References

and Bartlett Publishers, Sudbury, Massachusetts, 562 pp. Clark, L.J., Shawe, K.G., Hoffmann, G., Stewart, G.R., 1994. The effect of Striga hermonthica (Del.) Benth. infection on gas-exchange characteristics and yield of a sorghum host, measured in the field in Mali. Journal of Experimental Botany 45, 281-283. Combari, A. Pineou, R., Schiavon, M., 1990. Influence du degré de décomposition de produits organiques sur la germination de graines de Striga hermonthica (Del.) Benth. Weed Research 30, 29-34. Dawoud, D.A., Sauerborn, J., Kroschel, J., 1996. Transplanting of sorghum: a method to reduce yield loss caused by the parasitic weed Striga. In: Advances in Parasitic Plant Research. Eds M.T. Moreno, J.I. Cubero, D. Berner, D. Joel, L.J. Musselman, C. Parker, Proceedings of the Sixth International Parasitic Weed Symposium, Cordoba, Spain, pp. 777-785. Debrah, S.K., 1994. Socio-economic constraints to the adoption of weed control techniques: the case of Striga control in the West African Semi-Arid Tropics. International Journal of Pest Management 40, 153-158. Debrah, S.K., Defoer, T., Bengaly, M., 1998. Integrating farmers’ knowledge, attitude and practice in the development of sustainable Striga control interventions in southern Mali. Netherlands Journal of Agricultural Science 46, 65-75. Defoer, T., 2000. Moving Methodologies. Learning about integrated soil fertility management in sub-Saharan Africa. PhD Thesis Wageningen University, Wageningen, The Netherlands, 189 pp. Delft, G.J. van, 1997. Root architecture in relation to avoidance of Striga hermonthica infection. PhD Thesis, University of York, York, UK. Delft, G.J. van, Graves, J.D., Fitter, A.H., Pruiksma, M.A., 1997. Spatial distribution and population dynamics of Striga hermonthica seeds in naturally infested farm soils. Plant and Soil 195, 1-15. Delft, G.J. van, Graves, J.D., Fitter, A.H., Ast, A. van, 2000. Striga seed avoidance by deep planting and no-tillage in sorghum and maize. International Journal of Pest Management 46, 251-256. Dembélé, B., Raynal-Roques, A., Sallé, G., Tuquet, C., 1994. In: Plantes parasites des cultures et des essences forestières au Sahel. Ed. J. Libby. Montrouge, France, 43 pp. Doggett, H., 1965. Striga hermonthica on sorghum in east Africa. Journal of Agricultural Science 65, 183-194. Doggett, H., 1988. Witchweed (Striga). In: Ed. H. Doggett Sorghum. London, Longmans, pp. 368-404. Dörr, I., 1996. New results on interspecific bridges between parasites and their hosts.

119

References

In: Advances in Parasitic Plant Research. Eds M.T. Moreno, J.I. Cubero, Proceedings of the Sixth International Symposium on Parasitic Plants. Cordoba, Spain, pp. 195-201. Drennan, D.S.H., El Hiweris, S.O., 1979. Changes in growth regulating substances in Sorghum vulgare infected by Striga hermonthica. In: Proceedings Second International Symposium on Parasitic Weeds. Eds L.J. Musselman, A.D. Worsham, R.E. Eplee. University Graphics, North Carolina State University, Raleigh, pp. 144- 155. Ejeta, G., Butler, L., Babiker, A.G.T., 1992. New approaches to the control of Striga: Striga research at Purdue University. Agricultural Experiment Station, Purdue University, West Lafayette, IN. 27 pp. Eplee, R.E., Norris, R., 1987. Chemical control of Striga. In: Parasitic weeds in agriculture. Ed. L.J. Musselman, Volume I, Striga. Boca Raton, CRC Press Inc., pp. 173-182. Eplee, R.E., Westbrooks, R.G., 1990. Movement and survival of witchweed seeds in the soil. In: Witchweed Research and Control in the United States. Eds P.F. Sand, R.E. Eplee, R.G. Westbrooks, WSSA, Champaign, IL, pp. 81-84. Eplee, R.E., Norris, R., 1995. Control of parasitic weeds. In: Parasitic plants. Eds M.C. Press, M.C., J.D. Graves. Chapman & Hall, London, pp. 256-277. Estabrook, E.M., Yoder, J.I., 1998. Plant – plant communications: rhizosphere signalling between parasitic angiosperms and their hosts. Plant Physiology 116, 1-7. FAOSTAT, 2005. http:// faostat.fao.org. Frost, D.L., Gurney, A.L., Press, M.C., Scholes, J.D., 1997. Striga hermonthica reduces photosynthesis in sorghum: the importance of stomatal limitations and a potential role of ABA? Plant, Cell and Environment 20, 483-492. Gacheru, E., Rao, M.R., 2005. The potential of planted shrub fallows to combat Striga infestation on maize. International Journal of Pest Management 51, 91-100. Gbèhounou, G., Pieterse, A.H., Verkleij, J.A.C., 1996. The decrease in seed germination of Striga hermonthica in Benin in the course of the rainy season is due to a dying-off process. Experientia 52, 264-267. Gbèhounou, G., Adango, E., 2003. Trap crops of Striga hermonthica: in vitro identification and effectiveness in situ. Crop Protection 22, 395-404. Gbèhounou, G., Pieterse, A.H., Verkleij, J.A.C., 2003. Longevity of Striga seeds reconsidered: results of a field study on purple Witchweed (Striga hermonthica) in Benin. Weed Science 51, 940-946. Gbèhounou, G., Adango, E., Hinvi, J.C., Nonfon, R., 2004. Sowing date or transplanting as components for integrated Striga hermonthica control in grain cereal crops. Crop Protection 23, 379-386.

120

References

Gebremedhin, W., Goudriaan, J., Naber, H., 2000. Morphological, phenological and water-use dynamics of sorghum varieties (Sorghum bicolor) under Striga hermonthica infestation. Crop Protection 19, 61-68. Genty, B., Briantais, J.M., Baker, N.R., 1989. The relationship between the quantum yield of photosynthetic electron transport and quenching of chlorophyll fluorescence. Biochemica et Biophysica 990, 87-92. Goudriaan, J., Laar, H.H. van, 1994. Modelling Potential Crop Growth Processes. Textbook with Exercises. Kluwer Academic Publishers, Dordrecht, The Netherlands, 238 pp. Graves, J.D., Press, M.C., Stewart, G., 1989. A carbon balance model of the sorghum- Striga hermonthica host - parasite association. Plant, Cell and Environment 12, 101- 107. Graves, J.D., Wylde, A., Press, M.C., Stewart, G.R., 1990. Growth and carbon allocation in Pennisetum typhoides infected with the parasitic angiosperm Striga hermonthica. Plant, Cell and Environment 13, 367-373. Gressel, J., Hanafi, A., Head, G., Marasas, W., Obolana, A.B., Ochanda, J., Souissi, T., Tzotzos, G., 2004. Major heretofore intractable biotic constraints to African food security that may be amendable to novel biotechnological solutions. Crop Protection 23, 661-689. Gurney, A.L., Press, M.C., Ransom, J.K., 1995. The parasitic angiosperm Striga hermonthica can reduce photosynthesis of its sorghum and maize hosts in the field. Journal of Experimental Botany 46, 1817-1823. Gurney, A.L., Press, M.C., Scholes, J.D., 1999. Infection time and density influence the response of sorghum to the parasitic angiosperm Striga hermonthica. New Phytologist 143, 573-580. Gurney, A.L., Taylor, A., Mbwaga, A., Scholes, J.D., Press, M.C., 2002. Do maize cultivars demonstrate tolerance to the parasitic weed Striga asiatica? Weed Research 42, 299-306. Harbinson, J., 1995. Detection of stress in pot plants. Acta Horticultura 405, 320-334. Haussmann, B.I.G., Hess, D., Welz, H.G., Geiger, H.H., 2000. Improved methodologies for breeding Striga-resistant sorghums. Field Crop Research 66, 195-211. Haussmann, B.I.G., Hess, D.E., Reddy, B.V.S., Mukuru, S.Z., Kayentao, M., Welz, H.G., Geiger, H.H., 2001. Quantitative-genetic parameters of sorghum growth under Striga infestation in Mali and Kenya. Plant Breeding 120, 49-56. Havaux, M., Lannoye, R., 1985. Drought resistance of hardy wheat cultivars measured by a rapid chlorophyll fluorescence test. Journal of Agricultural Sciences 104, 501- 504.

121

References

Hess, D.E., Ejeta, G., Butler, L.G., 1991. Research into germination of Striga seeds by sorghum root exudates. In: Proceedings of the Fifth International Symposium of Parasitic Weeds. Eds J.K. Ransom, L.J. Musselman, A.D. Worsham, C. Parker, Nairobi, Kenya, CIMMYT, pp. 217-222. Hess, D.E., Ejeta, G., 1992. Inheritance of resistance to Striga in sorghum genotype SRN39. Plant Breeding 109, 233-241. Hess, D.E., Ejeta, G., Butler, L.G., 1992. Selecting sorghum genotypes expressing quantitative biosynthetic trait that confers resistance to Striga. Phytochemistry 31, 493-497. Hibberd, J.M., Barker, E.R., Scholes, J.D., Quick, W.P., Press, M.C., 1996. How does Orobanche aegyptiaca influence the carbon relations of tobacco and tomato? In: Advances in Parasitic Plant Research. Eds M.T. Moreno, J.I. Cubero. Cordoba, Spain, Junta de Andalucia, Consejería de Agricultura y Pesca, pp. 311-318. Hibberd, J.M., Quick, W.P., Press, M.C., Scoles, J.D., Jeschke, W.D., 1999. Solute fluxes from tobacco to the parasitic angiosperm Orobanche cernua and the influence of infection on host carbon and nitrogen relations. Plant, Cell and Environment 22, 937-947. Hirsch, A.M., Bauer, W.D., Bird, D.M., Cullimore, J., Tyler, B., Yoder, J.L., 2003. Molecular signals and receptors – controlling rhizosphere interactions between plants and other organisms. Ecology 84, 858-868. IITA (International Institute of Tropical Agriculture), 1992. Sustainable food production in sub-Saharan Africa 1. IITA’s contributions. IITA, Ibadan, Nigeria, 208 pp. Inskeep, W.P., Bloom, P.R., 1985. Extinction coefficients of chlorophyll a en b in N,N dimethylformamide and 80% acetone. Plant Physiology 77, 483-485. Jayne, T.S., Da, J.C. and Dregne, H.E., 1989. Technology and Agricultural Productivity in the Sahel. Agricultural Economic Report No. 612. Resources and Technology Division, Economic Research Service, U.S. Department of Agriculture, USA. Kanampiu, F.K., Ransom, J.K., Gressel, J., 2001. Imazapyr seed dressings for Striga control on acetolactate synthase target-site resistant maize. Crop Protection 20, 885- 895. Kebreab, E., Murdoch, A.J., 1999. A quantitative model for loss of primary dormancy and induction of secondary dormancy in imbibed seeds of Orobanche spp. Journal of Experimental Botany 50, 211-219. Ker, A., 1995. Farming systems of the African Savanna. A continent in crisis. International Development Research Centre, Ottawa, Canada. Keyes, W.J., Taylor, J.V., Apkarian, R.P., Lynn, D.G., 2001. Dancing together. Social

