International Journal of Tropical Insect Science Vol. 24, No. 1, pp. 79–86, 2004 q ICIPE 2004 DOI: 10.1079/IJT20046 Pollinators of the bottle gourd (Lagenaria siceraria) observed in Kenya Yasuyuki Morimoto1,*, Mary Gikungu2 and Patrick Maundu3 1 International Plant Genetic Resource Institute (IPGRI), Sub-Saharan Africa Regional Office, c/o ICRAF, PO Box 30677, Nairobi, Kenya: 2 Department of Invertebrate Zoology, and 3Kenya Resource Centre for Indigenous Knowledge (KENRIK), The National Museums of Kenya, PO Box 40658, Nairobi, Kenya (Accepted 19 November 2003) Abstract. The study reported examined mechanisms of pollen transfer within the genus Lagenaria, including insect vectors in the natural and cropped field environment. The major flower visitors to Lagenaria were observed for 10 months, from June 2001 to April 2002 at five sites, namely the Botanic Garden of the National Museums of Kenya, three farmers’ crop fields in Dungicha village in Kilifi District, Katitika and Wayowani villages in Kitui District and the home garden of one of the authors, in Westlands, Nairobi, Kenya. Twenty-two species of insects represented in four orders belonging to 10 families were found to be the major visitors of L. siceraria flowers. From the foraging patterns of the flower visitors, visiting frequencies and development of fruits, four groups of flower visitors comprising hawkmoths (Hippotion celerio, Agrius convolvuli), moths A – D (Noctuidae spp.), skipper butterfly (Gorgyra johnstoni) and honeybee (Apis mellifera) were considered active flower visitors. Hawkmoths were suspected to be the major pollinators of this plant in the locations surveyed. Key words: Kenya, bottle gourd, Lagenaria siceraria, pollinators, hawkmoths Résumé. Cette étude examine les mécanismes de transfert du pollen dans le genre Lagenaria, faisant intervenir les insectes vecteurs dans les habitats naturels et cultivés. Les plus importants visiteurs de fleurs de Lagenaria ont été observés pendant 10 mois, de juin 2001 à avril 2002 dans le jardin botanique du Muséum National de Nairobi, dans une parcelle paysanne du village de Dungicha dans le district de Kilifi, dans les villages de Katitika et Wayowani dans le district de Kitui et dans le jardin de l’un des auteurs, à Westland à Nairobi. Vingt deux espèces d’insectes appartenant à quatre ordres et 10 familles ont été trouvées. A partir du comportement de recherche, de la fréquence des visites et du développement des fruits, quatre groupes de visiteurs de fleurs comprenant les sphingides (Hippotion celerio, Agrius convolvuli), les papillons de nuits A –D (Noctuidae spp.), l’hespéride (Gorgyra johnstoni) et l’abeille (Apis mellifera) sont considérés comme des visiteurs de fleurs actifs. Les sphingides semblent être les pollinisateurs les plus importants de cette plante dans les localités étudiées. Mots clés: Kenya, arbre bouteille, Lagenaria siceraria, pollinisateurs, sphingides *Email: yasunbk@hotmail.com Y. Morimoto et al. 80 Introduction Lagenaria siceraria (Molina) Standley, commonly known as the ‘bottle gourd’ is an edible, medicinal and otherwise utilitarian domesticated cucurbit with an ancient pan-tropical distribution (Heiser, 1979). The gene centre of L. siceraria is believed to be Africa but the wild progenitor has not been confirmed. The plant is an annual, strong-growing climber or trailer with a monoecious habit; i.e. separate pollenbearing male flowers and seed-bearing female flowers occur on the same plant. Flowers are white, opening at dusk and closing the following day. Female flowers have three spreading stigmatic lobes. Petals and sepals are fused. The spiny, sticky pollen is not windborne and the plants thus require pollinators to move pollen from male to female flowers. The use of the gourd as a container is common among most African cultures. The young, tender fruits of some forms are also eaten in some communities. The most well known of edible types are the forms with round or egg-shaped fruits which normally have a warted surface. Some edible forms also produce large fruit, which can be used as containers at maturity. The diversity in shapes, sizes, shell colour, surface texture and shell thickness of L. siceraria fruit in Kenya is amazingly high. Interestingly, fruit appear linked to culture, as some forms are found only in certain communities where the cultivars