DIVERSITY, ECOLOGY AND POPULATION DYNAMICS OF LEPIDOPTERAN STEM BORERS IN KENYA By ONG' AMO, GEORGE OTIENO 184/15666/05 B. Envt. Studies (Se), MSe - Kenyatta University (Kenya) Ong'amo. Gcorge . Diversity. ecology and population IIII IIRIII 111111 2011/359847 AR A THESIS SUBMITTED IN FULFILMENT OF THE REQUIREMENTS FOR THE AWARD OF THE DEGREE OF DOCTOR OF PHILOSOPHY IN APPLIED ENTOMOLOGY IN THE SCHOOL OF PURE AND APPLIED SCIENCES OF KENYATTA UNIVERSITY July 2009 11 DECLARA TION Candidate This thesis is my original work and has not been submitted for a degree in any other university or any other award. Ong'amo, George Otieno (MSc.):i;~~:~~~.~~~.~~~~~Date ?: ~ .~? ~:. Supervisors We confirm that the candidate under our supervision carried out the work reported in this thesis. Date.!.q.~.':?l .. -!!:~.~.J ological Sciences, School of Pure and Applied Sciences, Kenyatta University, Nairobi, Kenya Prof. Elizabeth D. Kokwaro Signature.~O) . Department of Zoological Sciences, School of Pure and Applied Sciences, Kenyatta University, Nairobi, Kenya Dr. Bruno Le Ru 1_ Signatuesx ,*'4~............... Date J.7dtd:-4J.'1 . Unite de Recherche IRD 072, Noctuid Stem Borer Biodiversity Project (NSBBP), Institut de Recherche pour le Developpernent (IRD/ icipe), Nairobi, Kenya Dr.Jean-Franc~ Signature f. ::::::::::...... Date '2:i!/.df:,/!)j. . Unite de Recherche IRD 072, CNRS, Laboratoire Evolution, Genornes et speciation, UPR 9034, 91198 Gif-sur-Yvette cedex, France and Universite Paris-Sud 11, 91405 Orsay cedex, France. Jll DEDICATION I dedicate this thesis to my beloved wife, Esther Adhiambo Abonyo, and our son, Fidel Jones Ong'amo; for their understanding and patience during research period which subjected them to lonely family life. L A IV ACKNOWLEDGEMENTS I am grateful to the Insect-plant interaction sub-program (072) of Institut de Recherche pour le Developpement (IRD) for granting the opportunity to conduct this study and the Departement Soutien et Formation des Communautes Scientific de Sud (DSF) for finacial support. I am very grateful to Kenyatta University (KU), especially the department of Zoological Sciences for accepting my registration as a post graduate student. I feel honoured to have worked with Professor(s) Callistus K. P. O. Ogol and Prof. Elizabeth D. Kokwaro both of the Department of Zoological Sciences (Kenyatta University) as my supervisors. I am grateful for their excellent supervision, encouragement and advice that contributed greatly to the completion and production of this thesis. I am sincerely grateful to Dr. Bruno Le Ru of Institut de Recherche pour le Developpernent (IRD), France for his tireless effort in supervision, valuable advice, encouragement during my studies and special thanks for organising research funds through Noctuid Stem Borer Biodiversity Project (NSBBP) which allowed unlimited access to research facilities at the International Centre of Insect Physiology and Ecology (icipe). You have always motivated me throughout the seven years of my post graduate life. I will never forget your influence in my academic life. Besides being my supervisor, you are good friend I always counted on during the hour of need. My sincere gratitude is extended to Dr Jean-Francois Silvain, Director of IRD's Insect-plant interaction sub-program (IRD) for his supervision and encouragement. Dr Jean-Francois Silvain cordial collaboration with Laboratoire Evolution, Genornes and Speciation (LEGS) in Centre National de la Recherche Scientifique (CNRS) provided vadvanced facilities for molecular diagnostics in France. I wish to thank Antoine Branca, Magally Tores and Claire Capdeviile-Dulac of (IRD-LEGS) France for their invaluable help during molecular studies. I wish to thank Dr(s) Paul Andre-Calatayud and Fritz Schulthess for their support and advice during the study. I extend my appreciation to Mr. Leonard Ngala, Boaz Musyoka and Antony N. Kibe for their technical assistance during both field collection and rearing, not to forget Mr. Simon Mathenge of East African Herbarium (Nairobi) for the assis.tance with identification of plant materials. My special thanks go to local farmers in the four localities who unconditionally allowed access to their fields through out the study period. Without forgetting to mention my colleagues, Gerald Juma, Meshak Obonyo, Duna Mailafiya, Benjamin Muli, Bruce Anani, David Bugeme, Benjamin Muli, Eric Koum, Bonaventure Aman, Anderson Kipkoech, Ivan Rwornushana, Lorna Migiro, Susan Sande, David Mburu and all the ARPPIS students for their friendship and support. Many thanks to Capacity Building, library and Information technology office especially Lisa Omondi, Lilian Igweta, Margaret Ochanda, Wellington, Eddah, John Mwangi, John Masiwe and Glenn Sequeira for their constant support. I thank members of my family for their patience, encouragement and prayers. My dear wife Esther and our son Fidel patiently perservered the many days I spent out in the field and foreign countries. This work would not have been completed without the moral support and encouragement from my colleagues in the African Regional Postgraduate Programme (ARPPIS) and Dessertation Research Internship Programme (DRIP); to all I am grateful. Last but not least, I am grateful to my mother Consolata Odinga; brother, VI Michael Juma; and sisters Benter Achieng' and lane Awuor, for their support and encourangement during the study period. It is not possible to reach everybody, but I want to thank all people who contributed to this thesis through their support and invaluable friendship. God bless you all. I hope that this document provide useful insight for future work on the diversity, ecology and management of stemborers in Africa. l(FNVATTA UNIV R~ITV lJBRAPV ADC AMOVA BSU Bt CNRS CYMMIT Cytb DNA DnaSP dNTP DSF GMO HWE icipe IRD KARI LEGS MNHN NSBBP PCR SAS VII ABBREVIATIONS AND ACRONYMS Agricultural Development Co-operation Analysis of Molecular Variance Biosystematics Unit Bacillus thuringiensis Centre National de la Recherche Scientifique International Maize and Wheat Improvement Center Cytochrome b Deoxyribonucleic acid Deoxyribonucleic acid Sequence Polymorphism Deoxyribonucleotide triphosphate Departement Soutien et Formation des Communautes Scientific de Sud Genetically Modified Organism Hardy- Weinberg equilibrium International Centre of Insect Physiology and Ecology Institut de Recherche pour le Developpement Kenya Agricultural Research Institute Laboratoire Evolution, Genomes et Speciation Museum National d'Histoire Naturelle Noctuid Stem Borer Biodiversity Project Polymerase chain reaction Statistical Analysis Software NVAITA UNIV SIT II RA VIII TABLE OF CONTENTS DECLARATION ii DEDICATION iii ACKNOWLEDGEMENTS iv ABBREVIATIONS AND ACRONyMS vii TABLE OF CONTENTS viii LIST OF TABLES xiv LIST OF FIGURES xv LIST OF PLATES xvi ABSTRA.CT xviii CHAPTER ONE 1 1.1 General introduction 1 1.2 Statement of the problem and justification 5 1.3 Research questions 6 1.4 Null hypothesis 6 1.5 Objectives of the study 6 1.5.1 General objective 6 1.5.2 Specific objectives 7 CHAPTER TWO 8 LITERATURE REVIEW 8 2.1 Evolution of stem borers as pests 8 - E Y ITA U IVEP IB A IX 2.2 Cultivation and economic importance of maize and sorghum 8 2.2.1 Domestication and introduction of maize in Africa 9 2.2.2 Domestication and expansion of sorghum cultivation 10 2.3 Factors limiting production of maize and sorghum 10 2.4 Biology and damage symptoms of stem borer pests l l 2.5 Management of stem borer pests 12 2.5.1 Chemical control 12 2.5.2 Cultural control 13 2.5.3 Biological control. 17 2.5.4 Plant resistance 18 2.6 Insect population carry-over between seasons 18 2.7 Importance of wild habitat in pest dynamics .20 2.8 Genetic structure of pest population .20 2.9 Diversity of non-economic stem borers .21 CHAPTER THREE 23 DIVERSITY AND ECOLOGY OF LEPIDOPTEROUS STEM BORERS IN WILD AND CULTIVATED HABITATS IN SELECTED VEGETATION MOSAICS IN KENYA 23 3.1 Introduction 23 3.2 Materials and methods 25 3.2.! Description of study areas 25 3.2.1.1 Muhaka 26 x3.2.1.2 Mtito Andei 26 3.2.1.3 Kakamega 27 3.2.1.4 Suam 27 3.2.2 Sample size determination 28 3.2.2.1 The total number of plants per locality in each sampling session 28 3.2.3.2 Number of fields, and number of maize and / or sorghum plants per field 29 3.2.4 Sampling of stem borers 30 3.2.4.1 Cultivated fields 30 3.2.4.2 Uncultivated habitats 32 3.2.5 Rearing and identification of stem borers 32 3.2.6 Stem borer species diversity .34 3.3 Results 35 3.3.1 Stem borer species diversity and composition in the cultivated fields 35 3.3.2 Distribution and the importance of host plant species 37 3.3.3 Diversity and distribution of stem borer species .40 3.3.4 Stem borer species distribution and host range .42 3.3.5 Stem borer faunistic similarities among wild host plants .43 3.4 Discussion 45 CHAPTER FOUR 49 DYNAMICS AND MANAGEMENT OF STEM BORERS IN KENYA .49 4.1 Introduction 49 4.2 Materials and methods 51 Xl 4.2.1 Data management and analysis 51 4.3 Results 52 4.3.1 Stem borer pest infestations and the associated seasonal variations 52 4.3.2 Stem borer species composition and seasonal density fluctuations 54 4.3.4 Stem borer pest management practices 58 4.3.5 Management of crop residues 60 4.4 Discussion 62 CHAPTER FIVE 65 HOST-PLANT DIVERSITY OF SESAMIA CALAMISTIS HAMPSON (LEPIDOPTERA: NOCTUIDAE): CYTOCHROME B GENE SEQUENCES REVEAL LOCAL GENETIC DIFFERENTIATION 65 5.1 Introduction 65 5.2 Materials and methods 67 5.2.1 Survey sites and processing of specimens 67 5.2.2 DNA extraction and sequence analysis 69 5.2.3 Evaluation of reproductive parameters 71 5.2.4 Statistical analysis 72 5.3 Results 73 5.3.1 Diet breadth of Sesamia calamistis 73 5.3.2 Differentiation of S. calamistis populations in Kenya 75 5.3.3 Genetic differentiation in host utilization (Mtito Andei) 78 5.3.4 Reproductive and life trait parameters 80 Xli 5.4 Discussion 82 CHAPTER SIX 86 GENETIC DIVERSITY AND POPULATION STRUCTURE OF BUSSEOLA PHAIA SSP. PHAIA BOWDEN (LEPIDOPTERA; NOCTUIDAE) IN WILD AND CUL TIV ATED HABITATS 86 6.1 Introduction 86 6.2 Materials and Methods 89 6.2.1 Description of the study area 89 6.2.2 Sampl ing, rearing and identification of stem borers ,..90 6.2.3 DNA extraction and sequence analysis 91 6.2.4 Data management and statistical analysis 92 6.3 Results 93 6.3.1 Distribution and utilization of potential host plants 93 6.3.2 Genetic diversity and differentiation in host utilization 95 6.3.2.1 Diversity between habitats 95 6.3.2.2 Seasonal variations in haplotype composition 95 6.3.3 Stem borer carry-over between habitats and seasons 98 6.4 Discussion 100 CHAPTER SEVEN 104 GENERAL DISCUSSION, CONCLUSIONS AND RECOMMENDATIONS 104 7.1 General discussion 104 XIII 7.2 Conclusions and recommendations 111 7.3 Future prospects 112 REFERENCES 114 ~E VATTA UNIVE TV L BRAR' XIV LIST OF TABLES Table No. Page 3.1: Number of cultivated fields sampled during different sampling sessions 31 3.2: Average composition of stem borer pest community among cultivated host plants in different localities 35 3.3: Relative annual cover (%) of different host species in different localities 39 3.4: Stem borer species diversity in different localities .41 3.5: Kulczynski coefficient (KC) as calculated between the different localities surveyed in Kenya 44 4.1: Stem borer annual and seasonal infestations (%) in different cultivated fields 53 4.3: Seasonal pest densities in the wild habitats in different localities 57 5.1: List of plant species from which Sesamia calamistis larvae were collected in different 74 5.2a: Genetic diversity of the Cytochrome b gene in Sesamia calamistis populations from four localities in Kenya 76 5.2b: Genetic diversity of the Cytochrome b gene in Sesamia calamistis populations from wild and cultivated hosts in Mtito Andei 78 5.3: Reproductive parameters of the Clade I and 11 populations of Sesamia calamistis as recorded under laboratory conditions 81 6.1: Infested plant species in the surveyed agricultural landscape in Kakamega during long and short rain growing seasons 94 6.2: Genetic diversity of the Cytochrome b gene in Busseola phaia ssp. phaia populations from different habitats across seasons 97 xv LIST OF FIGURES Figure No. Page 3.1 Map of Kenya showing distribution of the four sites surveyed during the study on stem borer species diversity and ecology 25 5.1 Map of Kenya showing the four surveyed sites and the estimated spatial distribution of Sesamia calamistis, Clades I and H 68 5.2 TCS mitochondrial haplotype network of 195 Sesamia calamistis specimens processed from seven African countries 77 5.3 TCS mitochondrial haplotype network of Sesamia calamistis individuals collected from different host plants in Mtito Andei 79 6.1 TCS mitochondrial haplotype network of Busseola phaia ssp. phaia individuals collected from different host plants in Kakamega 96 6.2 Summary of assumed population movement between habitats and seasons 99 XVI LIST OF PLATES Plate No. Page 2.1 Wild grass habitats neighbouring the fields of cultivated Graminae; Pennisetum pupureum in Kakamega '" 14 2.2 Old maize stalks left in the cultrivated fields (a) old maize stems stack on trees within the field (b) Old maize stalks left in the prepared field ready for planting 16 3.1 Stem borer sampling process in Kakamega (a) Sampling of young maize plants (b) Sampling wild plants, Euclaena mexicana along the banks of river Yala 31 3.2 Insect rearing set up in the laboratory, (a) Glass vials with artificial diet in which larvae were maintained until pupal formation (b) Wild plants in which larvae were maintained and plastic vials with soft paper towel where the pupae are maintained until emergence 33 3.3 a Stem borer species in the Busseola genus; Busseola fusca, and Busseola phaia 36 3.3 b Stem borer pest species (i) Sesamia calamistis (ii) Chilo partellus (iii) Chilo orichalcociliellus (iv) Eldana saccharina 36 4.1a Bulldock