VARIETY SCREENING FOR STRIGAHERMONTHICA RESISTANCE IN UPLAND RICE IN EASTERN UGANDA [301593]

VARIETY SCREENING FOR STRIGAHERMONTHICA RESISTANCE IN UPLAND RICE IN EASTERN UGANDA

BY

KAYONGO NICHOLAS

2012/HD02/105U

208006518

SUPERVISORS

DR. JENIPHER BISIKWA (PhD)

PROF: JULIE .D.SCHOLES (PhD)

DR. SSEBULIBA JAMES (PhD)

RESEARCH THESIS SUBMITTED TO THE SCHOOL OF AGRICULTURAL SCIENCES IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE AWARD OF MASTER OF SCIENCE DEGREE IN CROP SCIENCE OF MAKERERE UNIVERSITY

2016

DECLARATION

I Kayongo Nicholas do hereby declare that the work presented in this thesis is original and has not been submitted to any institution of learning by any one for Academic award.

Signed…………………………………………Date…………………………………………

This thesis has been submitted to the college of Agriculture and Environmental Sciences and School of Graduate Studies with approval from the supervisors

Signed………………………………………………Date …………………………………….

DR. JENIPHER BISIKWA (PhD)

Senior Lecturer College of Agriculture and Environmental Sciences

School of Agricultural Sciences

Signed……………………………………………… Date………………………………………

DR. JAMES SSEBULIBA (PhD)

Senior Lecturer College of Agriculture and Environmental Sciences

School of Agricultural Sciences.

DEDICATION

To people of Namutumba district, I hope the research findings will help you improve your livelihood through rice production. Also to my parents and friends thank you for the moral support and prayers.

ACKNOWLEDGEMENTS

In the first place, I [anonimizat].

I [anonimizat]. Jenipher Bisikwa and Dr. [anonimizat] U.K for the technical advice as well as the wonderful supervision. Am very grateful that you continued supervising me despite all your other commitments.

Also special thanks to Dr. [anonimizat], thank you so much l would not have made it without your help. I also thank Dr. Mamadou Cissoko for helping in setting up of field experiments in Namutumba district. I have never met a [anonimizat]. Am very grateful for your help. I would also like to give thanks to members of C45 lab at the University of Sheffield U.K, Department of Plant and Animal sciences for all the assistance during my Laboratory experiments. More thank to Moses and all members of Abendewoza farmers group for providing land and participating in the field experiments in Namutumba.

Lastly am also very grateful to the Seohyun Foundation that provided me the scholarship for the two academic years’ [anonimizat] (DfID) and Biotechnology and Biological Sciences Research Council (BBSRC) for allowing me do this research. Am very grateful for all the financial assistance during the course of the research.

[anonimizat] 1: List of varieties to be screened in Nsinze, Namutumba District, Uganda 29

Table 2: Initial soil characteristics of the field trial for season 2014 A and 2015 B in Eastern Uganda. 35

Table 3: Variance component analysis (F-stat and F-prob) and standard error of difference of means (SED) of varieties; effects on plant height, tiller number and productive tiller number for 2014 and 2015. 36

Table 4: Variance component analysis (F-stat and F-prob) and Standard errors of differences of means (SED) of varieties; effects on rice grain dry weight (Rice grain DW), rice straw dry weight (Rice straw DW) and above-ground Striga biomass dry weights (Striga DW) at harvest for season of 2014A and 2015B 42

Table 5: Variety effect on Striga hermonthica emergence and flowering days. 43

Table 6: Farmer participant characterization: gender ratios, age-groups and rice farming experiences of farmers in Namutumba district Eastern Uganda in 2015. 59

Table 7: Reasons for rice variety selection among the least-liked, as indicated by farmers in Eastern Uganda in 2015. 62

Table 8: Criteria rice farmers use in selection of the best upland rice varieties in Eastern Uganda in 2015 62

Table 9: Ranking of the nine varieties of upland rice based on farmer selection criteria in Eastern Uganda in 2015. 63

LIST OF FIGURES

Figure 1: Life cycle of Striga 21

Figure 2: Schematic representation of the field trial in Nsinze, Namutumba district, Uganda with the indication of the path, the homesteads and the water pump. 31

Figure 3: The field trial in Nsinze, Namutumba district, Uganda in the first season (2014) at 25days after sowing. 32

Figure 4: Height of rice plants per variety for 2014 (a) and 2015 (b) 37

Figure 5: The number of tillers per rice variety 2014a and 2015c and productive tillers per rice variety 2014b and 2015d.. 40

Figure 6: Above-ground biomass dry weight of rice varieties grown under Striga hermonthica infested field conditions for 2014 (a) and 2015 (b). 41

Figure 7: Maximum number of emerged Striga plants m-2 (NSmax) per variety for two seasons (6a) and cluster analysis of rice varieties (6b). 45

Figure 8: Above ground dry weight of Striga per variety for 2014 and 2015.. 46

Figure 9: Rice grain dry weight per variety for 2014 (a) and 2015 (b) and cluster analysis for rice grain dry weight among rice varieties for 2014 (c) and 2015 (d).. 47

Figure 10: Farmer participatory preference of upland rice varieties in Namutumba district Eastern Uganda in 2015 60

Figure 11: Rice variety selection by gender of rice farmers in Eastern Uganda (Namutumba district) in 2015. 61

Figure 12: Height of the infected and uninfected control rice plants per variety. 73

Figure 13: Stem width of the infected and uninfected control rice plants per variety. 74

Figure 14: Number of tillers of the infected and uninfected control rice plants per variety. 75

Figure 15: Maximum number of leaves of the infected and uninfected control rice plants per variety. 76

Figure 16: Figure 5: Evaluation of post-attachment resistance of different varieties of rice, number of Striga plants taken 21 days after Striga hermonthica infection 77

Figure 17: Parasitic plants attached on the host plant roots 21 days after infection with Striga hermonthica. 78

Figure 18: Above-ground rice dry biomass of infected rice plants and their respective uninfected control plants. 79

Figure 19: Relationship between percentage losses in total rice biomass of infected plants compared with control plants and the amount of parasite biomass dry weight on roots of rice plants. 80

ABSTRACT

Two sets of experiments and one participatory variety selection study were conducted. The first experiment on field for two seasons in Namutumba district starting first season of 2014 and the second was laboratory experiment conducted at Department of Plant and Animal Sciences University of Sheffield U.K. The main objective was to screen a selection of rice varieties for Striga hermonthica resistance in Eastern Uganda. The field experiment had two studies i.e. one study evaluated the resistance of twenty five upland rice varieties sourced from different areas of Africa against Striga hermonthica under field conditions and the second study identified Striga resistant upland rice varieties that could be adopted by farmers in Eastern Uganda. The laboratory study was aimed at determining the post attachment resistance of farmer selected upland rice varieties against Striga hermonthica under controlled conditions. These were tested in min rhizotron root observation system. In the field experiment, a 5×5 lattice design was used, with 25 varieties constituting the treatments in this design. A total of 150 plots with six replicates were involved. Prior to planting the plots were supplied with 0.92g of Striga seed mixed with 200g of white sand to homogenize Striga levels in each plot. This study indicated variations (P<0.001) in resistance of rice varieties to S.hermonthica, the most resistant varieties were WAB928, Blechai, WAB935, SCRID090, WAB935, NERICA-10,-2, -17, IRGC, IR49, WAB181-18, CG14, Anakila, WAB880 and the most susceptible varieties were IAC165 and Superica. In addition WAB56-50, WAB56-104, NARC3 (ITA 257), were also susceptible in this study. There was also significant variation (P<0.001) in yield across rice varieties. Varieties such as NERICA-1,-2,-17, -10, WAB181-18, WAB880, SCRID90, CG14, BLECHAI, Anakila, ACC and AGEE produced the highest yield under Striga hermonthica infested conditions. Varieties IAC 165, Superica, WAB928, WAB935, UPR, NARC3 (ITA 257) produced the lowest yield. It was also shown that some varieties that were resistant did not necessarily have high yields attributed to their yielding potential and adaptability to these field conditions.

In the second study for experiment one, farmers selected the most appropriate varieties they would wish to sow in their fields. Out of the 25 varieties, farmers, selected varieties SCRID090, WAB880, Blechai, WAB181-18 and NERICA-17 as the most preferred varieties. These were liked for their good qualities such as yield, Striga resistance, good tillering ability and being tall. Also farmers indicated a list of traits as being undesirable in their rice verities such as Striga susceptibility, lodging, short suture, low yields and drought susceptibility. There was a significant variation (P<0.05) in variety selection among the women and men.

In the second experiment, in the mini rhizotron system we sought to understand whether the resistance of the farmers selected varieties was replicated under controlled conditions. Following a complete randomized design, a total of 10 plants, 6 infected and 4 uninfected control plants were used. A total of 14 varieties were evaluated under controlled conditions. Each rice plant was infected with 12mg of sterilized Striga seeds and number of destructive and nondestructive measurements were taken. For the first time post-attachment resistance of varieties Blechai, WAB880, SCRID090, WAB928 and IRGC were tested. Additionally, the resistance detected in these varieties elsewhere in field experiments was replicated under laboratory conditions. However varieties WAB880 and SCR1D090 were for the first time documented both in field and under the post attachment study. This study also showed significant variations (P=0.001) in post attachment resistance across rice varieties. Varieties WAB880, Blechai, SCRID090, WAB928, CG14, IRGC, WAB56-50, WAB181-18 and WAB56-104 had an excellent post-attachment resistance to the ecotype of Striga from Namutumba Uganda. The local variety (Superica and IAC 165) had significantly (P=0.001) highest number parasites per root system compared to other varieties. Also varieties NERICA-1and NERICA-17 had a moderate number of Striga plants attached per root system. There was also a reduction in rice biomass with respect to infected plants compared to the uninfected control plants. However some varieties such as NERICA-17 and NERICA-1 with high Striga numbers compared to CG14, WAB928, IRGC, WAB880, and WAB181-18 had lower reduction in rice biomass. The latter however had a high reduction in rice biomass. These can only be described as tolerant varieties producing yield irrespective of the high Striga plants supported by these varieties. These results are highly relevant to rice breeders, agronomists molecular biologists working on Striga resistance. Additionally those Striga resistant varieties combining high yields and excellent adaptability to field conditions can be recommend to farmers in Striga prone areas elsewhere in Uganda.

CHAPTER ONE

1.0 INTRODUCTION

1.1 Background

Rice belongs to the genus Oryza, sub-family Oryzoidaea of family gramineae. It is a small genus with approximately twenty two species (Vaughan, 1994). Twenty species of genus Oryza are wild species and there are only two cultivated species Oryza sativa L (Asian rice) and Oryza glaberrima Steud (African rice). Oryza sativa is the most widely grown of the two cultivated species. Oryza glaberrima can be distinguished from Oryza sativa by differences in ligule shape, lack of secondary branches in the panicle and an almost glabrous glume. In many parts of Africa, Oryza sativa and interspecific crosses between Oryza glaberrima and Oryza sativa are replacing Oryza glaberrima (Linares, 2002). Among such interspecifics is NERICA upland rice, which is as a result of backcrossing between the Asian and African rice species (Jones et al; 1997; Wopereis et al; 2008). This is spreading in most parts of Africa because of its resistance to drought and its ability to produce high yields.

It is thought that African and Asian rice were domesticated independently (Khush, 1997). It is suggested that Oryza sativa originated in India from Oryza perennis whereas cultivation may have been earlier in China (Purseglove, 1975). Oryza sativa was first cultivated in south-east Asia, India and China between 8000 and 15000 years ago (Normile, 2004). Oryza glaberrima is believed to have been cultivated in the primary area of diversity in West Africa since 1500 BC while in the secondary areas cultivation begun 500-700 years later (Porteres, 1956).

Rice plays an important role in many ancient customs and religious magical rites in the East, a sign of antiquity of the crop. Its significance is connected with fecundity and plenty for example rice cakes are eaten at certain festivals in Asia to symbolize long life, happiness and abundance (Purseglove, 1975). Rice straw can be fed to livestock, although it is not nutritious as compared to straw of other cereals. It can also be used for the manufacture of strawboards, for thatching and brading as well as making of mats and hats. In some countries like China and Thailand rice straw is used for making mushroom culture (Purseglove, 1975).

1.2 Overview of rice production in the world

Rice is the most important staple food for about half of the people in the world. Rice has been cultivated worldwide for more than 10,000 years, longer than any other crop (Kenmore, 2003). Globally, the area under rice cultivation is estimated at fifteen million hectares with an annual production of about 500 million metric tons (Tsuboi, 2004). This accounts for about 29% of the total output of the grain crop worldwide (Xu and Shen, 2003). South and Eastern Asia produce about 90% of world’s rice crop. China and India are the leading producers and consumers of rice in the world. In Africa, rice is becoming a major staple food among the diet of many people. It is the staple food for about ten African countries. Rice is grown in more than 75% of the countries in the African continent. In 2008, Africa produced an estimated quantity of 23 million metric tons of unmilled rice on 9.5 million hectares (FAOSTAT, 2010). The major producing regions in Africa were Western, Northern, and Eastern with an estimated production of 10.2 million metric tons, 7.3 metric tons and 5 million metric tons respectively. These were harvested on 5.8 million hectares, 0.8 million hectares and 2.4 hectares respectively (FAOSTAT, 2010). In East Africa, rice production is considerably increasing due to establishment of upland rice varieties.

1.3 Rice production in Uganda and importance

Rice is becoming a major food crop in Uganda. Changes in consumption trends and increased population have led to increased production of rice over the years. The per capita consumption of rice is estimated at 8 kg producing a total consumption of 224,000 metric tons by a population of 28-30 million with an annual growth rate of 3.2%. Rice production was introduced by Indian traders in Uganda in 1904 (MAAIF, 2009). Rice production only became economically relevant in the late 1940’s when the government included rice-based food rations in the diet of soldiers (Wilfred, 2006). The establishment of rice schemes in Kibimba in 1960 and Doho in 1976 led to production of lowland rice in Eastern and Northern parts of the country. In the 1980’s the production area under rice increased tremendously up to date. Since 1997, the annual unmilled rice production has increased from 80,000 metric tons in 1997 to over 210,000 metric tons in 2010, representing an annual growth rate of 12% (FAOSTAT, 2010). This increase in production is partly attributed to the introduction of upland rice varieties particularly NERICA varieties in 2002 (Kijima et al., 2006). Since the introduction of the upland NERICA varieties rice, area under cultivation has increased estimated at 72000 hectares in 2000 (Uganda Bureau of Statistics 2002) and currently estimated at 90000 hectares (Uganda Bureau of Statistics 2012). About 80% of rice farmers are small-scale farmers with less than two hectares with women playing a big role in rice production.

Rice production in Uganda has had a positive influence on the livelihoods of farmers. It has also contributed greatly to the development of the country by providing people with income through the selling of rice as grains. About 60% of the rice produced in Uganda is sold (PMA, 2009). This has improved the livelihoods of rice farmers. The demand for rice especially in the urban centers has increased tremendously as a result of population increase. It is estimated that the population growth rate in Uganda increases by 3.2% annually (PMA, 2009; MAAIF, 2009). This has acted as potential market for the rice that is produced by the farmers. According to PMA (2009) and MAAIF (2009), Uganda’s rice imports dropped from 77,600 tonnes in 2000 to 33,000 tonnes in 2010. This reduction has increased domestic market supply by the farmers increasing their incomes.

1.4 Problem statement

In 2003, the government of Uganda encouraged the adoption of NERICA high yielding upland rice varieties, as one of the strategies to eradicate poverty and increase food security. Since the adoption of NERICA varieties, rice production has risen both in acreage and volume of production (MAAIF, 2011). However, despite the rise in volumes, production per unit area is currently declining in Uganda and one of the major constraints that has been identified as contributing to this decline is the increase in Striga infestation. Most cereals including rice are attacked by species of Striga from the family Orobanchaceae namely; Striga hermonthica (Del.) Benth and Striga asiatica, (L.) Kuntze. Striga hermonthica is however the most important biological constraint in upland rice production in Uganda. It is estimated that 62,000 hectares of farmland in Uganda is infested with Striga (AATF, 2006) and this infestation can cause around 40-100% yield loss if not checked in the field (Oswald, 2005).

The problem of Striga is accentuated under conditions of low soil fertility especially low nitrogen and moisture levels. The recent increase in Striga infestation in Eastern Uganda is attributed to two main reasons; the decline in cotton production, which acted as a trap crop decreasing infestations since it was not a host crop for Striga and declining soil fertility as a result of continuous cultivation of the soil without any replenishment of the removed nutrients from the soils. This is partly due to an increase in the population with many of the areas which should be under fallow being put to cultivation.

Previous studies on rice show that some rice varieties exhibit resistance to either Striga hermonthica or Striga asiatica or both (Johnson et al., 1997; Harahap et al., 1993; Cissoko et al., 2011; Gurney et al., 2006; Rodenburg et al., 2015). Recently 18 upland NERICA varieties have been assessed for pre- and post- attachment resistance to Striga hermonthica and Striga asiatica under controlled environmental conditions (Cissoko et al., 2011; Jamil et al., 2011). According to these studies, NERICA varieties including NERICA1, 10, 17 and 2 exhibited very good post- attachment resistance to several ecotypes of S.hermonthica (Cissoko et al., 2011). Additionally, NERICA varieties produce different amounts and types of strigolactones in their root exudates which will alter pre-attachment resistance (Jamil et al., 2011). Inspite of the above successes under controlled environmental conditions, there has been no published information about the effect of the environment on the expression of resistance apart from Rodenburg et al., (2015) who evaluated 18 NERICA cultivars and their parents under field conditions in Kyela, Tanzania and Mbita Kenya. They showed variation in field resistance among the NERICA cultivars and their parents to different Striga ecotypes. In Uganda, NERICA varieties have not been evaluated for Striga resistance since their introduction in farmer’s fields. Also, there are a number of upland rice varieties being reported to be resistant to either S.hermonthica or S.asiatica in different areas of Sub Saharan Africa (Harahap et al, 1993; Johnson et al., 1997). Such varieties and NERICA varieties should be evaluated in different Striga infested agro-ecosystems to determine the level of resistance in the field.

1.5 Justification

Control of Striga hermonthica presents a challenge and is complex because of the possible interactions between ecotypes of the parasitic weed, the host and the environment (Oswald, 2004). A number of strategies have been proposed in the management of Striga in rice fields for example application of nitrogenous fertilizers (Riches et al., 2005), use of chemical herbicides such as 2, 4-D (Carsky et al., 1994a), crop rotations and improved fallow management (Oswald, 2004) and hand weeding. However, none of these control approaches has proven very effective in Striga weed management in rice production systems. Use of resistant varieties is an important approach in Striga management due to its genetic nature. It is also cost effective as farmers do not have to invest a lot of time and resources in implementing this management technology. Because of the genetic variability existing among different species and ecotypes of the parasite, resistance found in some varieties may be overcome by a small subset of Striga individuals within the seed bank leading to development of a virulent population of Striga overtime (Rodenburg and Bastiaans, 2011). Also some studies have shown high levels of variability existing within and between Striga populations in Kenya, Mali and Nigeria (Gethi et al., 2005). This means that resistance of a variety in one place with a particular ecotype of Striga does not guarantee resistance in another place with another parasitic ecotype. Therefore varieties have to be tested at multiple /different agro-ecosystems. Likewise farmers in one place may have different variety preferences compared to farmers in another place. Also farmers need to have a wide range of variety options so that they choose varieties that are not only resistant/ tolerant but also have desired characteristics such as grain yield, grain colour/ size, taste etc. This study is therefore aimed at screening a selection of upland rice varieties sourced from different areas in Africa for S.hermonthica resistance and identify rice varieties that can be adopted by farmers in Eastern Uganda.

1.6 Objectives of the study

The overall objective of this study was to screen a selection of upland rice varieties for Striga hermonthica resistance in Eastern Uganda. The specific objectives include the following;

To evaluate the resistance of twenty five upland rice varieties sourced from different areas of Africa against Striga hermonthica under field conditions

To identify Striga resistant upland rice varieties that can be adopted by farmers in Eastern Uganda.

To determine the post attachment resistance of farmer selected upland rice varieties against Striga hermonthica under controlled conditions.

1.7 Hypotheses

There is a significant variation in resistance among rice varieties to Striga hermonthica.

The resistance of rice varieties to Striga hermonthica expressed under field and controlled environmental conditions does not differ.

