PIP: Researchers applied longitudinal data on 2884 women who were at least 6 months pregnant between May 1983-April 1984 and living in the Cebu metropolitan area, the Philippines to examine assumptions which may explain the association between lower breast feeding levels and dwindling social support for mothers who breast feed in urban areas. Extended family households were much more common in urban areas than rural areas (46.3% vs. 27.5%). This finding did not support the 3rd assumption of falling extended family patterns with urbanization. Further as the levels of complexity and extension of the household rose so did infant care support regardless of location. For example, the number of hours of help with infant care for mothers in nuclear families in urban areas was 1.23 which climbed from 2.53 for horizontal extended families to 2.9 for vertical extended families and to 3.09 for horizontal and vertical extended families. This result supported the 2nd assumption of much more social support in extended families than nuclear families. Yet social support did not translate into promotion of breast feeding--the 1st assumption. Mothers in horizontally extended families had a lower likelihood of breast feeding than those in nuclear families. Further only adult female relatives of the same generation as the mother has a negative influence on breast feeding in urban mothers, but not rural mothers. In fact, female children of at least 7 years old, female servants, and yayas all negatively affect breast feeding in urban mothers, but not rural mothers. In rural households, young female relatives (at least 7 years old) significantly increased the probability of breast feeding whereas young male relatives decreased the probability.