Kenyan landraces of the cultivated white-flowered gourd Lagenaria siceraria have highly variable morphology. In order to reveal the inter-and intra-specific variation in fruit and seed morphology in L. siceraria and its wild relatives in Kenya, various traits were examined in a total of 425 strains from L. siceraria (269), Lagenaria sphaerica (124), Lagenaria abyssinica (27) and Lagenaria breviflora (5). Data analysis revealed the following patterns: (1) L. siceraria is more diverse than its wild relatives in both qualitative and quantitative and fruit traits are more variable than seed traits within each species. (2) Principal component analysis of L. siceraria with 15 quantitative traits showed a continuous variation among strains, in which general size factor of fruit and seed, shape factor of fruit and shape factor of seed were the principal causes of variation.(3) No correlation was found between fruit and seed shape, or between size and shape. (4) Image analysis with elliptic Fourier descriptors revealed continuous shape variation in the landraces of L. siceraria. Fruit shape features such as the contrast between a wide base with a distinct handle, and a slender base with an indistinct handle and the degree of bulge of the elongated handle (bilobate shape) were evaluated quantitatively. (5) Analysis of variance of 12 quantitative traits based on the progeny test demonstrated that the degree of heterozygosity is considerably low in the white-flowered gourd existing in the natural environment in Kenya. (6) The quantitative evaluation of the intra-specific variation in fruit and seed in L. siceraria was possible, but it was difficult to classify the landraces into distinct groups. Most of the variation observed in the cultivated L. siceraria, including differences in fruit size and shape, shell thickness and handle development, probably resulted from selection by the local human population.

Soil fertility is vital for agricultural productivity, yet poor soils and erosion remain a management challenge in many parts of sub-Saharan Africa. One challenge is that soil scientists and farmers often evaluate soil fertility using different knowledge systems and the implications have not been clearly reconciled within the literature. In particular, whether farmers are observing similar aspects of structure and function as classified in soil science. If so, what can we learn about how soil fertility is evaluated and communicated in terms of developing a hybrid approach that improves communication of ideas between different stakeholders. This paper addresses this challenge by examining the similarities and differences between farmers' qualitative evaluation and soil science quantitative analysis for soil fertility classification, and how location of soils influence farmers' evaluation of soil fertility. Empirical fieldwork was carried out in two villages in Kitui County, Kenya with 60 farmers using semi-structured interviews and focus group discussion. Based on farmer perception, 116 soil samples of the best and worst soil fertility taken and analysed for physiochemical factors. Farmers had a consistent classification system and primarily relied on texture and colour as indicators for good soil fertility and texture alone for poor soils.Soils with fine texture under the local semi-arid climate were associated with higher pH, TOC and WHC and fertile black and red soils were associated with pH, TOC, WHC and AP based on differences in bed rock. Poor soil fertility was associated with sandy soils and soils with no colour in their local name. Spatial location is an important consideration in farmers' evaluations, reflecting awareness of local diversity in soil and historical social or environmental factors. Local historical narratives reveal the importance in changes to humus, consistent with technical knowledge about the role of soil organic matter for soil fertility. The paper provides better understanding of farmers' soil classification, evaluation processes and perspectives that help to inform scientists working with alternative frameworks for assessment and, in doing so, supports the development of local tailor-made soil assessment systems.

No abstract