Farmer preferred characteristics in Striga resistant upland rice varieties differs from researchers.

CHAPTER TWO

2.0 LITERATURE REVIEW

2.1 Major weeds of rice

Worldwide, weeds are estimated to account for 32% potential and 9% actual yield losses in rice (Oerke and Dehne, 2004). In Africa, research has shown that there are 130 most common weed species of which 61 species are found in upland rice fields, 31 species in hydromorphic and 74 species in lowland rice. The most common weed species in upland rice include; Rottboellia cochinchinensis (Lour.) W. Clayton, Digitaria horizontalis Willd, Ageratum conyzoides L; and Tridax procumbens L., while Ageratum conyzoides L., and Panicum laxum S.W are common in hydromorphic areas, Cyperus difformis L., Echinochloa colona (L). Link, are most common in lowland rice (Rodenburg and Johnson, 2009). In general, most weed species in lowland rice are from families Graminaea 43%, Cyperaceae 37% and upland rice with species from Graminaea 36% and Compositae 16% (Rodenburg and Johnson, 2009). Also important in upland rice production are parasitic weeds of genus Striga which include; Striga hermonthica (Del) Benth and Striga asiatica (L.) Kuntze. Agronomic factors such as inadequate land preparation, rice seed contamination, broad cast seeding in lowlands, inadequate water and fertilizer management, mono cropping, delayed herbicide application and use of poor quality seed are responsible for the major weed problems in rice.

2.2 Striga

The genus Striga is a major limiting biotic constraint in production of cereal crops. Members of this genus are obligate hemi parasites, they are chlorophyllous but require a host to complete their life cycle (Musselman, 1987). The genus Striga includes over 40 species and of these eleven species are considered parasitic on agricultural crops. Striga species infest an estimated two-thirds of cereals and legumes in Sub-Saharan Africa, causing annual crop losses estimated at US$7 billion and negatively affecting the livelihood of people in the region (Berner et al., 1995). Based on host preference, Striga species can be split into two groups. The first group parasitizes Poaceae, including major food and forage grains such as maize (Zea mays), sorghum (Sorghum bicolor), rice (Oryza sativa) and millet. The second group preferentially attacks legumes including cultivated and wild species (Mohamed et al., 2001). According to Ramiah et al. (1983), the most important species of Striga in Africa include; Striga hermonthica (Del.) Benth; Striga asiatica (L.) Kuntze; Striga gesneriodes (Willd) Vatke; Striga aspera (Willd) Benth; and Striga forbesii Benth. All species except Striga gesneriodes, which parasitizes cowpea and other wild legumes, parasitize the major cereal crops in Africa, including rice (Rodenburg et al., 2010).

2.3 Origin and distribution of Striga

The genus Striga is predominantly African in origin and distribution and about 30 species are endemic to Africa (Mohamed et al., 2001). The prevalence and extent of genetic diversity of species in a particular geographic area are often indicators of the place of origin (Gebisa and Jonathan, 2007). The vast tropical savannah between the Semien Mountains of Ethiopia and the Nubian hills of Sudan has the greatest biodiversity of sorghum and millet which are infested by Striga. These regions are believed to be the origin of Striga hermonthica (Del.) Benth and Striga asiatica (L.) Kuntze (Gebisa and Jonathan, 2007). Striga gesneriodes is thought to have originated from West Africa, Striga hermonthica has had the largest geographical distribution with obligate out-crossing behaviour and it is found in Sub-Saharan Africa with most prevalence in western, central and eastern Africa and some parts of southwestern part of the Arabian Peninsula across the Red sea. The red-flowered and weedy ecotype Striga asiatica (L.) Kuntze is mostly found distributed in eastern and southern Africa, whereas the yellow-flowered ecotype is found in West Africa (Mohammed et al., 2001; Rodenburg et al., 2010). The latter ecotype has no importance as a weed.

2.4 Taxonomy and Botany of Striga

The genus Striga traditionally included in the Scrophulariaceae family is now included in Orobanchaceaae (Olmstead et al., 2001). This was based on molecular evidence with a result that both genera Striga and Orabanche are in the same family. According to Mohammed et al. (2001) of the approximately forty described species of Striga, thirty species are endemic to Africa. The hemi-parasite Striga is a succulent, greenish yellow annual herb up to 35 cm tall, usually branching from the base, glabrous or minutely puberulent. Each plant has a single, large, tuberous primary haustorium 1–3 cm in diameter. The species have numerous adventitious roots emerging from subterranean scales; stem square but obtusely angled. Leaves are opposite, appressed to the stem, scale-like, 5–10 mm × 2–3 mm. The inflorescence is a terminal bracteates spike. Flowers are bisexual, zygomorphic, 5-merous, not fragrant, sessile; the calyx is tubular with 5 teeth at apex, 4–6 mm × 2 mm; corolla tubular, 2-lipped, up to 15 mm long, bent in upper part of tube, pale blue to dark purple, upper lobes 2, fused, sharply recurved, up to 2.5 mm long, lower lobes 3, spreading, 3 mm long; stamens 4, 2 longer and 2 shorter; ovary superior, tubular, 2-celled, style terete, stigma 2-fid. The fruit an ovoid capsule 1–2 mm × 3 mm, many-seeded. Seeds are very small, dust-like, with prominent encircling ridges.

Striga species are distinguished from other root parasites by unilocular anthers and bilabiate corollas with pronounced bend in the corolla tube. In some species such as Striga hermonthica, the corolla is bent within the calyx teeth, while other species have their corolla bent above the calyx as in Striga asiatica. Other distinguishing features used include; indumentum type that is either pubescence, hirsute and whether hairs are ascending or retorse (pointing backward), stem shape with some species having either round, obtusely square and square in cross section; leaf lobbing and dentations whether leaf margins are lobed, serrate or smooth; inflorescence types with some species having spike, raceme and the length of the inflorescence relative to the vegetative stem; number of ribs on the calyx tubes and length of the calyx teeth relative to the tubes, corolla color and type of indumentum on corolla.

2.4.1 Life Cycle of Striga

Figure 1: Life cycle of Striga (Mweze et al., 2015)

Striga species have a complex life cycle. Seeds are the sole source of inoculum, they are produced in abundance ranging from 10,000- 100,000 or more per plant (Pieterse and Pesch, 1983). First dormancy of the seeds has to be broken; seed dormancy can persist for about six months (Vallance, 1950). The seeds require warm and humid conditions for a period of one-two weeks for them to germinate (pre-conditioning or conditioning). The pre-germination requirement of exposure to moisture, combined with temperatures above 200C for a period of one week or more is probably a survival adaptation which prevents the seed from germinating before rainy season is well established. Before the rains the host roots are not well established and expanded to allow successful Striga attachment. The next step is a signal by a specific biochemical germination stimulant from the host roots (Vallence, 1950; Worsham, 1987). These groups of biochemicals have been identified as germination stimulants for parasitic weeds; dihydrosorgoleones, sesquiterpelactones and strigolactones. The latter are the most important biochemicals with respect to cereals (Bouwmeester et al., 2003). Production of the stimulant by the host roots directs the Striga radicle to grow directly towards the source of the stimulant; this growth is chemotropic (Chang and Lynn, 1986). After germination, a series of chemical signals direct the radicle to the host root where it attaches and penetrates. However, when the seedling does not attach to a host root within three to five days, the seedling dies (Worsham, 1987). Once penetration has taken place, an internal feeding structure (haustorium) is formed. Through the haustorium, the parasitic weed establishes a xylem to xylem connection with the host roots (Worsham, 1987). This is also thought to require biochemical triggers (Yoder, 2001). Through the xylem to xylem connection with the host root, the parasite obtains metabolites, water and amino acids from the host plants (Press et al., 1987a). As the host matures, the parasite emerges and begins to produce chlorophyll and starts to photosynthesize (Saunders, 1933). Reproductive strategies range from autogamy to obligate allogamy depending on the species (Musselman, 1987). Species known to be obligatory allogamous are Striga hermonthica and Striga aspera (Safa et al., 1984; Mohammed et al., 2007). These require insect pollinators and can hybridize producing viable pollen and virulent offspring. The other remaining species of Striga are autogamous (Musselman et al 1982). Following reproduction, seeds are dispersed and then the cycle begins again.

2.5 Ecology of Striga

Problems of Striga appear to be associated with degraded environments and freely draining soils of rain fed cereal production systems and are most severe in subsistence farming systems with little option for addition of external inputs. Striga is prevalent under conditions of low soil fertility especially low nitrogen and moisture levels. Striga species have been described as indicators of low soil fertility and their infestation is linked to low nutrient conditions (Oswald, 2005). Striga is common in tropical and subtropical areas of Africa and some parts of India except in extremely cold climates. Striga species require a suitable temperature of 200-400C under moist conditions for the seeds to germinate coupled with host derived signals. Soil type and soil pH are probably not critical for growth, as Striga occurs in all types of soils from sandy acidic soils to alkaline clay soils.

2.6 Nature of Striga damage

Host damage due to Striga infestation occurs already before the emergence of the parasite, and continues thereafter. Initial symptoms occur while the parasite is still subterranean; they appear as water soaked leaf lesions, chlorosis and eventual leaf and plant desiccation and necrosis, severe stunting and drought like symptoms such as leaf margin curling also occur (Kim et al., 1991). Parasitism by Striga species reduces the yields in upland rice in two main ways;

The parasite directly derives water and mineral nutrients from the root vascular system retarding the host growth and development (Press and Stewart, 1987; Rodenburg et al., 2006; Atera et al., 2011). Although Striga is chlorophyllous, its rate of photosynthesis is low; therefore as much as 60% of its carbon is host derived (Graves et al., 1989). The nutrients flow from the host’s vascular system to the parasites via the haustorium; this flow is facilitated by the high rate of transpiration in Striga that exceeds transpiration in the host. One of the most important causes for Striga damage is its effective competitive ability in depriving the host plant of carbon, nitrogen and inorganic salts. This happens while at the same time inhibiting the growth and impairing photosynthesis of its host (Khan et al., 2006).

The parasite pathologically affects the growth and development of upland rice. This is associated with the phytotoxic effects of Striga within days of attachment. It is hypothesized that the parasite produces phytotoxic substances that affect the crop’s growth with even low levels of infection resulting in host dehydration. The parasitic weed also alters the hormone balance of the host plant (Frost et al., 1997). High concentration of abscisic acid inhibits growth by reducing the rate of photosynthesis. It was observed that introduction of abscisic acid to the xylem stream of plants affects photosynthesis by suppressing ribulose biphosphate carboxylation (Fischer et al., 1986). Some reports show a decrease in the concentrations of cytokinins and gibberellins with an increase in the concentration of abscisic acid in Striga infested hosts (Drennan and El Hiweris, 1979).

2.7 Control of Striga

Management of Striga presents a challenge because of the competitive ability of the weed, high seed production, mechanisms of dormancy and possible interactions between the parasitic ecotype, host and environment. Varying degrees of control has been achieved through cultural methods such as crop rotations, intercropping, hand weeding, trap cropping and fertilizer application. Use of biological control, chemical control and use of resistant cultivars have also been applied to control Striga species (Rodenburg et al., 2010).

2.7.1 Cultural control

There are a number of methods described in the management of Striga in the field. These are described below;

2.7.1.1 Hand weeding

This is the most used method of Striga management by farmers in Africa. Weeding removes emerging Striga shoots, preventing them from flowering and seed production, though the damage normally occurs before the parasitic weed emerges. For this reason the benefit of hand weeding may not be realized in the current season, but rather in the subsequent seasons as it can reduce the Striga seed bank over the long term. Use of hand weeding reduced Striga infestation by 12% in the maize fields in the one season, when applied for four consecutive seasons it reduced infestations by 26.6% in the fourth year (Ransom et al., 1997). Timing of weeding operations is essential in management of Striga, for example Parker and Riches, (1993) recommend that hand pulling should be done before or right after Striga hermonthica begins to flower to avoid seed setting. The effectiveness of hand weeding can be increased in combination with other methods of Striga control. Ransom and Odiambo (1994) reported improved maize yields after integrating soil fertility amendment measures with hand weeding.

2.7.1.2 Crop rotation and trap crops

Under suitable conditions successful control can be obtained by use of trap crops. Various grass species parasitized by Striga can be shown to stimulate germination of the seed and then ploughed back before the seed reaches maturity which can reduce the soil seed bank if continuously carried out. For example Sudan grass (Sorghum sudanense) has been used in management of Striga hermonthica and ploughed 5 weeks after sowing (Ivens, 1989). Other trap crops like cowpea (e.g. Carsky et al., 1994b) and pigeon pea (e.g. Oswald and Ransom, 2001) are possible means of lowering Striga spp. infestations in cereal production systems. Crop rotation cycles disrupt the parasitic weed life cycles by reducing the soil seed bank. Rotations including trap crops such as cow pea (Vigna unguiculata), soy bean (Glycine max), groundnut (Arachis hypogeal) and pigeon pea (Cajanas cajan) have all proved effective, by stimulating and contributing greatly to reduction in soil seed bank (Eplee and Langston, 1991). Rotations with cow pea resulted in reduced infestation of Striga asiatica in upland rice fields in Tanzania (Riches et al., 2005). Use of the trap crops encourages suicidal germination of Striga seeds (Ramaiah, 1983) therefore use of the crops can be important to control Striga.

2.7.1.3 Use of fertilizers

Several studies have reported success with the use of fertilizers in the control of Striga. Striga species are more prevalent in low fertility soils. Soil fertility technologies and mineral nutrients stimulate the growth of the host, at the same time affecting the germination of the weed (Aberyena and Padi, 2003). Use of nitrogen fertilizers can reduce Striga infections, for example application of Urea three weeks after sowing reduced Striga asiatica infestations on upland rice in Tanzania (Riches et al., 2005). Also Adagba et al., (2002a) reported delayed and reduced Striga infection in upland rice fields in Nigeria after using 90-120kgN ha-1. Reduction in Striga infestation as a result of use of fertilizer is believed to be partly due to decreased biosynthesis of germination stimulants strigolactones (Yoneyam et al., 2007a; 2007b). Therefore use of nitrogen and phosphorous fertilizers can greatly reduce parasitic weed germination (Lopez- Raez et al., 2009)

2.7.2 Chemical control

There are few herbicides that have been proved effective in control of Striga. Herbicides such as 2, 4-D or MCPA can be used to kill established weeds (Ivens, 1989). Herbicide 2, 4-D has been used against Striga hermonthica (Carsky et al., 1994a), Striga asiatica (Delassus, 1972). The herbicides control the weed post-emergence, but germinations continues after the residues loose effectiveness (Ivens, 1989). Also the weed exerts its harmful effects before it emerges above the ground. Therefore coating of seeds with herbicides can be effective at improving the chemical control, for example use of herbicide coated Imazapyr maize seeds in East Africa has been used as an effective method to manage Striga hermonthica and Striga asiatica (Kanampiu et al., 2003). Imazapyr is an Imidazolinone herbicide, based on ALS inhibition with broad spectrum and residual effects. Use of such herbicides can have an effect on the soil biology and suppression of Striga species (Ahonsi et al., 2004). To date, no herbicide-tolerant rice varieties are identified for the rain-fed uplands that can be used in combination with these ALS inhibiting chemicals as seed coating (Rodenburg and Demont, 2009)

2.7.3 Use of resistant varieties

Use of resistant crop cultivars is probably the most economically feasible and environmentally friendly means of Striga control. In general cultivars of the African rice species (Oryza glaberrima) show more Striga resistance than Oryza sativa genotypes (Riches et al., 1996; Johnson et al., 1997). An example is CG14 a cultivar of African rice that showed resistance against Striga hermonthica and Striga aspera (Johnson et al., 2000). Mechanisms of resistance in cereals can be categorized as post-attachment or pre-attachment resistance. Pre-attachment resistance involve mechanisms that prevent the parasitic weed from attaching on the host, such mechanisms include; the absence or reduced production of germination stimulants (Hess et al., 1992), this was observed by Jamil et al., (2011) who assessed eighteen NERICA cultivars and their parents, Oryza sativa L. (WAB 56-50, WAB 56-104 and WAB 181-8) and Oryza glaberrima Steud parent CG14.The results of his study showed a significant variation among NERICA cultivars and their parents for strigolactone production and Striga germination confirming feasibility of this approach in rice. NERICA cultivars 7, 8, 11 and 14 represented the top five highest strigolactone producers (Jamil et al., 2011) and NERICA cultivars -6, -10, -15, -2, -9 and -5 showed an intermediate production levels and NERICA-1 and CG14 produced the smallest amount of strigolactones meaning highly resistant cultivars (Jamil et al., 2011). Other mechanisms are; inhibition of germination by the host (Rich et al., 2004), inhibition or reduction in the formation of the haustorium (Rich et al., 2004), thickened host root cell walls resulting in mechanical barrier to infection by the weed (Maiti et al., 1984; Olivier et al., 1991). Post–attachment mechanisms involve mechanisms that prevent attachment of the parasitic weed, these mainly include; failure of the Striga to establish the xylem-xylem connections with the host as a result of blockage in vascular continuity an inability to penetrate through the endodermal barrier (Gurney et al., 2006) or in some cases due to a blockage in vascular continuity e.g. as observed in NERICA 10 following infection with S. hermonthica (kibos isolate) (Cissoko et al., 2011). Other mechanisms are; hypersensitivity reactions resulting in the death of host root tissue around the point of attachment discouraging parasite penetration (Mohammed et al., 2003); also plants can have natural ability to release phenolic compounds such as the rice phytoalexins in infected host cells. Cissoko et al. (2011) studied post-attachment resistance of eighteen NERICA cultivars and their parents, NERICA cultivars and their parents exhibited a range of susceptibility to the Striga species. They showed that some NERICA cultivars such as NERICA-7, -8, -9, -11 and -14 were very susceptible supporting between 100 and 150 parasites per root system, while NERICA-1, 2 and 10 exhibited the greatest resistance supporting a smaller number of parasites per root system (Cissoko et al., 2011). It is clear, as with many host-pathogen systems, that different cultivars of rice may be resistant to some ecotypes of Striga but susceptible to others. This is because of genetic variation in both Striga and the host; different ecotypes of S.hermonthica with different levels and mechanisms of virulence exist and may overcome the genetic resistance on some cultivars but not others (Mohamed et al., 2007; Scholes and Press, 2008). However, Cissoko et al, (2011) discovered that NERICA-1 and -10 exhibited a broad spectrum of resistance when subjected to different ecotypes of Striga hermonthica and Striga asiatica; this shows that broad-spectrum resistance exists and may play an important role in Striga control in rice. Also recently Rodenburg et al. (2015) evaluated the resistance of 18 NERICA cultivars and their parents under field conditions. Interestingly nine of the 18 NERICA cultivars (NERICA-1, -5, -10, -12, -13 and -17) and two of the NERICA parents (WAB181-18 and CG14 showed excellent resistance to S. hermonthica ecotype from Mbita. These same varieties had a good post attachment resistance to S. hermonthica ecotype from Kibos Western Kenya (Cissoko et al., 2011) and pre-attachment resistance to S. hermonthica ecotype Medani Sudan (Jamil et al., 2011). This is a clear indication that these varieties have a broad-spectrum of resistance to different ecotypes of Striga hermonthica. Most cultivars of the African rice species Oryza glaberrima show yield advantages under weedy conditions due to vigorous growth, high tillering ability and large phanophile leaves (Johnson et al., 1998; Saito et al., 2010). Combining

CHAPTER THREE

RESISTANCE OF UPLAND RICE VARIETIES AGAINST STRIGA HERMONTHICA UNDER FIELD CONDITIONS

3.1 Introduction

Resistant varieties present a cost- effective Striga control measure since their cultivation does not require a lot of resources and time. However genetic variability existing among different species and ecotype of Striga presents a great challenge in the use of resistant varieties. Striga hermonthica is an obligate out-crossing species (allogamous reproduction) (Safa et al., 1984). In allogamous species, variation is expected to be maintained from one generation to the next resulting in individuals that are heterozygous at many different loci (Huang et al., 2012). The resistance against such species is therefore weak because the genetic variability of S. hermonthica seed bank is so high, some individuals are likely to contain the virulence loci that over comes resistance in some varieties (Rispail et al., 2007; Scholes et al., 2008; Huang et al., 2012). Musselman et al., (1991) reported significant variations within and between populations of S. hermonthica in different areas of Africa. Resistance of sorghum varieties in one location and susceptibility in another location is an indication of the existence of different ecotypes of S. hermonthica in East Africa (Doggett, 1952). Similar results were reported by Ramaiah, (1987) in West Africa. This implies that crop varieties have to be screened in different agro-ecological zones to establish their resistance or susceptibility to the different ecotypes of Striga. For example when Johnson et al, (1997) screened a selection of upland rice cultivars, he reported that rice cultivars IR38547-B-B-7-2-2 and IR47097-4-3-1 remained parasite-free at two sites in Kenya but supported more parasite emergence in Ivory Coast. Also in this same study a number of cultivars showed greater susceptibility to Striga hermonthica in West Africa than when observed in field trails in Kenya by Harahap et al., 1993. This is a clear indication that these two areas have different ecotypes of Striga hermonthica. Therefore resistance of a variety in one place with a particular ecotype of Striga does not guarantee resistance in another place with another parasite ecotype. This study was therefore aimed at screening selected upland rice varieties reported to be resistant to Striga hermonthica or S. asiatica elsewhere to the ecotype of Striga hermonthica in Uganda.