122

References

controls in parasitic plant development. Plant Physiology 127, 1508-1512. Kieft, H., Keita,, N., van der Heide, J., 1994. Engrais fertiles? Vers une fertilité durable des terres agricoles au Mali. ETC, Leusden, The Netherlands, 99 pp. Kropff, M.J., Schippers, P., 1986. Simulation of the growth of faba beans (Vicia faba L.) infested with broomrapes (Orobanche crenata Forsk.). In: Biology and Control of Orobanche. Ed S.J. ter Borg, Proceedings of a Workshop. Pudoc, Wageningen, The Netherlands, pp. 70-79.

Kropff, M.J., 1989. Quantification of SO2 effects on physiological processes, plant growth and crop production. PhD Thesis, Wageningen Agricultural University, The Netherlands. Kropff, M.J., Spitters, C.J.T., 1991. A simple model of crop loss by weed competition from early observations on relative leaf area of the weeds. Weed Research 31, 97- 105. Kropff, M.J., Weaver, S.E., Smits, M.A., 1992a. Use of eco-physiological models for crop-weed interference: relations amongst weed density, relative time of weed emergence, relative leaf area, and yield loss. Weed Science 40, 296-301. Kropff, M.J., Spitters, C.J.T., Schnieders, B.J., Joenje, W., de Groot, W., 1992b. An eco-physiological model for interspecific competition, applied to the influence of Chenopodium album L. on sugar beet. II. Model evaluation. Weed Research 32, 451- 463. Kroschel, J., 1999. Analysis of the Striga problem, the first step towards joint action. In: Advances in Parasitic Weed Control at On Farm Level, vol. 1. Eds J. Kroschel, H. Mercer-Quarshie, J. Sauerborn. Margraf Verlag, Weikersheim, pp. 3-25. Kuijt, J., 1969. The Biology of Parasitic Flowering Plants. University of California Press, Berkeley, USA, 246 pp. Kuijt, J., 1977. Haustoria of phanerogamic parasites. Annual Review of Phytopathology 17, 91-118. Kuiper, E., Verkleij, J.A.C., Pieterse, A.H., 1996. Differences in the primary dormancy pattern of Striga species; an ongoing study. In: Advances in Parasitic Plant Research. Eds M.T. Moreno, J.I. Cubero. Proceedings of the Sixth International Symposium on Parasitic Plants. Cordoba, Spain, pp. 441-450. Kuiper, E., 1997. Comparative studies on the parasitism of Striga aspera and Striga hermonthica on tropical grasses. PhD Thesis, Vrije Universiteit, Amsterdam. Labrousse, P., Arnaud, M.C., Serieys, H., Bervillé, A., Thalouarn, P., 2001. Several mechanisms are involved in resistance of Helianthus to Orobanche cumana Wallr. Annals of Botany 88, 859-868. Lane, J.A., Bailey, J.A., Terry, P.J., 1991. An in vitro growth system for studying the parasitism of cowpea (Vigna unguiculata) by Striga gesnerioides. Weed Research

123

References

31, 211-217. Langer, R.H.M., Hill, G.D., 1982. Agricultural plants. Cambridge University Press, Cambridge, UK, 344 pp. Lendzemo, V.W., Kuyper, Th.W., Kropff, M.J., van Ast, A. van, 2005. Field inoculation with arbuscular mycorrhizal fungi reduces Striga hermonthica performance on cereal crops and has the potential to contribute to integrated Striga management. Field Crops Research 91, 51-61. Lim, L.G., Gaunt, R.E., 1986. The effect of powdery mildew (Erysiphe graminis f. sp. hordei) and leaf rust (Puccinia hordei) on spring barley in New Zealand. I. Epidemic development, green leaf area and yield. Plant Pathology 35, 44-53. Louwerse, W., Oorschot, J.L.P. van, 1969. An assembly for routine measurements of photosynthesis, respiration and transpiration of intact plants under controlled conditions. Photosynthetica 3, 305-315. Manschadi, A.M., 1999. Modelling the growth and development of Faba bean (Vicia faba L.) infested with the parasitic weed Orobanche crenata Forsk. Agroecology 1, 1-128. Matusova, R., van Mourik, T., Bouwmeester, H.J., 2004. Changes in the sensitivity of parasitic weed seeds to germination stimulants. Seed Science Research 14, 335 344. Maxwell, K., Johnson, G.N., 2000. Chlorophyll fluorescence – a practical guide. Journal of Experimental Botany 51, 659-668. Mohamed, A.H., Ejeta, G., Butler, L.G., Housley, T.L., 1998. Moisture content and dormancy in Striga asiatica seeds. Weed Research 38, 257-265. Moran, R., 1982. Formulae for determination of chlorophyllous pigments extracted with N,N-dimethylformamide. Plant Physiology 69, 1376-1381. Musselman, L.J., 1987. Taxonomy of Witchweeds. In: Parasitic Weeds in Agriculture. Volume I. Striga. Ed L.J. Musselman. CRC Press Inc., Boca Raton, Florida, 317 pp. OIA, (Office of International Affairs), 1996. Lost crops of Africa: Volume I: Grains. The National Academic Press, Washington DC. Oijen, M. van, 1991a. Identification of the major characteristics of potato cultivars which affect yield loss caused by late blight. PhD Thesis, Wageningen Agricultural University, The Netherlands. Oijen, M. van, 1991b. Leaf area dynamics of potato cultivars infected by Phytophthora infestans. Netherlands Journal of Plant Pathology 97, 345-354. Okonkwo, S.N.C., 1991. The germination of Striga – a review. In: Proceedings of the 5th International Symposium of Parasitic Weeds. Eds J.K. Ransom, L.J. Musselman, A.D. Worsham, C. Parker. Nairobi, Kenya, pp. 144-154. Olivier, A., Benhamou, N., Leroux, G.D., 1991a. Cell surface interactions between sorghum roots and the parasitic weed Striga hermonthica: cytochemical aspects of

124

References

cellulose distribution in resistant and susceptible host tissues. Canadian Journal of Botany 69, 1679-1690. Olivier, A., Ramaiah, K.V., Leroux, G.D., 1991b. Selection of sorghum (Sorghum bicolor [L.] Moench) varieties resistant to the parasitic weed Striga hermonthica (Del.) Benth. Weed Research 31, 219-226. Orr, A., Ritchie, J.M., 2004. Learning from failure: smallholder farming systems and IPM in Malawi. Agricultural Systems 79, 31-54. Oswald, A., Ransom, J.K., Kroschel, J., Sauerborn, J., 2001. Transplanting maize and sorghum reduces Striga hermonthica damage. Weed Science 49, 346-353. Oswald, A., Ransom, J.K., 2002. Response of maize varieties to transplanting in Striga-infested fields. Weed Science 50, 392-396. Oswald, A., 2005. Striga control – technologies and their dissemination. Crop Protection 24, 333-342. Pageau, K., Simier, P., Naulet, N., Robins, R., Fer, A., 1998. Carbon dependency of the hemiparasite Striga hermonthica on Sorghum bicolor determined by carbon isotopic and gas exchange analyses. Australian Journal of Plant Physiology 25, 695- 700. Parker, C., 1965. The Striga problem – a review. Pest Articles and News Summaries 11, 99-111. Parker, C., 1991. Protection of crops against parasitic weeds. Crop Protection 10, 6- 22. Parker, C., Riches, C.R., 1993. Parasitic Weeds of the World: Biology and Control. CAB International, Wallingford, UK, 332 pp. Pieterse, A.H., Pesch, C.J., 1983. The Witchweeds (Striga spp.) – a review. Abstracts on Tropical Agriculture 9, 9-37. Pieterse A.H., Verkleij, J.A.C., 1991. Effect of soil conditions on Striga development: a review. In: Proceedings of the Fifth International Symposium on parasitic weeds. Eds J.K. Ransom, L.J. Musselman, A.D. Worsham, C. Parker. Nairobi, Kenya, pp. 329-339. Press, M.C., Stewart, G.R., 1987. Growth and photosynthesis in Sorghum bicolor infected with Striga hermonthica. Annals of Botany 60, 657-662. Press, M.C., Tuohy, J.M., Stewart, G.R., 1987. Gas exchange characteristics of the sorghum-Striga host parasite association. Plant Physiology 84, 814-819. Press, M.C., Graves, J.D., 1991. Carbon relations of parasitic angiosperms and their hosts. In: Progress in Orobanche Research. Eds K. Wegmann, L.J. Musselman. Tübingen, Germany, Eberhard-Karls-Universität, pp. 55-65. Press, M.C., Graves, J.D. (Eds), 1995. Parasitic Plants. London, UK, Chapman and Hall, 292 pp.