are maintained by the local customs. The impressive diversity of bottle gourds has, however, received little attention by researchers, as the crop is of little commercial significance. Furthermore, the plant’s immense cultural value in many African countries is rapidly waning as plastics and modern foods take over the role of the bottle gourd. It is quite likely that the rich diversity of bottle gourds is largely attributable to diverse cultural practices in selection of fruits and seeds. These cultural practices throughout Africa in turn, exist in diverse physical and natural environments. Since Lagenaria depends upon animal vectors for pollination, it makes sense to believe that insects are also crucial in maintaining this diversity, and that the suite of pollinators must either change or adapt to the diverse environments in which bottle gourds are grown. Unfortunately, very little is documented about the reproduction mechanism of the bottle gourd in Africa. A study was therefore designed with the aim of studying floral visitation patterns in natural and crop field environments, as a key element to understanding the unique biological diversity of the bottle gourd. The study identified major flower visitors, and inter-varietal differences such as variation in male and female flower ratio, flowering time of the day and floral longevity, all potentially important factors affecting gene flow. Materials and methods Field experiment Seeds of 36 representative fruits of L. siceraria and three accessions of its wild relatives (L. sphaerica, L. abyssinica and L. breviflora) were selected, taking into consideration significant differences in fruit shape, use and the location collected and were sown in the botanic garden field of the National Museums of Kenya (NMK) on 13 March 2001, the beginning of the long rainy season in Nairobi. Three plants for each accession were grown to maturity and used for the survey of flowering time, male/female ratio, number of the fruits per stem and monitoring flower visitors. Number of the fruits per stem, planting distance, stem numbers per farmer’s crop field and general management information of the Lagenaria plants were also recorded in the three framer’s crop fields in Dungicha, Katitika and Wayowani villages. Monitoring and collecting flower visitors For the purposes of comparison, flower visitor surveys were carried out in five different locations, including at the botanic garden NMK, one author’s (YM) home garden in Nairobi, Dungicha village in Kilifi District (80 km northwest of Mombassa, Kenya), and Katitika and Wayowani villages, Kitui District (150 km southeast of Nairobi, Kenya), between 26 June 2001 and 23 April 2002. A total of 41 days of observations were logged between these five locations. To record the occurrence of insect visits to flowers, a digital video camera (Sony digital Handycamw DCR-TRV20) was set up at dusk, when flowers start to open, and pointed at one or two flowers. The camera was kept running until the flower closed the following day. A built-in infrared facility enabled filming at night. Insect specimens found visiting flowers—primarily during the daytime—were also captured with a locally made insect net (ring diameter: 37 cm, net length: 85 cm, handle length 45 cm). The collected and video-recorded insects were identified at the Department of Invertebrate Zoology, National Museums of Kenya (NMK). Results Growth and development of Lagenaria plants in the experimental gardens In the Botanic Garden, Nairobi, emergence of seedlings occurred 5 –7 days after direct sowing. Pollinators of the bottle gourd Ninety percent of germination appeared 14 days after sowing. Seed dormancy was observed in three out of 24 accessions germinated. One of the accessions germinated 109 days after sowing. The creeping stage of growth began 5 to 6 weeks after emergence and was characterized by rapid elongation of lateral stems and tendrils. Flowering behaviour and characteristics Early- and late-flowering varieties were identified. In the early-flowering variety, the male flower blossomed 77 days after sowing while its female flower blossomed 99 days after sowing. In the lateflowering variety, the male flower blossomed 131 days after sowing while the female flower bloomed 152 days after sowing. Female flowers appeared about 1– 4 weeks after male flowers. The flowering