granules in a container during field application and granules. as seen on the leaf whorl immediately after application in Muhaka 59 4.1b Tractor mounted sprayer app lying Bulldock suspension on young maize plants in Suam 59 XVll Plate No. Page 4.2: Management of old maize stalks in Muhaka (a) Old stalks cut and used for mulching (b) Stalks left in the field after harvest for natural decomposition 61 4.3: Management of old maize stalks in Suam (a) Stalks cut and burnt in the old fields after harvest Cb) Remaining stalks buried in the soil during land preparation 61 5.1: Pictorial comparison of different larvae belonging to Sesamia and Sciomesa genera and dismembered genetalia of S. calamistis used during identification 69 XVlll ABSTRACT Stem borers are important field insect pests of maize [Zea mays L.] and sorghum [Sorghum bicolor (L.) Moench] in Africa. They account for more than 30% yield losses depending on fhe composition of the pest community. A total of 21 pest species have been reported in sub-Saharan Africa all of which are indigenous to the continent except for Chilo partellus (Swinhoe), which was accidentally introduced from Asia. Stem borers are susceptible to environmental fluctuations and the pest species are thought to have experienced changes in physiology and behaviour after close association with highly nutritive crops. Recent studies indicate that in addition to the pest species, there are non- economic stem borer species among wild hosts in the uncultivated fragments. Owing to susceptibility of stem borer species, continued habitat fragmentation and degradation may ultimately result in host range expansion and eventual emergence of "new" pests. Unfortunately, previous studies have been geared towards reducing populations of pest species in the cultivated fields with few attempts to understand possible evolution of less known species to pest status. This research was therefore designed to gather information on stem borer species diversity, host range and ecology in selected agricultural landscapes in Kenya. Surveys were conducted during 2005/2007 growing seasons in and around selected cultivated fields in four localities, Muhaka, Mtito Andei, Kakamega and Suam, representing different altitudinal gradients across the country. A total of 29 stem borer species were identified from 9,771 larvae collected. The identified stem borer species were grouped into 10 different known genera (Acrapex, Busseola, Carelis, Manga, Poecopa, Sciomesa, Sesamia, Eldana, Chilo and Ematheudes) while the unknown species belonged to five different families (Crambidae, Peoriinae, Pyralidae, Schoenobiinae and Tortricidae). There was evidence of variation in both distribution and dominance among the surveyed localities with majority of the species belonging to the Noctuidae family found in Kakamega and Suam, while species belonging to Crambidae and Pyralidae were mainly found in Muhaka and Mtito Andei. The wild stem borer species were identified from 38 different plant species belonging to three different families (Cyperacea [27], Poaceae [10] and Typhaceae [1]), while pest species, Busseola fusca (Fuller), Sesamia calamistis Hampson, Chilo partellus Swinhoe, Chilo orichalcociliellus (Strand) and Eldana saccharina Walker were mainly found on maize and sorghum. Sesamia calamistis and B. phaia ssp. phaia occurred among both wild and cultivated hosts and provided good models for studying exchange of stem borer pest populations between the wild and cultivated habitats. Cytochrome b gene sequences, through the existence of strong genetic structuration, revealed evidence of limited exchange of S. calamistis populations between the habitats. However, genetic analyses of the same gene of Busseola phaia ssp. phaia Bowden populations revealed weak differentiation with respect to host use in different habitats (FsT = 0.016; P = 0.015). Observed variations in the distribution of pest and non-pest stem borer species coupled with differences in genetic structure among model species (S. calamistis and B. phaia ssp. phaia) suggest two things; i) no single management strategy would apply across different landscapes and ii) continued habitat fragmentation 1 degradation would affect ecosystem stability resulting in host range expansion or local species extinctions. Similar intensive studies need to be extended to other areas as it will form the basis upon which different integrated pest management (IPM) packages could be developed. 1CHAPTER ONE 1.1 General introduction Tropical ecosystems support remarkably rich arthropod assemblages (Janz et al., 2001). Despite this wealth of biodiversity, our knowledge of these systems is based primarily on faunistic survey records and only little information is available on what determines arthropod abundance, community structure and species interactions (Erwin, 2004). The primary goal in ecology and evolutionary biology studies is to understand factors that explain patterns of arthropod diversity and associations among interacting species (Janz & Nylin, 1997; Janz et al., 2006). An approach to investigate such multi- species interactions requires focus on the dominant group or a "key assemblages" which is potentially critical to food-web dynamics of the local community. One suitable group for such studies is the herbivorous insects as insect-plant interactions are common and incredibly diverse. The great diversity of insect-plant interactions reflects the fact that herbivores are highly host specific which leads to numerous questions regarding factors that generate, maintain and constrain these associations (Jaenike, 1990; Bernays & Chapman, ]994; Janz & Nylin, 1997; Janz et al., 2001). Herbivorous insects are remarkably species-rich, making up at least one-quarter of all described species (Jaenike, 1990). Explaining the mechanisms behind the diversification and interaction of these groups will thus go a long way towards the understanding of global biodiversity. The earliest possible link between insect diversification and feeding on plants was reported by Ehrlich & Raven (1964) in their paper on the eo-evolution between butterflies and plants. Since then, it has been clearly demonstrated that herb ivory has repeatedly led to rapid diversification of insects, though mechanisms behind this diversification remain uncertain. 2In an attempt to generate information in the diversification process, insect ecologists have been manipulating aspects of plant dispersion and patterning for decades, and recording impacts on herbivore diversity and densities. Reviews of progress in this area over the past two decades indicate that vegetation attributes such as diversity or "heterogeneity" sometimes have effects and other times do not (Andow, 199 I). More recent attempts to quantify the effects of plant diversity using meta-analysis have indicated that diversity has at most a low to moderate effect on herbivores (Tonhasca & Byrne, 1994), highlighting the difficulties inherent in predicting how habitat transformations will affect herbivore communities. Although this does not bode well for the development of the general theory, it is important to note that these experiments have been carried out with a wide array of insects under a variety of field conditions. However, one unexplored dimension of these interactions is the influence of landscape structure at which plant diversity is manipulated. The modem human-dominated landscapes are typically characterized by intensive land-use and high levels of habitat destruction. Most agricultural landscapes consist of a mosaic of wild and cultivated patches that provide the herbivore community with a large tapestry of habitats and resources (Janzen, 1994; Banks, 1998). Fragmentation of remaining habitats is a major threat to biodiversity and an important issue in landscape management (Banks, 1998). Several characteristics of habitat fragments such as size, isolation, proportion of edges, and habitat quality, as well as characteristics of the surrounding landscape are known to influence abundance of populations and diversity of communities (Banks, 1998). The relative importance of each characteristic remains unknown since species are differentially affected by characteristics of habitat fragmentation such as isolation, area and habitat quality (Janzen, 1994). Nonetheless, 3significance of the resulting changes in community structure, interspecific interactions and ecological functions remain unknown. Fragmentation is usually considered in the context of conservation, but is also related to the efficiency of biological control in the agricultural landscape (Manel et al., 2003). Species richness and the strength of interactions of populations between habitats often vary though it is still controversial as to which mechanisms create a positive relationship between biodiversity and ecological functions (Janz et al., 2006). More information may be gathered through ecological studies based on model phytophagous insects of which lepidopterous stem borers forms the best candidate. Stem borers are among the most speciose and taxonomically tractable group (Manel et al., 2003), and due to their important functional roles as selective herbivores, they constitute an important fauna for understanding how changes in habitat quantity and quality interact to influence species richness and community structure. Stem borers feed on a wide range of graminaceous plants including crops (Bowden, 1954; Khan et al., 1997; Gounou & Schulthess, 2004). Some of the species have narrowed their diet breadth and currently feed mainly on cultivated crops where they have attained pest status. Such species include Busseola fusca (Fuller), Sesamia calamistis Hampson, Chilo partel/us (Swinhoe), Chilo orichalcociliellus (Strand) and Eldana saccharina Walker. These species however vary in their importance across the continent with S. calamistis dominating pest community in Western and Central Africa (Bosque-Perez & Schulthess, 1998), while B. fusca and C. partel/us dominate the community in Eastern and Southern Africa (Overholt et al., 1997; De Groote, 2002). Despite variations in pest composition across the continent, severe yield losses have been 4reported from countries in Southern (van Den Berg et al., 2001), Western (Bosque-Perez & Schulthess, 1998) and Eastern Africa (Songa et al., 1998; De Groote, 2002). In addition to cultivated crops, the pest species have for along time been known to occur among the non-cultivated graminaceous plants hereafter referred to as "wild" host plants (Ingram, 1958; Nye, 1960). Persistence and build up of pest populations early in the season is thus blamed on the movement of individuals from wild hosts growing in the uncultivated fragments (Polaszek & Khan, 1998). In order to get a better insight in ecology and the way of controlling these pests, studies in wild environments have been recommended for a long time (Bowden, 1954). This is projected to provide understanding on the infestation dynamics, the possibilities of survival of both introduced and indigenous parasitoids as biological control agents (Gounou & Schulthess, 2004) and estimating the risk of species shift from wild host plants to cultivated crops (Le Ru et al., 2006a). An approach of this kind has been carried out in Eastern Africa (polaszek & Khan, 1998; Le Ru et al., 2006a) and in Western and Central Africa (Schulthess et al., 1997) which led to a better knowledge of wild host plants, mainly for known pest species.. Studying stem borer species diversity and populations in the natural landscapes require proper identification of species which may be difficult because of morphological similarities between species, and intra-specific variability (Tarns & Bowden, 1953; Holloway, 1998). However, stem borers other than the pest species have been reported from surveys carried out in natural habitats in Eastern and Southern Africa (Ingram, 1958; Nye, 1960; Khan et al., 1997; Haile & Hofsvang, 2001; Mazodze & Conlong, 2003; Le Ru et al., 2006a). Recent surveys in East and Southern Africa countries yielded 136 stem borer species from 75 wild host species (Le Ru et al., 2006b), exceeding stem borer species and host plants recorded in the earlier studies. Recorded increase in the 5number of borer species and host plants in this region suggest that these lists are far from being exhaustive. Unfortunately, habitats with some host plants of these stem borer species are currently undergoing destruction thus imposing potential danger of host range expansion or local extinction of some unknown species. It is therefore necessary to catalogue stem borer species diversity in less disturbed habitats as this would provide insight on the role of wild habitats. in pest ecology and dynamics that could be useful in designing future management strategies. 1.2 Statement of the problem and justification Indigenous African stem borers have been associated with a wide range of wild hosts belonging to Poaceae, Cyperaceae and Typhaceae families for millions of years (Harris, 1962; Harris & Nwanze, 1992). However, stem borers appear to be susceptible to environmental fluctuations and the pest species are thought to have experienced changes in physiology and behaviour after close association with highly nutritive crops (Haile and Hofsvang, 2001). Continued habitat fragmentation and degradation as demand for more agricultural land to feed the growing human population increases may ultimately result in host range expansion and eventual emergence of "new" stem borer pests, or local species extinction. Distribution of stem borers (economic and non-economic species), however varies with respect to their ecological requirements (Le Ru et al., 2006a) suggesting that impacts of habitat degradation on species diversity and ecology may vary among different vegetation mosaics. This study was designed to gather information on stem borer species diversity, host range and ecology in selected agricultural landscapes, with a view to get insight on species diversity and exchange of stem borer pests between wild and cultivated habitats. The findings are projected to provide understanding on potential NYAITA UNI· ERSi Y LI.,; 6ecological consequences of changes in the spatial structure of uncultivated fragments on stem borer species diversity and ecology. 