3.2 Materials and Methods

3. 2.1 Experimental site

The study was carried out in Nsinze, a small village in Namutumba district in the eastern region of Uganda. It is bordered by Pallisa district to the north, Butaleja district to the east, Bugiri district to the south, Iganga district to the South east and Kaliro district to the North West. The district headquarters at Namutumba are located approximately 88 kilometers by road North West of Jinja, the largest city in the sub-region. It is located 000 51’ N, 330 41’ E. Namutumba district has one county and six sub-counties, thirty six parishes and 233 villages. The district has a total population of 167,691 people with a population density of 208 persons per km2. The district receives a bimodal type of rainfall, which averages at 1,250mm. The topography ranges from 1,167 m a.s.l to 1,249 m a.s.l (Namutumba Census report, 2007).

3.2.2 Rice varieties used

The experiment involved twenty- four varieties of upland rice obtained from Africa Rice Center, Tanzania and one local variety used as a control. These varieties were selected because of their putative different levels of resistance and yield (Table 1). The local check, Superica 1 (WAB165) is a high yielding but Striga-susceptible variety.

Table 1: List of varieties to be screened in Nsinze, Namutumba District, Uganda

3.2.3. Experimental design

The field experiment was laid out as a 5 × 5 lattice square design replicated six times. The twenty- five varieties constituted the treatments. The above design (a lattice square) and the high number of replicates (6) was used to account for the inherent high heterogeneity of natural Striga infestation in a farmer’s field (Rodenburg et al., 2015). Such spatial variation affects the correct interpretation of Striga resistance screening based on the number of emerged Striga plants.

3.2.4. Experimental layout

The experimental area measured 50 m × 15 m. The upland rice cultivars (treatments) constituted the sub-plots of the experiment. Each net sub-plot, containing an individual cultivar, was 1.25 m × 2.75 m (3.44 m2) and was separated from the adjacent sub-plot by an open row (0.50 cm) to avoid neighbor effects and to allow easy access. A distance of 1 m was left between each replicate for access and separation. These plots were maintained throughout the two seasons of 2014 and 2015 (Figure 1).

Figure 2: Schematic representation of the field trial in Nsinze, Namutumba district, Uganda with the indication of the path, the homesteads and the water pump.

Figure 3: The field trial in Nsinze, Namutumba district, Uganda in the first season (2014) at 25days after sowing.

3.2.5. Cultural practices

Field preparation was done two weeks before the beginning of the rains. The field was prepared using an oxen plough. Field preparation was done on 2 March 2014 for the first season and on 3 March 2015 for the second season.

Striga seed collected from farmers’ fields were used for artificial infestation of each sub-plot to supplement and homogenize the existing Striga seed bank of the soil. Each sub-plot received a similar amount of seed (~0.92g) mixed with 200 ml of white construction sand. This sand- seed mixture was applied to designated sub-plots and incorporated in the upper 5-10 cm of soil using a hoe. Only the part where rice is sown 1.25 × 2.75 m, hence 3.44m2 was infested.

Sowing was done at the onset of rains, on 12 March 2014 and on 13 March 2015 at a rate of six seeds per hill to a depth of 2-3 cm following a spacing of 0.25 m between rows and 0.25 m between plants in the row. A total of 55 hills were maintained per sub-plot constituting 5 rows and 11 hills of plants per row. Thinning was done at two weeks after sowing leaving 3 plants per hill. Weeding was done regularly every 10 days after Striga infestation to remove all weeds other than Striga. A basal application of (17:17:17) NPK fertilizer, at a rate equivalent of 50kg ha-1, was applied at 35 days after planting. Each sub-plot received 69g of NPK fertilizer.

The major pests of upland rice such as stalked- eyed flies, sting bug and rice bug were controlled using contact insecticide Dursban (Chlorpyrifos), every after 2 weeks starting 30 days after planting until grain formation, nematodes and termites were controlled using Furadan (Carbofuran) applied in plant holes before sowing.

3.4 Data collection

3.4.1 Soil data

Soil of the upper 20 cm was sampled prior to sowing in each block. Three randomly selected but distinct points were sampled per replicate and then combined into one sample equivalent to 200 g. This was used for the analysis of nitrogen, phosphorous, potassium, organic carbon and soil texture. Soil analysis was done at Makerere University, in the soil science laboratory;

Organic carbon content of the soil were determined by reduction of potassium dichromate by organic carbon compounds and oxidation-reduction titrations with ferrous ammonium sulphate solution. Soil texture was determined by a hydrometer method (differential settling within the water) using particles less than 2 mm diameter (Jou, 1979; FAO, 2008). This procedure measures percentage of sand (0.05-2.0 mm), silt (0.002-0.05 mm) and clay (<0.002 mm) fraction in soils. Total nitrogen content was determined by the Kjeldahl method (Dewis and Freitas, 1975). Soil pH was determined using a ratio of 1:2.5 soil to water ratio (Okalebo et al., 2002) and read from a glass electrode attached to a digital pH meter after shaking samples for 1 hour. Potassium was determined using Ammonium acetate extraction method and flame photometer as described in Okalebo et al., (2002). Available phosphorous was determined using the Bray 1 method since the pH of the soil was below 7 (Oisen and Sommers, 1982).

3.4.2 Crop data

Data collected on rice included; plant height, above ground biomass rice dry weight at harvest, number of tillers and grain yield (expressed at 14% grain moisture content). This was collected on a random sample of 9 hills per sub-plot as described; Tiller numbers were assessed starting 29, 43 and 57 days after rice sowing and at physiological grain maturity. Plant height was determined with a tape measure at 29, 43 and 57 days after sowing and at grain maturity. The height of the tallest plant was measured from the soil surface to the tip of the tallest leave or the tip of the tallest panicle. Above-ground rice biomass dry weight was determined at harvest. For each subplot the rice straw of the plants of 9 hills was harvested, air dried for about 1-2 weeks and then oven dried for 72 hours at 700C. Directly after oven drying the weights were assessed using a weighing scale and recorded.

Grain yield at grain maturity was determined from a random sample of 27 rice hills per sub-plot, this was expressed in kg per hectare, corrected at 14% grain moisture content. However, the panicles of the central 9 hills and the panicles of the remaining 18 central hills were put (and kept) in separate (cotton) bags. Panicles were cut and collected for each plot and air-dried. After 1-2 weeks of air-drying the panicles were weighed. The panicles were then threshed and the grains obtained after threshing were weighted. Then the grains were winnowed (removing all empty grains) and weighted again, immediately followed by grain moisture content measurements which will enable correction of all grains to a standard 14% moisture content. All measurements were recorded.

3.4.3 Striga data

Data collected on Striga include; Striga emergence, days to Striga flowering, Striga dry weight at rice harvest and above-ground Striga numbers as described below. Above-ground Striga numbers were assessed by close examination of rice varieties in the field. Regular Striga counts are done at 2-weekly intervals, starting after Striga emergence, to assess maximum above-ground Striga numbers (NSmax) and Area under the Striga Number Progress Curve (ASNPC) following Rodenburg et al. (2005). These counts included the number of living plants within the central observation area containing 27 hills (hence excluding the border rows) and the number of dead plants, recorded separately. Striga counts were taken at 29, 43, 57, 71, 85 days after rice sowing and at rice maturity. In each sub-plot, days to Striga emergence were monitored and recorded every two days starting at 28 days after sowing. Once Striga had emerged in a plot, it was marked with a coloured stick to exclude it from the next round of observations. Days to Striga flowering were estimated based on observations done every 3 days, starting at 21 days after first Striga emergence in each sub-plot. For assessment of Striga biomass dry weight, in each subplot, all above-ground Striga plants were collected from the central observation area containing 27 hills. During the season Striga plants that died before rice harvest were collected and put in an envelope designated to the associated sub-plot. At harvest all above-ground parts of all of the collected and remaining Striga plants (dead or living) were bulked and oven-dried for 72 hours at 70°C after which dry weights were obtained using an electronic weighing scale.

3.4.4 Soil characteristics of the field trial for 2014A and 2015B seasons

Table 2: Initial soil characteristics of the field trial for season 2014 A and 2015 B in Eastern Uganda.

Results of the chemical and texture analysis of the soil of the experimental field at the onset of the season in 2014 and 2015, is shown in Table 2. The pH was 6.1, and within range of optimum soils for upland rice established by Somado et al. (2008). Available nitrogen was highest in 2015, phosphorous was highest in 2014 while potassium was highest in 2015. According to Gerakis and Baer. (1999) the soil can be described as sand clay loam. The remaining micronutrients were highest in 2014 (Table 2). Organic matter of the soil was above the minimum >2% in 2014, but below this threshold in 2015. The nitrogen content was below the minimum threshold (<1.0N/kg) established by Oikeh et al. (2008) in both years. Oikeh et al. (2008), described nitrogen levels below 1.0g N/kg as low, between 1.0 and 2.0 N/kg as medium and above 2.0g N/kg as high. The soil phosphorus level was above the minimum of 3 mg/kg, suggested by Wopereis et al. (2009) and Oikeh et al. (2008), in 2014 but below this minimum in 2015. The potassium content of the soil was below the minimum of 0.2 cmole/kg in 2014, however it was above the minimum in 2015 (Oikeh et al., 2008). The remaining micronutrients except Calcium, Sodium and Magnesium recorded low values in both seasons.

3.5 Data analysis

Prior to analysis data was subjected to tests of normality and homogeneity of variance (Sokal and Rohlf, 1995). After these tests, data were analyzed using a Linear Mixed Model. Using Liner Mixed Model, cultivar was considered as a fixed effect and block, nested into replicate, and replicate as random effects. For analysis of above-ground Striga numbers (NSmax), data were transformed using Natural logarithms, log (x+1) to meet the requirement of normality of data distribution for classical ANOVA (Sokal and Rohlf, 1995). Using hierarchical cluster analysis, a squared Euclidean distance matrix basing on means was computed. This assisted in dendrogram plotting, clustering varieties into similarity groups basing on Striga resistance parameters (e.g. numbers and biomass) and rice grain yield under Striga infested conditions. Spearman rank correlations were also calculated between Ls means of NSmax and rice yields, NSmax and tiller number and NSmax and Striga dry weight. All the data were subjected to statistical analysis of variance (ANOVA) using GenStat computer software (V12).

3.6 Results

3.6.1 The growth parameters of upland rice varieties grown under Striga infested fields.

3.6.1.1 Plant height per variety

For both seasons, variety x season interaction effects on plant height was highly significant (P<0.001, F=4.97), and therefore data were analyzed separately for each season. Rice varieties had a highly significant effect (P<0.001) on plant height (Table 3). In both seasons, Blechai produced the tallest stems across varieties (Figure 1a, b). Plant height ranged from 129 cm for Blechai to 53 cm for Narc3 in 2014 (Figure 4a). In 2015 it ranged from 122 cm for Blechai and 36 cm for Superica (Figure 4b). In 2014, Narc3 and UPR had the shorter stems than any other variety, while in 2015 Superica had the shortest stem.

Table 3: Variance component analysis (F-stat and F-prob) and standard error of difference of means (SED) of varieties; effects on plant height, tiller number and productive tiller number for 2014 and 2015.

Figure 4: Height of rice plants per variety for 2014 (a) and 2015 (b), data presented as means and standard errors of means, height per variety were significantly different at p<0.001

In 2015, varieties differed significantly (P<0.001) in plant height (Table 3). Five classes can be identified, (1) varieties with tall plants (100-125cm), i.e. Blechai, Makassa, ACC and MG12 , (2) varieties with medium-sized plants (90-99cm), i.e. WAB880, NERICA-17, CG14, Anakila, IR49, NERICA-4 and NERICA-2, (3) varieties with medium to short plants (80-87 cm), i.e. NERICA-10, SCRID090, IRGC, WAB50, AGEE, NERICA-1 and WAB181-18, (4) varieties with short stature (70-75 cm), i.e. WAB935, WAB928 and WAB56-104 and (5) those with very short stems (34-50 cm), i.e. UPR, NARC3, IAC165 and Superica (Figure 4b).

Similarly in 2014, the varieties varied significantly in heights (P<0.001) (Table 3). Five classes were also identified, (1) varieties with the tallest plants (101-130cm), i.e. Blechai and Makassa, (2) varieties with medium stem height (90-100cm) i.e. WAB880, ACC, NERICA-4, SCRID090, NERICA-17, CG14, Anakila, NERICA-2 and WAB56-104, (3) varieties with medium to short stature ( 81-88cm) i.e. Agee, IR49, NERICA-10, NERICA-1, WAB181-18, and WAB56-50, (4) varieties with short stems 70-80cm i.e. IAC 165, IRGC, MG12, Superica, WAB928 and WAB935, (5) varieties with very short stems (50-58cm) i.e. NARC3 and UPR (Figure 4a).

3.6.1.2 Number of tillers and productive tillers per variety

Variety x season interaction effects on tiller number and productive tiller number was highly significant, (tiller number: (P<0.001 F=2.36; Productive tiller number: P<0.001 F=5.54). Therefore data were analyzed separately for each season. There was a significant difference (P<0.001) in the number of tiller and productive tillers produced per variety for both seasons (Table 3). There was a highly positive correlation between the number of tillers and the number of productive tillers in both years, though in 2014 the correlation was weaker (Spearman correlation r2015= 0.78, P<0.001; r2014=0.45, P=0.006).

In 2014, variety CG14 had the highest number of tillers per plant while varieties Blechai and IAC had the lowest number of tillers per plant (Figure 5a). Additionally with respect to the number of tiller produced, varieties can be grouped into high tillering varieties (AGEE, Anakila, UPR and WAB928), intermediate high tillering (ACC, MG12, NARC3, WA935) to intermediate low tillering (IR49, IRGC and Makassa ). (Figure 5a) other varieties were low tillering (NERICA-17, NERICA-4, WAB56, WAB181-18, WAB880, NERICA-10, SCRID090, NERICA-2 and Superica) and lastly very low tillering (WAB50, NERICA-1, Blechai and IAC) (Figure 5a).

In 2015, varieties Narc3 and CG14 had the highest number of tillers compared to other varieties (Figure 5b). Additionally local check (Superica) and IAC had the lowest number of tillers compared to other varieties (Figure 5b). According to the number of tillers produced in 2015, varieties can be grouped into group 1 (NARC3 and CG14) with an average of 27 tillers, group 2 (MG12, AGEE, Anakila, Makassa, UPR, ACC and WAB928) with an average of 17 tillers, group3 (WAB935 and 1RGC) with an average of 14 tillers, group 4 (NERICA-17 and IR49) with an average of 11 tillers (Figure 5b). Group5 (NERICA-2, WAB880, NERICA-4, NERICA-10 and WAB181-18) with an average of 9 tillers, group 6 (SCRID090, Blechai, NERICA-1, WAB50 and WAB56) with mean tiller of 7 tillers and lastly group 7 (IAC165 and Superica) with average of 3 tillers (Figure 5b).

There were significant (P<0.001) differences among rice varieties with respective to productive tillers (Table 3). In 2014 variety AGEE had the highest number of productive tillers while in 2015, variety CG14 had the highest number of productive tillers. Varieties MG12 and WAB935 had the lowest number of productive tillers as compared to other varieties in 2014 (Figure 4b and d). Contrary in 2015 varieties IAC and Superica had the lowest number of tillers compared with other varieties (Figure 5b).

In 2014, with respective to mean productive tillers, varieties were grouped into group 1(CG14, Anakila, AGEE) group 2 (Makassa and UPR), group 3 (ACC, NERICA-10, NERICA-2, WAB181-18 and WAB928) group 4 (NERICA-17, NERICA-1, NERICA-4, SCRID090, WAB56 and WAB880), group 5 ( Blechai, IR49, IRGC, Superica and WAB50) and lastly group 6 ( IAC, MG12, NARC3 and WAB935) (Figure 5c).

On the other hand in 2015, based on mean productive tillers varieties were categorized into group 1 (AGEE and CG14), group 2 (ACC, Anakila, Makassa, MG12, UPR and WAB935), group 3 (IRGC, NERICA-17 and WAB928), group 4 (IR49, NERICA-10, NERICA-4, NERICA-2, WAB56 and WAB880), group 5 (Blechai, NERICA-1, NARC3, SCRID090 and WAB50) and lastly group 6 (IAC and Superica) (Figure 5d). The mean maximum above ground Striga numbers (NSmax) per variety correlated negatively with tiller number per variety for both 2014 and 2015 (Spearman correlation coefficients r2014= -0.265 and r2015= -0.077).

Figure 5: The number of tillers per rice variety 2014a and 2015c and productive tillers per rice variety 2014b and 2015d. Data presented as means and standard errors per variety.

3.6.1.3 Rice biomass of upland rice varieties under Striga hermonthica infested field conditions.

Figure 6: Above-ground biomass dry weight of rice varieties grown under Striga hermonthica infested field conditions for 2014 (a) and 2015 (b). Data presented as means and standard errors of means per variety.

For both seasons, variety x season interaction effects on rice biomass was highly significant (P<0.001; F=2.61), therefore data were analyzed separately for each season. Rice biomass was significantly different (P<0.001) among different varieties of rice for 2014 and 2015 seasons (Table 4). In 2014, varieties WAB928, Makassa and ACC had the highest rice biomass compared to other varieties (Figure 6a). Additionally the local check (Superica) had the lowest rice biomass in the same season. However for 2015 season, varieties WAB935, WAB928 and IRGC had the highest rice biomass while IAC and local check (Superica) had the lowest rice biomass per rice plant (Figure 6b).

Table 4: Variance component analysis (F-stat and F-prob) and Standard errors of differences of means (SED) of varieties; effects on rice grain dry weight (Rice grain DW), rice straw dry weight (Rice straw DW) and above-ground Striga biomass dry weights (Striga DW) at harvest for season of 2014A and 2015B

a Rice grain DW and rice straw DW for season 2014 A had 23 degrees of freedom, because the field was cultivated by farmers before MG12 was harvested.

Based on the rice biomass produced by each variety, varieties can be categorized into four groups: (1) high rice biomass producers, i.e. WAB935, WAB928, ACC in 2014 and Makassa and WAB928, WAB935 and IRGC in 2015, (2) intermediate rice biomass producers, i.e. WAB880, CG14, Blechai, IR49, IRGC and NERICA-4 (2014) and ACC, Blechai, MG12, Makassa, WAB880, CG14 and IR49 (2015), (3) intermediate low rice biomass producers, i.e. SCRID090, NERICA-2, NERICA-17, NERICA-1, NARC3, Anakila and AGEE (both 2014 and 2015) and WAB56-50 (2015) and (4) the low rice biomass producers, i.e. WAB181-18, WAB56-104, UPR, NERICA-10, IAC and local check Superica (2014) and Superica and IAC (2015) (Figure 6a, b)

3.6.1.4 Days to Striga emergence and flowering of the different rice varieties.

For both seasons, variety x season interaction effects were significant (SFED; P=0.004, F=2.04; SFFD; P=0.021, F=1.18). Data was therefore analyzed separately for both seasons.

Table 5: Variety effect on Striga hermonthica emergence and flowering days.

There was a significant difference in first Striga emergency P=0.005 among rice varieties for 2014 and P<0.001 for 2015 (Table 5). Similarly, rice varieties varied significantly in the time to Striga flowering for season 2014 at P=0.002 while in 2015 at P<0.001 (Table 5).