period was normally shorter in female flowers than in male flowers. For female flowers, the flowering period was 3 – 12 days in each plant. The sex ratio (F:C) was highly skewed; the estimated ratio from all plants studied was 26:1. Both male and female flowers opened and closed at approximately the same time of the day, but male flowers sometimes opened 30 to 60 min earlier and closed 30 to 60 min later than female ones. Flowers did not bloom after a cold rainy day, when nighttime temperatures dropped to below 88C. Flowering normally started between 1730 and 2300 h taking almost 60 to 90 min to open fully. The flowers closed 8 –20 h after flowering. Generally the early-opening varieties closed early and late-opening ones closed late, although a large inter-varietal variation could be detected on this characteristic. Flowering time differed depending on the location. In the Botanic Garden, Nairobi, flower opening started between 1730 and 2300 h and closed between 1200 and 1700 h the following day. In Dungicha village, Kilifi District, opening started at around 1830 to 2300 h and closed between 1000 to 1300 h the following day. In Katitika and Wayowani villages in Kitui District, flower opening took place between 1800 and 2300 h and closing between 0700 and 0900 h the following day. The female flowers were normally formed at the 1st to 3rd leaf axils of the auxiliary vines surrounded by male flowers. They also occurred towards the tip of the main and auxiliary vines. Female flowers appearing on the early leaf axils (early nodes) had a higher chance of developing fruits than those appearing later. Later, female flowers developing towards the end of the creeping branches usually developed smaller fruits or sometimes dried up. In the NMK Botanic Garden, a few hermaphroditic flowers were observed on Acc.11 and 54. These flowers were usually larger than male or female flowers. 81 In both the Botanic Garden and the author’s home garden in Nairobi, three wild species grew well and were characterized by early lateral stem development and rapid internode elongation. They showed stronger creeping characteristics than L. siceraria. Lagenaria abyssinica dried in late September 2001 without flowering. Lagenaria sphaerica and L. breviflora were still growing in the NMK Botanic Garden and author’s home garden, Nairobi, though no flowers had developed, 14 months planting. It was observed that small round edible, often warted types produced more female flowers compared to the dipper, bilobal and club-shaped types. Cultivation by local people As there are no commercial varieties available, farmers carefully select the seeds from their previous harvest. Several different types are chosen depending on their needs, and in each type one or two fruits are selected according to their criteria. The seeds are taken out from the intact fruits at the time of seeding and sown. Therefore the chance that the seeds of a fruit are planted in the same spot or adjacent sites seems high. However, the seeds are sometimes mixed with the seeds of other landraces or vegetables. In the field, the bottle gourd is mostly mix-planted with other crops, especially pumpkins. In the farmer’s crop field in Dungicha village in Kilifi District, the average planting distance was 2 – 15 m. Occasionally, vines got entangled with those of neighbouring plants (see Fig. 1). The number of plants per farmer’s crop field was 1 – 70 depending on the size of the crop field. Normally bottle gourds were planted near the homestead or animal enclosures where the soil is considered fertile. It was also observed that round or elongated edible types were planted more frequently than container types. This was commonly observed in both Kilifi and Kitui Districts. Weak or diseased seedlings were thinned out 3– 4 weeks after emergence and 1 –2 plants retained. Planting normally took place during the long rainy season in March to April around Nairobi and October to November in Kilifi and Kitui Districts. In Dungicha village, it was observed that L. siceraria demonstrated very good ability to suppress weeds. Fruits developed rapidly. For use as a vegetable, harvesting started 2 –2.5 months from the time of sowing in Dungicha village in Kilifi District and 3 months from sowing in Katitika and Wayowani villages in Kitui District. Young fruits of edible types were eaten