1.3 Research questions a) How does the diversity and ecology of lepidopteran stem borers vary among different habitats? b) How does stem borer pest community vary III agricultural systems within a season? c) What is the role of wild habitats in stem borer pest build-up between the growing seasons? 1.4 Null hypothesis a) Diversity and ecology of lepidopteran stem borers in both wild and cultivated habitats are the same in different vegetation mosaics. b) Stem borer pest community in agricultural systems does not vary with seasons. c) Wild habitats are not responsible for stem borer pest population build-up between the cropping and non-cropping seasons. 1.5 Objectives of the study 1.5.1 General objective To assess the diversity, ecology and population dynamics of lepidopteran stem borers in the cultivated and wild habitats, and determine exchange of model species between the habitats. 71.5.2 Specific objectives a) To estimate diversity and ecology of lepidopteran stem borers in both wild and cultivated habitats in different vegetation mosaics. b) To evaluate stem borer pest community in the cultivated habitats and monitor seasonal variation in population dynamics. c) To determine the genetic structure and movement of model pest species, S. calamistis and B. phaia ssp. phaia, between wild and cultivated habitats A 8CHAPTER TWO 2 LITERATURE REVIEW 2.1 Evolution of stem borers as pests Insects may become pests due to several factors. Some previously harmless insects became pests after their accidental or intentional introduction to areas outside their native range, where they escaped from their natural enemies (Strauss, 1997). Such range extensions have allowed many previously innocuous phytophagous insects to flourish as pests, usually following the deliberate spread of their hosts through human cultivation (Strauss, 1997). Additionally, some native insects may become pests if they move from native plants to introduced ones; such host switching is common among polyphagous and oligophagous insects (Magurran, 1988). Stem borer pests are among the phytophagous insects that remained among wild hosts before domestication of sorghum and introduction of maize in Africa (Harris, 1962; Harris & Nwanze, 1992). Their pest status has been elevated by the provision of abundant food resources that occur in simplified or virtually monocultural ecosystems in which maize and sorghum crops create dense aggregations of predictably available resources (Overholt et al., 2001). 2.2 Cultivation and economic importance of maize and sorghum Maize originally domesticated in Central America is currently among the most widely cultivated cereal crop in the world (Seshu Reddy, 1998). Its successful distribution may be attributed to high productivity and availability of many varieties developed for diverse ecological conditions (Schulthess et al., 1997). In tropical Africa, it is used for many purposes such as human food, feed for animals and raw materials for many industrial products. Similarly, sorghum is another widely grown cereal crop in 9tropical Africa. After initial domestication about 5000 years ago probably In the savannahs West of Ethiopia and East of Chad, cultivation of sorghum effectively spread to other parts of the world due to its good resistance to drought (Haile & Hofsvang, 2001). In Eastern Africa, sorghum is the staple food for millions of people and also is grown to feed livestock in form of grain, forage and fodder (Seshu Reddy, 1983). However, production of these crops has not kept pace with the ever-increasing human demand for food supply. 2.2.1 Domestication and introduction of maize in Africa Domestication of maize is thought to have started from 7,500 to 12,000 years ago though it is not known what precipitated its domestication sine the edible portion of the wild variety is too small and hard to obtain to be eaten directly. However, studies of the hybrids readily made by intercrossing teosinte and modern maize suggest that this objection is not well-founded (Benz, 2001). Archaeological remains of early maize cobs, found at Guila Naquitz Cave in the Oaxaca Valley of Mexico, date back roughly to 6,250 years. Little change occurred in cob form until ea. 1100 BC when great changes appeared in cobs from Mexican caves: maize diversity rapidly increased and archaeological teosinte was first deposited (Benz, 2001). Perhaps as early as 1500 BC, maize began to spread widely and rapidly (lItis, 2006). Currently, maize is widely cultivated throughout the world, and a greater quantity is produced each year than any other grain (Benz, 2001). Its successful distribution may be attributed to high productivity and availability of many varieties developed for diverse ecological conditions. As it was introduced to new cultures, new uses were developed and new varieties selected to better serve in those preparations. 10 2.2.2 Domestication and expansion of sorghum cultivation Although wild species of sorghum were attested as early as 8000 years ago in the Nilotic regions of southern Egypt and Sudan, the location of true domestication within East Africa is still speculative (Harris & Nwanze, 1992}. It is widely held that genetic separation of domesticated S. bicolor from the wild progenitor did not occur much before 2000 years ago somewhere in Eastern Africa, possibly the Ethiopian highlands (Seshu Reddy, 1989). The presence of true domesticated S. bicolor is claimed much earlier than this (3700-4900 years ago) in India, Oman, and Yemen, although the identity of the remains as full domesticated plants is still disputed (Haile & Hofsvang, 2001). Sorghum requires an average temperature of at least 25°C to produce maximum grain yields in a given year though it is well adapted to grow in hot, arid or semi-arid areas (Seshu Reddy, 1989). 2.3 Factors limiting production of maize and sorghum Notable factors limiting production of maize and sorghum In tropical Africa include poor climatic conditions, low soil nutrients, weeds, diseases and insect pests (Bowden, 1976). Some of these limitations have been solved through improved farm management and development of suitable crop varieties (Overholt et al., 1997). Important among these limitations are the crop damages caused by field insect pests. Outstanding among field insect pests are larval stages of stem boring lepidopteran moths belonging to Pyralidae and Noctuidae families (van den Berg et al., 1998). Over 17 stem borer species belonging to these two families. have been reported to attack cultivated cereals in Africa, and account for about 20-50% yield losses in East Africa (Khan et al., 1997). In Kenya, 11 B. fusca and C. partellus constitute the major proportion of the community and strongly limit yields of maize and sorghum crops from 20 - 80 % depending on region, borer population density and crop phenology during infestation (Seshu Reddy, 1983; Khan et al., 1997). With the exception of C. partellus, which was accidentally introduced from Asia (Tarns, 1932), all other pest species are indigenous to the African continent where they have been associated with wild hosts belonging to Poaceae, Cyperaceae and Typhaceae families for millions of years (Harris & Nwanze, 1992). Evolutionary changes that ended in host shifts may have involved both host plants and some indigenous stem borer pest species (Magurran, 1988). Cultivated crops probably lost their resistance because of long selection by agronomists that allowed for a better expression of stem borers' fitness (polaszek & Khan, 1998). At the same time, economically important species are thought to have experienced changes in physiology and behaviour after close association with highly nutritive crops, including reduced resistance to plant secondary compounds (Haile & Hofsvang, 2001). Like other phytophagous insects, it is thought that there might have been some specific aspects in the ancestral population of species currently specific to cultivated crops that enhanced their ability to switch and / or to adapt genetically to cultivated crops (Nylin & Gotthard, 1998; Hawthorne & Via, 2001). 2.4 Biology and damage symptoms of stem borer pests Stem borer species often associated with maize and sorghum in Kenya includes B. fusca, S. calamistis, C. partellus, C. oricalchociliellus and E. sacharina (Overholt et al., 2001). Other species such as S. nonagrioides and S. cretica are also present in some regions (In gram , 1958; Overholt et al., 2001). All stem borer species characteristically 12 undergo four development stages; egg, larva, pupa and adult. Life cycle begins with emergence of adults and mating follows immediately after emergence by males finding females with the help of pheromones released by the female moths (Overholt et al., 2001). Gravid moths oviposit on suitable young leaves that are later attacked by first instar larvae after hatching (Omwega et al., 2006). Larval attack on the meristematic tissue generally affects the translocation of nutrients in the plant resulting in reduced growth and in some cases death of the plant either through "dead heart" or breakage of the stem. 2.5 Management of stem borer pests Different control methods have been used in Africa to reduce losses associated with field insect pests (Nwanze & Mueller, 1989; Saxena, 1990; Kfir, 1991). In Kenya, chemical, biological, cultural as well as planting of resistant crop varieties have been used in the management of stem borers. 2.5.1 Chemical control Chemical control of stem borers is difficult as pesticides are expensive, often toxic and at times relatively ineffective since target larvae often burrow inside the meristematic tissue (Kumar, 1984; Seshu Reddy, 1998). Despite these limitations, large- scale farmers in Trans-Nzoia district (Kenya) still apply insecticides such as carbofuran, carbaryl and endosulfan (Has san et al., 1998) for the control of the first generation of stem borer population. In addition to economic constraints, other problems associated with this management option that may follow routine application include the need for 13 reapplication, effect on non-target organisms, problems of residues and eventual development of resistance among the target pests (Kumar, 1984). 2.5.2 Cultural control Habitat management, early planting, removal or destruction of crop residues are among the cultural practices associated with stem borer management as they reportedly disrupt stem borer population build-up (Randriamananoro, 1996; Khan et al., 1997). 2.5.2.1 Habitat management Habitat management strategy has been developed in maize-based farming systems for small- and medium-scale farmers of eastern Africa (Khan et al., 1997; Polaszek and Khan, 1998). This strategy involved selection of plant species that could be employed as trap crops to attract (pull) stem borer colonization away from the cereal plants, or as intercrops to repel (push) the pests. The two most successful trap crops Napier grass, Pennisetum purpureum Schumaker (Plate 2.1), and Sudan grass, Sorghum vulgare var. sudanensis Hitchc attract greater oviposition by stem borers, than cultivated maize. Intercrops giving maximum repellent effect are molasses grass, Melinis minutiflora Beauv. and a legume species, silverleaf, Desmodium uncinatum (Jacq.). Adoption of this management option has relatively been slow as proper groundwork to educate farmers on its potential has not been undertaken as this effort is still on experimental trials. However, recent field surveys indicate that majority of stem borer species found in P. purpureum are actually not pest species (Le Ru et al., 2006a) and thus it may not pull target pest like B. fusca in high potential areas in Kenya. 14 Plate 2.1: Wild grass habitats neighbouring the fields of cultivated Graminae; Pennisetum pupureum in Kakamega. 2.5.2.2 Timely planting as a stem borer management option Stem borer management based on timely planting follows the principle of growing target crops when the pest is not present or when the pest is least abundant (Overholt et al., 1997). Field survey results indicate that infested young plants are slow in their recovery unlike older plants that recover fast and compensate for the attack (Kumar, 1984). Through field studies, flying periods of most species (B. fusca and S. calamistis) have been determined and this knowledge is being used in drawing planting schedule (Overholt et al., 1997). Most farmers plant early in the season to avoid early sorghum and maize infestations (Khan et al., 1997). Alternatively, maize or sorghum may be planted later in the cropping season after the flying period of most stem borer moths. This enables 15 plants to escape infestation from the first generation and grow big enough to withstand infestations from populations of the second-generation. Utilization of this knowledge in stem borer management is very low particularly among subsistence farmers who depend on rainfed agriculture (Okech et aI., 1994). Only few farmers may accept to delay planting during the beginning of growing seasons due to unreliable weather pattern. 