Children are the most vulnerable among the malnourished population of Bangladesh. Child and maternal care practices are now being considered as important complements to increasing household income or targeted food interventions to address child growth needs. In Bangladesh, as elsewhere, many children, even in poor households, do well nutritionally, whereas others do not. This study attempts to identify characteristics of the existing child and the maternal care environment that could be used as a basis for designing policies and programs to improve the nutritional status of children. For the present study, all children between 6-18 months of age were selected from a nutrition survey of a cross section of 741 households conducted by the IFPRI Bangladesh Food Policy Project in February-March 1992. Households of 111 children thus were revisited in May-June 1993 to obtain, retrospectively, information from mothers or alternative primary caregivers about selected child care practices and related indicators. Information was obtained on feeding practices of infants and mothers, indicators of psychosocial care, and health and hygiene practices. In this study, information on child care practices obtained together with information from the original nutrition survey on maternal and child nutrition, individual food consumption, and household demographic and socioeconomic status was used. Children who exhibited the best growth status, holding age and income level constant, compared to the others in the same environmental setting, are identified as positive deviants. Those with the worst growth are categorized as negative deviants. Children falling in-between positive and negative deviants are labeled as median growers. Even though an increase in income was found to be associated with improving child nutrition, on average, this association was not very evident at the two tails of the nutrition status distribution, with household income of negative deviant children higher than for both the positive deviants and median growth children, implying a limited access or allocation of household income by mothers in these households, and the relevance of non-income factors. Also, in the sample as a whole, gender differences in child nutrition were not found to be very significant. There was, however, unmistakable evidence of differential treatment of children by gender. There were three times as many male children in the positive deviants group as compared with female children. Even though there were an equal number of male and female children in the negative deviants group, there is evidence of differential child mortality by gender, with evidence of large numbers of "missing" female children in this group, who were, on average, only one year old. A selection of caring practices and indicators were identified for infant feeding, complementary feeding, maternal diet and health, psychosocial care, and health and hygiene practices. Descriptive and multivariate analyses were conducted to identify key caring practices and indicators associated with well and poorly growing children. The analysis supports earlier work that indicates that determinants of child nutrition are not exactly the same for different groups of children, even in the same population. Two key factors that were important across the board were hygiene practices and mothers' access to knowledge (listening to radio programs on child health and nutrition). Important factors contributing to negative deviance were found to be an early introduction of complementary food (before four months), restricting maternal diet for longer periods after the child's birth, and the absence of specially prepared food items in the child's diet. Care factors of the caretaker were also found to be important: a mother's expression of "satisfaction with her family life," which was used as one of the indicators of psychosocial care, was found to be statistically significant. Many local practices were identified that programs and policies could support and build upon to facilitate the participation and empowerment of local communities, families, women, and men in Bangladesh for better child nutrition.