The most susceptible varieties such as local check (Superica) and IAC had Striga emerging and flowering faster than any other variety in the trial (Table 5). Also the most resistant such as NERICA-2, NERICA-10 and WAB928 had Striga emerging later in the growing season especially in the season of 2015 (Table 5).

3.6.1.5 Maximum above ground Striga plants and Striga dry weight among different rice varieties

For both seasons, the variety x season interaction effects were not significant (P=0.298), therefore data were combined and analyzed together. Rice varieties had a highly significant effect (P<0.001) on the maximum number of emerged Striga hermonthica plants (NSmax) in both seasons. Rice varieties also significantly affected (P<0.001) Striga dry weights at harvest for the two seasons (Table 4). In all seasons, mean maximum above-ground Striga numbers (NSmax) per variety correlated positively and highly significantly (P<0.001) with the mean Striga dry weights at harvest per variety (Spearman correlation coefficients were r2014=0.89 and r2015=0.88).

Based on the maximum above- ground S. hermonthica numbers (NSmax) observed in the field in both seasons, varieties can be categorized into five statistically separate clusters: (1) very resistant , (2) moderately resistant, (3) intermediate, (4) susceptible, (5) very susceptible. Varieties NERICA-2, NERICA-10 and WAB928, with a mean NSmax of 0.30 per m2, were classified as very resistant while varieties Anakila, CG14, IR49, NERICA-17, NERICA-4, WAB880, Blechai, IRGC, WAB935, SCRID090, WAB181-18, AGEE, with a mean NSmax of 0.85 per m2, were moderately resistant (Figure 7a,b). Other varieties such as ACC, Makassa, UPR, MG12 and NERICA-1 with NSmax of 1.27 per m2 were classified as intermediate (between resistant and susceptible). Varieties WAB56-50, WAB56-104 and NARC3 (ITA 257) with an NSmax of 1.75 per m2 were classified as susceptible. The last cluster consisted of very susceptible varieties with NSmax of 2.24 per m2 i.e. IAC165 and local check (Superica) (Figure 7a, b).

Figure 7: Maximum number of emerged Striga plants m-2 (NSmax) per variety for two seasons (6a) and cluster analysis of rice varieties (6b). Data presented as means and standard error of means. NSmax data was transformed using (logx+1)

For both seasons, variety x season interaction was significant (P<0.001; F=2.21), therefore data were analyzed separately for each season. Striga dry weight varied significantly (P<0.001) among the rice varieties for the two seasons. In both seasons, IAC and local check (Superica) produced the highest Striga dry weight at harvest (Figure 8). These were followed by WAB56-50, NARC3 andWAB56-104 for both seasons. Also varieties NERICA-1 and UPR for both seasons produced moderate Striga biomass compared to other varieties though not as high as the above mentioned varieties. The rest of the varieties had a lower Striga biomass at harvest when compared to the above varieties in both seasons (Figure 8).

Figure 8: Above ground dry weight of Striga per variety for 2014 and 2015. Data presented as means and standard error of means.

Figure 9: Rice grain dry weight per variety for 2014 (a) and 2015 (b) and cluster analysis for rice grain dry weight among rice varieties for 2014 (c) and 2015 (d). Data presented as means and standard error of means.

For both seasons, variety x season interaction effect was highly significant (P<0.001), therefore data were analyzed separately for each season. Rice grain dry weight varied significantly (P<0.001) among rice varieties for both seasons (Table 4). In 2014 SCRID090 produced the highest yield while NARC3 and WAB928 produced the lowest yield (Figure 9a). In 2015, NERICA-2 and NERICA-4 produced the highest yield while NARC 3 produced the lowest yield (Figure 9b). Yield per variety in 2015 ranged from 0.4 t ha-1 to 3.7t ha-1 with an average yield of 2.4 t ha-1. In 2014 yield per variety ranged from 0.1 t ha-1 to 3.5 t ha-1 with an average yield of 2.1 t ha-1 (Figure 9a, b). In 2015, there was a significant correlation between the above-ground emerged Striga plants (NSmax) and grain yield (r2015= -0.435; P=0.03) with more resistant varieties showing the highest grain yield (Figure 8b). In 2014, the infestation level being low, the correlation between NSmax and grain yield though negative was very weak and not significant (r2014= -0.182; P= 0.383).

Based on the yield produced per variety, varieties can be categorized into five statistically separate clusters: (1) high yielding, (2) intermediate-high yielding, (3) intermediate-low yielding, (4) low yielding and (5), the lowest yielding varieties. In 2014, variety SCRID090 with an average of 3.6 t ha-1, was classified as the high yielding variety while in 2015 this cluster consisted only of NERICA-4 and NERICA-2 with an average of 3.6 t ha-1 (Figure 9c, d). Varieties classified as intermediate-high yielding for 2014 were NERICA-2, NERICA-4, NERICA-10, NERICA-17, WAB880, WAB181-18, NERICA-1, Anakila, WAB56-104, with an average yield of 3.0 t ha-1, while in 2015 this cluster consisted of NERICA-17, WAB880, AGEE, CG14, SCRID090, ACC, WAB181-18, NERICA-10, Blechai, Makassa and NERICA-1, with an average yield of 3.2 t ha-1 (Figure 9c, d). Intermediate-low yielding varieties in 2014 were AGEE, WAB56-50, Blechai, Makassa ACC, IAC, Superica with an average yield of 2.0 t ha-1, while in 2015 this cluster consisted of MG12, IRGC, Anakila, IR49, WAB56-50, and WAB 56-104, with an average yield of 2.0 t ha-1. The low yielding in 2014 cluster were CG14 and UPR variety with an average yield of 1.3t ha-1 while in 2015 this cluster was consisted of UPR, WAB935, WAB928, with 1.2 t ha-1 in 2015 (Figure 9c, d). The cluster with the lowest yielding varieties, producing less than 1 t ha-1, contained IR49, IRGC, WAB935, NARC3 and WAB928 in 2014 and Superica (local check), IAC and NARC3 in 2015 (Figure 9c, d).

3.7 Discussion

3.7.1 Effect of Striga hermonthica on the yield and growth parameters when rice varieties are grown under Striga infested fields.

It was observed that the tallest varieties were also the most resistant and some of the shortest varieties were among the most susceptible varieties. This could be partly attributed to differences in Striga infestation levels among varieties. For better assessment of the effects of Striga on stem stunting of rice plants, you would need Striga-free control plants which is provided in chapter five. Varieties IAC 165, Superica had short stems probably due to high Striga infestation. Similar results were reported by Atera et al. (2012) who also showed a reduction in plant height of the susceptible rice varieties in Western Kenya. Johnson et al. (1997) also reported high levels of Striga infestation reflected in increased levels of stunting in rice plants of the susceptible cultivars, including IAC 165. Other varieties such as Makassa, IR49, ACC, with the highest plant height as per this study were reported by Johnson et al. (1997) as having the lowest stunted levels in Ivory coast. The other reason for the shortness of the rice plants could be related to the genetic make-up of the varieties. For example varieties WAB928 and WAB935 inspite of being short were highly resistant to Striga in both seasons. The Oryza glaberrima varieties, on the other hand, with intermediate resistance levels, were all relatively tall. Previous studies have shown that O. glaberrima cultivars, have a different plant type than O sativa cultivars and are therefore often more competitive to weeds (Johnson et al., 1995). NERICA varieties and Other O.sativa varieties were tall mainly because of reduced Striga infestation or simply because they are genetically tall varieties.

3.7.1.1 Total tiller and productive tiller number

Tiller number and productive tiller number varied highly across varieties. Tiller number and productive tiller number correlated positively. However the correlation was weak in r2014= 0.451 and stronger in r2015=0.78. This suggests that there was wasteful production of tillers with a few physiologically mature grains in 2014 than in 2015. This difference is more likely related to stress levels and timing this stress sets in. Normally the plant will adjust the number of tillers it produces with available resources. Studies have indicated that stress occurring in the initial stages will not tremendously affect yield as stress in following stages (Boonjung and Fukai, 1996). Stress occurring at the vegetative stage is expressed in decreased number of tillers (Boonjung and Fukai, 1996). However if the Striga attaches on the host later after tiller initiation, it is more likely that because of nutrient diversion by the parasite (Parasitism), the plant will not be able to produce panicles on all the tillers.

In respect to the tiller production, it has been shown that strigolactones have an effect on tiller production as well as on Striga infection (Jamil et al., 2012). Strigolactones inhibit tiller/shoot branching (Umehara et al., 2010) and also triggers Striga germination (Jamil et al., 2011). This will therefore mean that varieties that produce high amounts of strigolactones are more likely to produce less tillers. Also such varieties will be susceptible to Striga. This same reason would explain the lower number of tillers produced by varieties IAC165, Superica, WAB56-50 and WAB56-104 since they were the most susceptible to Striga as per this study. Additionally, WAB56-50 and IAC 165 were ranked by Jamil et al. (2011) and Jamil et al. (2012) as the highest strigolactone producers. Studies by Jamil et al. (2011) ranked variety WAB56-104 as the lowest strigolactone producer. It is however not clear why it has a low tiller number given a lower strigolactones produced. The study has shown that O. glaberrima varieties (ACC, CG14, Makassa, Anakila, AGEE and MG12) produced higher number of tillers per plant. Varieties Agee, Anakila and CG14 have also been reported to produce a high number of tillers (Jamil et al., 2011; Jamil et al., 2012). This high number of tillers among these varieties could be partly related to low strigolactone production. This is consistence with studies by Jamil et al. 2012 who also showed varieties AGEE, Anakila and CG14 as low strigolactone producers.

The differences in tiller production among rice varieties cannot entirely be accounted by Striga infestation due to increased strigolactone production. Genetic differences across varieties can also explain this occurrence. Similarly Fageria et al. (1997) acknowledged tillering characteristics to be related to genetic characteristics of the variety. For example O.glabberima varieties tend to produce more tillers than O.sativa varieties (Jamil et al., 2012). This would explain the low tiller numbers on NERICA varieties, and O.sativa varieties SCRID090, WAB880, WAB181-18, IRGC and IR49 compared with O.glabberima varieties with similar or high resistance levels. The high tiller production in O.sativa varieties WAB928 and WAB935 could be related to low strigolactones since these varieties were highly resistant to Striga or simply high tillering potential. This study also observed a negative correlation between NSmax and the number of tillers in 2014 and 2015. This would indicate a negative effect of Striga on tiller production. However this is across varieties with some varieties more affected than others. This also suggest that lower tillering varieties are higher strigolactone producers and are more susceptible to Striga.

3.7.1.2 Rice biomass

Results from the current study indicate differences in rice biomass production across varieties. Varieties such as IAC 165 and Superica with lower rice biomass producers were also the most susceptible varieties. It is therefore possible that parasitism lowered their biomass. Additionally studies by Johnson et al. (1997) have reported Striga-inflicted biomass reduction in IAC 165. Also the high rice biomass among the O.glabberima, NERICA varieties and O.sativa varieties IRGC, IR49, WAB928, WAB880, SCRID090 and Blechai is partly related to lower parasitic load on these varieties. It has also been proposed that a grass plant is a collection of tillers (Nelson et al., 1992). This could imply that those varieties with high tillering capacity could also have high rice biomass. This would also explain the difference in rice biomass among O.glabberima varieties, NERICA varieties and O.sativa varieties because former produce a lot of tillers. Another observation from the study was some varieties such as NARC3 (ITA 257), WAB56-50, O.glabberima varieties ACC, Makassa, MG12 and Varieties NERICA-1 and UPR categorized as susceptible and intermediate respectively had a high rice biomass. Similarly Johnson et al., (1997) also reported a high rice biomass in O. glaberrima varieties Makassa, ACC with relatively high levels of Striga hermonthica infection. Varieties combining high Striga numbers with high host biomass are deemed tolerant. Tolerance is the ability of a variety to withstand Striga infection with minimum yield losses (Rodenburg and Bastiaans, 2011). It is also true that highly resistant varieties can also have parasites developing on them. Therefore a combination of tolerance and resistance is a very important strategy to improve crop yields (Rodenburg and Bastiaans, 2011). Tolerance has also been reported in sorghum (Gurney et al., 1995; Van Ast et al., 2000; Rodenburg et al., 2006). However tolerance cannot be easily assessed in the field as it would require infected and uninfected control plants which was not been provided in this study (see Chapter five).

3.7.2 Resistance levels of rice varieties grown under Striga infested field conditions

Rice varieties varied significantly in days to first Striga emergency and days to Striga flowering. The average time taken by a Striga plant to complete its cycle from emergence to flowering varied greatly among seasons from 73 days for 2014 and 94 days for 2015. Studies by Atera et al. (2012) reported the first Striga plant emerging 42 after rice emergence with minimum of 56 days to complete life cycle. The current study however reports a shorter time to first Striga emergence (38 days) in 2014, and a longer time to first emergence in 2015 (50 days). The difference in emergence days between Atera et al. (2012) study and this study could be attributed in differences in rainfall, Striga seed bank and genotype. Several studies have reported variation in first Striga emerging for example 21 days for sorghum in Sudan (Bebawi, 1981), 54 days for sorghum in Mali (Clark et al., 1994), and 35 days for sorghum and maize in Kenya (Gurney et al., 1995). Low soil moisture at critical stages, caused by low or rainfall levels, can prevent the germination of Striga seed (Ransom and Njorge, 1991). This same reason can account for the late emerging of Striga plants in 2015, the rains appeared about two weeks later when compared in 2014. This partly explains the higher Striga numbers in this season because even after rice physiological maturity, new Striga plants were emerging. Generally even the most susceptible cultivars in 2015 had their first Striga plants emerging later than in 2014.

There were also differences in Striga emergence and Striga flowering across rice varieties with more resistant varieties i.e. WAB928, NERICA-2 and NERICA-10 having the Striga plants emerging and flowering later. In comparison with the susceptible varieties i.e. Superica, IAC 165, WAB56-104 and WAB56-50 with earlier emergence and flowering Striga plants. This can be attributed to differences in susceptible to resistance levels among rice varieties. Similarly, Gebremedhin et al. (2000) reported earlier emergence of Striga in susceptible sorghum cultivar compared to the resistant one.

3.7.2.1 Rating the resistance of rice varieties against Striga hermonthica under Striga infested field conditions.

Results from the current field study indicate a highly significant difference in the number of emerged Striga plants (NSmax) and Striga dry weight among the 25 varieties screened in both seasons. All the NERICA varieties (NERICA-10, NERICA-17, NERICA-2 and NERICA-4) except NERICA-1, seven of O.sativa varieties (WAB928, WAB935, IR49, IRGC, Blechai, WAB880 and SCRID090), one O.sativa parent (WAB181-18) and three O. glaberrima (CG14, AGEE and Anakila) showed an excellent resistance to the S. hermonthica ecotype from Namutumba, Uganda. All the NERICA varieties, one O. glaberrima variety CG14 and one O. sativa parent WAB181-18 in this study have also been reported to resistant to Striga hermonthica ecotype from Mbita, in western Kenya approximately 211 km south east of Namutumba, under field conditions (Rodenburg et al., 2015). Similarly rhizotron experiments by Cissoko et al. (2011) have also ranked these same varieties as having an excellent post-attachment resistance to S. hermonthica ecotype found in Kibos, western Kenya. Similar results reported by Jamil et al. (2011) in his pre-attachment resistance study on an ecotype of Striga hermonthica from Medani (Sudan). This implies that these varieties have excellent resistance levels to various ecotypes of Striga. Broad spectrum resistance of rice varieties to Striga have also been reported by Cissoko et al. (2011). Additionally varieties AGEE and Anakila categorized as resistant as per this study were also reported by Jamil et al. (2012) as having a lower number of Striga plants emerging on them. Varieties WAB928 and WAB935 with an excellent Striga resistance as found in this study have also been reported to have an excellent resistance to Striga hermonthica by Johnson et al. (2000) in Ivory Coast. We also carried out a rhizotron study on a selection of varieties, whereby WAB928 showed an excellent post attachment resistance to Striga hermonthica (see Chapter 5). On the other hand resistance found with varieties SCRID090 and WAB880 have not been reported before.

IAC 165 and Superica were the most susceptible varieties in this study. For IAC 165, it was used as a susceptible check variety in this study and it exhibited high susceptible levels confirming studies by Johnson et al. (1997); Gurney et al. (2006); Cissoko et al. (2011); Jamil et al. (2012); Rodenburg et al., (2015) where it was grouped as very susceptible to Striga. Superica, the local check variety, was very susceptible probably due to continuous cultivation of the same variety in the region. Therefore it is likely that the virulence levels of local Striga population against this variety increased in time. Similar insights were provided by Rodenburg et al. (2015) on variety Supa India which was very susceptible in Kyela but resistant at Mbita where it had not been grown before. The O .glaberrima varieties (Makassa, MG12, and ACC) had an intermediate level of resistance to S. hermonthica. Similar results were reported by Johnson et al. (1997) who also showed varieties ACC and Makassa having a partial resistance to S. hermonthica. Studies have shown that O. glaberrima have an excellent competitive ability over weeds (Johnson et al., 1995). This could explain the lower Striga plants on these varieties. Also Johnson et al. (1997) reported later attachment of Striga on O.glabberima varieties and because they are vigorous growers, it is possible that they can outcompete Striga. Similarly Johnson et al. (1997) reported that O.glabberima varieties are less affected by Striga compared to O.sativa varieties. Other varieties such as Blechai and UPR have been reported in Harahap et al. (1993) to be Striga resistant. This study has however demonstrated partial resistance of UPR. Varieties NARC 3 (ITA 257), WAB56-50 and WAB56-104 were classified as susceptible in this study. Studies by Harahap et al. (1993) also classified variety NARC3 (ITA 257) as being susceptible to S. hermonthica ecotype in Kenya. On the other hand, WAB56-50 was ranked by Jamil et al. (2011) as the highest strigolactone producers. Since strigolactones trigger the germination of Striga, it is therefore not surprising that these supported a high number of Striga plants as per this study.

3.7.2.2 Striga dry weight

Results showed that there was a highly significant positive correlation in NSmax and Striga dry weight at harvest, confirming Rodenburg et al. (2015). The highly susceptible varieties such as IAC 165 and Superica did not only have the highest Striga numbers but also had the highest Striga dry weight at harvest. This can only be explained by earlier infection Striga as a result of increased strigolactone production. Similarly Van Ast and Bastiaans. (2006) reported an increase in Striga dry weight when sorghum plants were infected with in Striga 7 days after sowing compared to 21 days after sowing. However this only holds true for the most susceptible varieties. Varieties WAB56-50, WAB56-104 and NARC 3 also with high numbers of emerged Striga plants had a high Striga dry weight at harvest though not as high as IAC 165 and Superica. NERICA-1 and UPR also had relatively high Striga dry weight compared to the rest of the varieties though not as high as the earlier mentioned varieties. The rest of the varieties had a lower Striga dry weight at harvest. This could be attributed to only a lower number of Striga plants but also to the small sized parasitic plants on these varieties.

3.7.3 Yield performance of rice varieties under Striga infested conditions

Rice varieties combining excellent resistance to Striga hermonthica with high yields and environmental adaptability would be very useful to rice farmers growing in Striga prone areas. Results of this study showed varying levels of yield among the 25 rice varieties and a negative correlation between number of emerged Striga plants (NSmax) and grain yield in 2015. Normally it is in conditions of high Striga infestation that such correlation occurs (Rodenburg et al., 2005; Rodenburg et al., 2015). This would imply that highly resistant varieties provide a yield advantage under conditions of high Striga infestation.

The results also reported high yield performance among the NERICA’s varieties and one O. sativa parent, WAB181-18. The latter variety was reported before by Rodenburg et al. (2015) as high yielding under Striga-infested conditions. The relative high yields among the NERICA’s can be attributed to the relative low Striga hermonthica infection levels and good yield potential and environmental adaptability. Previous studies have demonstrated NERICA’s high yielding ability (Gridley et al., 2002; Saito et al., 2012). All the O. glaberrima varieties, CG14, Makassa, AGEE, ACC and Anakila were categorized as intermediate high to low yielding. These varieties are not only high tillering (Jamil et al., 2012) but also have a greater competitive ability with weeds (Johnson et al., 1995). This characteristic, combined with lower Striga numbers (NSmax) recorded per variety could be the reason for this yield. Studies by Johnson et al. (1997) also reported Striga hermonthica number among O. glaberrima varieties though the rice biomass was not reduced highly. This could imply a reduced effect of Striga on these varieties compared to O.sativa varieties IAC 165 and Superica. These varieties can be regarded as tolerant to Striga, a situation where varieties with high parasitic load still produce high yields. Varieties SCRID090, WAB880, Blechai produced a higher yields under high Striga pressure in 2015 and low Striga pressure in 2014 compared to other varieties. This could be attributed to relatively low Striga infestation.