2.5.2.3 Removal and destruction of crop residues During the dry conditions, stem borers that are unable to complete their development in time enter diapause in stubbles / stalks of maize or sorghum left in the fields in anticipation of limited food resource or oviposition sites (Plate 2.2). This strategy enables various stem borer species to continue with normal development during favourable ecological conditions. Kumar (1984) recommended that burning or spreading residues / stalks in the field to expose larvae to the full effect of adverse weather conditions would limit stern borer carry-over between seasons. These recommendations are only practical in areas where crop residues / stubbles are neither used as animal feed nor fuel. In many parts of Kenya, stalks / residues are carried to different homes where they are used to feed animals and this encourages translocation of stem borers (Bonhof, 2000). L A 16 Plate 2.2: Old maize stalks left in the cultrivated fields (a) old maize stems stack on trees within the field (b) Old maize stalks left in the prepared field ready for planting. 17 2.5.3 Biological control Biological control is often used to include any biologically based methods of pest suppression (Overholt et al., 1997). In the traditional sense, biological control means the manipulation of natural enemies of pests to reduce pest populations to levels where economic losses due to their attack are tolerable. The range of naturally occurring biological control agents, such as parasitoids, predators and diseases have been reported for different growth stages of stem borers (Overholt et al., 1997; Schulthess et al., 1997). Few studies on their effectiveness as well as host/parasite relationships have been made. Biological control agents of interest to many researchers particularly in Kenya include the egg parasitoids, such as Telenomus and Trichogramma species, larval parasitoids including Cotesia sesamiae (Cameron) and pupal parasitoids eg Pediobius furvus Gahan among others (Overholt et al., 1997). Extensive distribution of the exotic C. partellus and the resultant losses associated with its infestations in Coastal low altitude areas evoked the search for effective classical biological control agent from their area of origin (Overholt et al., 2001). Natural enemies of Chilo were collected in different ecological zones in India, and cocoons of C. flavipes Cameron, pupae of Sturmiopsis inferens Townsend and Xanthopimpla stemmator were shipped to (International Centre of Insect physiology and ecology (icipe) Kenya for use in the management of C. partellus in East Africa. The braconid C. flavipes after release has established in most of the countries in East Africa where C. partellus dominates (Songa, 1999). There has been recorded reduction of losses by about 10% due to biological control of cereal stem borers by C. flavipes, which annually caused 10 to 40% loss in grain yield (Seshu Reddy, 1983; Zhou et al., 2002). 18 Despite the recorded impacts of C. flavipes on C. partellus population, research work on the management of other stem borer species are at their lowest (Omwega et al., 2006). Cotesia flavipes is only reported to be effective in C. partellus but do parasitise other species including certain biotypes of B. fusca which encapsulate C. flavipes rendering them ineffective, as they do not emerge (Gitau et al., 2006). Busseola fusca and S. calamistis that dominate high potential areas are given little attention though they cause higher maize losses compared to low potential areas (Eastern and Coastal) where C.partellus dominates (De Groote, 2002). 2.5.4 Plant resistance Plant resistant strategy so far tried in the management of lepidopteran stem borers consists of introducing genetically engineered Bt-maize (Bacillus thurigiensis - Bt). The African pyralid stem borers are close relatives to the European corn borer Ostrina nubilalis Hubner against which the Bt-maize was constructed (Overholt et al., 2001). The Swiss company Novartis in co-operation with the Kenya Agricultural Research Institute (KARI) and the Latin-American CYMMIT introduced Bt-maize in Kenya: in 2000, in a 5-year program. This is still in pilot phase despite oppositions from anti-GMO crusaders. 2.6 Insect population carry-over between seasons Chemical, biological and cultural control methods disrupt stem borer population build-up (Randriamananoro, 1996; Khan et al., 1997). However, these strategies only focus on the management of borer populations within the agricultural systems while ignoring the potential role of wild habitats in pest out breaks. Stem borers feed on one or more closely related plant families in addition to cultivated host crops (Polaszek & Khan, 'KtNYAlTA UNIVERSITY LI. RA \ 19 1998; Haile & Hofsvang, 2001). During cropping seasons, stem borers occur in large numbers in maize and sorghum plants (Songa et al., 1998). After harvest, gravid moths oviposit in alternative wild hosts where their populations survive during the crop free periods (Ingram, 1958; Nye, 1960). The presence of alternative hosts and crop residues in or near a field can increase survival of stem borers, thereby increasing the population that colonise maize and sorghum crops in subsequent growing season. However, surveys in the forest zones of Cameroon, Cote dIvoire and Ghana showed that higher wild host abundance in the surrounding fields was correlated with a lower pest incidence on maize (Schulthess et al., 1997). Oviposition preference and life table studies revealed that some wild host species namely grasses, were highly attractive to ovipositing female moths, although survival of immature stages and adult moth fecundity were mostly close to 70 - 80% against 100% on maize (Shanower et al., 1993). In addition, relatively high parasitism of S. calamistis eggs by Telenomus spp (Hymenoptera: Scelionidae) was found during the dry season on wild hosts in Cameroon (Ndemah et al., 2001). Although Schulthess et al. (1997) showed that at the local scale, wild host plants can attract pests and reduce damage on cultivated plants, there is a possibility of increase in damage in the following season since stem borers have higher survival on some alternative host plants (Polaszek & Khan, 1998). These views appear to differ either because generated hypotheses have not been fully tested or because species in question are different. 199%) upon emergence for DNA extraction. 3.2.6 Stem borer species diversity Identified stem borer species were recorded with respect to host plants and localities. Species recorded from different host plants and localities during the study period (2005/2007) were analyzed for species richness (S) and diversity using Shannon's diversity index (H') (McAleece, 1997)). The hypothesis that stem borer species vary in distribution with respect to their ecological requirements was tested by calculating faunistic similarities. Similarities were compared using Kulczynski coefficient (KC), which measures the percentage similarity among communities between two areas (Price, 1982). This index was chosen because of its advantage over other similarity indices as it takes into account the disproportionate number of species. The Kulczynski coefficient is given by the following formula, 1 [s s 1KC::;::- ( )+-( -) xl002 s+u s+v Where s is the number of species common to area A and B; u is the number of species found in area A and absent from B and, v is the number of species found in area B but absent in area A. 3.3 Results 35 3.3.1 Stem borer species diversity and composition in the cultivated fields Seven stem borer species were identified from the collections made III the cultivated fields (Table 3.2; See Plates 3.3 a and b). Distribution of these species varied among localities with Busseola species occurring in Kakamega and Suam, while Chilo species occurred in Muhaka and Mtito Andei. The Chilo species (c. partellus and C. orichalcociliellus) co-existed with S. calamistis in Muhaka, a community dominated by C. partellus (65%). The dominance of C. partellus was also observed in Mtito Andei (78%) where it co-existed with only S. calamistis. In Suam, only two pest species, B. fusca and S. calamistis were identified with B. fusca constituting about 97% of the community. Both species (B. fusca and S. calamistis) were found in Kakamega where they co-existed with three other species, B. phaia ssp. phaia, S. piscator and E. saccharina. Pest community in Kakamega was generally dominated by B. fusca (40%) followed by B. phaia ssp. phaia (30%) with species richness enhanced by the presence of less known pest species, B. phaia ssp. phaia and S. piscator. Table 3.2: Average composition of stem borer pest community among cultivated host plants in different localities Stem borer species Composition of the stem borer community (%) Muhaka Mtito Andei Kakamega Suam Busseolafusca (Fuller) Busseolaphaia Bowden Sciomesapiscator Fletcher Sesamia calamistisHampson Chilo orichalcociliellus (Strand) Chilopartellus (Swinhoe) Eldana saccharinaWalker 40 97 30 2 22 22 26 3 13 65 78 2 36 Plate 3.3 a: Stem borer species in the Busseola genus; Busseola fusca, i [Male], iii [Larva]; Busseola phaia, ii [Male], iv [Larva] Plate 3.3 b: Stem borer pest species (i) Sesamia calamistis (ii) Chilo partellus (iii) Chilo orichalcociliellus (iv) Eldana saccharina. NV 11 37 3.3.2 Distribution and the importance of host plant species A total of 34 plant species were found infested during this study. These plants however, cover different areas in the surveyed localities as summarized in table 3.3, a table based on the average seasonal plant species cover as established by Otieno et al. (2006 & 2008). Ten of these host plant species were identified in Muhaka with P. maximum covering the largest proportion of the study area (1.61%) followed C. dactylon that covered 0.32%. Cynodon aethiopicus covered the largest area (2.31%) in Mtito Andei, followed by P. maximum (0.67%) and R. cochinchinensis (0.32%), dominating the other 11 host plants in the area. In Kakamega where there were a total of 20 plant species, the largest area was covered by C. dactylon (5.49%) followed by S. megaphylla (2.04%) and S. arundinaceum (0.73%). In Suam, the largest area was covered by C. nardus (7.34%), followed by C. dactylon (7.22%), s. incrissata (1.06%) and C. dives (1.53%). Stem borer host plants generally varied in their importance (Table 3.3). The highest number of stem borer species were found on P. purpureum (10) followed by P. maximum (9), S. arundinaceum (6), C. involucratus (6), P. macrourum (5) and E. haploclada (5). The other host plants were infested by between one and four stem borer species. However, importance of these hosts varied among the localities. Pennisetum purpureum which was infested by 7 stem borer species in Kakamega, was only infested by two stem borer species in both Mtito Andei and Suam, and only one in Muhaka. The other important host, P. maximum, was infested by 5 different stem borer species in Muhaka, three in both Mtito Andei and Suam, and two in Kakamega. Infestation of some hosts were however limited to certain localities. For example infestation in S. arundinaceum was observed only in Muhaka, Mtito Andei and Suam, while C. 38 aethiopicus and S. megaphylla were infested only in Mtito Andei and in Kakamega respectively. K NVATT ITVl 39 Table 3.3: Relative annual cover (%) of different host species in different localities. Host plant species Average area cover (%) Muhaka Mtito Andei Kakamega Suam 1 Cymbopogon nardus (L.) Rendle [3] 00.00 00.00 2 Cynodon aethiopicus Clayton & J.R. Harlan [3] 00.00 02.31 [3] 3 Cynodon dactylon (L.) Pers. [2] 00.32 00.00 4 Digitaria milinjiana (Endle) Staff(!] 00.01 [1] 00.00 5 Echinochloa colonum (L.) Link [I] 00.00 00.01 [1] 6 Echinochloa haploclada (Stapt) Stapf[5] 00.06 [5] 00.13 7 Echinochloa pyramidalis (Lam ark) Hitchcock & Chase [I] 00.00 00.00 8 Eleusine corocana L. * [3] 00.00 00.00 [2] 9 Eleusinejaegeri Pilg. [I] 00.00 00.02 [1] 10 Euclaena mexicana Schrader [2] 00.00 00.00 11 Panicum deustum Thunb. [2] 00.00 00.01 [2] 12 Panicum infestum Andersson. [4] 00.01 [4] 00.00 13 Panicum maximum Jacquin [9] 01.61 [5] 00.67 [3] 14 Pennisetum cladestinum Hochst ex Chiov. [I] 00.00 00.00 15 Pennisetum macrourum Trinius [5] 00.00 00.00 16 Pennisetumpurpureum Schumach§YO] 00.00 [1] 00.04 [2] 17 Pennisetum trachyphyllum Pilg. [2] 00.00 00.00 18 Pennisetum unisetum (Nees) Benth. [I] 00.00 00.00 19 Rottboella cochinchinensis (Lour.) Clayton [4] 00.14 [4] 00.32 [2] 20 Saccharum officinarum L.*[2] 00.00 00.00 21 Setaria megaphylla (Steud.) T. Durand & Schinz [3] 00.00 00.00 22 Setaria incrassata (Hochst.) Hack. [2] 00.00 00.00 23 Setaria verticil/ata (L.) P. Beauv. [3] 00.00 00.00 [2] 24 Sorghum arundinaceum (Desv.) Stapf[6] 00.02 [4] 00.09 [3] 25 Cyperus articulatus L.(!] 