during all stages of development, including the ovary of the flower in Dungicha village, Kilifi District. Farmers confirm the non-bitterness by scratching the fruit surface or chewing the leaves. When bitter plants are found, they are instantly removed. The seeds of 82 Y. Morimoto et al. Surveys of flower visitors Fig. 1. Seeds are often mixed and planted with other vegetables or grains, especially pumpkins. Here the vines are grown among cassava and coconut palms. Dungicha village, Kilifi District, August 2001 variant types that appeared unexpectedly are not used for the next seeding. For use as containers and seed production, fruits are permitted to mature on the vine, typically for 2– 3 months. Immature fruits often rot very fast when harvested. Maturity is reached when the shell hardens and outer and inner layers become yellow. The specific shape of the fruits depends on the variety while the size of the fruits depends on amount of rain, and position along the vine. The crop grows without needing special attention. The fruits are sometimes placed upright on a flat ground for the bottom to form and hence ensure stability later when in use, or hung straight. Fruits are stored and dried as intact fruit with other types of harvests in the house, granary, under trees or elsewhere in the home garden away from rain. Under optimum conditions, approximately 1 to 20 fruits are harvested per plant depending on the size of individual fruit. Plants that produce small fruits can produce 20 or more, and those with large or extra-large ones produce only a few. A total of 22 species of insects represented in four orders and 10 families were found to be the major flower visitors of L. siceraria in the locations surveyed (Table 1). Ten species of insects representing six families were found in the Botanic Garden of NMK during an 18-day survey between 26 June and 2 September 2001. Moth A (Noctuidae) and honeybee (Apis mellifera) were the most frequent visitors. The hawkmoth (Hippotion celerio) and cucumber beetle (Hyperacantha semipalliata) followed, respectively. In this specific survey, honeybees collecting pollen were observed between 1000 and 1100 h (the warm morning hours) on 25 August and 2 September. Moth A visited the male flower twice at about 0530 h on 8 July and stayed for 60 s and 80 s respectively. In the first visit, the moth was captured clearly by the video camera as it sucked nectar. A hawkmoth (Hippotion celerio) also appeared twice for 3s duration around 0620 h on 31 July and 6 August 2001. A high species diversity of insects was found (12 species in eight families) in local farmer’s crop fields in Dungicha village, Kilifi in a 3-day survey in August 2001. Hawkmoth (Hippotion celerio) was particularly active between 0500 and 0600 h on 16th and 18th August. The infrared camera captured clearly the activity of the hawkmoth (Hippotion celerio) hovering and flying between male and female flowers to collect nectar, hovering over the flowers with wings vibrating, while sucking the nectar from the blooms with its long mouth parts (Fig. 2). The average time spent on each flower was about 2 s. Skipper butterflies (Gorgyra johnstoni) were also seen visiting the flower particularly between 0600 and 0800 h, sucking the nectar from the blooms. Unlike the hawkmoths, the butterflies landed on the flower to suck the nectar. Their average stay per visit was 22 s. Besides the skipper butterfly, three other butterflies (Anthene lunulata, Belonois creona and a Pieridae) were identified as flower visitors in Dungicha and this contrasted with the results obtained in Nairobi surveys where no butterfly visitors were observed. The leaf cutter beetles (Coryna species) were also captured but it was observed that they ate the flower petals. It was observed that a fly sp. A (Syrphidae) visited a female flower and stayed for 163 s, but it stayed mainly on the petals, moving on the flower and touching the surface of the stigma with its mouthparts repeatedly for a few moments. Seven species of insects representing four families were found in Katitika and Wayowani villages, Kitui District in the 12-day survey of 15 to 20 March and 18 to 23 April 2002. Hawkmoths, including Hippotion celerio and Agrius convolvuli were observed between 1900 and 0600 h almost Table 1. Observed major flower visitors of L.agenaria siceraria in the