Results show that NERICA-1, WAB56-50 and WAB56-104 which had high Striga numbers produced a high yield under these conditions. This could imply that these varieties are tolerant to Striga hermonthica producing significantly high yields even under high Striga levels. Tolerance has also been reported in sorghum (Gurney et al., 1995; Rodenburg et al., 2006). Understanding of this trait can help plant breeders to incorporate it in the most resistant varieties in order to improve yield (Rodenburg et al., 2008; Rodenburg and Bastiaans, 2011). However to further exploit this phenomenal, one needs to have Striga-free plants of the same varieties looking at relative yield losses (Rodenburg et al., 2006; Rodenburg et al., 2008). Another reason could be that they have a high yielding potential. The lower yields of local check (Superica) and IAC165 can be attributed to high Striga infection levels since these two supported the highest number of parasitic plants.

The yielding ability of the varieties cannot be solely attributed to Striga infection. For example inspite of the excellent resistance levels of varieties WAB928, WAB935, IR49 and IRGC, these produced a lower yield. This could be attributed to limited adaptability to the prevailing weather conditions or simply because of an inherent lower yield potential. These varieties progress to flowering when the rains are over and therefore are affected by too much sunlight. This significantly affected their yield potential. Also the differences in yield ranking across the varieties in 2014 and 2015could be partly attributed to rainfall levels in these years. For example, varieties Makassa, AGEE, ACC, Blechai and CG14 ranked as intermediate low yielding and low yielding respectively in 2014 were ranked as intermediate high yielding in 2015.

3.8 Conclusions

The main aim of this study was to evaluate the resistance of upland rice varieties against the Striga hermonthica ecotype of Namutumba (Uganda). The study has reported variations in resistance among rice varieties and showed that NERICA varieties (NERICA-2, NERICA-10, NERICA-17 and NERICA-4) have excellent resistance levels. The study has also documented resistance of varieties SCRID090, WAB880 to Striga hermonthica for the first time. Additionally it has demonstrated variations in resistance and tolerance levels to Striga hermonthica of O. glaberrima varieties ACC, CG14, Makassa, Anakila, AGEE and MG12.

The study has also demonstrated differences in yield potential among rice varieties. The highly resistant varieties did not necessarily guarantee good yields. We also show that adaptability to the prevailing weather conditions is an important yield determining characteristic. Therefore the use of Striga resistant rice varieties that combine this resistance with high yields and adaptability to prevailing conditions is thought to be a key element in sustainable rice production in Striga prone areas. The identified Striga resistance/tolerance can be incorporated in high-yielding or locally adapted breeding lines. Some of the lines that already combine Striga resistance with good yield levels and adaptability could be readily promoted among rice farmers in this area.

CHAPTER FOUR

FARMER PARTICIPATORY VARIETY SELECTION OF UPLAND RICE VARIETIES

4.1 Introduction

Rice is the world’s most important cereal considering the area under cultivation and the people who depend on it; it is becoming an increasingly important food and cash crop in Africa (Seck et al., 2012). In response to increasing demand for rice in Eastern Africa, farmers have to increase or expand the area under rice production (NPA, 2010). This has led to farmers encroaching on wetlands threatening their existence. In Uganda a few upland rice varieties have been cultivated including NERICA- 1, -4, -10 (Kijima et al., 2011) and local varieties such as Superica. Therefore utilization of improved varieties that are high yielding, Striga resistant and better responding to inputs is the first step in increasing rice production. However adoption of new varieties of rice necessitates that these varieties possess traits that are acceptable to farmers. The participation of farmers at the beginning of variety development or, in the case of established varieties, in the selection and dissemination, is a means to incorporate their preferences. This is essential for breeding or variety disseminations programmes, if real impact is to be achieved (Nanda, 2001). There is also a need to avail information on improved rice varieties that are adapted to the local farmer conditions. Therefore the objective of this study was to identify, in a participatory manner, upland rice varieties that could be disseminated among farmers in Eastern Uganda.

4.2 Materials and methods

4.2.1 Participatory variety selection trial

This study was undertaken at the grain maturity stage in the variety screening trial in Nsinze, Subcounty, Bulagala village, Namutumba district in the main rainy season of 2015. The field screening trial was set up as described before in Chapter three. An open-ended questionnaire was set up to avail information on variety preferences and traits that are important to farmers.

4.3 Data collection

Participants were asked to select the five best varieties among 25 varieties, in terms of their appearance in the field. In order to select the five best varieties, scores were assigned from 1 to 10; with 10 meaning ‘superior’ and score 1 meaning ‘inferior,’ these were used to rank the most preferred varieties. Farmers were also asked to provide the five worst performing varieties in the field trial. To avoid biased choices, participants were not allowed to communicate to each other during the selection exercise. After making the choices and considerations, participants were engaged in a one-on one discussion with extension officers and researchers to avail data that would explain why the choices were made. This was aimed at getting diverse views on preferences and selection criteria by farmers. Farmers were also asked to provide information on the most preferred traits in rice varieties and rank them from 1 to 5, with 1 indicating ‘not important at all’ and 5 indicating ‘very important’. Farmers evaluated the expression of these traits in the selected varieties, to avail whether these varieties would correspond well enough to these criteria to disseminate them among Ugandan farmers.

4.4 Data analysis

Data obtained from questionnaires and interviews was coded, entered in a database and analyzed using descriptive analysis procedures of the statistical package for social scientists (SPSS, 2000) version 16 computer package. Graphs, frequency tables, means and percentages generated were used to summarize responses from respondents. To find whether there was a significant difference in variety selection with respect to gender, data was restructured keeping sex as a fixed variable and varieties as target variables. Then an independent sample t-test was performed.

4.5 Results and discussion

4.5.1 Results

4.5.1.1 Social demographic factors and Rice farming practices

A total of 21 (55%) men and 17 (45%) women were involved in the rice selection exercise (Table 6). The largest share of these farmers was in the age-group 42-54 years (37%) followed by the age-group 31-41 (28%), 18-30 (24%) and lastly 55 and above (11%) (Table 6). Out of 38 farmers, 17 farmers (45%) had between 4 and 6 years’ of experience in growing upland rice, 11 farmers (29%) had 1-3 years, 9 farmers (24%) had 7-9 years and only one farmer (3%) had between 10-12 years of experience (Table 1).

Table 6: Farmer participant characterization: gender ratios, age-groups and rice farming experiences of farmers in Namutumba district Eastern Uganda in 2015.

4.5.1.2 Rice variety selection; the most preferred varieties by farmers in Eastern Uganda

The most preferred variety of rice was SCRID090 (SCRID090-60-1-1-2-4) (18%), this was followed by NERICA-17 (16%) and an equal share of farmers (11%) preferred Blechai (IRGC78281) and WAB 880 (WAB880-1-32-1-1-P2-HB-1-1-2-2). About 13% of the farmers selected NERICA-10, 10% selected NERICA-4, 5% NERICA-2 and only 2% selected NERICA-1 as the most preferred variety (Figure 10)

Figure 10: Farmer participatory preference of upland rice varieties in Namutumba district Eastern Uganda in 2015, bars indicate the percentage (%) of farmers preferring a certain variety.

4.5.1.3 Rice variety selection by gender; the most preferred varieties by male and female farmers in Eastern Uganda

A total of 18% and 16% of women selected SCRID090 and NERICA-10 respectively as the best two most preferred varieties, accounting 1/3 of the all the most-preferred varieties by women (Figure 11). Among men, 18% selected NERICA-17 as their most-preferred variety and 17% selected SCRID090 as their second best variety (Figure 11). Generally there were significant differences (t= 4.803 P<0.001) in variety preferences between men and women, for example, 12% of women preferred NERICA-4, compared to 9% of men selecting this variety (Figure 2). NERICA-4, was preferred by 13% of the men and only 10% of the women, Blechai was preferred by 13% of the men and 15% of the women, while WAB181-18 was preferred by 13% of the men and only 9% of the women. (Figure 11). Sorted from the most to the least frequent choices by women, the most-preferred varieties are SCRID090, NERICA-10, WAB181-18, NERICA- 17, NERICA- 4, Blechai, WAB 880, NERICA- 2 and NERICA-1. For men the order was: NERICA-17, SCRID090, Blechai, WAB880, WAB181-18, NERICA-10, NERICA-4, NERICA- 2 and NERICA-1 (Figure 11).

Figure 11: Rice variety selection by gender of rice farmers in Eastern Uganda (Namutumba district) in 2015, data presented as percentage responses by farmers.

4.5.1.4 The worst performing varieties of rice according to farmers.

The worst performing varieties according to farmers were Superica and Anakila, these two were selected by 20% of farmers as least-preferred. Superica the local variety, according to farmers was very susceptible to Striga, while Anakila was low yielding with weak stems hence being susceptible to lodging, this variety also had very small seeds (Table 7). Variety UPR (UPR-103-80-1-2) was the also selected among the least-preferred varieties (19%) due to its late maturity, short stature – making it difficult to harvest -, small seed sizes and susceptibility to Striga. Variety IAC 165 (17%) was also among the least-preferred varieties due to its high susceptibility to Striga and consequently low yields. NARC 3, selected by 16% of the farmers as ‘least-preferred’ was judged by farmers too short (difficult to harvest) and too susceptible to Striga, with too small seeds and high susceptibility to drought (Table 7). Variety Makassa (8%) was disliked for being late maturing and producing small grains and panicles (Table 7).

Table 7: Reasons for rice variety selection among the least-liked, as indicated by farmers in Eastern Uganda in 2015.

4.5.1.5 Farmer selection criteria for upland rice varieties.

Striga resistance and maturity period, with the highest mean score 4.86 and 4.81 respectively, were the most important characters for farmers to select rice varieties (Table 8). Other characters such as yield (4.62), tillering ability (4.42), drought resistance (4.38) and height (4.33) were also considered by farmers as very important attributes in selection (Table 8).

Table 8: Criteria rice farmers use in selection of the best upland rice varieties in Eastern Uganda in 2015

*Scores 1-5 where1 = not important 2 = not important at all 3 = more or less important 4= important, and 5= very important. Ranking was performed from most important-least important trait, with 1 indicating the most important trait. Data was presented as mean scores for different characters.

4.5.1.6 Variety ranking based on farmer selection criteria

Table 9 shows farmers’ evaluation on the expression of traits in the most preferred varieties. Results indicated SCRID090 with the highest phenotypic acceptability among rice farmers and with mean score of 2.76 ranked as the first variety. This was followed by WAB880 with (2.75) and NERICA-17 (2.63) which were ranked as second and third respectively (Table 9). BLECHAI with mean score of 2.60 and WB 18-18 with mean score of 2.58 were ranked as fourth and fifth respectively. NERICA-4 with (2.54) and NERICA- 1 with (2.52) were ranked as sixth and seventh respectively. NERICA-10 with mean score of 2.43 and NERICA- 2 with mean score of 2.37 were ranked as eighth and ninth respectively (Table 9).

Table 9: Ranking of the nine varieties of upland rice based on farmer selection criteria in Eastern Uganda in 2015.

*Scale (1-3) where 1= Bad, 2=Good/sufficient and 3=Very good. Ranking was performed from most preferred -least preferred among the most selected varieties, with 1 indicating the most preferred variety. , data was presented as mean scores

4.5.2 Discussion

This study revealed that the largest number of upland rice farmers who turned up for the participatory variety selection (PVS) were men (Table 6), these results are similar to Nanfumba et al. (2013) and Adekunle et al. (2013) who also found a lower number of women who turned up for the variety selection for rain fed lowland ecologies and upland rice respectively. The lower involvement of female farmers is attributed to social economic constraints including resource endowment, capital and land (Adekunle et al., 2013). Addison et al. (2014) who also found a lower number of women rice farmers attributed it to rice being labour intensive considering other roles by women. It was also revealed that a high number of farmers in the PVS were in age-group 42-54 and 31-41 years, these two groups constituted the largest percentage share of farmers. (Table 6). This suggests that there is low involvement of youth and elderly in rice farming. These findings were similar to Addison et al. (2014) who also reported a lower number of youth and elderly involved in rice farming in low land rice ecologies in Ghana. On other hand, the study indicated, out of 38 farmers in the PVS, the largest percentage of farmers had experience of 4-6 years and 1-3 years (Table 6), this confirms a survey carried out in 2005 by Advanced studies on international development (FASID) and Makerere University which also revealed that in Eastern Uganda household experience in rice production was highest for years 1-3 (Kijima and Serunkuma 2008). The lower number of farmers with low experience in rice farming could be attributed to high dropout rate of NERICA’S which the most cultivated upland rice varieties in Uganda and the lack of functioning seed distribution system in these areas (Kijima et al., 2011). According to Kijima et al. (2006) for areas which had upland rice introduced for first time, these got seed supply from NGO’S such as Africa 2000 Network, such areas lacked seeds even in input stores. From the discussion with farmers, they claim to have got their seeds from NGO’S and that they lacked seeds for cultivating in their fields.

The study showed that out of the 25 varieties of rice in the trial, the best preferred varieties of rice in ascending order were SCRID090, NERICA-17, NERICA-10, Blechai and WAB880 (Figure10). These constituted the five best preferred verities of rice, interestingly all the varieties except NERICA-10 were introduced for the first time in Uganda and they demonstrated suitability to local Ugandan conditions. Nanfumba et al. (2013) also reported high preferences for improved rice varieties compared to local varieties by farmers in PVS for rain-fed lowland ecologies in Uganda. Other varieties selected by farmers were WAB181-18, NERICA-4, NERICA-2 and NERICA-1 (Figure 10), these also except WAB181-18 and NERICA-2 have been grown in Uganda. From the study all the NERICA varieties featured among the preferred varieties by farmers, these results are similar to Gridley et al. (2002) who also found out in PVS (Participatory Variety Selection) in 2000 that NERICA’s were highly preferres in Ivory coast. Among the 58 varieties in they screened in their PVS trial in 2000 in West Africa, the most frequently selected varieties were designated by code WAB, with WAB 18-18 among the most preferred varieties which is also the case in our study. NERICA varieties normally show stable yields under low and high input conditions and therefore are expected to reduce risks and increase productivity in farmers’ fields (Gridley et al., 2002), this could be the reason why they were highly selected by farmers.

The results of the current study also revealed differences in variety selection among men and women. Men selected NERICA-17 as their best rice variety whereas women selected SCRID090 as their best variety. In ascending order the five best varieties selected by women were SCRID090, NERICA-10, WAB181-18, NERICA-17 and NERICA-4 while by men were NERICA-17, SCRID090, Blechai, WAB880 and WAB181-18 (Figure 11). There were mainly three common varieties among the two groups i.e. SCRID090, NERICA-17, WAB181-18, it was not surprising to see that the first two constituted the best varieties of rice selected by farmers. The selection could be related to the women’s and men’s roles in crop valve chain. For example since men take a leading role in marketing (Wanyonyi et al., 2008), they prefer varieties that are high yielding with big grain size which are characteristics of the chosen varieties in their group. Dorward et al. (2007) also found similar selection criteria used by men in PVS in Ghana. On the other hand, women are often involved in farm activities contributing over 50% of agricultural labour besides other reproductive roles (FAO, 2011), and they prefer a variety that is early maturing to reduce on farm work load. It is therefore not surprising that they picked SCRID090, WAB181-18 and NERICA varieties since these varieties are all early maturing. These results are similar to Dorward et al. (2007), who also stated early maturity and yield as the most traits women used in selection of upland rice varieties in Ghana. Also Addison et al. (2014) reported early maturity as the most important trait in varietal preferences among women in low land rice ecologies in Ghana.

The study also revealed the five worst varieties, which included Anakila, Superica, Makassa, IAC165, UPR and NARC3 (Table 7). Superica (local variety) and IAC165, were very susceptible to Striga. These would therefore give low yields especially in these areas which are highly infested with Striga. On the other hand Anakila and Makassa were low yielding with weak stems hence could be attached by soil pests when they lodge. These two varieties also have small seeds, this could explain the low yields. However Anakila was early maturing while Makassa was late maturing. Other varieties like UPR (UPR-103-80-1-2) and NARC3 were not only judged for being late maturing with short stature but also had small seed sizes and were susceptible to Striga. The short stature makes it difficult to harvest these varieties especially in areas where farmers use knives for harvesting. NARC3, despite having a good aroma, was negatively judged for being susceptible to drought. These results show that farmers dislike low yielding, late maturing, rice varieties with weak stems, small grains and very susceptible to Striga, drought and lodging. These results are similar to those by Namufumba et al. (2013) who also indicated similar reasons for farmers disliking rice varieties for low land ecologies except Striga resistance.

Results from farmer evaluation indicated that when farmers get new varieties, they often compare them with those currently grown on the basis certain characteristics. These were mainly Striga resistance, maturity period, yield, height, tillering, drought tolerance and grain size (Table 8). Among these traits, Striga resistance, early maturity, tillering and yield were the most important criteria for rice variety selection in Eastern Uganda. Farmers prefer high yielding, Striga resistant rice varieties that mature early with good tillering ability to provide high incomes even in short rainy seasons. All the selection criteria as per this study, except Striga resistance, were similar and consistent amongst farmers in 17 countries PVS in 1999 in West Africa (Gridley et al., 2002). Surprisingly, in most studies yield is seen as the most important criteria in rice variety selection (Lamo et al., 2010; Nanfumba et al., 2013; Gridley et al., 2002; Kimani et al., 2011). In the current study, Striga resistance emerged as an important trait and this could be due to this area being infested with Striga. A survey done in 2008 indicated Striga as a serious problem in upland rice in Eastern Uganda (Pittchar and Mbeche, 2008 unpublished information). The local rice variety (Superica) inspite of being high yielding was rejected by farmers for being susceptible to Striga. Therefore new sources of resistance were seen as the only way to solve the problem.

In relation to farmer selection criteria, farmers’ ranked varieties SCRID090, WAB880, NERICA-17, Blechai and WAB181-18 as the best varieties in ascending order, these were ranked as the five best rice varieties in this study (Table 9). Other varieties also ranked by farmers were NERICA-4, NERICA-1, NERICA-2 and NERICA-10, they were ranked sixth’s, seventh’s, eighth’s and ninth’s respectively. Farmers ranked SCRID090 and WAB800 as their best varieties owning to their unique attributes, these varieties were not only early maturing but also with good grain size and tillering ability which are all traits related to yield. These constitute the most sought attributes in rice varieties (Girdley et al., 2002; Dorward et al., 2007; Efisue et al., 2008; Nanfumba et al., 2013). The inherent ability of these varieties to resists or perform despite Striga infestation is another reason for selecting them (see Chapter three). Farmers selected NERICA-17 because of its big grains as well as being drought tolerant and Striga resistant. Blechai was chosen due to its height and Striga resistance. It has been reported by several studies that farmers prefer tall varieties because they reduce the burden of bending when harvesting (Efisue et al., 2008; Kimani et al., 2011). WAB 181-18 was chosen among the five best varieties owning to its big sized grains and early maturity, early maturity varieties are desired because they are seen as drought escaping options (Nanfumba et al., 2013; Hill, 2004). On the other hand NERICA-4 and NERICA-1 were ranked sixth and seventh respectively due to both being high yielding and drought resistant, all of which are characteristic traits in NERICA varieties (Gridley et al., 2002; Rodenburg et al., 2015), however NERICA-1 had a better tillering ability compared to NERICA-4. NERICA-10 and NERICA-2 ranked eighth and ninth respectively because both were early maturing and thus could be cultivated in short rainy season which are characteristics of the Ugandan conditions, however NERICA-10 was high yielding though not as resistant as NERICA-2 to Striga.