125

References

Press, M.C., Gurney, A.L., Frost, D.L., Scholes, J.D., 1996. How does the parasitic angiosperm Striga hermonthica influence host growth and carbon relations? In: Advances in Parasitic Plant Research. Eds M.T. Moreno, J.I. Cubero, D. Berner, D. Joel, L.J. Musselman, C. Parker, Proceedings of the Sixth International Parasitic Weed Symposium, Cordoba, Spain, pp. 303-310. Press, M.C., Gurney, A.L., 2000. Plant eats plant: sap-feeding witchweeds and other parasitic angiosperms. Biologist 47, 189-193. Ramaiah, K.V., Childley, V.L., House, L.R., 1991. A time course study of early establishment stages of the parasitic angiosperm Striga asiatica on susceptible sorghum roots. Annals of Applied Biology 118, 403-410. Ramlan, M.F., Graves, J.D., 1996. Estimation of the sensitivity to photoinhibition in Striga hermonthica-infected sorghum. Journal of Experimental Botany 47, 71-78. Ransom, J.K., Odhiambo, G.D., 1994. Long-term effects of fertility and hand weeding on Striga in maize. In: Biology and management of Orobanche. Eds A.H. Pieterse, J.A.C. Verkleij, S.J. ter Borg, Proceedings of the Third International Conference on Orobanche and Related Striga Research. Royal Tropical Institute, Amsterdam, The Netherlands, pp. 513-519. Ransom, J.K., Odhiambo, G.D., Gressel, J., 1995. Seed dressing corn with imidazolinone herbicides to control Striga hermonthica (Del.) Benth. Weed Science Society of America Abstracts 35, 15 pp. Ransom, J.K., 2000. Long-term approaches for the control of Striga in cereals: field management options. Crop Protection 19, 759-763. Rascher, U., Liebig, M., Lüttge, U., 2000. Evaluation of instant light-response curves of chlorophyll fluorescence parameters obtained with a portable chlorophyll fluorometer on site in the field. Plant, Cell and Environment 23, 1397-1405. Reda, F., 2002. Striga hermonthica in Tigray (Northern Ethiopia). Prospects for control and improvement of crop productivity through mixed cropping. PhD Thesis, Vrije Universiteit, Amsterdam, The Netherlands, 119 pp. Reda, F., Butler, L.G., Ejeta, G., Ransom, J.K., 1994. Screening of maize genotypes for low Striga asiatica stimulant production using “agar gel technique”. African Crop Science Journal 2, 173-177. Reda, F., Verkleij, J.A.C., Ernst, W.H.O., 2005. Intercropping for the improvement of sorghum yield, soil fertility and Striga control in the subsistence agriculture region of Tigray (northern Ethiopia). Journal of Agronomy & Crop Science 191, 10-19. Riches, C.R., Parker, C., 1995. Parasitic plants as weeds. In: Parasitic Plants. Eds M.C. Press, J.D. Graves. Chapman and Hall, London, UK, pp. 226 – 255. Riopel, J.L., Baird, W.V., 1987. Morphogenesis of the early development of primary haustoria in Striga asiatica. In: Parasitic Weeds in Agriculture. Volume I. Striga.

126

References

Ed. L.J. Musselman, CRC Press, Inc. Boca Raton, Florida, pp. 107-125. Riopel, J.L., Timko, M.P., 1995. Haustorial initiation and differentiation. In: Parasitic Plants. Eds M.C. Press, J.D. Graves, Chapman and Hall, London, pp. 39-79. Rodenburg, J., 2005. The role of sorghum genotype in the interaction with the parasitic weed Striga hermonthica. PhD thesis, Wageningen University, Wageningen, The Netherlands. Rodenburg, J., Bastiaans, L., Weltzien, E., Hess, D.E., 2005. How can field selection for Striga resistance and tolerance in sorghum be improved? Field Crops Research 93, 34-50. Rodenburg, J., Bastiaans, L., Kropff, M.J., Ast, A. van, 2006a. Effects of host plant genotype and seed bank density on Striga reproduction. Weed Research (in press). Rodenburg, J., Bastiaans, L., Kropff, M.J., 2006b. Characterization of host tolerance to Striga hermonthica. Euphytica (accepted). Rodenburg, J., Bastiaans, L., Schapendonk, A.H.C.M., Putten, P.E.L. van der, Ast, A. van, Kropff, M.J., 2006c. Can host plant tolerance to sorghum, Striga hermonthica be detected by photosynthesis measurements? Annals of Botany (submitted). Rolando, T.C., Wayne, R.J., Malcolm, C.D., 1992. Structural changes and associated reduction of hydraulic conductance in roots of Sorghum bicolor L. following exposure to water deficit. Plant Physiology 99, 203-212. Sallé, G., Dembéle, B., Raynal-Roques, A., Hallais, M.F., Tuquet, C., 1987. Biological aspects of Striga species, pest of food crops. In: Parasitic Flowering Plants. Eds H. Chr. Weber, W. Forstreuter. Marburg, Germany, Druckerei Kempkes, Offset + Buchdruck GmbH, 3554 Gladenbach, pp. 719-731. Sand, P.F., 1987. The American Witchweed quarantine and eradication program. In: Ed L.J. Musselman, Parasitic Weeds in Agriculture. Volume I. Striga. CRC Press, Inc. Boca Raton, Florida. pp. 207-223. Sauerborn, J., 1991. The economic importance of the phytoparasites Orobanche and Striga. In: Proceedings Fifth International Symposium on Parasitic Weeds. Eds J.K. Ransom, L.J. Musselman, A.D. Worsham, C. Parker. Nairobi, Kenya, June 24 - 30, pp. 137-143. Sauerborn, J., Sprich, H., Mercer-Quarshie, H., 2000. Crop rotation to improve agricultural production in sub-saharan Africa. Journal of Agronomy & Crop Science 184, 67-72. Saunders, A.R., 1933. Studies in phanerogamic parasitism with particular reference to Striga lutea Lour. South Africa Dept of Agriculture, Science Bulletin 128, 1-56. Schans, J., 1993. Population dynamics of potato cyst nematodes and associated damage to potato. PhD Thesis, Wageningen Agricultural University, The Netherlands.

127

References

Schapendonk, A.H.C.M., Putten, P.E.L. van der, Dolstra, O., Haalstra, R., Tonk, W.J.M., 1992. Chlorophyll fluorescence: a non-destructive method for detecting damage in the photosynthetic apparatus in plants. Acta Horticultura 304, 61-70. Schreiber, U., 1986. Detection of rapid induction kinetics with a new type of high- frequency modulated chlorophyll fluorometer. Photosynthesis Research 9, 261-272. Shah, N., Smirnoff, N., Stewart, G.R., 1987. Photosynthesis and stomatal characteristics of Striga hermonthica in relation to its parasitic habit. Physiologia Plantarum 69, 699-703. Shew, H.D., Shew, B.B., 1994. Host resistance. In: Epidemiology and Management of Root Diseases. Eds C.L. Campbell, D.M. Benson, Springer Verlag, Berlin, pp. 244- 275. Sinebo, W., Drennan, D.S.H., 2001. Vegetative growth of sorghum and Striga hermonthica in response to nitrogen and the degree of host root infection. European Journal of Plant Pathology 107, 849-860. Sivakumar, M.V.K., 1992. Climatic change and implications to agriculture in Niger. Climatic Change 20, 297-312. Smith, L.H., Keys, A.J., Evans, C.W., 1995. Striga hermonthica decreases photosynthesis in Zea mays through effects on leaf cell structure. Journal of Experimental Botany 46, 759-765. Smith, M.C., Webb, M., 1996. Estimation of the seedbank of Striga spp. (Scrophulariacea) in Malian fields and the implications for a model of biocontrol of Striga hermonthica. Weed Research 36, 85-92. Stewart, G., 1990. Witchweed: a parasitic weed of grain crops. Outlook on Agriculture 19, 115-117. Sunderland, N., 1960. The production of the Striga and Orobanche germination stimulants by maize roots. II. Conditions of synthesis in the root. Journal of Experimental Botany 11, 356-366. Taylor, A., Martin, J., Seel, W.E., 1996. Physiology of the parasitic association between maize and Witchweed (Striga hermonthica): is ABA involved? Journal of Experimental Botany 47, 1057-1065. Taylor, A., Seel, W.E., 1998. Do Striga hermonthica-induced changes in soil matric potential cause the reduction in stomatal conductance and growth of infected maize plants? New Phytologist 138, 67-73. Tennant, D., 1975. A test of a modified line intersect method of estimating root length. Journal of Ecology 63, 995-1001. Vallance, K.B., 1950. Studies on the germination of the seeds of Striga hermonthica. I. The influence of moisture treatment, stimulant dilution, and after-ripening on germination. Annals of Botany 14, 347-363.

128

References

Vogt, W., 1993. Entwickelung der Wirt-Parasit Beziehung Sorghum bicolor-Striga hermonthica unter dem Einfluß verschiedener Stickstofformen und Standortsfaktoren. PLITS 11 (3). Koch Publ. Stuttgart, 166 pp.

Watling, J.R., Press, M.C., 1997. How is the relationship between the C4 cereal

Sorghum bicolor and the C3 root hemi-parasites Striga hermonthica and Striga

asiatica affected by elevated CO2? Plant, Cell and Environment 20, 1292-1300. Watling, J.R., Press, M.C., 2000. Infection with the parasitic angiosperm Striga

hermonthica influences the response of the C3 cereal Oryza sativa to elevated CO2. Global Change Biology 6, 919-930. Watling, J.R., Press, M.C., 2001. Impacts of infection by parasitic angiosperms on host photosynthesis. Plant Biology 3, 244-250. Wayne Smith, C., Frederiksen R.A., Eds, 2000. Sorghum: Origin, History, Tech- nology, and Production. Wiley Series in Crop Science, 840 pp. Webb, M., Smith, M.C., 1996. Biology of Striga hermonthica (Scrophulariaceae) in Sahelian Mali: effects on pearl millet yield and prospects of control. Weed Research 36, 203-211. Weber, G., Elemo, K., Lagoke, S.T.O., Awad, A., Oikeh, S., 1995. Population dynamics and determinants of Striga hermonthica on maize and sorghum in savanna farming systems. Crop Protection 14, 283-290. Williams, C.N., 1960. Growth movements of the radicle of Striga. Nature 188, 1043- 1044. Williams, C.N., 1961. Tropism and morphogenesis of Striga seedlings in the host rhizosphere. Annals of Botany 25, 407-415. Wylde, A., 1991. Qualitative and quantitative studies on the effect of Striga hermonthica upon the quality and quantity of grain yield from millet (Pennisetum typhoides). In: Progress in Orobanche Research. Eds K. Wegmann, L.J. Musselman. Tübingen, Germany, Eberhard-Karls-Universität, pp. 322-328. Yoder, J.I., 2001. Host-plant recognition by parasitic Scrophulariaceae. Current Opinion in Plant Biology 4, 359-365.

