This study assesses the empirical evidence for the presence and the pa
ttern of heterogeneity in child mortality between families in a rural
area of Senegal that is apparently homogeneous with regard to the know
n determinants of mortality. The presence of heterogeneity is assessed
by comparing the distribution of child deaths for women with a chance
distribution. To weigh the relative contribution of causes of within-
and between-family heterogeneity in the risk of child death to the obs
erved pattern of clustering, detailed simulations in which empirical c
hild-based data from the study are used, are undertaken. There are imp
ortant variations in child mortality between families in this communit
y, and the familial component in child mortality is almost entirely co
rrelated with the size of the family. High-risk women experience repea
ted child deaths and repeated short birth intervals; they therefore ha
ve larger families. Heterogeneity in mortality and selective fertility
, however, fail to explain the observed association between low mortal
ity risk and small family size.