4.6 Conclusion

The main objective of this study was to identify Striga resistant varieties that could be adopted by farmers in Eastern Uganda. The study demonstrated that there are other traits than Striga resistance alone which farmers look for in newly introduced varieties. These other traits such as yield, tillering, height, drought tolerance, seed size and maturity should also be given attention in breeding programmes. Normally farmers will appreciate and select varieties that perform better than their local varieties. In the current study all varieties in the top-five varieties were introduced in the study site for the first time. They were all preferred over the local varieties. In fact farmers rejected their local variety Superica judging it too susceptible to Striga and consequently producing low yields in these areas. When we interacted with farmers, all of them were willing to adopt the newly introduced varieties if the seeds were available. Therefore the preferred newly introduced varieties stand a high chance of being adopted when they will be officially released.

CHAPTER FIVE

EVALUATION OF POST ATTACHMENT RESISTANCE OF UPLAND RICE VARIETIES TO STRIGA HERMONTHICA UNDER CONTROLLED ENVIRONMENT CONDITIONS

5.1 Introduction

Plant cells are not able to move, so they have evolved a number of strategies to defend themselves against attack by antagonistic organisms such as herbivores and parasitic plants. In cereals mechanisms of resistance against parasitic weeds can be categorized as pre-attachment and post-attachment resistance. Pre- attachment includes mechanisms that prevent attachment of the parasitic weed to the host plant. These include absent or reduced production of germination stimulants (Hess et al., 1992; Rich, 1996; Jamil et al., 2011), germination inhibition and inhibition or reduction in haustorium formation (Rich et al., 2004; Gurney et al., 2003). Another mechanism involves thickened host root cell walls resulting in a mechanical barrier to infection (Maiti et al., 1984; Olivier et al., 1991). Post-attachment mechanisms comprise different incompability reactions such as the failure of the parasite to establish xylem-xylem connections with the host plants due to blockage in vascular continuity (Cissoko et al., 2011). Other mechanisms include hypersentivity reactions resulting in death of host root tissue around the point of attachment and discouragement of parasite penetration (Mohammed et al., 2003; Cissoko et al., 2011). Plants can also produce phenolic compounds such as phytoalexins in infected host cells (Oliver et al., 1991). Post-attachment resistance to Striga in rice has been reported in rice to S. hermonthica and S. asaitica. For example, in the cultivar Nipponbare, S. hermonthica (Kibos ecotype) site penetrates the host root cortex but is unable to penetrate the endodermis to form a vascular continuity with the host and eventually dies (Gurney et al., 2006). Recently Jamil et al., (2011); and Cissoko et al., (2011) studied pre- and post-attachment resistance of NERICA varieties and their parents respectively against Striga hermonthica and Striga asiatica. In the post attachment study NERICA and other varieties varied greatly in their resistance to Striga hermonthica and Striga asiatica. Mechanisms of resistance ranged from intense necrosis at the site of attachment to the host plant, failure to form a vascular continuity with host after penetration of the host root cortex and dense staining as the parasite penetrates the endodermis making the parasite to remain small (Cissoko et al., 2011). In the pre-attachment study it was found that different varieties produced different amounts and types of strigolactone germination stimulants, which correlated with the resistance level of the cultivar (Jamil et al., 2011).

Evaluation of post attachment resistance under field conditions is a challenge as resistance is usually measured as the number of parasite attachments that emerge above ground over time (Rodenburg et al., 2015). This reflects the combination of pre-and post-attachments mechanisms that are operating in each cultivar but it is not possible to separate these. In addition, the expression of resistance in the field can be affected by environmental factors. For example the nutrient status of the soil, particularly nitrogen and phosphorus levels will affect the exudation of strigolactones and thus germination of parasite seeds (Yoneyama et al., 2007a, b). The distribution of Striga seeds in the soil can also be (Haussman et al., 2000) and environmental conditions including temperature and rainfall (Haussman et al., 2000) can affect the resistance levels observed. Thus in the field the resistance level of a cultivar results from both Genetic and Environmental factors, often referred to as the G x E interaction (Rodenburg et al., 2015). To clearly evaluate post-attachment resistance, cultivars have to be grown under controlled environment conditions so that the resistance level observed can be attributed to genetic factors rather than to environmental variables. Many studies to evaluate post-attachment resistance of different cereal hosts to Striga have been carried out using soil-less rhizotron that allow easy access to the host root system for quantification of host resistance reactions (Gurney et al., 2006; Swarbrick et al., 2008; Cissoko et al., 2011). These growth systems also allow roots to be inoculated with germinated Striga seeds, which by-passes the differences in production of strigolactones by the roots of different cultivars (Gurney et al., 2006; Cissoko et al., 2011). Thus the aim of this study were was to quantify, under controlled environment conditions, post-attachment resistance levels of selected rice varieties that were grown in the field in Namutumba to (1) determine whether the resistance ranking of the cultivars was the same or different and (2) to assess the influence of environmental conditions of the expression of resistance.

5.2 Materials and methods

5.2.1. Plant materials

An evaluation of resistance of upland rice varieties against Striga hermonthica under controlled environmental conditions was carried out at the University of Sheffield, Department of Animal and Plant Sciences. This was intended to understand the Striga resistance of a selected group of upland rice varieties using the Striga hermonthica ecotype found in Namutumba district, Uganda. Rice seeds were provided by the Genetic Resources Unit of Africa Rice. Striga hermonthica (Del.) Benth seeds were obtained from farmer fields in Namutumba district in Eastern Uganda.

5.2.2 Growth and infection of rice plants with Striga hermonthica

Rice seeds were germinated between blocks of moistened horticultural rock wool for six days after which a single rice seedling was transferred to a root observation chamber (rhizotron) as described previously by Gurney et al. (2006). Each rhizotron consists of a 25 × 25 × 2 cm3 perspex container packed with vermiculite onto which a 100µm polyester mesh was placed. Roots of the rice seedlings grew down the mesh, and openings at the top and bottom of the rhizotron allowed for shoot growth and water drainage respectively. Rhizotrons were covered with aluminum foil to prevent light from reaching the roots. Rhizotrons were supplied with 25ml of 40% long Ashton (Hewitt, 1960), nutrient solution containing 2mM ammonium nitrate three times each day via an automatic watering system.

Striga seeds were sterilized in 10% bleach, washed thoroughly with dH2Oand then incubated on moistened glass-fiber filter paper (Whatman), in petri-dishes for 12-15 days at 300C (Gurney et al., 2006). Eighteen hours before infection of rice seedlings, 1 ml of 0.1ppm solution of an artificial germination stimulant GR24 was added to Petri dishes containing conditioned Striga seeds to stimulate germination. Two weeks after sowing, rice plants were infected with 12 mg of germinated Striga seeds by aligning them along the roots using a paint brush (Gurney et al., 2006). Infection of rice roots with germinated Striga seeds eliminates differences due to variation in production of germination stimulants by the different rice cultivars (Cissoko et al., 2011; Jamil et al., 2011; Gurney et al., 2006). Uninfected rice cultivars acting as the control were treated in a similar way as described above without the infestation step. A total of 10 plants were sown per variety with four controls (uninfected) and six infected plants. Plants were grown in a temperature controlled growth chamber with 60% relative humidity and day and night temperatures of 280C and 240C respectively. The irradiance at plant height was 500 μmol m-2 s-1

5.2.3 Non-destructive measurements of growth

A series of non-destructive growth measurements were made each week from the day of infecting rice seedlings with germinated Striga hermonthica seeds including the height of the main stem determined using a ruler measured from the stem base until the attachment of the new leaf, number of leaves on the main stem, stem diameter, using a digital Vernier caliper (Mitutoyo England UK), number of tillers and tiller leaves

5.2.4 Destructive measurement of above ground rice biomass

Above-ground rice biomass was determined three weeks after infection of the rice cultivars. Plant materials was separated into main stem, tiller stem, main stem leaves and tiller stem leaves. Plant material was dried at 700C for 7 days and then weighed to determine dry biomass.

5.2.5 Quantification of post-attachment resistance of the rice cultivars

Post-attachment resistance was quantified 21 days after infection of the rice roots. Before harvest, the root system of each rhizotron was photographed using Canon EOS 300D digital camera. Striga seedlings growing on the roots of each infected plants were harvested and placed in Petri dishes and photographed using a Canon EOS 300 digital camera. The number and length of Striga seedlings from each rice plant were determined from the Petri dishes photographs using Image- J. Striga plants were then dried at 48 0C for 2 days and the amount of dry biomass per host was determined.

5.2.6 The phenotype of resistance

The phenotype of resistance was investigated by photographing parasites developing on the root systems of each cultivar at different stages after infection using a Leica MZFLIII stereo microscope and a diagnostic instruments camera Model 7.4 and by cutting small sections of root with attached parasite and mounting on a glass slide in water. The root tissue was observed using an Olympus BX51 microscope employing differential interference contrast microscopy and photographed using a digital camera.

5.2.7 Statistical analysis

Data were subjected to Analysis of Variance (ANOVA) using GenStat computer software (v12), mean values were separated using LSD at (P=0.05). To meet the requirement of normality of data distribution for classical ANOVA (Sokal and Rohlf, 1995), Striga emergence counts were transformed using Natural logarithms, log (x+1), and stem width was square-root transformed.

5.3 Results

5.3.1 Effect of Striga hermonthica on the growth of the rice cultivars

5.3.1.1 Plant height

Plant height of the infected rice plants were compared with those of the uninfected plants to evaluate the impact of Striga hermonthica on the growth of the rice plants. Rice plant height was significantly (P<0.001) affected by Striga for varieties IAC 165, Superica (local variety), WAB181-18 and WAB880, these had significantly shorter stems compared to their respective controls. The rest of the varieties were not significantly reduced (P>0.001) in height with respect to their controls (Figure 12). Varieties Superica, IAC 165, WAB181-18 and WAB880 had a great reduction in stem height 44.6%, 33.9%, 24.8% and 23.6% respectively when infected with Striga hermonthica (Figure 12). There was moderate reduction in stem height among varieties IRGC (12.1%), WAB928 (8.8%), NERICA-17(8.3%) while varieties WAB56-104, CG14, NERICA-4 and Blechai had <6% reduction in stem height (Figure 12). Additionally the remaining varieties (SCRID090, NERICA-1 and WAB56-50) registered no reductions in plant height when infested with Striga (Figure 12).

Figure 12: Height of the infected and uninfected control rice plants per variety. Numbers next to the bars indicate percentage loss in plant height of S.hermonthica infected plants compared to the uninfected control. (**) indicates that control and S.hermonthica infected plants were significant P<0.01

5.3.1.2 Stem width

The stem width per variety of rice were significantly reduced (P<0.05) for varieties IAC 165, Superica, WAB 181-18, Blechai and WAB 880. These varieties had significantly thinner stems compared to their respective control plants (Figure 13). There was no significant difference in stem width (P>0.05) of the infected plants of the remaining varieties with respect to their non-infected controls. Also there was a reduction in stem width when rice plants were infected with Striga. However varieties varied greatly with some having significantly reduced stem diameters such as WAB880 (22.3%), IAC 165 (20.0%) and Superica (26.3%) (Figure 13). Other varieties with moderate reductions in stem diameter following infection were Blechai (16.2%), WAB181-18 (14.9%), IRGC (13.4%) and CG14 (13.0%). The rest of the varieties had low reduction in stem diameter (<10%) with WAB56-50 having registered no reduction in stem width (Figure 13).

Figure 13: Stem width of the infected and uninfected control rice plants per variety. Numbers next to the bars indicate percentage loss in stem width of S.hermonthica infected plants compared to the uninfected control. (*) indicates that control and S.hermonthica infested were significant P=0.05

5.3.1.3 Number of tillers

There was no significant difference (P> 0.05) in number of tillers produced by infected plants compared to the non-infected controls when analyses across the varieties. However after Striga infection the local variety (Superica) produced a significantly (P< 0.05) lower number of tillers compared to the uninfected control plants (Figure 14). Number of tillers for the control plants ranged from ~3tillers for WAB 181-18 to ~1 tiller for Blechai. For infected plants, tiller number ranged from highly susceptible varieties with no tillers for the local variety to highly resistant with 2 tillers per plant for CG14 (Figure 14). The results show that except the local variety Superica, which produced no tillers within 21days after infection with Striga hermonthica, the rest of the varieties’ control plants produced more tillers though they were not significantly different (p>0.05) from their respective infected plants.

Figure 14: Number of tillers of the infected and uninfected control rice plants per variety, data presented as means and LSD bar, (*) indicates that control and S.hermonthica infected plants differ significantly (P<0.05) using LSD (0.05) =1.098.

5.3.1.4 Number of leaves

Number of leaves was significantly different (P<0.05) between the infected and uninfected control plants for local variety (Superica). However there was no significance difference (P>0.05) in the number of leaves produced per plant for the infected and uninfected controls of all the remaining varieties of rice (Figure 15). For the infected plants varieties WAB 181-18 and CG14 produced the highest number of leaves per plant (12 leaves) while Superica produced the lowest number of leaves per plant (6 leaves). On average control plants produced 10 leaves while infected plants produced 9leaves.

Figure 15: Maximum number of leaves of the infected and uninfected control rice plants per variety. Data presented as means and Standard error of means. (*) indicates that control and S.hermonthica infected plants differ significantly (P<0.05) using LSD (0.05)

5.3.1.5 Number of Striga attachments and Striga dry biomass per variety

There was a significant difference (P< 0.001) in the number of parasites attached per rice plant among the rice varieties. Parasitic attachments per host plant ranged from 30- 45 on the very susceptible varieties Superica and IAC 165, supporting the highest number of Striga plants, to <1 on the most resistant varieties IRGC and CG14 (Figure 16a, b). Also cultivars such NERICA-1, NERICA-4 and NERICA-17 supported relatively large number of Striga plants (6-15) parasites per host root (Figure 17c, d). The remaining cultivars exhibited a good level of post-attachment resistance supporting 1-5 parasitic plants per host (Figure 16a). Generally varieties Superica and IAC 165, had on average the largest well developed parasites on root system compared to the rest of the varieties (Figure 17 a, b, c, d, e and f).

Striga biomass was significantly (P<0.001) higher for varieties IAC 165 and SUPERICA (21-24mg), compared to other varieties. Varieties NERICA-17, -1 and WAB56-50 had moderate biomass of Striga ranging from 3-5mg per plant on average (Figure 16b). There was no significant difference (P>0.001) in Striga biomass between the remaining varieties (Figure 16b). These varieties had a relatively small Striga biomass ranging from 0.02mg-1.58mg per plant (Figure 16b). The results characterize varieties into susceptible (IAC165, SUPERICA) intermediate (NERICA-1, -17 and WAB56-50), resistant (WAB800, WAB56-104, NERICA-4, SCRID090, WAB181-18) and very resistant (WAB928, BLECHAI, IRGC and CG14) (Figure 16 a, b)

Figure 16: Figure 5: Evaluation of post-attachment resistance of different varieties of rice, number of Striga plants taken 21 days after Striga hermonthica infection (a); data was first transformed before analysis log (x+1). Striga dry biomass data (b), all data (b), all data a and b were presented as means+ S.E for all the varieties. Means both Striga attachments and biomass were significantly difference (p<0.001), means were separated by LSD

Figure 17: Parasitic plants attached on the host plant roots 21 days after infection with Striga hermonthica. These show differences in Striga attachments with IAC165 and Superica having the highest number of attachments followed by NERICA-1 and -17 and lastly WAB56-50 and WAB928 where parasites failed to attach on the host plants.

5.3.2 Effect of Striga hermonthica on above ground rice biomass of upland rice varieties with respect to the uninfected rice plants.

There was a significant difference (P<0.05) in above-ground dry biomass of infected rice plants for varieties WAB181-18, WAB 880, CG14, local variety (Superica) and IAC165 compared to their control plants. Rice biomass of the remaining varieties was not significantly (P>0.05) affected by Striga infection (Figure 18). With respect to control plants, WAB181-18 had the highest mean biomass (1.144g) and NERICA-17 had the lowest rice biomass (0.347g). Infected rice varieties ranged from very susceptible cultivars, such as Superica having biomass of 0.254g compared to the non-infected control plant biomass of 0.983g, to resistant varieties such, as IRGC with 0.721g biomass of infected plants compared to 0.922g of uninfected control plants (Figure 18).

Figure 18: Above-ground rice dry biomass of infected rice plants and their respective uninfected control plants data presented as means + S.E for all the varieties. Numbers next to the bars indicate percentage loss in rice biomass of S.hermonthica infected plants compared to the uninfected control. (*) indicates that control and S.hermonthica infected plants differ significantly (P=0.012) using LSD (0.05) =0.257

5.3.3 Impact of Striga hermonthica on the biomass of rice varieties

Generally there was a loss in biomass of all infected plants especially for the most susceptible varieties such as Superica (74.2%) and IAC 165 (60.2%) (Figure 18). Others varieties with high loss in biomass included WAB800 (54.2%), WAB181-18 (39.2%) CG14 (33.3%) and IRGC 28.2% (Figure 18). The remaining varieties except SCRID090 and WAB56-50 with 10-20% loss in biomass, the rest had <10% loss in rice biomass.

Figure 19: Relationship between percentage losses in total rice biomass of infected plants compared with control plants and the amount of parasite biomass dry weight on roots of rice plants.

There was a significant negative linear relationship between the effect of S.hermonthica on the host biomass and the amount of parasite biomass (R2=0.417, P=0.007). The most susceptible varieties Superica and IAC 165 were the most affected of all the varieties (Figure 19). This negative relationship would indicate that as you increase the Striga there is a significant effect on the biomass of the host (Figure 19). However this looks peculiar, for example NERICA-1 and WAB880 with similar levels of infection had 5.4mg and 1.6mg of Striga biomass respectively (Figure 16a). However the percentage loss in biomass with respect to uninfected control plants was 0.4% and 52.4% respectively (Figure 18). This only show variety differences in tolerance between the two varieties. This similar trend is depicted between NERICA varieties and varieties WAB928, CG14, WAB181-18 and IRGC which had small Striga biomass but with high reduction in rice biomass (Figure 16b, 18). This however would not be surprising because of the lower percentage of R2=0.417.

5.7 Discussion

5.7.1 How does Striga hermonthica alter the morphology of host

5.7.1.1 Plant height

The study has indicated highly significant (P<0.001) reduction in stem height of infected rice plants when compared to uninfected plants across varieties (Figure 12). Varieties Superica, IAC 165, WAB880 and WAB181-18 had a significantly high reductions in stem height following infection with respect to control plants. These results are consistent with studies by Cechin and Press, 1994; Walting and Press, 2000; Swarbrick et al., 2008; Atera et al., 2012 who also indicated that cereals such as rice when infected with Striga species exhibit characteristic changes in plant morphology and architecture as compared to uninfected plants including stunting /reduction in stem length of infected plants. Stunting could be as result of lack of internode elongation rather than increase in internode numbers. Internode elongation is based on increased cell elongation in well-delineated zones of the internode (Kende et al., 1998).

Alteration of growth regulators metabolism is one hypothesis that may account for this change in stem height following infection. A number of growth regulators are involved in elongation of stems in plants including Gibberellins, cytokinins and auxins (Kende et al., 1998; Sakamoto et al., 2006; Ikeda et al., 2001). Siliva and Jorge, (2010) reviewed a number of studies on the effect on the growth regulators on tillering, these concluded that existence of separate gibberellin-mediated pathways control tillering and plant height. Gibberellins are involved in cell wall expansion by induction of cell wall loosening enzymes (Kende et al., 1998). Many rice mutants have been described to a lack of gibberellins or to be insensitive to this hormone (Ishikawa et al., 2005). Therefore any alteration in the synthesis of gibberellins will eventually lead to reduced height of infected plants. According to study by Swarbrick et al. (2008) many genes involved in auxin and gibberellin signaling are down regulated following Striga infection. Therefore reduction in internode extension in Striga infected plants could be the result of an alteration in gibberellin and auxin signaling, hormones important in stem elongation.

Another explanation for stunting of rice plants after infection has been suggested to be related to translocation of toxic compound from S.hermonthica to the host (Musselman and Press, 1995). Also some studies have suggested the involvement of secondary metabolites toxic to cereals to have an effect on host morphology (Ejeta and Butler, 1993). These together with Rank et al. (2004) revelation of irioid glucosides and their suppression of cell division can also explain this impact of Striga on plant’s height. However such toxins have not been identified and evidence of this mediated effect of this host has not been presented.