00.00 00.00 26 Cyperus distans L. [3] 00.00 00.02 [3] 27 Cyperus dives Delile [4] 00.02 00.00 28 Cyperus involucratus Rottb. [6] 00.00 00.01 [6] 29 Cyperus papyrus L. [I] 00.00 00.00 30 Cyperus rotundus L. [3] 00.01 [3] 00.00 31 Scirpus inclinatus (Delile ex Barbey) Boiss. [I] 00.00 00.00 32 Schoenoplectus confusus (NEBr.) Lye [3] 00.00 00.00 33 Scleria,racemosa Poiret [I] 00.00 00.00 00.51 [2] 00.00 05.49 00.00 00.00 00.00 00.03 [1] 00.01 [2] 00.00 00.01 [2] 00.00 00.00 00.61 [2] 00.00 00.32 [3] 00.63 [7] 00.35 00.00 [1] 00.00 00.01 [2] 02.04 [3] 00.00 00.00 00.73 00.00 00.19 00.06 [2] 00.00 00.00 [1] 00.01 00.01 [1] 00.00 00.23 [1] 07.34 [1] 00.00 07.22 [2] 00.00 00.00 00.00 00.00 00.01 [2] 00.20 00.00 00.00 00.00 00.26 [3] 00.01 [1] 00.17[2] 00.50 [2] 00.66 [1] 00.52 00.39 00.00 00.09 01.01 [3] 00.30 01.06 [2J 00.00 00.22 01.53 [3] 00.00 00.00 00.00 00.00 00.01 [3] 00.00 34 Typha domingensis Pers. [2] 00.00 00.14 [2] 00.0100.15 Cultivated crops included in the list are marked with stars (*) while P. purpureum as a domesticated grass is marked §. In bracket [S] is the number of borer species recovered from different host plants in corresponding localities. KENY ITA U ~IVERS~IYu RA 40 3.3.3 Diversity and distribution of stem borer species A total of 29 stem borer species were identified from the 3,203 larvae collected. These species however varied in distribution among the localities as reflected in the diversity indices (Table 3.4). The highest Alpha (a) diversity was recorded in Suam (a = 3.0), followed by Mtito Andei (a = 2.77) and Kakamega (a = 2.5). In contrast, more larvae were collected in Muhaka (1551) followed by Kakamega (683), Mtito Andei (627) with least materials collected in Suam (342). The highest number of species were recorded in Kakamega (15) followed by both Mtito Andei and Suam with 13 species each. The least number of species was recorded in Muhaka where only 11 were identified. The total larvae and diversity indices recorded in different localities suggested variations in composition of stem borer communities. Stem borer species constituting communities ll1 different localities could be grouped in different categories but species constituting less that 1% of the total collection in respective localities are hereafter referred to as rare. Stem borer community in Muhaka was dominated by C. orichalcociliellus (Berger-Parker Dominance, d% = 46.68) followed by C. partellus and Ematheudes spp, with Acrapex sp as a rare species. In Mtito Andei, stem borer community was dominated by S. nonagrioides (d% = 33.81) followed by S. calamistis and C. partellus. Several species were rare in Mtito Andei and these included S. piscator, C. orichalcociliellus, Chilo sp. nr orichalcociliellus and E. saccharina. Stem borer communities in Kakamega and Suam were dominated by Sesamia nov sp 9 (d% = 20.80) and S. nyei (d% = 32.75) respectively. Several rare species including M melanodonta, P. mediopuncta, S. peniseti and E. saccharina, were found in Kakamega, while Busseola si nov sp 3 and S. poephaga were found in Suam. 41 Table 3.4: Stem borer species diversity in different localities. Stem borer species Muhaka M. Andei Kakamega Suam Acrapex sp 001 [6] 000 000 000 Busseola fusca 000 000 000 018 [15.17,24] Busseola phaia 000 000 135 [1,10,12,13,15,16,20]001 [14] Busseola si nov sp 3 000 000 000 005 [1,3,16] Busseola si nov sp 1 000 000 111 [21] 000 Caleris nov sp 3 000 000 058 [16,27,31,33] 000 Manga melanodonta 000 000 003 [13] 098 [13] Manga nubifera 089 [13] 000 000 000 Poecapa mediopuncta 000 000 008 [21] 000 Sciomesa nov sp 3 000 075[2] 000 000 Sciomesa nyei 000 000 010[15,16] 112 [13,17] Sciomesa piscator 000 003 [2,28] 022 [1,7,10,13,15,16,27] 028 [3,15,16,27,32] Sciomesa venata 000 000 000 026 [15,25,32] Sesamia calamistis 011 [6,13,24] 110 [5,8,9,11,13,16,23,24,26]010 [8,16] 003 [24,27] Sesamia peniseti 000 000 005 [16] 000 Sesamia nonagrioides 000 330 [2,5,23,26,28,34] 000 000 Sesamia nov sp 5 051 [6] 000 000 000 Sesamia nov sp 9 000 000 142 [29] 000 Sesamia poephaga 019 [4,12,13] 066 [13] 000 010 [13] Noctuid (unknown spp) 133 [10,11,16] 000 138 [21] 000 Chilo orichalcociliellus 724 [6,12,13,16,19,24]001 [13] 000 000 Chilo partellus 390 [12,13,19,24] 108 [8,16,24] 000 000 Chilo sp.nr orichalcociliellus 000 001 [11] 000 000 Crambidae (unknown) 000 000 001 [I] 004 [8] Eldana saccharina 000 002 [28] 005 [16,20] 000 Emautheudes spp 253 [6,12,19,20,24] 023 [19] 000 001 [22] Pyralidae 000 002 [24,28] 023 [6,24,29] 004 [22,27] Schoenobinus spp 004 [27] 013 [28,34] 000 000 Tortricidae sp. 011 [30] 007 [26,28] 006 [24] 003[32] Total Individuals 1551 627 683 342 Total Species (S) 11 13 12 13 Alpha (a) 1.43 2.77 2.50 3.00 Berger-Parker Dominance (d%) 46.68 33.81 20.80 32.75 In superscript are numbers for host plant codes from which different borer species were collected [refer to Table 3.3] 42 3.3.4 Stem borer species distribution and host range Stem borer species generally varied in distribution among the four localities, with species in the family Noctuidae occurring mainly in Kakamega and Suam, while the non- noctuids were found mainly in Muhaka and Mtito Andei (Table 3.4). Species in the Busseola genera (B. fusca, B. phaia ssp. phaia, Busseola sl nov sp 1 and Busseola sl nov sp 3) were found in both Kakamega and Suam. However, among the Busseola species, only B. phaia ssp. phaia was common to the two localities as B. fusca and Busseola sl nov sp 3 were found only in Suam, and Busseola sl nov sp 1 only in Kakamega. Species in the Sciomesa genus were found in three localities (Mtito Andei, Kakamega and Suam) and like the Busseola species, they varied in distribution among these localities. Sciomesa nov sp 3 was found only in Mtito Andei where it infested C. aethiopicus while S. venata was found only in Suam where it infested P. macrourum, C. articulatus and S. confusus. Among the Sciomesa species, only S. piscator was common to the three localities infesting a total of 11 plant species with seven of them occurring in Kakamega. The majority of the noctuid species identified belonged to the Sesamia genus. These species however showed a wide range of distribution, from the low altitude (Muhaka) to high altitude locality (Suam). Of these species, S. calamistis was found in all the four localities infesting a total of 11 plant species. The majority of infested plants (9) were found in Mtito Andei (9) followed by Suam (5). The second Sesamia species with a relatively wide distributional range was S. poephaga which was found in Muhaka, Mtito Andei and Suam. Unlike S. calamistis that was found on several plants species, S. poephaga was found on only 3 plant species with P. maximum being the most important host in all the three localities. The other two host plants included D. milinjiana and P. infestum, which were only infested in Muhaka. Other Sesamia species appeared localized K YATTA UNIV A 43 in their distribution using varied hosts in respective localities. Both S. peniseti and Sesamia nov sp 9 were found only in Kakamega infesting P. purpureum and C. papyrus respectively. Sesamia nonagrioides on the other hand was found only in Mtito Andei where it infested six different plant species though majority of the larvae were found on T domingensis, while Sesamia nov sp 5 was found only in Muhaka where it infested E. haploclada. The non-noctuids species belonged to Crambidae, Pyralidae, Schoenobinae and Torticidae families with species in the Chilo genera forming bulk of the materials. The Chilo species, C. partellus, C. orichalcociliellus and Chilo sp. nr orichalcociliellus were found only in Muhaka and Mtito Andei. These species infested a wide range of plants though the majority of C. partellus and C. orichalcociliellus were found on P. maximum and S. arundinaceum respectively. The pyralid, E. saccharina was found in both Mtito Andei where it infested C. involucratus, and Kakamega where it infested P. purpureum and S. officinarum. The only non-noctuid species found in four localities was the unknown species in the Tortricidae family. Though it occurred m low numbers, this species was found on C. rotundus in Muhaka, C. distans and C. involucratus in Mtito Andei, S. arundinaceum in Kakamega and S. confusus in Suam. 3.3.5 Stem borer faunistic similarities among wild host plants Cross-wise comparison of faunistic similarity (expressed as the second Kulczynski coefficient) yielded a matrix reflecting variation in distribution of species (Table 3.5). For instance when Mtito Andei was compared with Muhaka, a Kulczynski coefficient (KCMtitoAndeilMuhaka)of 60.0 was found, similarly the KCMtitoAndei/Suamand KCSuam/Kakamegar mained high (48.7 and 42.0 respectively) indicating that these localities 44 share considerably high number of species. However, the results of KCKakamegaIMuhaka = 17.7, KCSuamIMuhaka = 38.2 and KCKakamegaIMtito Andei = 29.7 suggested that there was considerable variation in species composition among these pairs. Table 3.5: Kulczynski coefficient (KC) as calculated between the different localities surveyed in Kenya. KC Muhaka (11) Mtito Andei (13) Kakamega (12) Suam (13) Muhaka Mtito Andei Suam 60.0 17.7 38.2 29.7 48.7 42.0 Kakamega The numbers between brackets in the leading row represent species counts in different localities. 45 3.4 Discussion This study reveals variation in species diversity and distribution among different localities. Despite variations in species diversity, B. fusca, S. calamistis, C. partellus and C. orichalcociliellus were the main pest species, corroborating earlier reports by Seshu Reddy (1983), Overholt et al. (2001) and Zhou et al. (2002). Busseolafusca dominated high altitude locality, Kakamega (planetary Guineo-Congolian rain forest) and Suam (Afromontane vegetation mosaic), while C. partellus dominated low altitude localities, Muhaka (Zanzibar Inhambane vegetation type). Though this study focused on species composition in different localities, the observed distribution pattern corroborates findings of previous studies conducted along different agro-ecological zones in Kenya (Overholt et al., 2001; Zhou et al., 2002). In addition to the four widely reported pest species III Kenya (B. fusca, S. calamistis, C. partellus and C. orichalcociliellus), other stem borer species, B. phaia ssp. phaia and S. piscator were found in cultivated crops in Kakamega, along the Guineo- Congolian rain forest. For a long time, these species have been associated with only wild plants in East Africa (Nye, 1960). This was however contradicted by results of the recent surveys in Kenya in which Le Ru et al. (2006b) found these species among cultivated crops. They indicated that the presence of these species in the cultivated fields could be as a result of accidental oviposition or the gradual shift in response to habitat modification. Continued presence of these species among the cultivated crops suggests the emergence of "new" stem borer pests in Kenya. This however is not the first time host shift among stem borers is reported in Africa as similar shifts have been reported of E. saccharina from sedges to sugarcane in South Africa (Atkinson, 1980). ITA UNIVERS LtBRAR 46 In addition to host range expansion and eventual emergence of "new" pest species, habitat transformation may affect the general species richness and composition. There have thus been increasing interest to understand stem borer species diversity among wild habitats with an aim of averting potential consequence of habitat transformation (Le Ru et al., 2006a; Matama-Kauma et al., 2008). In this study, 29 stem borer species were collected. Out of the 29 species collected, 15 of them are known, while the rest are unknown and could only be identified to either genus or family level. The identified species varied in distribution among the surveyed localities with the highest diversity recorded in Kakamega (wet and hot guineo-congolian mosaic). This corroborates the findings of a study conducted by Le Ru et al. (2006b) in which they found higher diversity of noctuids in wet and hot guineo-congolian mosaic in western Kenya. In Suam, Mtito Andei and Muhaka, stem borers were found mainly among the hosts growing along the riverines and swamps. Similar distribution pattern was observed by Nye (1960) when he found stem borer larvae mainly from wetter parts of different vegetation mosaics. The stem borer pests, B. fusca, S. calamistis, C. partellus, C. oricholcociliellus and E. saccharina were among the known species collected though their populations were lower compared to the collections made in the cultivated fields. This however is not the first time such a low pest population has been reported in the uncultivated fragments within agro-ecosystems. Similar results were observed by Le Ru et al. (2006a) during their surveys in Kenya and Eastern Africa respectively, and Gounou & Schulthess, (2004) in western Africa. Laboratory studies have shown low larval survival and development on several wild grasses (Shanower et al., 1993; Ofamata et al., 2000; van den Berg et al., 2001), which may be attributed to high silica contents in the epidermis of leaves 47 (McNaughton et al., 1985; Setamou et al., 1993). The other species collected; S. penniseti, S. poephaga, S. nonagrioides, S. venata, S. piscator, S. nyei, P. mediopuncta, M nubifera, M melanodonta and B. phaia ssp. phaia, varied in distribution among the surveyed localities. Each locality harbored dominant borer species accounting at least for 20% of the specimens collected. Populations of rare or unknown species in these localities were generally low, though they exerted a strong influence in the overall species assemblage. However, their role and importance in structuring broad community patterns among regions is not well understood (Price et al., 1995; Novotny & Basset, 2000). Some of the rare or unknown species, Sesamia nov sp 9, Sciomesa nov sp 3 and P. mediopuncta, appeared localised in their distribution, results that could be attributed to seasonality and ecological preference. Species like Sciomesa nov sp 3 and P. mediopuncta were earlier reported from C. aethiopicus and S. megaphylla in Mtito Andei and Kakamega respectively (Le Ru et al., 2006b). The presence of unknown and some rare species confirms earlier assertions that stem borer species and host list in East Africa is far from complete (Le Ru et al., 2006a). Stem borer species generally vary in their annual flying periods (Holloway, 1998) and thus the single or few sampling sessions that characterised previous studies may not have captured the seasonal species. The presence of unknown and some rare species in this collection can therefore be attributed to regular sampling intervals that allowed recovery of seasonal and conspicuous stem borer species. Majority of the rare and unknown stem borer species were found on isolated host batches growing along wetlands and riverines. However, these habitats are targeted for agricultural expansions as they constitute suitable areas for both irrigation and horticultural farming. Transformation of wetlands as well as other habitats poses a major KEN "ATIA U V' Y LI A ) 48 threat to stem borer species diversity. Of immediate concern to agriculture is the resultant effect of these transformations to stem borer pest dynamics. Will these have direct or indirect effect on pest populations in the cultivated fields? Will some non-economic species expand their host range and adapt to cultivated crops? 