locations surveyed Location/duration surveyed Common name Days of insect visits observed Time of insect visits observedþ Chrysomelidae Hyperacantha semipalliata 5 0900– 1100, 1600– 1700 1.4 Halictidae Syrphidae Syrphidae Syrphidae Noctuidae Syrphidae Noctuidae Sphingidae Apidae spp. spp. D spp. C spp. B spp. A spp. A spp. B Hippotion celerio Apis mellifera 3 1 1 1 1 2 1 2 2 1100, 1600, 1800 1000 0900 1100 0500 0900– 1000 1000 0600 1000– 1100 1.3 1 1 1 2 1 1 1.5 2 70.0 14.4 Family Species 3.0 Westlands Nairobi/ 06 December 2001– 26 January 2002 (total of 8 days) Honeybee Fly C Wild bee Apidae Syrphidae Apidae Apis mellifera spp. C ?Anthophora spp 3 1 1 0800– 0900, 1100, 1300 1700 1000 1.7 Dungicha villadge, Kilifi District/16 – 18 August 2001 (total of 3 days) Hawkmoth Skipper Beetle A Beetle B Fly A Fly B Butterfly Butterfly Wild bee Butterfly Honeybee Wild bee Sphingidae Hesperiidae Meloidae Meloidae Syrphidae Syrphidae Pieridae Lycaenidae Apidae Pieridae Apidae Halictidae Hippotion celerio Gorgyra johnstoni Coryna spp. A Coryna spp. B spp. A spp. B spp. Anthene lunulata Ceratina spp Belonois creona Apis mellifera spp. 2 3 1 2 1 2 1 1 1 1 1 2 0100, 0500– 0600 0600– 0800 0800 0900 0900 0600, 0800, 1100 0900 1100 1100 1000– 1200 0900 0700 6.5 5 1 2 1 1.5 1 1 1 2 1 2 2.0 22.5 4.0 33.5 163.0 11.3 1.0 1.0 36.0 1.0 48.0 22.6 Katitika and Wayowani village, Hawkmoth Kitui District/15 March – 23 Hawkmoth April 2002 (total of 13 days) Hawkmoth Moth C Moth D Skipper Fly A Sphingidae Sphingidae Sphingidae Noctuidae Noctuidae Hesperiidae Syrphidae spp. Hippotion celerio Agrius convolvuli spp. C spp. D Gorgyra johnstoni spp. A 3 3 1 1 1 1 2 1900– 0600 0500– 0600 2000 2200– 2300 0100 0800 0600– 0800 2.7 1 1 2 1 1 1.5 2.3 3.2 1.5 159.0 120.4 32.6 340.0 þ 2 5.5 Pollinators of the bottle gourd Botanic garden NMK, Nairobi/ Cucumber beetle 26 June – 2 September 2001 (total of 18 days) Wild bee Fly D Fly C Fly B Moth A Fly A Moth B Hawkmoth Honeybee Average number of visits /day Average (frequency duration of visits) of visit (s) 24 hour time. 83 84 Y. Morimoto et al. Fig. 2. Hippotion celerio (Sphingidae), hovering above female flower and sucking nectar at around 0530 h Dungicha village, Kilifi District, August 2001 surface of the anther. Interestingly, these flower visitors were all observed in the April survey. Thirty-three days after the April survey (on 26 May 2002), we noticed that only three of the six female flowers monitored had developed fruits. Only the flowers visited by hawkmoths had developed fruits (Table 2); during all 3 days they were monitored, hawkmoths were recorded more than twice a day. Three species of insects representing two families were found in the author’s home garden in Westlands, Nairobi, in a total of 8-day survey from December 2001 to January 2002. A fly species C and a few types of bees were observed but none of the nocturnal visitors and butterflies were monitored. A few types of bugs, e.g. Crytacrus comes (Pentatomidae) were found frequently on the flowers during the day and night in all the locations surveyed. The maximum recorded number of visits was 103 a day at the author’s home garden, where no fruits were produced. Thus, these small bugs are considered unimportant in the pollination of L. siceraria in spite of their frequent visits. Discussion General observations Fig. 3. The honeybee (Apis mellifera) visiting a Lagenaria sphaerica (wild species) flower, Gigiri, Nairobi, January 2001 every day in the April survey. The average time spent on each flower was as short as 3 s. The average number of visits per day on Sphingidae averaged 2.17. During the survey, identified Hippotion celerio were normally observed around 0600 h, roughly the same time as Dungicha village, at 0530– 0600 h. Besides the hawkmoths, two other moths (Noctuidae spp. C and D) were identified as flower visitors in Wayowani village. A skipper butterfly (Gorgyra johnstoni) was observed at 0830 h only on 18 April in Wayowani village, roughly the same time as Dungicha village, at 0800 h. A fly spp. A (Syrphidae) was also observed on 19 and 22 April. On 22 April, it visited only male flowers twice at around 0630 h and stayed for 594 and 412 s, respectively, moving on the flower and touching the Wide differences in male:female flower ratios, flowering times, and date of first and last flowering were evident among