129

Summary

Striga hermonthica (Del.) Benth., often known as witchweed, is an obligate root parasitic flowering plant of the family Orobanchacea. The species is native to Africa and is one of the most important growth-reducing factors in crops in vast areas of the Savannah zone in Africa. S. hermonthica negatively affects the growth of its host plant like sorghum (Sorghum bicolor [L.] Moench). Problems with S. hermonthica become worse with declining soil fertility, a consequence of agricultural intensification by resource-poor farmers. Seeds of the parasite germinate after stimulation by exudates from the roots of a host plant and the germinated seeds then attach to the host plant root system. Striga negatively affects its host by subtracting assimilates, water and nutrients. Furthermore, the parasite exerts negative influences on the host through a so-called pathological effect. This pathological effect expresses itself through physiological and metabolic disturbances, including disruption of carbon fixation and partitioning and changes in nitrogen metabolism and the plant hormone balance. Most effects occur almost immediately following infection, indicating that already shortly after attachment the parasite interferes with host metabolism. The extent to which the parasite reduces the growth of its sorghum host is highly variable and depends on factors like host plant genotype, parasite infestation level and environmental conditions. Several Striga control options have been developed but none of these measures on its own is both sufficiently effective as well as practical and accessible for the resource poor farmer. It is becoming increasingly clear that integration of various control options, based on a good understanding of the interaction between host and parasite, is the only possible approach to control Striga. In integrated Striga management the role of crop varieties with improved Striga defence mechanisms is essential. The sorghum - Striga relationship is characterized by mutual interactions. Infection by the parasite affects the host through several effects of which withdrawal of resources and a negative effect on photosynthetic rate of the host are the most important. These effects are known to have severe effects on growth and yield of the sorghum host. At the same time, these Striga-related effects have direct implications for the quality of the sorghum plant as a host and, as a consequence, for the further development of the parasite. This latter might have its consequences at the individual (growth of attached parasites) and at the population level (the number of attached parasites). This mutual interaction between sorghum and S. hermonthica, which is characterized by feedback mechanisms, is not yet completely understood. Because of the great importance of the relationship for both crop yield and parasite reproduction,

131

Summary the main objective of the current research was to gain more insight into the host plant - parasite relationship. In the first study of this thesis (Chapter 2), the objective was to gain more insight into the tolerance mechanisms of sorghum. For this purpose, the sensitive sorghum cv. CK60-B and the tolerant sorghum landrace Tiémarifing were grown in pots in a greenhouse in Wageningen in the presence and absence of Striga. It was found that both genotypes responded to parasite infection, but it was obvious that CK60-B was more strongly affected than Tiémarifing. In the sensitive cultivar, plant height, green leaf area and kernel yield were strongly reduced as a result of Striga infection. The strong reduction in kernel yield (95%) in this cultivar resulted, to a large extent, from the fact that only 40% of the infected plants produced kernels. No effect of Striga infection on the phenological development of Tiémarifing was observed and yield reduction was relatively small (35%) compared to that of the sensitive cultivar. Another interesting observation was that the tolerant landrace showed significantly lower and delayed emergence of the parasite than the sensitive cultivar. This delay in emergence could be explained by a delay in the onset of attachment which, most likely, resulted from differences in root architecture, with relatively more roots in deeper soil layers in the tolerant landrace. From this study it was concluded that the early infection was at least partly responsible for the stronger effects of the parasite on the sensitive cultivar. Chapter 3 describes the results of two pot experiments that were conducted to study the effects of S. hermonthica on the photosynthetic rate of the host. Earlier studies indicated that the ability of tolerant sorghum genotypes to maintain high rates of photosynthesis when infected with Striga is an important characteristic of tolerant genotypes. This phenomenon was further studied and for this purpose a comparison was made between the response of the sensitive cultivar CK60-B and the tolerant cultivar Tiémarifing. With CK60-B, the photosynthetic rate at light saturation as well as the initial light use efficiency was significantly reduced (48% and 27%, respectively) following parasite infection. With the tolerant cultivar Tiémarifing, Striga infection did not affect both parameters, supporting the earlier observations that this attribute is an important mechanism behind tolerance. In a further experiment the relationship between Striga infection level and the magnitude of the reduction in photosynthetic rate of the host was determined. For this purpose the photosynthetic response of the sensitive cultivar was studied at a wide range (8 - 2000 seeds dm–3 soil) of Striga seed infestation levels. This study revealed that already a few Striga attachments were sufficient to bring about the maximum reduction in leaf photosynthetic rate of the host, which was shown to be 40-50% of the photosynthetic rate of uninfected control plants. In most instances, this maximum

132

Summary reduction was already obtained before emergence of the Striga plants. Consequently, the relation between number of emerged Striga plants and the reduction in leaf photosynthetic rate could simply be represented by a horizontal line that reflected the maximum reduction level. However, even though the number of emerged Striga plants was found not to have a direct influence on the magnitude of the reduction in photosynthesis, Striga infestation level still showed to have a clear indirect effect on the extent of photosynthesis-related damage caused by the parasite. It was found that Striga infestation level also affected the time of infection, with earlier infection at higher infestation levels, and, consequently, an earlier onset of the negative effect of Striga on the photosynthesis of its host. As Striga is a root-parasitic plant, it is difficult and laborious to determine the actual infection level. For practical reasons, the number of emerged Striga plants is therefore often used as an indication of infection level. This number is however a reflection of two successive processes: attachment and emergence. For this reason, the objective of the research described in Chapter 4 was to study in more detail the relation between Striga seed infestation level of the soil, the infection level of the host and the performance of host plant and parasite. The experiment clearly demonstrated that yield reduction of sorghum due to Striga-infection could be attributed to three processes: a reduced dry matter production of the host, the withdrawal of host resources by the parasite and a reduced harvest index, which was associated with an altered allocation of assimilates within the host plant. Yield reduction of sorghum was found to increase with Striga infestation level, and this increase was associated with an earlier time of first Striga emergence and a higher maximum number of emerged Striga plants

(NSmax). To further clarify the relation between infestation level and kernel yield loss, the relationship was dissected into two separate components: the relation between infestation level and NSmax, reflecting host resistance, and the relation between NSmax and kernel yield loss, reflecting host tolerance. The relation between soil infestation level and NSmax developed according to a rectangular hyperbola, with a gradually lower increase in number of emerged Striga plants at higher infestation levels. This density-dependence hints at intra-specific competition during the subterranean phase. With the highest Striga infestation levels, intra-specific competition was also observed during the aerial phase of the parasite and reflected in a decrease in number of emerged Striga plants towards the end of the growing period of sorghum. A linear relation was found between NSmax and sorghum kernel yield loss, until at the highest infection levels a nearly complete loss was attained. The initial increase in sorghum kernel yield loss per emerged Striga plant showed to be 4.3%. Despite the reduction in host plant vigour at higher infestation levels, the parasite was able to increase (moderate infestation levels) or at least maintain (higher

133

Summary infestation levels) its dry matter production. This was mainly because the fraction Striga of the total dry matter production of the host-parasite association significantly increased with infestation level, reaching a value of nearly 70% at the highest infestation level. Despite the difference in emergence time of Striga plants as a result of different Striga seed densities in the soil, flowering of the parasite was always obtained around 71 days after sorghum emergence, indicating a synchronization of flowering time of Striga plants, which can be considered as a useful characteristic for S. hermonthica, because this species appears to be an obligate outcrosser. Though the results of the earlier chapters suggest a clear effect of Striga infection time on sorghum production, time of Striga infection was either confounded with genotype (Chapter 2) or Striga infection level (Chapters 3 and 4). For that reason experiments, using CK60-B and Tiémarifing, were conducted to further resolve the influence of time of S. hermonthica infection on the interaction between host and parasite. In these experiments, of which the results are presented in Chapter 5, delays in first Striga attachment of one and two weeks were artificially established. Only with CK60-B the interaction between host and parasite was significantly affected by delayed first Striga attachment. The parasite appeared most sensitive for delayed infection; biomass of the Striga plants was already significantly reduced after a one week delay in infection time. With a further week delay, an additional reduction in parasite biomass was accompanied by a strong increase in plant and panicle dry weight of the host. With the tolerant cultivar Tiémarifing, in which even under normal circumstances first attachment was already relatively late, any further delay in first Striga infection did not result in significant effects on Striga biomass or host plant performance. The results of these experiments thus indicate that the influence of delayed infection on both host and parasite strongly depends on actual infection time and confirm that earlier observed differences in time of first infection between the two cultivars do contribute to the more tolerant response of Tiémarifing to Striga-infection. Based on the results of these experiments, it was concluded that new studies have to be directed at developing or improving control methods that are based on the principle of separating Striga seeds and the sorghum root system during early development stages of the crop. The objective of the experiments presented in Chapter 6 was to find out whether shallow-tillage, deep planting, the use of transplants or a combination of these measures would lead to a delay in Striga attachment of sorghum. A pot experiment conducted in a greenhouse in Wageningen showed that application of the combination of al three measures resulted in a four-week delay in first emergence of the parasite, a reduced number of emerged Striga plants and a decimation of Striga dry weight. With the sensitive cultivar CK60-B, a nearly complete yield loss (92%) under standard practice was reduced to 26% when all three measures were combined. With

134

Summary the tolerant cultivar Tiémarifing, the combination of all three measures completely cancelled out a kernel yield loss of 28%. Application of the proposed cultural control measures in a field experiment in Mali showed that, despite a strong reduction in Striga infection level, the desired delay in Striga infection time was not achieved. The measures resulted in severe reductions in Striga dry matter, but significant improvements in sorghum grain yield were not obtained. It was argued that the presence of a natural Striga soil seedbank throughout the tilth and the difficulty of maintaining planting holes in the field might be important reasons for the large discrepancy between the results obtained in the pot and the field experiment. Though controlling the production of new Striga seeds is an important element of any long- term Striga management programme, the absence of a short term beneficial effect makes farmers’ acceptance of the proposed measures far less likely. In Chapter 7 the overall conclusions of this study are discussed in the context of the Striga problem in the sub-Saharan zone of Africa. It was argued that the large potential of a delayed contact between host root and parasite to contribute to a reduced negative impact of the parasite on cereal production, as demonstrated in this thesis, deserves further investigation. The application of herbicide seed dressings combined with the use of herbicide resistant or tolerant cereal varieties, as explored in a recent collaborative study, was indicated to be one of the potential options in this direction.