49 CHAPTER FOUR 4 DYNAMICS AND MANAGEMENT OF STEM BORERS IN KENYA 4.1 Introduction Busseola fusca, Sesamia calamistis, Chilo orichalcociliellus and Chilo partellus are the most important stem borer pests of maize and sorghum in Kenya. The distribution and economic importance of these species vary among different regions across the country depending on their respective ecological requirements (Le Ru et al., 2006a). Busseola fusca and C. partellus are found mainly in high and low altitude areas respectively with S. calamistis occurring in all altitudinal gradients as a minor pest species. Based on the knowledge of their distribution, several control techniques have been developed. Among the tried management strategies are late planting, application of insecticides, planting of border rows with grasses serving as trap plants and/or refugia for pests and natural enemies (Khan et al., 1997) and biocontrol using Cotesia flavipes against C. partellus in low altitude areas. Despite these management initiatives, stem borer pests persist in the cultivated fields where they account for up to 14% cereal losses annually (Songa et al., 1998). There is therefore need to introduce other strategies to augment the existing stem borer management practices to further reduce losses associated with their infestations. One of such strategies is the proposed introduction of transgenic maize expressing insecticidal proteins from the bacterium Bacillus thuringiensis (Bt). Transgenic plants were first commercialised in 1996 amid concern from some scientists, regulators and environmentalists that the widespread use of Bt crops would inevitably lead to resistance and the loss of a public good, especially, the susceptibility of insect pests to Bt proteins (Gould et al., 2002). However, proponents of transgenic crops argued that the refuge 50 approach in the agro-ecosystems would adequately contain rapid evolution of resistance. The theory underlying the refuge strategy for delaying insect resistance is that most of the rare resistant individuals surviving on Bt crops will mate with abundant susceptible individuals from refuges of host plants without Bt toxins (Bourguet et al., 2000a). If inheritance of resistance is recessive, the hybrid offspring produced by such matings will be killed by Bt crops, markedly slowing the evolution of resistance. Stem borers, the target pests are thought to be polyphagous and their persistence in crop fields is blamed on the influx of diaspore population from wild hosts growing in the uncultivated fragments (Polaszek & Khan, 1998). Based on this assumption, wild habitats fit well in the management of resistance as a refuge should the proposed transgenic maize get into hands of small scale farmers. However, this assumption has been contradicted by the results of recent surveys in Kenya and Eastern Africa in which stem borer pests were found to be more specialized infesting limited host plant species (Le Ru et al., 2006a). Busseola fusca which is an important pest in high altitude areas was only reported from S. arundinaceum while C. partellus was reported mainly from S. arundinaceum and P. maximum in low altitude areas. With this revelation, it appears that there is no generally agreed source of stem borer pests found in the cultivated fields as these may vary among pest species and regions. Does stem borer pest community vary in agricultural systems within a season? Could these variations be attributed to the influx of the diaspore populations from the wild? How does this fit in the context of the proposed introduction of transgenic maize? Thus, the study was initiated to identify stem borer pests in the cultivated habitats and monitor seasonal variation in population dynamics in an attempt to establish the role wild host plants play in population dynamics and relate this finding to the proposed introduction oftransgenic maize in Kenya. 51 4.2 Materials and methods This study was carried out in the four localities, Muhaka, Mtito Andei, Kakamega and Suam, occurring in different agro-ecological zones. Description of the study areas, sampling of stem borers in both wild and cultivated fields, rearing and identification of stem borers are presented in Chapter 3. 4.2.1 Data management and analysis The average infestation was estimated in each field from the number of plants infested against the number of plants checked for infestation. The total number of stem borer larvea collected in each field (both wild and cultivated fields) was divided by corresponding number of plants checked for infestation to estimate larval densities. The effects of growing seasons on the general stem borer infestations (in the cultivated fields) and species densities (in both wild and cultivated fields) were analysed for different localities. Infestation data (%) was arcsine transformed and subjected to one way analysis of variance (ANOVA) to compare the general variation stem borer pest infestations (%) in different localities. Means were later separated using Student-Newman-Keuls (SNK) multiple range test (SAS, 1997). Average infestations (%) per season for different localities were compared separately for each season using Students' t test. The average proportions of different stem borer species in the pest community (%) and their respective densities in different growing seasons were compared for each locality using One-way ANOV A. The percentage data was arcsine transformed while the density data was log transformed (1 + 10glO x). Means were separated using Student- Newman-Keuls (SNK) multiple range test (SAS, 1997). Seasonal variation in species 52 densities for both wild and cultivated fields were compared separately for each locality using Students' t test. 4.3 Results 4.3.1 Stem borer pest infestations and the associated seasonal variations There was evidence of variation in the mean annual stem borer pest infestations among the surveyed localities (P < 0.05; Table 4.1). The highest infestation was recorded in Muhaka and Mtito Andei while the lowest levels were recorded in Kakamega and Suam. Variations were also observed within seasons in respective localities (P < 0.05). The highest infestation in the LR growing season was observed in Muhaka (19.16 ± 2.89 %) followed by Mtito Andei (12.00 ± 5.51 %) while low infestations were recorded in Kakamega and Suam (4.07 ± 0.67 and 9.71 ± 1.09 % respectively). Similar trends were observed in the SR growing season during which the highest infestation was observed in Muhaka (35.57 ± 3.81%) followed by Mtito Andei (30.67 ± 7.54%). Despite the differences in infestation levels among the localities, there was no evidence of variation in infestation levels between the growing seasons except in Muhaka where there was significantly higher infestation during SR growing season (t83 ::;::3.26; P = 0.002). 'KENVATTA UNIVE SITV Lt A 53 Table 4.1: Stem borer annual and seasonal infestations (%) in different cultivated fields Mean ± SE Seasonal average infestation (Mean ± SE) % Locality (%) Long rain Short rain t DJ P Muhaka 28.43 ± 2.63" 19.16 ± 2.89' 35.57 ± 3.81" 3.263 83 0.002* Mtito Andei 26.00 ± 6.04" 12.0Q±5.51,h 30.67 ± 7.54' 1.369 18 0.188 Kakamega 05.43 ± 0.63h 04.07 ± 0.67h 06.17 ± 0.89h 1.606 121 0.111 Suam 09.71 ± 1.09h 09.71 ± 1.09h F 39.54 13.56 40.02 ell.~.•....~ DJ 3,376 3,133 2, 140.•...Cl:!.•...en p < 0.0001 < 0.0001 < 0.0001 Means (± SE) within columns followed by the same letters are not significantly different (Student-Newman-Keuls multiple range tests, P ~ 0.05). 54 4.3.2 Stem borer species composition and seasonal density fluctuations Stem borer pest species varied in distribution among the surveyed localities (Table 4.2). See chapter 2 for an overview of stem borer pest composition and distribution in the cultivated fields. In addition to the variation in distribution and composition of pest communities, there was evidence of variation in densities of different species within seasons (P < 0.05). The highest density was recorded during the LR growing season in Muhaka (0.64 ± 0.08) followed by Mtito Andei (0.25 ± 0.07). Chilo partellus dominated the larval population in both localities where its density was recorded as 0.46 ± 0.06 and 0.18 ± 0.07 per plant respectively. The other species with relatively higher density was B. fusca though this was only observed in Suam where it co- existed with S. calamistis. Variations in larval densities were also observed during the SR growing season with relatively higher densities in Mtito Andei (0.35 ± 0.07) and Muhaka (0.34 ± 0.05). Student's t test on species numbers in different localities revealed significant variations in numbers among some species between LR and SR growing seasons (P < 0.05; Table 4.2). In Muhaka, significant variations in larval densities were observed among S. calamistis and C. partellus with higher densities recorded during the LR for both species. Variation in species numbers was also observed in Kakamega among the populations of B. fusca and B. phaia. High B. fusca density was recorded during the LR growing season while B. phaia density was higher during the SR growing season. The other species S. calamistis S. piscator, E. saccharina, found in Kakamega generally had low densities with no evidence of variation between the growing seasons (P > 0.05). Similar results were observed in Mtito Andei where no seasonal variation in densities was recorded among the two pest species, S. calamistis and C. partellus. 55 Table 4.2: Stem borer pest composition (%) and seasonal average densities (mean number per stem ± SE) in the different cultivated fields and localities Stem borer pest species % Composition Seasonal average density (Mean ± SE) (Mean ± SE) Long rains Short rains t P Muhaka Sesamia calamistis 18.1±3.3b 0.1l6 ± 0.027b 0.045 ± 0.012b 2.28 0.025* Chilo partellus 67.0 ± 5.8- 0.461 ± 0.063- 0.131 ± 0.023- 4.68 0.000* Chilo orichalcociliellus 19.8 ± 2.8b 0.082 ± 0.020b 0.141 ± 0.033" 1.55 0.123 F 43.76 34.61 5.41 Df 2,234 2,170 2,149 P < 0.0001 < 0.0001 0.005 Mtito Andei Sesamia calamistis 25.6 ± 6.6b 0.073 ± 0.024- 0.046 ± 0.022b 0.80 0.431 Chilo partellus 74.4 ± 6.6- 0.177 ± 0.066- 0.306 ± 0.073- 1.19 0.243 F 27.04 2.32 16.26 Df 1,51 1,21 1,37 P < 0.0001 0.143 0.0003 Kakamega Busseola fusca 41.0 ± 5.2' 0.038 ± 0.008- 0.015 ± 0.004b 2.13 0.036* Busseola phaia 26.9 ± 4.5'b 0.015 ± 0.004b 0.043 ± 0.010' 2.93 0.004* Sesamia calamistis 26.4 ± 4.6-b 0.018 ± 0.006b 0.035 ± 0.010' 1.49 0.139 Sciomesa piscator 07.0 ± 5.4b 0.003 ± 0.002b 0.001 ± 0.001 b 0.62 0.535 Eldana saccharina 11.0 ± 5.9b 0.001 ± 0.001 b 0.002 ± 0.001 b 0.49 0.624 F 4.76 8.86 8.57 Df 4,245 4,279 4, 164 P 0.001 < 0.0001 < 0.0001 Suam Busseola fusca 99.8 ± 0.3- 0.206 ± 0.023' Sesamia calamistis 00.5 ± 0.5b 0.004 ± 0.002b F 2.21 92.34 Df 1,43 1, 119 P < 0.0001 < 0.0001 Means (± SE) within columns followed by the same lowercase letters are not significantly different (Student-Newman-Keuls multiple range tests, P :::0.05) 56 4.3.3 Seasonal variation in densities of stem borers among wild plants Seven stem borer pest species, B. fusca, S. calamistis, B. phaia, S. piscator, E. saccharina, C. partellus and C. orichalcociliellus, identified from the cultivated fields were found among wild plants. These species, however, varied in distribution and densities among the localities (Table 4.3). Four of these pest species, S. calamistis, B. phaia, S. piscator and E. saccharina were identified in Kakamega, and S. calamistis, C. partellus and C. orichalcociliellus, in Muhaka. Only two pest species were found in both Mtito Andei (S. calamistis and C. partellus) and Suam (B. fusca and S. calamistis). Busseola fusca was only found among wild plants in Suam where it infested S. arundinaceum. The importance of these species varied within localities with some species showing variations in average densities between LR and SR growing seasons. Despite the changes in density levels, none of the species had statistically significant difference between the growing seasons as revealed by t test (P> 0.05; Table 4.3). Busseola phaia ssp. phaia had the highest density level among the species in Kakamega during both LR and SR growing seasons followed by E. saccharina and S. piscator. The average density of 13.phaia increased from 0.03 to 0.05 between LR and SR growing seasons contrary to both E. saccharina and S. piscator that showed a general reduction in density between LR and SR growing seasons. Chilo orichalcociliellus was the most important pest species followed by C. partellus in Muhaka. During the LR growing season, C. orichalcociliellus' density averaged 0.04 and increased to 0.06 larvae per plant during the SR growing season. In contrast, the density of C. partellus decreased slightly from 0.021 to 0.018 larvae per plant between LR and SR growing seasons. 'KENYATTA UNIVERSITY LJ ~ A 57 Table 4.3: Seasonal pest densities (number per stem ± SE) in the wild habitats in different localities . Stem borer pest species Seasonal average density (Mean ± SE),Long rain Short rain t P Muhaka Sesamia calamistis 0.008 ± 0.005b 0.001 ± 0.001 b 1.471 0.1443 Chilo partellus 0.021 ± 0.008,b 0.018 ± 0.007b 0.275 0.7834 Chilo orichalcociliellus 0.040 ± 0.010' 0.064 ± 0.011' 1.627 0.1058 F 3.82 17.67 Dj 2,203 2,183 P 0.024 < 0.0001 Mtito Andei Sesamia calamistis 0.003 ± 0.001' 0.034 ± 0.018' 1.158 0.2543 Chilo partellus 0.023 ± O.Ol3' 0.031 ± 0.015' 0.343 0.7341 F 3.17 0.01 Dj 1,21 1,47 P 0.09 0.94 Kakamega Busseola jusca 0.000 ± O.OOOb 0.000 ± O.OOOb Busseola phaia 0.031 ± 0.006' 0.045 ± O.OlQ" 1.32 0.191 Sesamia calamistis 0.001 ± 0.001 b 0.001 ± 0.001 b 0 1.000 Sciomesa piscator 0.014 ± 0.004,b 0.005 ± 0.004b 1.158 0.251 Eldana saccharina 0.026 ± 0.029' 0.003 ± 0.004b 0.572 0.583 F 9.86 12.79 Dj 4,196 4,90 P < 0.0001 < 0.0001 Suam Busseola jusca 0.014 ± 0.014' Sesamia calamistis 0.006 ± 0.002' F 0.94 Dj 1,22 P 0.344 Means (± SE) within columns followed by the same letters are not significantly different (Student-Newman-Keuls multiple range tests, P :s 0.05). :99%) ready for DNA extraction. ENYATT U VE LI RA 68 3 " L Turkana + .