different accessions growing under identical conditions at NMK, suggesting that these are genetically determined factors. An interesting finding was that the interval of about 1 – 4 weeks between the blooming of the male and female flowers was consistently observed in all accessions irrespective of whether they were early- or late-flowering. This asynchronous flowering of male and female flowers is often considered to be a mechanism to prevent selfpollination. Surveys of flower visitors Based on the foraging patterns of the flower visitors, visiting frequencies and development of the fruits, hawkmoths (Hippotion celerio, Agrius convolvuli), moths A – D (Noctuidae spp.), a skipper butterfly (Gorgyra johnstoni) and a honeybee (Apis mellifera) were considered to be the most active flower visitors. From the data on frequency of visits and fruit development, the hawkmoths (Hippotion celerio and Agrius convolvuli) appeared to be the major pollinators of this plant in the locations surveyed. During their floral visits, it was clearly observed that these insects were actively moving from flower to flower. According to Heiser (1997) Pollinators of the bottle gourd 85 Table 2. Observed flower visitors, duration of visit, visiting time, number of visitors per day and fruit development of Lagenaria siceraria in the locations surveyed in Kitui District Date 18 April 2002 19 April 2002 20 April 2002 21 April 2002 22 April 2002 23 April 2002 Flower visitor Duration of visit (s) Sex of flower monitored1 Visiting time2 Number of visitors per day Fruit development3 Noctuidae spp. C Noctuidae spp. C Noctuidae spp. D Hippotion celerio Syrphidae spp. A Hippotion celerio Agrius convolvuli Hippotion celerio ?Hippotion celerio? ?Hippotion celerio? Gorgyra johnstoni ?Hippotion celerio? Syrphidae spp. A Syrphidae spp. A Sphingidae spp. Sphingidae spp. Sphingidae spp. Sphingidae spp. Sphingidae spp. 99.2 188.0 120.4 0.5 7.6 4.6 1.5 4.2 3.0 3.4 32.6 1.1 594.2 412.6 2.2 3.2 3.0 1.2 2.5 F F F F F M&F M&F M&F M&F M&F M&F M&F M&F M&F F F F F F 2250 , 2252 2348 , 2351 0139 , 0141 0602 0834 0601 2021 0545 0545 0601 0833 , 0834 2159 0622 , 0632 0635 , 0642 1913 1939 2155 2155 2048 3 2 2 2 2 þ 4 þ 3 2 5 þ 1 M: Male flower, F: Female flower. 24 h time. 3 þ : Fruit developed, 2 : Fruit not developed. 2 nocturnal pollination by the large hawkmoths Hyles lineata and Manduca quinquemaculata was suspected to be the major route for pollination of the plant. It was observed that these hawkmoths were foraging among adjacent flowers, though they can fly long distances and cover a wide area while collecting flower nectar. It is known that the honeybee can also cover areas of up to 2 km in diameter. It can therefore be concluded that bottle gourd plants within 2 km or more in a local community can be cross-pollinated by these pollinators to produce hybrid fruits. Through our survey, it was identified that flowers opened around 1730 to 2300 h and closed from 0700 to 1800 h the following day. Among the flower visitors identified, about 70% of the flower visits were made during the day. According to Kurata (1989), longevity of L. siceraria pollen is relatively long and its germinability lasts for 26 h after flowering. This fact supports the possibility of cross-pollination by butterflies and bees during the daytime until the flowers close in the afternoon especially around the higher-elevation Nairobi area, where flowers stay open longer. Cheng (1990) reported that L. siceraria can interbreed fully within the species including selfcombination, and that it also interbreeds with its wild relative, L. sphaerica. Interestingly, however, Hayashi (2002) found that in the hybrids between L. siceraria and L. sphaerica, female flowers abort and the pollen tubes do not elongate properly. During our study in Kenya, the authors observed that L. siceraria is occasionally cultivated just next to its wild relatives. The authors also observed that L. sphaerica flowered at dawn (0600 to 0800 h) and closed late in the evening in Nairobi. Honeybees (Apis mellifera) were observed moving from flower to flower of these different species, collecting pollen (Fig. 3). We suggest that L. siceraria growing in farmers’ fields may be pollinated not just by pollen of its own species, but also by pollen from wild adjacent species, maintaining a large gene pool within the species. This