135

Résumé

Le Striga hermonthica (Del.) Benth., souvent plus connu sous le nom d’herbe des sorcières, est une plante florissante parasite obligatoire des racines de la famille Orobanchaceae. L’espèce est originaire d’Afrique et est l’un des facteurs les plus importants de diminution de croissance des cultures dans les vastes étendues de la zone de savane Africaine. Le S. hermonthica affecte négativement la croissance de sa plante hôte, par exemple le sorgho (Sorghum bicolor [L.] Moench). Les problèmes liés à S. hermonthica s’aggravent avec le déclin de la fertilité des sols, conséquence de l’intensification de l’utilisation des terres par les paysans pauvres. Les graines du parasite germent après stimulation des exsudats racinaires d’une plante hôte et ces graines germées s’attachent ensuite au système de racines de la plante hôte. Le Striga affecte négativement son hôte en lui soustrayant ses assimilats, son eau et ses nutriments. De manière encore plus importante, le parasite exerce une influence négative sur l’hôte par un effet ‘pathologique’. Cet effet pathologique s’exprime par des dérèglements physiologiques et métaboliques, incluant la réduction du taux photosynthétique, des changements dans la partition du carbone, dans le métabolisme du nitrogène et dans l’équilibre hormonal de la plante. La plupart des effets ont lieu immédiatement après l’infection, ce qui indique que le parasite interfère avec le métabolisme de son hôte dès la fixation. La mesure dans laquelle le parasite diminue la croissance de son hôte sorgho varie fortement et dépend de facteurs comme le génotype de la plante hôte, le niveau d’infestation parasitaire et les conditions environnementales. De nombreuses options de contrôle du Striga ont été développées mais aucune de ces mesures n’a à elle seule été suffisamment efficace aussi bien que pratique et accessible aux agriculteurs pauvres. Il est de plus en plus clair que l’intégration de mécanismes de contrôle variées, basée sur une bonne connaissance de l’interaction entre l’hôte et son parasite est la seule approche possible pour contrôler le Striga. Dans la gestion intégrée du Striga, le rôle des variétés de cultures ayant un mécanisme de tolérance ou résistance au Striga est essentiel. La relation Striga - sorgho est caractérisée par une interaction mutuelle. L’infection par le parasite affecte l’hôte par différents effets dont les plus importants sont la subtilisation de ses ressources et l’effet négatif sur le taux photosynthétique de l’hôte. Ces effets ont des conséquences graves sur la croissance et le rendement de l’hôte sorgho. En même temps, ils ont des implications directes sur la qualité du plant de sorgho hôte et, en conséquence, sur le développement du parasite lui-même. Ceci peut avoir des conséquences au niveau individuel (croissance des parasites fixés) et au

137

Résumé niveau de la population (nombre de parasites fixés). Cette interaction mutuelle entre le sorgho et le S. hermonthica, qui est caractérisée par des mécanismes de rétroaction n’est pas encore totalement comprise. A cause de la grande importance de cette relation pour le rendement en grains du sorgho et la reproduction du parasite, le principal objectif de cette recherche était de mieux comprendre la relation plante hôte - parasite. L’objectif de la première étude de cette thèse (Chapitre 2) était de mieux comprendre le mécanisme de tolérance du sorgho contre le Striga. Dans cette optique, le sorgho sensible cv. CK60-B et le sorgho variété locale tolérant Tiémarifing ont été cultivé en pots dans une serre de Wageningen en présence et en absence de Striga. Les deux génotypes ont répondu à l’infection du parasite, mais il était évident que CK60-B était plus fortement affecté que Tiémarifing. Dans le cultivar sensible, la taille des plantes, la surface des feuilles vertes et la production de graines ont été fortement réduits par l’infection de Striga. La forte réduction en production de grains (95%) de ce cultivar a résulté, en grande partie, du fait que seules 40% des plantes infectées ont produit des graines. Aucun effet de l’infection de Striga sur le développement phénologique de Tiémarifing n’a été observé et la réduction de production des grains a été relativement faible (35%) comparée à celle du cultivar sensible. Une autre observation intéressante a été que le variété locale tolérant a montré une émergence du parasite significativement plus basse et tardive que le cultivar sensible. Ce retard dans l’émergence pourrait être expliqué par un retard dans le départ de la fixation qui a probablement résulté de différences dans l’architecture des racines, avec relativement plus de racines dans les couches de sol en profondeur pour le variété locale tolérant. De cette étude, nous avons conclu que l’infection précoce était au moins en partie responsable des effets plus importants du parasite sur le cultivar sensible. Le Chapitre 3 décrit les résultats de deux expériences en pot conduites pour étudier les effets de S. hermonthica sur le taux photosynthétique des feuilles de l’hôte. Des études préalables ont indiqué que la capacité des génotypes du sorgho tolérant à maintenir des taux plus élevés de photosynthèse quand ils sont infectés au Striga est une caractéristique importante. Dans cette étude, ce phénomène a été étudié plus en profondeur et dans ce but une comparaison a été faite entre la réponse du cultivar sensible CK60-B et le variété locale tolérant Tiémarifing. Avec CK60-B, le taux photosynthétique à la saturation lumineuse ainsi que l’efficacité de l’utilisation de lumière initiale ont été réduites de manière significative (respectivement 48% et 27%) après l’infection parasitaire. Dans le cultivar tolérant Tiémarifing, l’infection de Striga n’a pas affecté les deux paramètres, confirmant les observations préalables que cet attribut est un mécanisme de tolérance important. Dans une expérience supplémentaire, la relation entre le niveau d’infection de

138

Résumé

Striga et la magnitude de la réduction du taux photosynthétique de l’hôte a été déterminée. Dans cette optique, la réponse photosynthétique du cultivar sensible a été étudiée à travers une grande variation de densité des graines de Striga (8 - 2000 graines dm–3 de sol). Cette étude a révélé qu’un petit nombre de fixations de Striga suffisaient déjà à occasionner la réduction maximum de taux photosynthétique des feuilles de l’hôte, qui a été déterminée à 40-50% du taux photosynthétique des plantes contrôlées non infectées. Dans la plupart des cas, cette réduction maximum était déjà obtenue avant l’émergence des plants de Striga. En conséquence, la relation entre le nombre de plants de Striga émergés et la réduction du taux photosynthétique des feuilles pouvait simplement être représentée par une ligne horizontale qui représentait le taux de réduction maximum. Toutefois, même s’il a été prouvé que le nombre de plants de Striga émergés n’a pas d’influence directe sur la magnitude de la réduction de la photosynthèse, le niveau d’infestation de Striga avait un effet indirect clair sur l’étendue de dommages relatifs à la photosynthèse causés par le parasite. Nous avons découvert que le niveau d’infestation de Striga affectait aussi le moment de l’infection, avec une infection plus précoce à des niveaux d’infection plus importants, et, en conséquence, un départ plus précoce des effets négatifs de Striga sur la photosynthèse de son hôte. Comme le Striga est un parasite des racines, il est difficile et laborieux de déterminer le niveau réel d’infection. Pour des raisons pratiques, le nombre de plants de Striga émergés est pour cela souvent utilisé comme indicateur du niveau d’infection. Ce nombre reflète deux processus successifs: la fixation et l’émergence. Pour cette raison, l’objectif de la recherche décrite dans le Chapitre 4 était d’étudier plus en détail la relation entre le niveau d’infection de graines de Striga dans le sol, le niveau d’infection de l’hôte et la réaction de la plante hôte et du parasite. L’expérience a clairement démontré que la réduction de la production du sorgho due à l’infection au Striga pouvait être attribuée à trois processus: une production réduite de matière sèche de l’hôte, la subtilisation des ressources de l’hôte par le parasite et un indice de rendement réduit, associé avec une distribution altérée des assimilats dans la plante hôte. La réduction de production du sorgho s’est avérée diminuer avec le niveau d’infestation au Striga, et cette diminution a été associée d’un côté a l’ émergence de Striga plus précoce et d’autre côté à un nombre maximum de plants de Striga émergés plus élevé (NSmax) par pot. Pour mieux comprendre la relation entre le niveau d’infestation et la perte en rendement des graines, la relation a été séparée en deux composants: (i) la relation entre le niveau d’infestation et NSmax, reflétant la résistance de l’hôte, et (ii) la relation entre NSmax et la perte en production de graines, reflétant la tolérance de l’hôte. La relation entre le niveau d’infestation des sols et NSmax s’est développée selon une hyperbole rectangulaire, avec une diminution graduelle de