••• Otherlocalitieso Maincites --Rivers Kilometerso 100 200 I I Figure 5.1: Map of Kenya showing the four surveyed sites and the estimated spatial distribution of Sesamia calamistis, Clades I and Il. Estimates on spatial distribution of the two clades were based on findings ofthe current study and results of the ongoing phylogeography study in Africa. YA RS~ V LI RARY 69 (a) Plate 5.1: Pictorial comparison of different larvae belonging to Sesamia and Sciomesa genera and dismembered genetalia of S. calamistis used during identification; (a) Male genitalia (i) characteristic centrally placed flask-shaped juxta (ii) aedeagus (b) female genitalia. 5.2.2 DNA extraction and sequence analysis The total genomic DNA was extracted from the thoracic muscles usmg commercial kit (DNeaslM Tissue Kit, Qiagen GmbH, Germany) protocol with Proteinase K digestion as recommended for animal tissues. The extracted DNA was stored at -20°C until required for amplification. Voucher specimens are housed at the International Centre of Insect Physiology and Ecology (icipe) Biosystematic unit, Kenya. Polymerase chain reaction (PCR) was used to amplify the 873 bp Cyt. b mitochondrial 70 fragment using the primers CPl (5'-GATGATGAAATTTTGGATC-3') (modified from Harry et al., 1998) and Tser (5'-TATTTCTTTATTATGTTTTCAAAAC-3') (Simon et al., 1994). The PCR was performed on a Biometra GeneAmp PCR System in a 25 III reaction mixture containing 1 III of the genomic DNA, 5X Green GoTaq® Flexi Buffer, 0.24 mM dNTPs, 3 mM MgCh, 0.4 IlM of each primer and 1 unit of Taq polymerase (GoTaq, Promega). After initial denaturation at 94°C for 5 min, PCR condition was 40 cycles of 94°C for 1 min of denaturation, 46°C for 1 min 30 s of annealing, nOc for 1 min 30 s of extension and a final extension period of 10 min at 72°C. The PCR products were visualised by means of electrophoresis in 1% agarose gel previously stained with ethidium bromide to verify amplification. Amplified products were purified with the Promega Wizard SV Gel and PCR Clean up System following the manufacturer's protocol. DNA sequencing reactions were performed using the ABI PRISM® Bigfrye" Terminator v3.0 Ready Reaction Cycle Sequencing Kit (Applied Biosystems), cleaned using ethanol/EDTA precipitation. Sequences were visualized on an ABI 3130 automated sequencer using Big-Dye fluorescent terminators. The consensus sequences obtained were aligned manually using Mac Clade 4.05 (Maddison and Maddison, 2001). Additional sequences of individuals from maize were obtained from P. Moyal of Laboratoire Evolution, Genomes et speciation - France, to enable comparison with collection made in other localities in Kenya and some African countries (South Africa, Uganda, Benin, Ghana, Nigeria and Togo) (P. Moyal, Unpublished data). All the sequences were deposited in the Genebank (Accession numbers EU305065-EU305228) '"ENYATTA U 71 5.2.3 Evaluation of reproductive parameters Fourth and fifth larval instars of S. calamistis were collected from sorghum plants in Kisumu (western part of Kenya) and from plant species belong to Cyperaceae family in Shimba Hills (eastern part of Kenya, Coastal region). Both populations were reared separately on artificial diet as described by Onyango and Ochieng' -Odero (1994) until pupae formation. Some individuals from each reared population were randomly sequenced for Cyt. b to establish their genetic status in relation with the genetic population analysis. Male and female pupae were kept in separate plastic boxes (30 cm long, 12 cm wide, 10 cm high) containing a moist cotton pad to maintain a relative humidity (r.h) at about 80 %, and monitored for adult emergence. Pupae and adults were maintained in a controlled chamber at 25.3 ± 0.9 "C, 68.6 ± 12.8% r.h (means ± SE) under reversed L12:D12 photoperiod with scotophase lasting from 7.00 to 19.00 h, hereafter referred to as night. The reversed photoperiod allowed all experiments to be carried out during the day. In each population, one-day old males and females (minimum 10 individuals of each sex per night) were released in a mosquito-net cage (40 x 40 x 63 cm) at the onset of scotophase. They were provided with a diluted honey solution impregnated in a piece of cotton. Mating behaviour was observed and recorded after every 30 min during the dark period. Mated pairs were transferred from the cage to a transparent plastic jar (16 cm high, 9 cm diameter), to facilitate measurement of copulation duration. The plastic jar contained a wet piece of cotton wool that maintained relative humidity at around 80%. One cylindrical surrogate stem made from a rectangular piece of nylon cloth (15 long, 5cm wide) rolled helicoidally from top to bottom was placed in each jar. This support had earlier been found to elicit good ovipositional response in S. calamistis (P-A. Calatayud, E 'l A U ~I 72 personal observation). The total number of eggs laid by each female was counted each night, renewing the surrogate stem every night throughout the female life. The life duration of each female was finally recorded and similarly, the time of emergence of the first neonate after egg laying for each female was recorded. 5.2.4 Statistical analysis Basic sequence statistics were calculated using DnaSP (Rozas et al., 2003). The following parameters were used to estimate genetic variability among populations between sites (Muhaka, Mtito Andei, Kakamega and Suam) and between host plants in Mtito Andei (wild and cultivated hosts): number of haplotypes (h), number of polymorphic sites (S), haplotype diversity (d) (Nei, 1987), nucleotide diversity (Pi) (Lynch and Crease, 1990) using the Jukes and Cantor correction (Jukes and Cantor, 1969), mean number of nucleotide differences (K) (Tajima, 1983). The extent of genetic differentiation between the populations (FsT) (Hudson et al., 1992) was performed with the Arlequin 2.000 software (Schneider et al., 2000). A maximum parsimony network was drawn using TCS 1.21 software (Clement et al., 2000). For reproductive parameters, means were computed and separated by Mann- Whitney U-test (rank analysis for a two- sample test). E RA 73 5.3 Results 5.3.1 Diet breadth of Sesamia calamistis Sesamia calamistis larvae were found on twelve different plant species (Table 5.1). For the purpose of this study, all plant species from which S. calamistis larvae were recovered have been considered as host plants without quantifying their relative contribution to S. calamistis population dynamics. The host list (see Table 1) contains both known and unknown hosts of S. calamistis as given by Khan et al. (1997), Gounou & Schulthess (2004) and Le Ru et al. (2006b) with their importance varying among surveyed geographic sites. This species had a limited number of host plants in Suam and Kakamega. In addition to the cultivated cereals (maize, sorghum and finger millet), its larvae were found on Cyperus dives Delile and Sorghum arundinaceum (Desvaux) Stapf in Suam and from Pennisetum purpureum Schumacher in Kakamega. Together with the cultivated cereals, the larvae were recovered from seven more host plants in Mtito Andei (Cyperus distans L., Eleusine jaegeri Pilg., Panicum deustum Thunb, Panicum maximum Jacquin, P. purpureum, Setaria verticillata (L.) P. Beauv. and S. arundinaceum) and three in Muhaka (Echinochloa haploclada (Stapt) Stapf, P. maximum and S. arundinaceum). 74 Table 5.1: List of plant species from which Sesamia calamistis larvae were collected in different No. of moths processed Infested plants species Muhaka M. Andei Kakamega Suam Cyperus distans L. 2 Cyperus dives Delile 2 Echinochloa haploclada (Stapf) Stapf, Eleusine corocana L. 6 2 4 Eleusine jaegeri Pilg. * 2 Panicum deustum Th unb Panicum maximum Jacquin 2 Pennisetum purpureum Schumacher 3 Setaria verticillata CL.)P. Beauv. * Sorghum arundinaceum (Desvaux) Stapf 3 11 Sorghum bicolor (L.) Moench 14 4 2 Zea mays L. 24 47 31 Marked with stars (*) are "new plants" (previously not recorded as host plants) Y LIBRAR 75 5.3.2 Differentiation of S. calamistis populations in Kenya Qualitative TCS maximum parsimony network of 194 sequences revealed 68 haplotypes. These haplotypes separated into two c1ades with an average divergence of 1.89 ± 0.24 % (Fig. 5.2). Clade 1 containing 33 haplotypes showed an average divergence of 0.36 ± 0.19 % while Clade II containing 35 haplotypes showed an average divergence of 0.44 ± 0.20 %. Except for three specimens (EU305074, EU305088 and EU305112), all other individuals from Muhaka and a part of collection from Mtito Andei grouped in Clade I together with individuals from South Africa, while individuals from Suam, Kakamega and the remaining part of the collection from Mtito Andei grouped together in Clade II with individuals from Uganda, Benin, Togo, Ghana and Nigeria. There was evidence of variation in the spatial distribution of the two clades in Kenya. Clade I was mainly found in Muhaka and Mtito Andei while Clade II was found in Mtito Andei, Kakamega and Suam. However, partial distribution of individuals from Mtito Andei into Clade I and II suggested greater genetic variability in that area. This was further reflected in genetic diversity parameters (S, h, d, Pi and K) that revealed higher variability in Mtito Andei relative to Muhaka, Kakamega and Suam (Table 5.2a). E VAIT UNIVERS TV LI RAR 76 Table 5.2a: Genetic diversity of the Cytochrome b gene in Sesamia calamistis populations from four localities in Kenya (Diversity values ± SD) Genetic diversity parameters Sampled sites Muhaka M. Andei Kakamega Suam Number of sequences, n 43 78 35 8 Number of segregating sites, S 30 35 23 11 Number of haplotypes, h 15 24 12 5 Haplotype diversity, d 0.792 ± 0.060 0.834 ± 0.034 0.677 ± 0.082 0.786±0.151 Nucleotide diversity, Pi 0.004 ± 0.001 0.009 ± 0.001 0.003 ± 0.001 0.004 ± 0.001 Mean number of nucleotide 3.508 ± l.822 7.740 ± 3.644 2.927 ± 1.563 3.214 ± l.852differences, K YIT Yl RAR 77 k:;:z~~;jBenin _ Ghanac:::::J Kenya ~Nigeria ~ South Africa I:mmmml Togo== Uganda Figure 5.2: TCS mitochondrial haplotype network of 195 Sesamia calamistis specimens processed from seven African countries. The area of the circle is proportional to the number of samples sharing each haplotype. Lines represent single nucleotide mutations and shaded circles represent haplotypes, which are not observed in the sample while different shading patterns represent different countries. 78 5.3.3 Genetic differentiation in host utilization (Mtito Andei) Sesamia calamistis was found on a wide range of cultivated and wild host plants in Mtito Andei. Values obtained from genetic diversity parameters (d, Pi and K; Table 5.2b) indicate that individuals from the wild plants are more genetically variable compared to individuals from the cultivated fields. However, there were more haplotypes among the cultivated hosts (h, 16) compared to the wild host plants (h, 13). Excerpt of haplotypes in Mtito Andei from the global network (Fig. 5.2) indicated that individuals in respective clades varied in terms of host plant preference suggesting differentiation with respect to host plants (FST = 0.4008; P < 0.001). The first group with 14 haplotypes corresponded to Clade I, while the second group with 10 haplotypes corresponded to Clade II (Fig. 5.3). The majority of individuals in Clade I (87.5%) came from the cultivated host plants (Z mays, 78.6%; E. corocana, 5.4%; S. bicolor, 3.5%), while Clade II was dominated by individuals mainly from wild host plants (63.6%). However, haplotypes of individuals from S. arundinaceum appeared in both Clade I and II accounting for 7 and 32% respectively. Table 5.2b: Genetic diversity of the Cytochrome b gene in Sesamia calamistis populations from wild and cultivated hosts in Mtito Andei (Diversityvalues ± SD). Genetic diversity parameters Mtito Andei Cultivated host Wild hosts Number of sequences, n Number of segregating sites, S Number of haplotypes, h Haplotype diversity, d Nucleotide diversity Pi Mean number of nucleotide differences,K 57 21 30 25 16 13 0.723 ± 0.076 0.891 ± 0.049 0.006 ± 0.001 0.010 ± 0.001 5.312 ± 2.603 8.381 ± 4.073 E v: A U IV LI RA- 79 Host plant species zeamaysE::::I Petmjsetum pUTpUreum W}j EfeuSi/JaCOJtX;ana ~ ~tghum bic~lor I!fHil-~;1SoTghum arundmaceum ~ Ran/cumdeustumI.~.-j Eleusine jaegeri E::J RaniCum mo>rimumIm Cyperus distafls ~ Selari~IIfJtticill(1/J;l (0ooo()o Figure 5.3: TCS mitochondrial haplotype network of Sesamia calamistis individuals collected from different host plants in Mtito Andei. The area of each circle is proportional to the number of samples in each haplotype. Lines represent single nucleotide mutations and small white circles represent haplotypes that are not observed in the sample. Different shading patterns represent the different sampled host plants. ENVATTA 1I IVF 80 5.3.4 Reproductive and life trait parameters Clade I and II populations appeared to vary in their mating times (hour) after the onset of scotophase (Table 5.3; Mann-Whitney's U-test, P < 0.0001). The mean mating times of Clade I and 11 populations were 5:30 and 7:12 respectively. In addition to variations in mating time, there were also differences in the duration of mating (Mann- Whitney's U-test, P = 0.04) with Clade I taking relatively longer time (129.6 min) compared to Clade II (99.6 min). Despite variations in both mating time and duration, all females oviposited only during the scotophase and no eggs were deposited during the photophase. In addition, there were no differences among the two populations in terms of the total number of eggs laid by each female (Mann- Whitney's U-test P = 0.79) and the time of first ec1osion (Mann- Whitney's U-test P = 0.64). Each female laid about 680 eggs during their life and the laid eggs took about 8 days before ec1osion. On average however, the females varied in their life durations (Mann- Whitney's U-test P = 0.009), with Clade II living slightly longer (7.7 days) compared to Clade I (6.9 days). VA Table 5.3: Reproductive parameters of the Clade I and II populations of Sesamia calamistis as recorded under laboratory conditions Biological Hour of mating (h after Mating duration Total number of eggs Female life Time of eclosion after parameters the onset of night) (min) laid per female duration (days) egg laying (days) (/) Clade I 5.4 ± 0.2 (37) b 129.6 ± 8.5 (16) a 685.0 ± 47.8 (21) 6.9 ± 0.3 (23) b 8.1 ± 0.2 (23)c00.g :;0.. Clade 110 7.1 ±O.I (32) a 99.6 ± 13.8 (12) b 679.5 ± 39.2 (17) 7.7 ± 0.2 (19) a 8.1 ± 0.1 (19)0.. U (values) 150.0 51.5 169.5 87.5 182.0 (/).