5.7.1.2 Stem width

Study results indicated reduction in stem diameter of rice plants infected with Striga hermonthica when compared to uninfected control plants across varieties (Figure 13). Varieties Superica, IAC165 and WAB880 supported a high parasite load as per this study except WAB880. This is only an indication of great effect of Striga on these varieties when compared to other varieties in the study. It appears that this high reduction in stem diameter of these varieties could be attributed to earlier infection of the parasite. Studies by Cechin and press, (1993) demonstrated that earlier attachments of parasites had greater effect on host growth than later attachments. Other varieties with less reduction in stem diameter such as WAB928, WAB56-50, WAB56-104, NERICA-1, SCRID090, NERICA-4 and NERICA-17 is attribute to less effect of the parasite on these varieties.

Another possible explanation for reduction in stem width of infected plants relative to the uninfected control plants is related to limited assimilate partitioning to other activities of the plant including stem enlargement. Striga hermonthica survives by diverting essential nutrients which could otherwise be used by plants (Rodenburg et al., 2006; Atera et al., 2011). In other words the parasite may act as an alternate sink for the assimilates (Gurney et al., 1999). These nutrients are responsible for all the essential processes of the plant, it could therefore be possible that the thinner stems in susceptible cultivars such as IAC 165 and Superica is due to diversion of nutrients to parasite which could be used for stem enlargement.

5.7.2 Effect of Striga hermonthica on tillering and leaf number of rice varieties

5.7.2.1 Tillering

This study has indicated less tiller production among rice plants infected with S.hermonthica compared to uninfected control plants (Figure 14). The results are similar to Cissoko et al., (2011) who showed that infection of rice plants with Striga hermonthica and Striga asiatica suppressed tillering of infected plants compared to the uninfected plants across rice varieties. Also since tillers play a major role in determining plant architecture and yield. These results could suggest reduction in yield of Striga infected rice varieties especially those with high Striga numbers such as IAC 165 and Superica which had highly reduced tiller number. Reduction in the number of tillers of infected rice plants compared with the control are attributed to both suppression of outgrowth of tiller buds and inhibition of the formation of tiller buds by Striga. Out of 14 varieties tested, only three varieties WAB56-104, NERICA-1, IRGC had a high number of tillers compared to the uninfected plants. The remaining varieties had lower number of tillers when compared with the uninfected plants. This is a clear indication that Striga had an effect on the tillering performance of these varieties.

One possible explanation for lower number of tillers in infected rice plants is attributed to growth regulators just like in plant height. Hormones such as auxins and cytokinins have been known to control tillering (Garba et al., 2007; Kamato et al., 2006). Auxins have an indirect inhibitory action on tillering while cytokinins directly promote tillering (Ongaro and leyser 2008). Therefore any changes in synthesis of these hormones will affect tillering of rice. It is therefore possible this can also relate to the lower tiller production of infected compared to their control plants. Further evidence has showed Striga infected sorghum tissues having greater amounts ABA and ethylene and lower amounts of Cytokinins (Drennan and ELhiweris, 1979) hormone that is important in tiller formation.

Another possible explanation for the lower tiller production for the infected plants could be due to Nutrition imbalances especially with respect to nitrogen. Cruz and Boval, (2000); Mckenzie, (1998) found a positive effect on nitrogen availability and tillering. Striga hermonthica also depends on host for all its nitrogen demands once a xylem connection has been established with the host (Pageau et al., 2003). It is therefore possible that tillering is lowered because the nitrogen is diverted to the parasite.

5.7.2.2 Number of leaves

There was a reduction in number of leaves between the infected rice plants and uninfected control plants across varieties (Figure 15). These results agree with Cissoko et al. (2011) who also indicated a lower leaf dry weight across varieties when they were infected with S.hermonthica and S.asiatica. All the rice varieties when infected with S.hermonthica produced lower number of leaves compared to their respective uninfected plants. However varieties NERICA-1 and WAB56-104 had a high number of leaves compared to the uninfected plants. The lower leaf number in these varieties except these two is related to the genetic resistance levels difference among varieties. This led to some varieties more significantly affected than others. The high number of leaves among the two varieties could related to late Striga attachment on host roots. It should however be noted that even though these produced a high number of leaves they were not significantly different from control plants. Leaves are the main photosynthetic organs of the plant, therefore their reduction can affect assimilate manufacture and consequently yield of the plant. This only implies that those varieties that are severely affected such as Superica and IAC 165 can ultimately have a low yield in the Striga prone areas.

Changes in the number of leaves and leaf area of infected as compared to the uninfected plants has also been reported (Frost et al., 1997; Cechin and Press, 1994; Atera et al., 2012). The reduction in leaf number of the infected compared to uninfected plants could be due to nutrient diversion from the host plant to the parasite, there is less dry matter partitioned in the growth of new leaves. Studies have also shown rates of photosynthesis are usually lower in infected plants than the uninfected plants (Gurney et al., 1997; Frost et al., 1997). It is possible that the reduced photo-assimilate production due to the change in plant architecture after infection is responsible for the reduced number of leaves in infected plants compared to the control plants.

5.7.3 Rice cultivars exhibit differential resistance to Striga hermonthica.

Post-attachment resistance levels varied significantly P<0.001 across varieties tested (Figure 16a).Varieties IAC165 and Superica were the most affected both during the course of the experiment and at Harvest. Similarly studies by Gurney et al. (2006); Swabrick et al. (2008); Cissoko et al. (2011) have all reported IAC 165 as susceptible and has been used as a susceptible check in these studies. Variety CG14 was the most resistant variety as per this study having on average 1-2 developed parasites. Field and pot experiment studies by Johnson et al. (1997) stressed the high Striga resistance potential of O.glabberima varieties compared to O.sativa varieties. This therefore agrees with results from this study indicating O.glabberima variety CG14 as highly resistant. Similarly Kaewchumnog and price (2008) reported CG14 as one of the most resistant variety in their study. On the other hand post-attachment resistance of some varieties from this study was similar to observations by Cissoko et al, (2011) who also found out that varieties NERICA-17, -4, WAB56-50, WAB181-18 and CG14 exhibited a good level of post-attachment resistance to Striga hermonthica (Sh-Kibos) and Striga asiatica (Sa-USA) and IAC165 was the most susceptible as per this study. This implies that these varieties are resistant to a wide number of ecotypes of Striga including the one used in this study. Varieties such as WAB880, WAB928, SCRID090, Blechai, and IRGC were evaluated for post-attachment resistance for the first time. These varieties exhibited an excellent post-attachment resistance to S.hermonthica. Field experiments by Harahap et al. (1993) also showed that Blechai and IRGC were highly resistant to Striga hermonthica in field trails in Kenya. Also studies by Johnson et al., (2000) reported WAB928 as a resistant variety to S.hermonthica parasitism. However no documented studies have indicated the reaction of varieties SCRID090 and WAB880 to Striga. It is therefore the first time we report the resistance of these varieties in both filed (study three) and under controlled conditions. Varieties such as NERICA-1 and -17 reported to be resistant to Striga hermonthica (Cissoko et al., 2011; Jamil et al., 2011; Rodenburg et al., 2015) supported a moderately number of Striga hermonthica in this study, though the Striga plants were relatively small thus contributing less to the total Striga biomass per plant. The small sized parasites could be attributed to time of emergency of the parasite as they could have emerged later or due to defense mechanisms in these varieties.

5.7.4 Effect of Striga hermonthica on the rice biomass

Infection of rice plants with Striga hermonthica altered the partitioning of assimilates to the leaves and stems of rice plants. This was revealed in the above ground rice biomass of infected rice plants (Figure 18, 19). Similar results were reported for sorghum parasitized by S.hermonthica (Cechin and Press, 1993; Frost et al., 1997), maize (Taylor et al., 1996) and rice (Cissoko et al., 2011). Rice biomass of infected rice plants was high for varieties IRGC, SCRID090, BLECHAI, NERICA-1,-4,-17 and WAB928 compared to their respective uninfected rice plants. This could indicated less effect of Striga on these rice varieties and is attributed to the fact that these varieties were very resistant to Striga hermonthica (Harahap et al., 1993; Johnson et al., 1997; Johnson et al., 2000; Cissoko et al., 2011). The effect of Striga on rice biomass can be ultimately understood looking at the percentage loss in biomass following Striga infection. Generally there was a reduction in rice biomass across varieties following infection, it was however very pronounced in varieties Superica, IAC165 and WAB880. These except WAB880 supported the highest number of Striga plants. It is also possible that alteration in plant morphology of Striga infected plants depended upon the variety genetic differences in resistance and tolerance. Surprisingly varieties such as CG14, WAB928, IRGC, WAB181-18 and WAB880 inspite of having lower number of parasites per root system registered tremendous reduction in host biomass. On contrally varieties NERICA-1, NERICA-17 with high number of parasites when compared to the above varieties registered a less reduction in host biomass. This can only be explained by variety differences in tolerance levels. Similar insights were reported in sorghum by Rodenburg et al. (2006); in rice Cissoko et al. (2011); Rodenburg et al. (2015). This can only imply that inspite of high Striga numbers, such varieties can still provide high yields. Also the fact that percentage reduction in host biomass was linearly related to the host Striga biomass per root system would suggest that the highly resistant varieties can produced high yield potentials in Striga prone areas.

5.8 Conclusion

The main objective of this study was to evaluate the post-attachment resistance of farmer selected varieties under controlled conditions. This experiment was in tandem with the field evaluation of these varieties to establish whether the variety ranking differs from variety to variety and from the field. The study demonstrated varieties evaluated showed significant variations in the number of Striga hermonthica plants attached per host. Varieties CG14, IRGC, WAB928, SCRID090 and BLECHAI demonstrated excellent levels of post attachment resistance to Striga hermonthica. The most susceptible varieties as per the present study were Superica (local check) and IAC165. This study has also shown for first time varieties such as WAB880, SCRID090, Blechai, IRGC, and WAB928 exhibited an exceptionally high level of post-attachment resistance to S.hermonthica. The present study has also shown NERICA-1, -17 and WAB56-50 had a moderate number of parasites attached on the host roots. However there was less reduction in rice biomass of these varieties especially NERICA-17 and NERIC-1 as compared to the most resistant varieties CG14, IRGC which recorded a high reduction in rice biomass. This concludes one thing, differences in tolerance to deleterious effects of infection by Striga hermonthica. If such mechanism responsible for tolerance can be clearly understand, it could important to incorporate such genes in the most resistant varieties and the moderate resistance. This can help them avail yield even in heavily Striga affected areas.

CHAPTER SIX

6.0 GENERAL DISCUSSION AND CONCLUSION

Striga is a major biotic constraint in cereal production systems of Sub-Saharan Africa (Atera et al., 2012). Despite all efforts to manage Striga, it has persisted and increased in number. A number of Striga management technologies have been developed for a long time (Oswald et al., 2005; Rodenburg and Johnson, 2009; Atera et al., 2011). However, farmers have not universally adopted these technologies because either they are too costly or laborious (Gressel et al., 2004). The development of tolerant and resistant varieties in upland rice growing areas is a viable option to manage Striga. It is also cost-effective as farmers don’t have to invest in this technology. This study therefore sought to evaluate the resistance of rice varieties against Striga hermonthica in Eastern Uganda. In this study two experiments were conducted on-farm and under controlled conditions. The first experiment had two studies, (1) Evaluated the performance of selected rice varieties under Striga infested field conditions and (2) Identified in a farmer participatory manner rice varieties that could be adopted by farmers. The second experiment evaluated the post-attachment resistance of farmer selected upland rice varieties to Striga hermonthica.

The results of the first study indicated variation in Striga resistance and yield across varieties. Out of the 25 varieties evaluated, three O.glabberima varieties AGEE, CG14, Anakila, seven O.sativa varieties IR49, IRGC, WAB935, WAB928, SCRID090, WAB880, Blechai, one O.sativa parent WAB181-18 and NERICA varieties NERICA-2,-10, -17, -4 had an excellent resistance to Striga compared to the locally grown variety Superica and IAC 165. All the above varieties except WAB880 and SCRID090 have been reported as having an excellent resistance to S.hermonthica elsewhere. For example Blechai and IRGC by (Harahap et al., 1993) in Kenya, IR49 (Harahap et al., 1993; Johnson et al., 1997) in Kenya and Ivory Coast, WAB928 and WAB935 (Johnson et al., 2000) in Ivory Coast. Jamil et al. (2012) reported O.glabberima varieties AGEE, Anakila and CG14 have having an excellent resistance to Striga. The NERICA varieties -2, -10, -17,- 4 and the O.sativa parent WAB181-18 was also reported by Rodenburg et al. (2015) as having an excellent resistance to S.hermonthica in Mbita, Kenya. This can only indicate resistance of the above varieties to a number of Striga ecotypes. The susceptibility of the local variety (Superica) can be attributed to continuous cultivation of this variety. Therefore virulence levels of local Striga population against this variety increased in time. Results also indicated partial resistance of O.glabberima varieties ACC, Makassa and MG12 confirming results by Johnson et al (1997) who also reported them as partially resistant to Striga hermonthica in Ivory coast. Varieties NERICA-1 and UPR reported to be resistant in Mbita (Rodenburg et al., 2015) and (Harahap et al 1993) respectively supported moderate number of Striga plants in this study. This could be related to differences in virulence levels of the Striga ecotype in this area as compared to ecotype in Mbita against these varieties. This study has indicated susceptibility of varieties WAB56-50, WAB56-104, NARC 3 (ITA 257) to Striga hermonthica. These results are consistent with Rodenburg et al. 2015 who also confirmed WAB56-50 as susceptible to Striga hermonthica in Mbita and Harahap et al. (1993) who confirmed NARC3 (ITA 257) as susceptible to S.hermonthica in Kenya. However variety WAB56-104 had intermediate levels of Striga in (Rodenburg et al, 2015) though it was susceptible in this study.

Varieties that combine Striga resistance and high yields is important to improve rice production. However Striga resistant varieties don’t necessarily produce high yields. The study has indicated variation in yield production across varieties under Striga infested field conditions. The best performing varieties yielded an equivalent of 3.0t ha-1-3.6t ha-1 i.e. SCRID090, NERICA-17, Anakila, WAB880, WAB56-104, NERICA-1, WAB181-18, NERICA-10 and NERICA-4 in 2014A. Additionally in 2015, the best performing varieties yielded between 3.2t ha-1-3.6t ha-1 i.e. NERICA-2, NERICA-4, ACC, NERICA-1, NERICA-10, Blechai, SCRID090, WAB181-18, CG14, AGEE, NERICA-17, WAB880 and Makassa. These results are similar to Rodenburg et al. (2015) who also indicated varieties WAB 181-18, WAB56-104 and all the NERICA varieties as high yielding varieties in Mbita, Kenya. Also some studies have reported the high yielding potential of these varieties (Gridley, 2002; Saito et al., 2012). O.sativa varieties WAB880 and SCRID090 evaluated for the first did not only have good resistance to Striga but also produced high yields. This could simply because of high yielding potential of these varieties. The results Also O.glabberima varieties Makassa, ACC, MG12, NERICA-1 O.sativa parents WAB56-104, WAB56-50 with partial resistance to Striga had high yield compared to varieties WAB928, WAB935, IRGC and IR49 which were very resistance to Striga. This suggests variety differences in tolerance to the deleterious effects of infection by S. hermonthica. Tolerance is the ability of a give variety to produce yield even under high Striga infestation levels (Rodenburg et al., 2006; Rodenburg et al., 2008; Rodenburg and Bastiaan et al., 2011). This has also been reported in Sorghum (Gurney et al., 1995; Rodenburg et al., 2006). However to clearly understand this concept one will need Striga free plants of the same varieties which is not easily achieved in the field. Understanding the expression of this trait can be important to improve yield in the most resistant varieties with low yields (Rodenburg and Bastiaans, 2011). The results also reported a significant negative correlation between NSmax and grain yield in 2015. This indicated a high Striga pressure in this year compared to 2014. This also implied that the highly resistant varieties produced high yields. The local variety Superica produced the lowest yield in this season implying that this increase in Striga affected the performance of this variety.

In the second study, we sought to understand a selection criteria farmers’ use in identification of the best preferred varieties. The study indicated Striga resistance, yield, tillering ability, drought tolerance/resistance, maturity period and height of the varieties. Consequently, farmers ranked SCRID090 as their best variety, this was followed by WAB880, NERICA-17, Blechai and WAB181-18 as the five best varieties out of 25. These varieties were as ranked among those with high yields and excellent resistance to Striga hermonthica in the first study. Other varieties selected by farmers were NERICA-4, NERICA-1, NERICA-10 and lastly NERICA-2 in ascending order of preference. All the NERICA varieties included in this study one were all selected by farmers. Similar results were shown by Gridley et al. (2002) where farmers highly selected these varieties in Ivory Coast. Also several studies have indicated high yields and good Striga resistance among these varieties (Gridley et al., 2002; Saito et al., 2012; Rodenburg et al., 2015)

A number of studies have always indicated yield as the most sought trait in rice variety selection (Namufumba et al., 2013; Gridley et al., 2002; Kimani et al., 2011). This study has however indicated Striga resistance as the most considered trait in variety selection. Having Striga resistance as the most sought trait in decision making to adopt a variety only qualifies the effect of Striga on rice production in this area. Also from study one, results indicated the local variety (Superica) as a highly susceptible variety to Striga. Results also indicated the worst preferred varieties of rice which included NARC 3 (ITA 257), IAC 165, Superica, Makassa, Anakila and UPR. Surprisingly, varieties Anakila had an excellent resistance to Striga as well as high yields. Varieties Makassa and UPR has partial resistance levels to Striga. The O.glabberima varieties Makassa and Anakila are susceptible to lodging, this negative trait led farmers to reject all the O.glabberima varieties. The O.sativa variety NARC3 (ITA 257) inspite of its good aroma was rejected by farmers for being short, susceptible to Striga and being low yielding. On other hand UPR was rejected for it short suture and low yields. It was also shown in some studies that farmers prefer tall varieties which reduce on the burden of bending while harvesting (Efisue et al., 2008; Kimani et al., 2011).

This study also indicated significant difference in variety selection among men and women. For example Men selected NERICA-17 and women selected SCRID090 as the best variety for each group. This is related to the roles of men and women in the crop life cycle. Women normally carry out the initial roles in the crop life cycle such as sowing, weeding while men normally involve in the marketing chain of the crop. Studies have shown differences in traits men and women sought in the newly introduced varieties (Dorward et al., 2007). Variety SCRID090 selected by women is tall, Striga resistant and early maturing. However though NERICA-17 is high yielding and Striga resistant it is not as earlier maturing as SCRID090. Addison et al. (2014) acknowledged early maturity as one of the most important trait women sought in selecting varieties in Ghana.

In study three, a selection of varieties in the first and second study were evaluated for their post attachment resistance under controlled conditions. Also the study investigated whether the resistance observed in the field can be replicated under controlled conditions. Results indicated significant effects of Striga on rice varieties when compared to control plants. This was manifested in the reduced rice biomass. Generally there was a loss in biomass across varieties, howver some varieties i.e. Superica, IAC165, WAB880, and WAB181-18, CG14, WAB928 and IRGC were affected when compared to varieties NERICA-1, WAB56-50, WAB56-104, SCRID090, Blechai, and NERICA when compared to control plants. It has been proposed that the loss of biomass of Striga infected plants to some extent is due to acquisition of carbohydrates, nitrogen and other solutes from the host. This results in change in host architecture leading to stunting of plants especially for the most susceptible varieties, lowering in tillering (Cissoko et al.. 2011; Jamil et al., 2011), leaf area, leaf number, thinning of stems and shortening of stem internodes (Swarbrick et al., 2008; Gurney et al., 1999; Ransom 2004) which ultimately reduces the total biomass of the infected plants. The second notion to the reduction in infected biomass could be due to lowering of photosynthesis in infected plants. Studies have shown rates of photosynthesis are usually lower in Striga infected leaves due to stomatal closure following infection (Gurney et al., 1997; Frost et al., 1997). Photosynthesis plays a greater role in the regeneration of new plant tissues that eventually increases on the plant biomass. Striga hermonthica alters the partitioning of dry matter to the different parts of the plant (Cechin and Press, 1994). Therefore reductions or interruptions in this process can reduce the biomass and ultimately the yield of the varieties.