139

Résumé l’augmentation du nombre de plants de Striga émergents aux niveaux d’infestation les plus hauts. La densité-dépendance fait allusion à une compétition intraspécifique entre les Striga qui s’attachent pendant la phase souterraine. Au niveaux les plus hauts d’infestation au Striga, la compétition intraspécifique a aussi été observée pendant la phase aérienne du parasite et s’est reflétée dans une diminution du nombre de plants de Striga émergés vers la fin de la période de croissance du sorgho. Une relation linéaire a été trouvée entre NSmax et la perte de production de graines du sorgho jusqu’à ce qu’une perte presque complète soit atteinte au deux niveaux les plus hauts d’infection. La diminution de l’augmentation initiale des pertes des rendements en grains par plant de Striga émergé fut de 4.3%. Malgré la perte de vigueur de la plante hôte aux niveaux d’infestation des sols plus hauts, le parasite a été capable d’augmenter (niveaux d’infestation modérés) ou du moins de maintenir (niveaux d’infection plus hauts) sa production de matière sèche. Ceci fut surtout attribué au fait que la part Striga dans la production totale de matière sèche de l’association hôte - parasite a significativement augmenté avec le niveau d’infestation, atteignant une valeur de presque 70% au plus haut niveau d’infection. Malgré la différence de moment d’émergence des plants de Striga résultant de différentes densités de graines de Striga dans le sol, la floraison du parasite était toujours obtenue 71 jours environ après l’émergence du sorgho, indiquant une synchronisation du temps de floraison des plants de Striga, qui peut être considéré comme une caractéristique utile pour S. hermonthica car cette espèce se révèle une espèce d’un croisement extérieur obligatoire. Bien que les résultats des chapitres précédents suggèrent un effet évident du moment d’infection de Striga sur la production du sorgho, le moment de l’infection de Striga a été confondu soit avec le génotype (Chapitre 2), soit avec le niveau d’infection de Striga (Chapitres 3 et 4). Des expériences ont ainsi été conduites avec CK60-B et Tiémarifing pour mieux comprendre l’influence du moment de l’infection de S. hermonthica sur l’interaction entre l’hôte et le parasite. Dans ces expériences en serre, dont les résultats sont présentés dans le Chapitre 5, des délais de première fixation de Striga d’une et deux semaines ont été établis artificiellement. L’interaction entre l’hôte et le parasite n’a été affecté significativement par le report du moment de la première fixation de Striga qu’avec CK60-B. Le parasite est apparu plus sensible aux infections tardives; la biomasse des plants de Striga était déjà réduite de manière significative après un report d’une semaine du moment de l’infection. Avec un report d’une semaine supplémentaire, une réduction additionnelle de la biomasse du parasite a été accompagnée par une forte augmentation de la masse sèche de la plante et de la panicule de l’hôte. Dans le cultivar tolérant Tiémarifing, dans lequel, même dans des circonstances normales, la première fixation avait lieu relativement tard, tout délai

140

Résumé supplémentaire de la première infection de Striga n’eut pas d’effets signifiants sur la biomasse du Striga ou sur la performance de la plante hôte. Les résultats de ces expériences indiquent toutefois que l’influence d’une infection retardée sur l’hôte et sur le parasite dépend fortement du moment réel de l’infection et confirment que les différences de moment de la première infection observées plus tôt entre les deux cultivars contribuent au fait que la réponse de Tiémarifing soit plus tolérante à l’infection de Striga. En se basant sur les résultats de ces expériences, on a conclu que de nouvelles études doivent être faites afin de développer ou d’améliorer les méthodes de contrôle basées sur le principe de séparation des graines de Striga du système des racines du sorgho dans le temps et l’espace, surtout pendant les premières étapes de développement de la récolte. L’objectif de l’expérience présentée dans le Chapitre 6 était de savoir si le labour superficiel du sol, le semis en profondeur, l’utilisation de transplants ou une combinaison de ces mesures pourraient résulter en un retard de la fixation du Striga sur le sorgho. Une expérience en pot conduite en serre à Wageningen a montré que l’application de la combinaison des ces trois mesures résultait en un report de quatre semaines de la première émergence du parasite, un nombre réduit de plants de Striga émergés et une décimation de la matière sèche du Striga. Dans le cultivar sensible CK60-B, une perte presque complète (92%) des récoltes avec les pratiques standards à été réduite à 26% quand les trois mesures sont combinées. Dans le cultivar tolérant Tiémarifing, la combinaison des trois mesures a totalement annulé la perte de récolte de 28%, observée avec les pratiques standard. L’application des mesures de contrôle des cultures proposées lors d’une expérience sur le terrain au Mali a montré, malgré une forte réduction du niveau de l’infection de Striga, que l’on n’atteignait pas le report désiré du moment de l’infection au Striga. Les mesures ont résulté en de sévères réductions de la matière sèche du Striga mais on n’a pas obtenu d’améliorations significatives de la récolte des grains de sorgho. On argua du fait que la présence de bancs naturels de grains de Striga tout au long du labour et que la difficulté à créer et maintenir des trous profonds de plantation dans les champs pourraient être des raisons importantes du large fossé entre les résultats obtenus dans le pot et lors de l’expérience sur le terrain. Bien que le contrôle de la production de nouvelles graines de Striga soit un élément important de tout programme de gestion du Striga sur le long terme, l’absence de bénéfice sur le court terme rend l’acceptation des mesures proposées par les fermiers beaucoup moins probable. Dans le Chapitre 7 les conclusions générales de cette étude sont discutées dans le contexte du problème du Striga dans la zone de l’Afrique subsaharienne. Nous soutenons le grand potentiel du report du contact entre la racine hôte et le parasite, de

141

Résumé contribue à l’amélioration de production de sorgho dans les sols infestés par le Striga. L’application de semence de sorgho enveloppé par un taux modeste d’herbicide, en combinaison avec l’utilisation des génotypes de sorgho qui sont tolérants ou résistants à cet herbicide, comme exploré dans d’autres études, pourrait être une option qui ralentit l’infection par le Striga.

142

Samenvatting

Striga hermonthica (Del.) Benth. is een obligaat parasitaire plant uit de Orobanchaceae familie. De soort komt van oorsprong voor in Afrika en hecht zich aan de wortels van belangrijke voedselgewassen als maïs, sorghum en gierst. In grote delen van de savannezone in Afrika heeft Striga zich ontwikkeld tot de belangrijkste groeireducerende factor. Veelvuldig is waargenomen dat de problemen met Striga toenemen wanneer de bodemvruchtbaarheid afneemt. Dit laatste treedt met name op ten gevolge van de toenemende intensiteit in landgebruik door arme boeren. In dit proefschrift is de interactie tussen Sorghum bicolor (L.) Moench en S. hermonthica onderzocht. De interactie tussen gastheer en Striga start met de kieming van het zaad van de parasiet, waarbij chemische verbindingen, uitgescheiden door de wortels van de gastheerplant, de kieming induceren. De kiemende zaden produceren een kiemworteltje met aan het uiteinde een haustorium dat de wortel van de waardplant binnendringt. Vervolgens wordt een verbinding tot stand gebracht met de xyleemvaten in de wortel. Striga remt de groei van zijn gastheer door de onttrekking van assimilaten, water en nutriënten. Echter, het belangrijkste effect van de parasiet op de waardplant is het zogenaamde pathologische of fytotoxische effect. Dit pathologische effect omvat de verstoring van een aantal essentiële processen van de waardplant, zoals een reductie van de fotosynthesesnelheid, een verandering van de drogestofverdeling en een ontregeling van het stikstofmetabolisme en de hormoonhuishouding. Vrijwel onmiddellijk na aanhechting van de parasiet begint de verstoring van het metabolisme van de gastheer. De mate waarin Striga de groei van de waardplant remt, is erg variabel en hangt af van diverse factoren, zoals de cultivar, de mate van infectie en van omgevingsfactoren. In de loop van de afgelopen decennia zijn er diverse beheersmaatregelen tegen Striga ontwikkeld, maar geen van deze maatregelen is zowel voldoende effectief als praktisch toepasbaar gebleken. Tevens zijn bepaalde maatregelen ontoegankelijk gebleken voor arme boeren. Het wordt in toenemende mate duidelijk dat integratie van verschillende beheersmaatregelen de enige optie is om tot een adequate beheersing van Striga te komen. Kennis van de interactie tussen parasiet en gastheerplant is hierbij van essentieel belang, waarbij duidelijk is dat cultivars met een verhoogde tolerantie en resistentie in ieder geval een wezenlijk onderdeel vormen van een geïntegreerd bestrijdingspakket. De sorghum-Striga associatie wordt gekarakteriseerd door wederzijdse beïnvloeding. Groei en ontwikkeling van de waardplant worden ernstig verstoord na infectie met de parasiet. Tegelijkertijd hebben de negatieve effecten op de gastheer

143

Samenvatting gevolgen voor de kwaliteit van de sorghumplant als waardplant en daarmee op de verdere ontwikkeling van de parasiet. Dit laatste kan zich uiten op zowel het individuele niveau (groei van de aangehechte parasiet), als op het niveau van de populatie (het aantal aanhechtingen). Van deze wederzijdse beïnvloeding van sorghum en S. hermonthica, die gekarakteriseerd wordt door terugkoppelingsmechanismen, en die van belang is voor zowel gewasopbrengst als reproductie van de parasiet, is nog onvoldoende bekend. Vandaar dat het belangrijkste doel van dit onderzoek was om het inzicht in de interactie tussen gastheer en parasiet te vergroten. Het doel van de in Hoofdstuk 2 beschreven experimenten was om meer inzicht te krijgen in de tolerantiemechanismen van sorghum tegen Striga. Voor dit doel werden planten van de Striga-gevoelige sorghumcultivar CK60-B en van het Striga-tolerante landras Tiémarifing opgekweekt in potten met en zonder S. hermonthica. Beide sorghumcultivars werden door Striga beïnvloed, maar het was evident dat de effecten van de parasiet op CK60-B veel sterker waren dan op Tiémarifing. Bij de gevoelige cultivar werden planthoogte, het groene bladoppervlak en de korrelopbrengst sterk gereduceerd ten gevolge van de Striga-infectie. De sterke reductie in korrelopbrengst (95%) werd voor een groot deel veroorzaakt door het feit dat slechts 40% van de geïnfecteerde sorghumplanten korrels produceerde. Bij Tiémarifing werd er geen effect van Striga-infectie op de fenologische ontwikkeling gevonden. Ook de reductie van de korrelopbrengst (35%) was gering in vergelijking met die van de gevoelige cultivar. Een andere interessante waarneming was dat in het tolerante landras significant minder Striga planten opkwamen en dat het opkomsttijdstip van deze Striga planten substantieel later was. Dit verlate opkomsttijdstip van Striga bij de tolerante cultivar kon worden verklaard door een relatief late aanhechting van de parasiet aan de wortels van de gastheer. Wellicht dat dit samenhangt met de verschillen in wortelarchitectuur tussen beide sorghumcultivars, waarbij van het tolerante landras relatief veel wortels in het onderste gedeelte van de pot werden aangetroffen. Op basis van deze proeven werd geconcludeerd dat de vroege infectie op zijn minst gedeeltelijk verantwoordelijk was voor de sterkere effecten van de parasiet op de gevoelige sorghumcultivar. De resultaten van twee potproeven die werden uitgevoerd om de effecten van S. hermonthica op de bladfotosynthesesnelheid van de gastheer te bestuderen, staan beschreven in Hoofdstuk 3. De bladfotosynthesesnelheid van tolerante cultivars blijkt vaak ongevoelig voor infectie met de parasiet. Dit fenomeen werd nader bestudeerd, waarbij opnieuw planten van de gevoelige cultivar CK60-B vergeleken werden met planten van het tolerante landras Tiémarifing. Bij CK60-B werden zowel de bladfotosynthesesnelheid bij lichtverzadiging (48%) als de initiële lichtbenuttings- efficiëntie (27%) significant gereduceerd ten gevolge van infectie met Striga. Bij het