~.•.... (/).~.•.... P (values) <0.0001 0.0376 0.6382(/) 0.7916 0.0085 Statistics: Mann Whitney U test (rank analysis for two-sample test). Means ± SE (number of replicates) 82 5.4 Discussion This study shows that populations of S. calamistis in Kenya are divided into two clades (Clade I and II); Clade I dominant in South East, Clade II in the South West and that the two clades co-exist in Central Kenya. The genetic distance between both clades (about 1.8%) suggests an ancient fragmentation. According to an arthropod 2.3xlO-8 mitochondrial substitution/site/year rate (Brower, 1994), fragmentation may have occurred about one million years ago. Individuals collected from other localities in Kenya confirmed the observed geographic differentiation. This was further supported by the results of collection made in other African countries (Uganda, Ghana, Benin, Togo, Nigeria and South Africa) which showed a similar pattern with Clade I found only in South Africa while clade II was found only in countries on the west of Kenya (Uganda, Nigeria, Benin, Togo and Ghana). From a more general African point of view, these clades can be classified as East clade (Clade I) and West clade (Clade II), both of which meet in central Kenya. Currently, S. calamistis appears to favor maize and sorghum as preferred hosts though it appears to have retained the close association with the original host plants. Irrespective of the dominant clade, S. calamistis larvae were found in both cultivated and wild host plants in all localities (Muhaka, Mtito Andei, Kakamega and Suam). Preference for maize and cultivated sorghum as suitable hosts is reflected in the relatively higher densities observed in the cultivated fields, results that confirm earlier observations (Le Ru et al., 2006b). Shanower et al. (1993) noted the same phenomenon under experimental conditions and attributed low numbers of larvae among wild host plants to the poor nutritive value of the latter. Evolution of host and/or oviposition choice by S. calamistis 83 like other phytophagous insects may have included allelochemicals, quantity and/or quality of plant resources (Jaenike, 1990). Interactions between these factors may favour oviposition dispersion and egg laying on sub-optimal host plants. This may ultimately lead to variation in performance and survival of progeny for eggs deposited on different hosts as observed by Gassmann et al. (2006) during their study on adaptation of Ophraella notulata (Fabricius) to feed on Ambrosia artemisiifolia (Heliantheae: Ambrosiinae ). Unlike in other localities that were dominated by individuals from either Clade I or II (Muhaka - Clade I; Kakamega and Suam - Clade II), there was evidence of genetic differentiation with respect to host plant use in Mtito Andei where both clades co-existed. These clades differentiated in Mtito Andei with almost all individuals belonging to Clade I found on maize, while larger proportion of individuals in Clade II were found on wild plants. However, some individuals belonging to Clade I were found on wild plants and this explained the high genetic variability observed in this c1ade. Differentiation of these populations with respect to host plants is not specific to S. calamistis alone since similar results have been recorded on another noctuid species, the fall armyworm Spodoptera frugiperda (1. E. Smith). Two populations of S. frugiperda inhabited the same geographical area and showed such patterns of differentiation including host races (Prowell et al., 2004). The observed separation among S. calamistis population in host use in Mtito Andei could be attributed to either low attraction of maize to Clade II or competitive advantage of Clade I on that host plant. Clade II could be considered as a recent population gradually invading cultivated fields after retaining wild graminaceous plants as preferred hosts in the expansive NI 84 Kakamega (Kenya) and Mabira (Uganda) forests for a long time. Even though there is no evidence to support this hypothesis, retaining wild plants as their preferred hosts may have been facilitated by low agricultural activities around Mabira forest (Uganda) and Kakamega forest (Kenya). These forests and their environs are colonised by a wide range of natural enemies associated with diverse noctuid stem borers that inhabit the forests (B. P. Le Ru, unpublished data) and this may have limited rapid population build-up and subsequent expansion in this area. In addition to acting as a reservoir of natural enemies, other noctuid stem borers such as Busseola fusca (Fuller) out competes S. calamistis in cultivated fields since they oviposit early in the season while S. calamistis moths which arrive later in the season show little preference for infested host plants (Seshu Reddy, 1983). This may explain the observed low S. calamistis densities in Kakamega (Clade II) compared to the observation made in Mtito Andei (Clade I). Similar competition may have excluded Clade II from the maize fields in Mtito Andei where both clades exist. Though Shanower et al. (1993) did not test the performance of the two S. calamistis clades on different host plants, they observed variations in the development time among different hosts. Because of the good nutritive value of maize plants, stem borers reared on the latter complete their development faster compared to the individuals reared on wild host plants. In Mtito Andei where the two clades exist, Clade I which is mainly found on maize plants probably completes its development well before Clade Il and re-infests the available hosts (both maize and wild plants) before the emergence of moths belonging to Clade n. This therefore excludes Clade Il from the cultivated fields and limits its population to few available wild host plants. Coupled with the observed biological differences, this structuration may strongly reduce interbreeding between the two c1ades. KENYATTA UNIVER' ITV , R 85 Applied entomologists across Africa are presently concerned with S. calamistis because of its pest status (Bosque-Perez &Schulthess, 1998; Gounou &Schulthess, 2004; Le Ru et al., 2006a). Current revelation of differentiation of two clades with respect to different geographic regions brings in fresh knowledge that may radically influence management initiatives. The level of differences among these clades suggests that the evolutionary mechanism that separated them may have taken place one million years ago explaining variations in pest status observed across Africa (Seshu Reddy, 1983; Bosque- Perez &Schulthess, 1998). For sustainable management of S. calamistis, there is need to adopt a region specific management approach based on the knowledge of the dominant clade. However, questions with practical implications may still be asked about evolutionary shift and subsequent host preference of S. calamistis clades before designing a sustainable management strategy. For example, has Clade I, which constitutes an important stem borer proportion in low altitude areas in Kenya, adapted fully to the new host plants (maize and sorghum)? Is the observed preference (rather than an expanded host range) evolutionarily favoured because of trade-offs in fitness on different plants, with adaptations to the new host reducing fitness on the original host (Futuyma & Moreno, 1988)? Further laboratory experiments as well as intensive field studies need to be done at other sites and in different geographical locations. This is the only way to unravel the relationships between host-plant colonization, particularly the attractivity of both clades to the different host plants, and spatial distribution. YL RA 86 CHAPTER SIX 6 GENETIC DIVERSITY AND POPULATION STRUCTURE OF BUSSEOLA PHAIA SSP. PHAIA BOWDEN (LEPIDOPTERA; NOCTUIDAE) IN WILD AND CUL TIV ATED HABIT ATS 6.1 Introduction Natural habitats in Africa have been subjected to diverse forms of modification over the past half century resulting in novel spatial patterns of organisms and resources (Bosque-Perez & Schulthess, 1998; Gepts, 2004). This, coupled with increased human pressure, has severed connections between once-continuous expanses of native habitats resulting in native habitat fragments interspersed with areas of degraded agricultural landscapes (Forman, 1995; Fahrig, 2003). Habitat loss and fragmentations are currently widespread and are likely the most serious threats to the earth's biological diversity (Laurance & Bierregaard, 1997; Summerville & Crist, 2001). In response to natural habitat loss and fragmentation, and the subsequent increase in contact areas between wild and cultivated habitats, some indigenous phytophagous insects that initially remained among indigenous plants in the natural habitats expanded their host ranges and are currently taking advantage of these human induced resource changes (Futuyma & Moreno, 1988; Thies et al., 2003; Gassmann et al., 2006). But depending on the degree of landscape heterogeneity and stability, organisms can either specialize on a particular host plant or successively or even simultaneously exploit a wide range of host plant species (Jonsen & Fahrig, 1997; Kennedy & Storer, 2000; Tischendorf et al., 2003). However, the extent of ecological specialization remains poorly known in many species, KENYATfA UNIVERSITV I ,tlQ 87 and the reciprocal influences of populations from different crops and/or from wild and cultivated areas through the exchanges of individuals or genes is little studied. Lepidopteran stem borers are among the indigenous phytophagous insects that expanded their host ranges, some of which later specialised and are currently dependent on cultivated plants where they remain important insect pests (Bosque-Perez & Schulthess, 1998; Polaszek & Khan, 1998). Like other phytophagous insects, stem borer pests regularly experience sudden destruction of their habitats that force them to migrate to other favorable hosts or they get locally extinct (Ingram, 1958; Nye, 1960; Khan et al., 1997; Schulthess et al., 1997; Haile & Hofsvang, 2001; Mazodze & Conlong, 2003). For example, Harris (1962) argued that the distribution of Busseolafusca (Fuller) was closely linked with human population density and the intensity of cultivations of cereal crops. Similarly, in South Africa the pyralid Eldana saccharina (Walker) switched from Cyperus papyrus to sugar cane, where it became the economically most important pest (Carnegie, 1974). Due to the temporary character of many crops (e.g., maize and sorghum), migration frequently implies a temporal shift between different host plant species. As a consequence, stem borer pests evolved complex life cycles that could involve the exploitation of different plant species more or less phylogeneticaly related in cultivated or non-cultivated habitats (Gounou & Schulthess, 2004; Le Ru et al., 2006a and b). Although alternative host plants may constitute a temporary or a permanent . source of pest migrants as well as their natural enemies (Thies et al., 2003), the source- sink role of cultivated and non-cultivated habitats in the life cycle of crop pests has received little attention (Manel et al., 2003; Vialatte et al., 2005). This is largely due to K NYATTA, 'Mn/t:OC'IT\I I , 88 the difficulty of tracking movements of small organisms In agricultural landscapes (Lushai & Loxdale, 2004). Cultivated fields where maize [Zea mays L.] and sorghum [Sorghum bieolor (L.) Moench] are currently grown were initially natural habitats (Schulthess et al., 1997) and remain surrounded by non-cultivated fragments. Attempts to reduce losses associated with stem borer pest infestation have been one of the strategies to increase maize and sorghum production in Africa (Harris, 1962). These efforts have been concentrated on reducing stem borer pest populations in the cultivated habitats (Overholt et al., 1994; Schulthess et al., 1997) with few studies focusing on the diversity of non-pest species, diet breadth and potential role of native hosts on pest population build-up (Gounou & Schulthess, 2004; Ndemah et al., 2000; Le Ru et al., 2006a). The recent recovery of Busseola phaia ssp. phaia Bowden on maize (Le Ru B., unpublished data), formerly known to exist among the non-cultivated plants, supports earlier reports that the list of borers and host plant species in eastern Africa is far from being exhaustive (Polaszek & Khan, 1998). The presence of B. phaia ssp. phaia on several hosts in both Kisii and Kakamega in Kenya (Le RH et al., 2006b), brings into question its potential to shift and become an important pest of cultivated cereals since the non-cultivated fragments that accommodate its alternative host plants may not persist for long due to human population pressure. The analysis of relationships between populations living in wild and cultivated habitats is thus important for pest management at the landscape scale, through enhancing naturally occurring control or resistance management in transgenic crops using wild host plants (Gould et al., 2002). This study was initiated to examine the genetic relationships of B. phaia ssp. phaia populations living in the wild and cultivated fragments within the r