Embryogenesis in cotton (Gossypium hirsutum L. M-8) proceeds in an orderly fashion. Following self pollination embryos are visible around 15 days later. Progressive development of the embryo occurs over the next 45 days to boll opening around 60 days after anthesis (DAA). Based on fine structural data sequential loading of protein bodies occurs between 30 and 50 DAA. To test the effects of nutritional stress, flowering plants with bolls at different stages of development were defoliated. Bolls were allowed to mature in control and defoliated plants growing under greenhouse or field conditions. The number of mature seeds and pups (immature seeds) present in bolls of different age at the time of defoliation were recorded. These data indicated a progressive increase in mature seeds. The increase in number of mature seeds in older bolls indicated that a limited amount of maturation occurred following hand defoliation. This was most likely due to a significant reduction in photosynthate partitioning in the absence of leaves. We suspect that whatever partitioning occurred was the result of photosynthesis in the chloroldlaylous tissue of the boll. Germination tests revealed that mature seeds from bolls 45 DAA or younger at the time of defoliation had less than 50% germination. This contrasted sharply with greater than 90% Germination for mature, seeds from older bolls. A similar break point about 45 DAA was observed for sensitivity of mature seed to cold stress during germination. Fine structural analysis provided evidence for incomplete filling and maturation of protein bodies, especially in radicles of mature seeds from bolls at 40 DAA or younger at the time of defoliation. The presence of partially and completely filled protein bodies within cortical cells of radicles in mature seeds from bolls at 30, 40 or 50 DAA in defoliation experiments indicated that there may be time-dependent or selective formation and filling of protein bodies. In addition, the number of partially filled protein bodies decreased with age of the boll at the time of defoliation. For example, cortical cells of radicles taken from 30 DAA seeds had several and often large vacuole-like structures containing a minimal amount of protein. We have observed that protein body formation in embryos of control plants involves vacuole partitioning to form an initial population followed by dilation of ER, providing a second or complementary population of protein bodies. These two populations begin development at separate points in time, the first being completed by 27 DAA and the second initiated at the end of the first and completed by 35 DAA. Our results from examination of mature seeds from control and defoliated samples suggest that not only are the normal processes of protein body formation arrested by defoliation, but the structure of ER and Golgi apparatus, which participate in protein body development, are altered. Six to eight long cisternae of RER in laminated array were quite common in cortical cells of all defoliated samples. Numerous small Golgi vesicles in association with disordered cisternae were also common. The cytoplasmic matrix in cortical cells of radicles of mature seeds from defoliated plants, which in controls was filled with free ribosomes, appeared to contain fewer ribosomes and more cytoplasmic ground substance. In contrast, mitochondria and plastics of both control and defoliated samples had a similar level of fine structure and organization. We conclude from these observations that nutritional stress when applied to developing seeds through defoliation of the entire plant adversely affects embryogenesis, particularly protein body formation and Saturation. The physiological effect of nutrient stress was seen in poorer germination and increased sensitivity to cold stress during germination. These effects were most pronounced in mature seeds from bolls that were 43 DAA or less at the time of whole plait defoliation. The period from 25 to 45 DAA is viewed as the most critical and sensitive period of embryogenesis and seed development. Even though photosynthesis in the chlorophylous tissue of the boll may provide nutrients for partitioning to the ovules, this photosynthate partitioning appears to be insufficient for meeting nutrient requirements of the developing embryos, particularly those at 45 DAA or younger at the time of application of nutritional stress.