144

Samenvatting tolerante landras Tiémarifing werd daarentegen geen effect van Striga gevonden op beide fotosyntheseparameters, hetgeen bevestigt dat deze eigenschap een belangrijk tolerantie-mechanisme is. In een aanvullende proef werd de relatie tussen het Striga-infectieniveau en de reductie in bladfotosynthesesnelheid vastgesteld. Voor dit doel werd de fotosynthesesnelheid van de gevoelige cultivar bij een brede range besmettingsniveaus van de grond (8 - 2000 Striga zaden per dm3) bepaald. Slechts een klein aantal Striga infecties, behaald bij lage besmettingsniveaus van de grond, bleek al voldoende om de maximale reductie (40-50%) in bladfotosynthesesnelheid te bewerkstelligen. Deze maximale reductie van de fotosynthesesnelheid werd in de meeste gevallen reeds bereikt voordat de Striga planten waren opgekomen. Dientengevolge kon de relatie tussen het aantal opgekomen Striga planten en de mate van reductie van de bladfotosynthesesnelheid simpelweg worden weergegeven door een horizontale lijn die de maximale reductie weergeeft. Hoewel het aantal opgekomen Striga planten dus niet van invloed bleek te zijn op de mate van reductie van de bladfotosynthesesnelheid, bleek het Striga-besmettingsniveau van de grond wel een indirect effect te hebben op de omvang van de fotosynthese-gerelateerde schade. Bij een hoger besmettingsniveau werd de waardplant eerder geïnfecteerd en zodoende trad het negatieve effect van Striga op de bladfotosynthesesnelheid in een vroeger stadium op. Doordat Striga een parasitaire plant is die zich aan de wortels van zijn gastheer hecht, is het, zeker onder veldomstandigheden, technisch moeilijk en arbeidsintensief om het actuele infectieniveau te bepalen. Om praktische redenen is het daarom gebruikelijk om het aantal opgekomen Striga planten als maat voor het infectieniveau te gebruiken. Dit aantal is echter een reflectie van twee opeenvolgende processen: aanhechting en opkomst. Het in Hoofdstuk 4 beschreven onderzoek bestudeert de relatie tussen het Striga-besmettingsniveau van de grond (zaden per dm3 grond), het infectieniveau van de waardplant (infecties per plant) en de reactie van sorghum en Striga hierop. Het experiment maakte duidelijk dat opbrengstreductie van sorghum ten gevolge van de Striga-infectie kan worden toegeschreven aan drie processen: een verminderde drogestofproductie van de gastheerplant, onttrekking van assimilaten en nutriënten door de parasiet en een verlaagde harvest index. Dit laatste bleek gekoppeld aan een gewijzigde drogestofverdeling binnen de waardplant. Het opbrengstverlies van de sorghum nam toe met een oplopend Striga-besmettingsniveau van de grond. Deze toename in opbrengstverlies bleek gerelateerd aan een vroeger opkomsttijdstip van

Striga en een groter aantal maximaal opgekomen Striga planten per pot (NSmax). Om de relatie tussen het Striga-besmettingsniveau van de grond en het verlies aan korrelopbrengst verder te verklaren, werd deze relatie onderverdeeld in twee afzonderlijke componenten: (i) de relatie tussen besmettingsniveau van de grond en

145

Samenvatting

NSmax, die het resistentieniveau van de gastheer weergeeft, en (ii) de relatie tussen

NSmax en het verlies aan korrelopbrengst, die het tolerantieniveau van de waardplant weerspiegelt. De relatie tussen het Striga-besmettingsniveau van de grond en NSmax ontwikkelde zich volgens een rechthoekige hyperbool met een geleidelijk steeds kleiner wordende toename in het aantal opgekomen Striga planten bij hogere besmettingsniveaus. Deze dichtheidsafhankelijke relatie wijst op intraspecifieke concurrentie tussen Striga-aanhechtingen tijdens de ondergrondse fase. Bij de hoogste Striga-besmettingsniveaus van de grond werd ook intraspecifieke concurrentie tijdens de bovengrondse fase van de parasiet waargenomen. Dit laatste uitte zich in een teruglopend aantal Striga planten tegen het einde van het groeiseizoen. Tussen NSmax en het verlies aan korrelopbrengst werd een lineaire relatie waargenomen. De toename in korrelopbrengstverlies van de sorghum bedroeg 4.3% per opgekomen Striga plant, totdat bij de twee hoogste infectieniveaus van de plant het opbrengstverlies bijna volledig was en dus niet verder toenam. Ondanks de terugloop in vitaliteit van de gastheerplant bij een toenemend Striga- besmettingsniveau van de grond, bleek de parasiet in staat om zijn drogestofproductie toe te laten nemen (bij gematigde besmettingsniveaus van de grond) of in ieder geval te handhaven (bij hogere besmettingsniveaus). Dit was vooral toe te schrijven aan het feit dat bij toenemend besmettingsniveau een steeds groter deel van de totale drogestofproductie van de gastheer - parasiet associatie ten goede kwam aan de parasiet. Bij het hoogste besmettingsniveau bedroeg de fractie Striga bijna 70% van de totale drogestofproductie. Ondanks een verschil in Striga-opkomsttijdstip bij verschillende Striga zaaddichtheden, bleek de aanvang van de bloei van de parasiet altijd plaats te vinden rond 71 dagen na opkomst van de sorghum. Dit wijst op een synchronisatie van het bloeitijdstip van Striga, hetgeen als een functionele eigenschap kan worden beschouwd, aangezien deze soort als obligate kruisbestuiver bekend staat. Hoewel de voorgaande resultaten een duidelijk effect van het infectietijdstip van Striga op de sorghumproductie suggereren, bleek het tijdstip van infectie verstrengeld te zijn met de sorghumcultivar (Hoofdstuk 2) of met het Striga-infectieniveau van de waardplant (Hoofdstukken 3 en 4). Om de invloed van infectietijdstip op de interactie tussen waardplant en parasiet verder uit te diepen, werden aanvullende kasproeven uitgevoerd. In deze proeven werden op kunstmatige wijze vertragingen van één en twee weken in de eerste aanhechting van Striga aangebracht (Hoofdstuk 5). Bij CK60- B werd de interactie tussen waardplant en parasiet significant beïnvloed door het tijdstip van eerste aanhechting. De parasiet bleek het meest gevoelig voor verlate infectie; de biomassa van Striga werd al significant gereduceerd bij een één-week verlate infectie. Bij een twee-weken verlate aanhechting ging een verdere reductie van de biomassa van de Striga gepaard met een sterke toename in planten korrelgewicht

146

Samenvatting van de sorghum. Bij het tolerante landras Tiémarifing, waar onder normale omstandigheden de eerste Striga-infectie al relatief laat plaats vindt, resulteerde een verdere vertraging van het aanhechtingstijdstip niet in significante effecten op zowel waardplant als parasiet. De invloed van een verlate eerste infectie van Striga hangt dus sterk af van het actuele infectietijdstip. Verder bevestigen de resultaten dat het eerder waargenomen verschil in eerste infectietijdstip tussen de beide sorghumcultivars (Hoofdstuk 2), bijdraagt aan het meer tolerante karakter van Tiémarifing. Gebaseerd op deze resultaten werd geconcludeerd dat het principe van het scheiden van Striga zaden en gastheerwortels in ruimte en tijd aanknopingspunten biedt voor de ontwikkeling van nieuwe beheersstrategieën. Experimenten werden opgezet om uit te zoeken of cultuurmaatregelen als oppervlakkige grondbewerking, het zaaien in een diep plantgat, of het overplanten van jonge sorghumplanten, kunnen leiden tot een vertraging van het eerste Striga-infectietijdstip. Een potproef uitgevoerd in Wageningen liet zien dat de combinatie van de drie genoemde cultuurmaatregelen tot duidelijke effecten leidde. De opkomst van de parasiet werd met vier weken vertraagd, het aantal opgekomen Striga planten nam beduidend af en de geproduceerde Striga biomassa werd gedecimeerd. Bij de gevoelige sorghumcultivar CK60-B leidde toepassing van een combinatie van de voorgestelde cultuurmaatregelen tot een opbrengstverlies van 26%, terwijl er onder standaard praktijkomstandigheden sprake was van een nagenoeg volledig opbrengstverlies (92%). Bij het tolerante landras Tiémarifing werd een verlies aan korrelopbrengst van 28% onder standaard praktijkomstandigheden volledig teniet gedaan. Toepassing van deze cultuurmaatregelen in een veldproef in Mali liet zien dat de gewenste vertraging in infectietijdstip niet tot stand werd gebracht, hoewel er wel sprake was van een sterke reductie van het Striga- infectieniveau. De maatregelen resulteerden in een sterke reductie van de Striga- biomassa, maar significante verbeteringen in opbrengst van de sorghum werden niet behaald. Wellicht dat de aanwezigheid van een natuurlijke Striga-zaadbank in de bouwvoor en problemen bij het creëren en handhaven van de plantgaten een belangrijke rol hebben gespeeld bij het achterblijven van de resultaten in het veld. Hoewel het voorkómen van de productie van nieuwe Striga zaden een essentieel onderdeel is van elke Striga-bestrijdingsstrategie, is een verbetering van de opbrengst op de korte termijn van wezenlijk belang voor acceptatie van de voorgestelde maatregelen door de praktijk. In Hoofdstuk 7 worden de conclusies van het onderzoek bediscussieerd in het licht van de Striga-problematiek in de semi-aride tropen van Afrika. Er wordt benadrukt dat een vertraagd contact tussen Striga-zaden en het wortelstelsel van de gastheer een belangrijke bijdrage kan leveren aan het verbeteren van de sorghumproductie op door Striga-besmette akkers en zodoende nadere aandacht verdient. Het gebruik van

147

Samenvatting herbicidentolerante - of herbicidenresistente - sorghumcultivars, waarvan de zaden zijn behandeld met een lage concentratie herbicide, zoals recentelijk onderzocht in een verkennende studie, wordt daarbij aangeduid als één van de potentiële opties op dit gebied.