maintenance of high diversity within the species must be attributed to insect pollen vectors. Inter-specific hybridization would occur, but hybrids do not seem to survive because of their sterility. In the five locations surveyed, Dungicha village in Kilifi District showed the best growth of fruits and plants. Katitika and Wayowani village, Kitui District followed and the NMK Botanic Garden was third. In the author’s home garden Westlands, the plants had not borne fruits 14 months after planting. This observation is interesting considering that it Y. Morimoto et al. 86 matches our survey data of visiting frequencies and species diversity of insects. Lack of fruit set could be attributed to lack of sufficient pollinators due to habitat disturbance. More recently, pumpkin crops in New York have declined for the same reason (Watanabe, 1994). Future research The efficiency of each flower visitor on fruit productivity needs to be investigated further. Careful analyses of how efficient and reliable different floral visitors are in accomplishing pollination and cross-fertilisation needs to be undertaken. This survey raises a number of questions that may need to be answered in future research work. A critical question is why the diversity in L. siceraria is so high, as seen in their size, shape, shell colour, texture and thickness and the diversity in fruit and seeds shape. To what extent is this diversity a product of human selection, and to what degree is it a function of pollinator visitation patterns? Another question is why some forms of bottle gourd are found only in certain community groups. Local beliefs are held that different types of the plant never interbreed, yet our observations, and the literature indicates the opposite may be true, and may contribute to diversity. Conclusions Twenty-two species of insects represented in 10 families of four orders were found to be the major flower visitors of L. siceraria. From the foraging patterns of the flower visitors, frequency of visits and development of the fruits, four groups of flower visitors, namely hawkmoths (Hippotion celerio, Agrius convolvuli), moths A –D (Noctuidae spp.), skipper butterfly (Gorgyra johnstoni) and honeybee (Apis mellifera) were considered the active pollen vectors. Through the data identified on frequency of visits and fruit development, hawkmoths (Hippotion celerio, Agrius convolvuli) were suspected to be the major pollinators of this plant in the locations surveyed. Monoecious L. siceraria existing in the natural and farm environments are entirely dependent upon insect vectors for pollination. This appears to contribute to a huge gene pool within the species. Acknowledgements The authors wish to thank Prof. H. Morishima (Tokyo University of Agriculture) and Dr B. Gemmill (Environment Liaison Centre International) for reading the manuscript and offering very useful suggestions. This work would not have been possible without the help provided by a number of researchers and community members. We are particularly indebted to the following: Dr G. Abungu (Director General, NMK), MS M. Imbumi (Kenya Resource Centre for Indigenous Knowledge NMK), Mr W. Wambugu and Mr D. Odhiambo (Botanic garden NMK), Mr M. Karisa and Mr J. Fondo (Coast Forest Conservation Unit, NMK), Mr K. Gohu and his family members (a farmer in Dungicha village, Kilifi District), Mr J. Kimonyi and Kyanika Adult Women Group members (farmers in Katitika village, Kitui District), Mr O. Mutuvi and his family members (a farmer in Wayowani village, Kitui District) and Mr F. Oundo (a gardener in Westlands, Nairobi). References Cheng Yau-Sin (1990) Ethno botanical study on Lagenaria siceraria. PhD thesis. Department of Agriculture, Tokyo University of Agriculture (in Japanese). Hayashi A. (2002) Studies on inter-specific hybridization on Lagenaria speices. BSc thesis, Department of Agriculture, Tokyo University of Agriculture (in Japanese). Heiser C. B. (1979) In The Gourd Book. University of Oklahoma Press, Norman, Oklahoma. 248 pp. Heiser C. B. (1997) Are bottle gourds moth pollinated? Cucurbit Network News 4, 3. Kurata H. (1989) Cucumis melo and Citrullus lanatus, pp. 247 – 333. Horticultural Encyclopedia No. 4. Rural Culture Association, Minato-ku, Tokyo, Japan (in Japanese). Watanabe M. E. (1994) Pollination worries rise as honey bee decline. Science 265, 1170.