This study has indicated excellent post-attachment resistance for varieties CG14, WAB928, IRGC, SCRID090, WAB181-18, WAB880, Blechai, WAB56-104 and NERICA-4. These same varieties except WAB56-104 had an excellent resistance to Striga under field conditions in study one. Also Cissoko et al. (2011) reported varieties WAB181-18, NERICA-4 and CG14 as having an excellent Post-attachment resistance to S.hermonthica ecotype (Sh.Kibos). The hypothesis set that the resistance under field and controlled conditions does not differ holds only true for the above varieties. However for varieties WAB56-104 and WAB56-50 considered susceptible under the field conditions edged out Striga by having a good resistance to Striga. This however is not surprising for variety WAB56-50 because it was reported as highly resistance under controlled conditions (Cissoko et al., 2011; Rodenburg et al., 2015). However when tested under field conditions in Kenya by Rodenburg et al. (2015) it was susceptible. This can only be expressed by lack of pre-attachment resistance mechanism, no wonder Jamil et al. (2011) reported it as the highest strigolactone producer. Results indicated varieties NERICA-17 and NERICA-1 producing moderate levels of Striga plants per root system. Similarly NERICA-1 as also categorised under partial resistance varieties in the field study. However NERICA-17 was among the resistant varieties in the field which was not replicated under controlled conditions. Results contradict with Cissoko et al. (2011) who reported these two varieties as having an excellent post- attachment resistance to Striga. This could be related to differences in the ecotypes of Striga used in these two studies.

This study also clearly examined the concept of tolerance, where some varieties perform better than others under similar parasitic load (Cissoko et al., 2011). Also the most resistant varieties can also have well developed Striga plants on their root system. Some of the most resistant varieties were also less affected by Striga i.e. Blechai, SCRID090 was 6.5% and 12.8% less than the biomass of uninfected control plants compared to Superica and IAC165 whose biomass was 74.1% and 60.2% less than their respective control plants. This study also indicated genetic variation for tolerance to Striga among rice varieties. For example varieties CG14, WAB880 with lower Striga attached on the host roots had a high reduction in rice biomass 33.3% and 54.2% respectively. However varieties NERICA-1 and NERICA-17 with a high parasitic number had less reduction in rice biomass 1.9% and 6.1% respectively compared to uninfected control plants. This concept can therefore be exploited such that tolerance genes can be incorporated in the resistant varieties.

From the results of the whole study, the following conclusions can be drawn

Rice varieties that combine high Striga resistance/tolerance levels and yield such as SCRID090, WAB800, BLECHAI, WAB181-18, and NERICA-17,-2, -4, -10 can be incorporated in integrated Striga management packages to improve rice production in these areas.

Rice varieties that were selected by farmers such as WAB880, SCRID090, NERICA-17, WAB181-18 and Blechai can be recommended for dissemination in all Striga prone areas in Uganda. This because these varieties also had an excellent resistance and high yields both under field and controlled conditions.

Use of invitro methods based on a multiple mechanisms either pre-attachment /post-attachment resistance is important in selection of resistance for Striga amongst a large accessions of rice.

Understanding the tolerance genes in rice varieties such as WAB56-50, NERICA-1 can be important to incorporate these genes in the most resistance varieties such as WAB880 and CG14 to improve their yielding potential.

The O.glabberima varieties Anakila, ACC, AGEE, CG14 and Makassa with high yields under Striga prone areas can be improved by incorporating lodging resistance genes to improve their suitability among Striga prone areas.

Varieties WAB928, WAB935, IRGC and IR49 with lower yield potential despite the excellent resistance levels to Striga can also be improved or crossed with other susceptible varieties to improve on these varieties.

REFERENCES

Abbasher, A.A., Kroschel, J., Sauerborn, J., 1995. Micro-organisms of Striga hermonthica in northern Ghana with potential as bio-control agents. Bio-control Science and Technology 5, 157–161.

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 and Development 14, 335-343.

Adagba, M.A., Lagoke, S.T.O., Imolehin, E.D., 2002a. Nitrogen effect on the incidence of Striga hermonthica (Del.) Benth in upland rice. Acta Agronomica Academiae Scientiarum Hungaricae 50, 145–150.

Africa Rice Center (WARDA). 2009. The growing NERICA Boom in Uganda. WARDA Publications: Cotonou, Benin

Ahmed, N.E., Sugimoto, Y., Babiker, A.G.T., Mohamed, O.E., Ma, Y., Inanaga, S., Nakajima, H., Ma, Y.Q., 2001. Effects of Fusarium solani isolates and metabolites on Striga germination. Weed Science 49, 354–358.

Ahonsi, M.O., Berner, D.K., Emechebe, A.M., Lagoke, S.T., 2004. Effects of ALS inhibitor

Analytical Report February 2007 and Tanzania. Asp. Appl. Biol. 75, 127–138.

Assuero, S. G., & Tognetti, J. A. 2010. Tillering regulation by endogenous and environmental factors and its agricultural management. Am. J. Plant Sci. Biotechnol, 4, 35-48.

Atera, E. A., Itoh, K., & Onyango, J. C. 2011. Evaluation of ecologies and severity of Striga weed on rice in sub-Saharan Africa. Agriculture and Biology Journal of North America, 2(5), 752-760.

Atera, E. A., Itoh, K., Azuma, T., & Ishii, T. 2012. Response of NERICA rice to Striga hermonthica infections in western Kenya. International Journal of Agriculture and Biology, 14(2), 271-275.

Baguma, 1996.Striga research activities in Uganda, 1994-1995. Striga Newsletter, 6, 5-6.

Becker, M., Johnson, D.999. The role of legume fallows in intensified upland rice based systems of West Africa. Nutrient Cycling in Agro-ecosystems 53, 71–81

Berner, D. K., Kling, J. G., and Singh, B. B. 1995.Strigaresearch and control – a perspective from Africa. Plant Disease 79:652-660.

Carsky, R.J., Singh, L., Ndikawa, R., 1994a. Effect of herbicide and hand weeding on current and subsequent season Striga hermonthica density on sorghum. International Journal of Pest Management 40, 111–116.

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(7), 925-930.

Cissoko M, Boisnard A, Rodenburg J, Pres M.C, Scholes J.D. 2011. New Rice for Africa (NERICA) cultivars exhibit different levels of post-attachment resistance against the parasitic weeds Striga hermonthica and Striga asiatica. New Phytologist 192: 952-963.

Cruz, P., & Boval, M. 2000. Effect of nitrogen on some morphogenetic traits of temperate and tropical perennial forage grasses. Lemaire, G.; Hodgson, J, 151-168.

Delassus, M., 1972. Methods of controlling Striga spp. Agronomie Tropicale 27, 249–254.

Drennan, D.S.H., El Hiweris, S.O., 1979. Changes in growth regulator substances in Sorghum vulgare infected with Striga hermonthica. In: Musselman, L.J., Worsham, A.D., Eplee, R.E. (Eds.), Proceedings of the Second International Symposium on Parasitic Weeds, pp. 144–155.

El Hiweris, S.O., 1987. Nature of resistance to Striga-hermonthica (Del.) Benth. Parasitism in some Sorghum vulgare (Pers.) cultivars. Weed Research 27, 305–312.

Eplee, R. and Langston, M. 1991. Intergrated control methodology for Striga in the USA. In: Ransom, J.K., Musselman, L.J., Worsham, A.D. and Parker, C. (Eds.), Proceedings of the 5th international symposium of parasitic weeds, Nairobi: CIMMYT. Pp 397-399.

FAO STAT, 2010 (http://faostat.fao.org).

Fisher, E., Raschke, K., and Stitt, M. 1986. Effects of Abscisic acid on photosynthesis in whole leaves: changes in carbondioxide assimilation levels of carbon-reduction- intermediates and activity of ribulose 1-5- phosphate carboxylase Planta 169,536-545.

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 for ABA. Plant Cell and Environment 20, 483–492.

Gebisa Ejeta and Jonathan Gressel. 2007. Integrating new technologies for Striga control: towards ending the witch hunt; World Scientific Publishing Co.Pte-Ltd.

Gitau, R., Mburu, S., Mathenge, M. and M. Smale. 2011. Trade and agricultural competitiveness for growth, food security and poverty reduction: A case study of wheat and rice production in Kenya. WPS 45/2011. Tegemeo Institute of Agricultural Policy and Development. Nairobi, Kenya. Pp. 69.

Graves JD, Press MC, Stewart GR. 1989. A carbon balance model of the sorghum–Striga hermonthica host–parasite association. Plant, Cell Environment 12: 101–107.

Greathead, D.J. and Milner, J.E.D. 1971. A survey of Striga species (Scrophulariaceae) and their insect natural enemies in East Africa with discussion of the possibilities of biological control. Tropical Agriculture, 48: 111-124

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(12), 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(3), 573-580.

Gurney, A.L., Slate, J., Press, MC., Scholes, J.D., 2006. A novel form of resistance in rice to angiosperm parasite Striga hermonthica. New Phytologist 169, 199-208

Hess, D.E., Ejeta, G., Butler, L.G., 1992. Selecting sorghum genotypes expressing a quantitative biosynthetic trait that confers resistance to Striga hermonthica. Photochemistry 31, 493-497

Ivens G.W. 1989. East African weeds and their control 2nd Edition, Oxford University press

Jamil M, Rodenburg J, Charnikhova T, Bouwmeester HJ. 2011. Pre-attachment Striga hermonthica resistance of NERICA cultivars based on low strigolactone production. New Phytologist 192: 964-975

Jimmy P, Mbeche I, 2008. A report on farmer’s evaluation of the Striga threat and potential of push-pull technology in controlling Striga weed in upland rice in East Africa; www.icipe.org

Johnson, D.E., Riches, C.R., Diallo, R., Jones, M.J., 1997.Striga on rice in West Africa; crop range and the potential of host resistance. Crop Protection 16, 153–157.

Johnson, D.E., Riches, C.R., Jones, M.P., Kent, R., 2000. The potential for host resistance to Striga on rice in West Africa. In: Haussmann, B.I.G., Hess, D.E., Koyama, M.L., Grivet, L., Rattunde, H.F.W., Geiger, H.H. (Eds.), Breeding for Striga Resistance in Cereals: Proceedings of a Workshop Held at IITA, pp. 139–146.

Jones, M. P., Dingkuhn, M., Aluko, G. K., & Semon, M. 1997. Interspecific Oryza sativa L. x O. glaberrima Steud. Progenies in upland rice improvement. Euphytica, 94(2), 237-246.

Kanampiu, F.K., Kabambe, V., Massawe, C., Jasi, L., Friesen, D., Ransom, J.K., Gressel, J., 2003. Multi-site, multi-season field tests demonstrate that herbicide seed-coating herbicide-resistance maize controls Striga spp. and increases yields in several African countries. Crop Protection 22, 697–706.

Kenmore P, 2003.Sustainable rice production, food security and enhanced livelihoods, in “Rice Science: Innovations and Impact for Livelihood, pg. 27-34.

Khan, Z. R., Midega, C. A. O., Hassanali, A., Pickett, J. A., Wadhams, L. J., and Wanjoya, A. 2006. Management of witch weed, Striga hermonthica, and stem borers in sorghum, Sorghum bicolor, through intercropping with green leaf desmodium, Desmodium intortum. Int. J. Pest Manag. 52(4), 297–302.

Khush, G. S. (1997). Origin, dispersal, cultivation and variation of rice. Plant Mol. Biol. 35, 25 34.

Kim, S. K. (Ed). 1991. Combating Striga in Africa. Proceedings, International Workshop organized by IITA, ICRISAT, and IDRC, 22-24 August 1988. IITA, Ibadan, Nigeria. 151 pp.

Lendzemo, V.W., Kuyper, T.W., Kropff, M.J., Ast, A., 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 Crop Research 91, 51–61.

Linare, O.F 2002. African rice (Oryza glaberrima) history and future potential. Proceedings of the National Academy of Science of United States of America 99: 16360-16365.

Lopez-Raez, J.A., Matusova, R., Cardoso, C., Jamil, M., Charnikhova, T., Kohlen, W., Ruyter-Spira, C., Verstappen, F., Bouwmeester, H., 2009. Strigolactones: ecological significance and use as a target for parasitic plant control. Pest Management Science 65, 471–477.

Maiti, R.K., Ramaiah, K.V., Bisen, S.S., Chidley, V.L., 1984. A comparative study of the haustorial development of Striga asiatica (L.) Kuntze on sorghum cultivars. Annales Botanici 54, 447-457

Marley, P.S., Shebayan, J., Aba D.A., Idem, N., 2004. Possibilities for control of Striga hermonthica in sorghum (Sorghum bicolar) using neem (Azadiractha indica) and Parkia (Parkia biglobosa) based products. International Journal of pest management 50, 291-296.

Ministry of Agriculture, Animal Industry and Fisheries (MAAIF). 2001. Agicultural Production Statistics, Ministry of Agriculture, Animal Industry and Fisheries: Kampala, Uganda.

Ministry of Agriculture, Animal Industry and Fisheries (MAAIF). 2009. Uganda National Rice Development Strategy. Second draft. Ministry of Agriculture, Animal Industry and Fisheries, Entebbe, Uganda

Ministry of Agriculture, Animal Industry and Fisheries (MAAIF). 2011. Statistical Abstract. Ministry of Agriculture, Animal Industry and Fisheries, Entebbe, Uganda.

Mohammed, A., Ellicot, A., Housley, T.L., Ejeta, G., 2003. Hypersensitive response to Striga infection in sorghum. Crop Science 43, 1320-1324.

Mohammed, K.I., Bolin, J.F., Musselman, L.J., Townsend Peterson, A., 2007. Genetic diversity of Striga and implications for control and modeling future distribution. In: Ejeta, G., Gressel, J. (Eds.), Integrating New Technologies for Striga Control-Towards ending the witch hunt World Scientific Singapore, Pp. 71-84

Mohammed, K.I., Musselman, L.J., Riches, C.R., 2001.The genus Striga (Scrophulariaceae) in Africa. Annals of the Missouri Botanical Garden 88, 60-103

Musselman, L.J. (Ed) 1987. Taxonomy of witch weeds. Page 3-12. In: Parasitic plants in Agriculture Striga, CRC Press. Boca Raton. USA 317Pp

Namutumba Census Report. 2007. Namutumba District 2002 Population and Housing Census

Normile, D, 2004. Yangtze seen as earlier rice site Science 275:309

Odhiambo, G.D., Ransom, J.K., 1994. Long term strategies for Striga control. In: Proceedings of the Fourth Eastern and Southern Africa Regional Maize Conference, 28 March-1 April 1994, pp. 263-266.

Olivier, A., Benhamou, N., Leroux, G.D., 1991. Cell surface interactions between sorghum roots and parasitic weed Striga hermonthica: cytochemical aspects of cellulose distribution in resistant and susceptible host tissues Canadian Journal of Botany 69, 1679-1690.

Oswald A., 2004. Striga control-technologies and their dissemination; Crop Protection Journal: 24, 333-342.

Pageau, K., Simier, P., Le Bizec, B., Robins, R. J., & Fer, A. 2003. Characterization of nitrogen relationships between Sorghum bicolor and the root‐hemiparasitic angiosperm Striga hermonthica (Del.) Benth. Using K15NO3 as isotopic tracer. Journal of Experimental Botany, 54(383), 789-799.

Parker, C., Riches, C.R., 1993. Parasitic Weeds of the World: Biology and Control. Cab International, Wallingford, Oxon.

Plan for Modernization of Agriculture (PMA) Secretariat. 2009. Rice Value Chain Study in Acholi and Lango Sub-regions. Plan for Modernization of Agriculture (PMA) Secretariat. Kampala, Uganda.

Porteres, R. 1956. Taxonomic agrobotanique der riz cultives O. sativa Linne et O. glaberrima Steudelo, J. Agri.

Press, M.C and Stewart, G.R, 1987. Growth and photosynthesis in Sorghum bicolar infected with Striga hermonthica. Annals of Botany 60: 657-662.

Purseglove, J.W. 1975. Tropical crops; Monocotyledons vol.1 and 2. Published by Longman Group UK Ltd.

Ramaiah, K.V., Parker, C., Vasudeva Rao, M.J., and Musselman, L.J., 1983. Striga identification and control handbook. Information Bulletin No. 15. Patancheru, A.P., India: International crop Research Institute for the Semi- Arid Tropics

Ransom, J.K., Odiambo G.D., Abayo G., and Oswald, A. 1997. An update on Striga control research in Africa In: Adipala, E., Tusiime, G., and Okori P (eds) Proceedings of the 16th Biannual Weed Science Society Conference for Eastern Africa Kampala Pp 215-219.

Rich, P., Grenier, C., Ejeta G., 2004.Striga resistance in the wild relatives of sorghum. Crop Science 44, 2221-2229.

Riches, C.R, Johnson, D.E., Jones, M.P., 1996. The selection of resistance to Striga species in upland rice. In: Moreno, M.T., Cubero, J.L., Berner, D., Musselman, L.J., Parker, C. (Eds.) Advances in Parasitic Research Proceedings of the Sixth International Parasitic Weed Symposium Cordon Pp. 673-680.

Riches, C.R., Mbwaga, A.M., Mbapila, J., and Ahmed, G.J.U. 2005. Improved weed management delivers increased productivity and farm incomes from rice in Bangladesh and Tanzania Asp. Appl. Biol. 75-127-128.

Rodenburg J; Charles R. Riches; Jumam K. 2010. Addressing current and future problems of parasitic weeds in rice Journal of Crop Protection; 29: 210-221.

Rodenburg, J., Cissoko, M., Kayeke, J., Dieng, I., Khan, Z. R., Midega, C. A., &Scholes, J. D. 2015. Do NERICA rice cultivars express resistance to Striga hermonthica (Del.) Benth. and Striga asiatica (L.) Kuntze under field conditions?. Field Crops Research, 170, 83-94.

Rodenburg, J., Saito, K., Kakaï, R. G., Touré, A., Mariko, M., & Kiepe, P. 2009. Weed competitiveness of the lowland rice varieties of NERICA in the southern Guinea Savanna. Field Crops Research, 114(3), 411-418.

Sakamoto, T. 2006. Phytohormones and rice crop yield: strategies and opportunities for genetic improvement. Transgenic research, 15(4), 399-404.

Shikawa, S., Maekawa, M., Arite, T., Onishi, K., Takamure, I., & Kyozuka, J. 2005. Suppression of tiller bud activity in tillering dwarf mutants of rice. Plant and Cell Physiology, 46(1), 79-86.

Swarbrick, P. J., Huang, K., Liu, G., Slate, J., Press, M. C., & Scholes, J. D. 2008. Global patterns of gene expression in rice cultivars undergoing a susceptible or resistant interaction with the parasitic plant Striga hermonthica. New Phytologist, 179(2), 515-529.

Tatsushi Tsuboi, 2005. Paper presented at the WARDA – NERICA rice Workshop, Ivory Coast, 8thOctober, 2005.

Uganda Bureau of Statistics (UBOS), 2002. Uganda National House hold Survey 1999/2000 report on the crop Survey module. Entebbe, UBOS.

Uganda Bureau of Statistics (UBOS), 2012. Statistics Abstract 2012, Kampala UBOS statistical house available at www.ubos.org (accessed May 2013).

Vaughan, D.A, 1994. The wild relatives of rice; A genetic hand book, International Rice Research Institute, Manila.

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(8), 919-930.

Wilfred, O. 2006. Final survey report on the status of rice production, processing and marketing in Uganda, a report submitted to the Embassy of Japan in Uganda through JICA and Sasakawa Africa Association –Uganda pp.90

Wopereis, M. C. S., Diagne, A., Rodenburg, J., Sié, M., & Somado, E. A. 2008. Why NERICA is a successful innovation for African farmers: A response to Orr et al from the Africa Rice Center. Outlook on AGRICULTURE, 37(3), 169-176.

Xu Kuangdi and Shen Guofang 2003.Promoting Chinese rice production through innovative science and technology, pg 11-18 of the Proceedings of the International Rice Research Conference, 16-19 September 2002, Beijing, China.

Yoneyama, K., Xie, X., Kusumoto, D., Sekimoto, H., Sugimoto, Y., Takeuchi, Y., & Yoneyama, K. 2007a. Nitrogen deficiency as well as phosphorus deficiency in sorghum promotes the production and exudation of 5-deoxystrigol, the host recognition signal for arbuscular mycorrhizal fungi and root parasites. Planta, 227(1), 125-132.

Yoneyama, K., Yoneyama, K., Takeuchi, Y., & Sekimoto, H. 2007b. Phosphorus deficiency in red clover promotes exudation of orobanchol, the signal for mycorrhizal symbionts and germination stimulant for root parasites. Planta, 225(4), 1031-1038.