148

Publications of the author

Refereed journals Ast, A. van, Bastiaans, L., 2006. The role of infection time in the differential response of sorghum cultivars to Striga hermonthica infection. Weed Research 46 (in press). Rodenburg, J., Bastiaans, L., Kropff, M.J., Ast, A. van, 2006. Effects of host plant genotype and seed bank density on Striga reproduction. Weed Research 46 (in press). Ast, A. van, Bastiaans, L., Katilé, S., 2005. Cultural control measures to diminish sorghum yield loss and parasite success under Striga hermonthica infestation. Crop Protection 24, 1023-1034. Lendzemo, V.W., Kuyper, Th. W., Kropff, M.J., Ast, A. van, 2005. Field inoculation with arbuscular mycorrhizal fungi reduces Striga hermonthica performance on cereal crops and has the potential to increase cereal production. Field Crops Research 91, 51-61. Ast, A. van, Bastiaans, L., Kropff, M.J., 2000. A comparative study on Striga hermonthica interaction with a sensitive and a tolerant sorghum cultivar. Weed Research 40, 479-493. Delft, G.J. van, Graves, J.D., Fitter, A.H., Ast, A. van, 2000. Striga seed avoidance by deep planting and no-tillage in sorghum and maize. International Journal of Pest Management 46 (4), 251-256. Dobben, W.H. van, Ast, A. van, Corré, W.J., 1984. The influence of temperature on morphology and growth of bean seedlings. Acta Botanica Neerlandica 33 (2), 85- 193. Dobben, W.H. van, Ast, A. van, Corré, W.J., 1981. The influence of light intensity on morphology and growth rate of bean seedlings. Acta Botanica Neerlandica 30 (1/2), 33-45.

Book chapters Ast, A. van, Borg, S.J. ter, 1990. Het groei-stimulerend effect van een holo-parasiet (Orobanche crenata Forssk.) op zijn gastheer (Vicia faba L.). In: Zonderwijk en VPO. Eds J.M. van Groenendael, W. Joenje, K.V. Sykora. LUW, vakgroep Vegetatiekunde, Plantenoecologie en Onkruidkunde, Wageningen, pp. 23-28. Ast, A. van, Groenendael, J.M. van, 1990. Het ontstaan van oecotypen binnen zuringsoorten onder invloed van strooizout. In: Zonderwijk en VPO. Eds J.M. van Groenendael, W. Joenje, K.V. Sykora. LUW, vakgroep Vegetatiekunde, Plantenoecologie en Onkruidkunde, Wageningen, pp. 59-64.

149

Ast, A. van, Nijs, L.J. de, Groenendael, J.M. van, 1990. Dispersieproblemen in een wegberm bij Barneveld. In: Zonderwijk en VPO. Eds J.M. van Groenendael, W. Joenje, K.V. Sykora. LUW, vakgroep Vegetatiekunde, Plantenoecologie en Onkruidkunde, Wageningen, pp. 167-172. Zweep, A., Ast, A. van, 1990. De invloed van de R/VR verhouding van het licht en van een bodembedekker op de kieming van enkele onkruidsoorten In: Zonderwijk en VPO. Eds J.M. van Groenendael, W. Joenje, K.V. Sykora. LUW, vakgroep Vegetatiekunde, Plantenoecologie en Onkruidkunde, Wageningen, pp. 11-16. Ast, A. van, Groenendael, J.M. van, Braakhekke, W., 1985. Ecotypic variation within Rumex spp. from a 13-year old road verge. In: Structure and functioning of Plant Populations. Eds J. Haeck, J.W. Woldendorp, North Holland Publ. Comp. Amsterdam.

Proceedings Ast, A. van, Bastiaans, L., Kropff, M.J., Pieterse, A.H., Sanogo, Z.L., 2001. Longevity of Striga hermonthica seeds under field and laboratory conditions. In: Proceedings of the 7th International Parasitic Weed Symposium. Eds A. Fer, P. Thalouarn, D.M. Joel, L.J. Musselman, C. Parker, J.A.C. Verkleij. Nantes, France, pp. 130. Ast, A. van, Bastiaans, L., Kropff, M.J., 2001. Effects of Striga hermonthica on photosynthesis and carbon allocation in sensitive and tolerant sorghum. Proceedings of the 7th International Parasitic Weed Symposium. Eds A. Fer, P. Thalouarn, D.M. Joel, L.J. Musselman, C. Parker, J.A.C. Verkleij. Nantes, France, pp. 151-154. Bastiaans, L., Ast, A. van, Kropff, M.J., 2001. Systems analysis and modeling in parasitic weed research. In: Proceedings of the COST 849-meeting on Parasitic plant management in sustainable agriculture. Bari, Italy, p. 44. Westerman, P.R., Ast, A. van, Bastiaans, L., 2001. Population dynamics of Striga hermonthica: analysis of preventative and control measures for long-term Striga control. In: Proceedings COST 849 meeting ‘Parasite plant management in sustainable agriculture. Bari, Italy, pp. 18-21. Bastiaans, L., Ast, A. van, Kropff, M.J., 1999. Utilizing a better understanding of crop damage due to pests, diseases and weeds in strategic pest management. In: Proceedings of the XIVth International Plant Protection Congress (IPPC), Jerusalem, pp. 145. Borg, S.J. ter, Schippers, P., Ast, A., van, 1994. A mechanistic model for dynamic simulation of carbohydrate production and distribution in Vicia faba and Orobanche crenata; problems and possibilities. In: Biology and management of Orobanche. Eds A.H. Pieterse, J.A.C. Verkleij, S.J. ter Borg. Proceedings of the Third Interna- tional Workshop on Orobanche and related Striga research. Amsterdam, pp. 57-66.

150

Woerden, I.C. van, Ast, A. van, Zaitoun, F.M.F., Borg, S.J. ter, 1994. Root exudates of resistant Faba Bean cultivars are strong stimulants of Broomrape germination. In: Biology and management of Orobanche. Eds A.H. Pieterse, J.A.C. Verkleij, S.J. ter Borg. Proceedings of the Third International Workshop on Orobanche and related Striga research. Amsterdam, pp. 169-172. Ast, A. van, Groenendael, J.M. van, 1993. The population dynamics of the pseudo- annual Calystegia sepium (Bindweed): a troublesome weed in the Netherlands. Med. Fac. Landbouw Univ. Gent, 58/3, 973-982. Borg, S.J. ter, Ast, A. van, 1991. Parasitic plants as stimulants of host growth. In: Proceedings Fifth International Symposium on Parasitic Weeds. Eds J.K. Ransom, L.J. Musselman, A.D. Worsham, C. Parker. Nairobi, Kenya, pp. 442-446. Borg, S.J. ter, Ast, A. van, 1990. Soil moisture, root architecture and Broomrape (Orobanche crenata) infestation in Faba bean (Vicia faba). In: Proceedings International Workshop on Orobanche Research. Eds K. Wegmann, L.J. Musselman. Eberhard-Karls-Universitat, Tubingen, pp. 278-292. Zweep, A., Ast, A. van, 1990. The influence of the red/farred ratios of light and of a soil cover on the germination of some weed species. Med. Fac. Landbouw. Univ. Gent 55(3b), 1209-1217.

Articles in professional journals Ast, A. van, Bastiaans, L., 2005. Vertraging van het tijdstip van Striga hermonthica - aanhechting: een optie voor opbrengstverhoging van sorghum? Gewasbescherming 36 (2), 116-120. Ast, A. van, Borg, S.J. ter, Groenendael, J.M. van, 1987. De oorzaak van de achteruitgang van Biggekruid in onze bermen. De Levende Natuur 88 (2), 88-93. Ast, A. van, Groenendael, J.M. van, Borg, S.J. ter, 1988. Strooizout veroorzaakt variatie binnen zuringsoorten. De Levende Natuur 89 (2), 51-54.

151

Curriculum vitae

Adriaan van Ast werd geboren op 12 oktober 1950 te Rotterdam. Na het behalen van het diploma MULO-B in 1968 aan de Chr. U.L.O. te Middelburg, begon hij zijn studie, met als specialisatie Veehouderij, aan de Hogere Landbouwschool te Dordrecht. In juni 1971 behaalde hij het diploma van deze studie. Aansluitend aan deze opleiding werkte hij tot oktober 1971 bij de Plantenziektenkundige Dienst te Rotterdam en daarna tot januari 1974 bij het Instituut voor de Bewaring en Verwerking van Landbouwproducten (IBVL) te Wageningen. Hier werkte hij als onderzoeksmedewerker in diverse projecten op het gebied van de gras- en hooiwinning. In januari 1974 trad hij in dienst bij de vakgroep Vegetatiekunde van de toenmalige Landbouwhogeschool, waar hij zich o.l.v. Prof. dr. W.H. van Dobben specialiseerde tot onderzoeks - en onderwijsmedewerker op het vakgebied van de Plantenoecologie. Deze functie bekleedde hij, later bij de vakgroep Vegetatiekunde, Plantenoecologie en Onkruidkunde, tot 1995. In deze periode vervulde hij tevens diverse bestuursfuncties binnen de Landbouwuniversiteit Wageningen, o.a. als lid van het Dagelijks Bestuur van de Vaste Commissie Wetenschapsbeoefening (VCW). In 1987 en 1988 volgde hij cursussen in de Natuurwetenschappen en Algemene Milieukunde aan de Open Universiteit. In 1995 werd het dienstverband bij de Landbouwuniversiteit Wageningen voortgezet bij de toenmalige vakgroep Theoretische Productie-ecologie en later bij zijn huidige leerstoelgroep Gewas- en Onkruidecologie. In 1996 werd hij in de gelegenheid gesteld, naast zijn taken als onderzoeks- en onderwijsmedewerker, onderzoek te doen binnen een door hem zelf geformuleerd onderzoeksproject aan het parasitaire onkruid Striga hermonthica. Het resultaat van dit onderzoek wordt in het voorliggende proefschrift beschreven.

153

Funding

Part of the work carried out for this PhD research was supported by funds of the European Union 5th Framework project “Improved Striga control in maize and sorghum” (ISCIMAS, Project No. ICA4-1999-30071).

154