Longitudinal weight growth data from three groups in West Sepik Province, Papua New Guinea, are compared. A form of analysis based on the four-parameter Jenss curve is adopted, to allow intelligible comparisons of parameter means despite irregular weighing schedules for individual children. Although many weighing records include notes of a child's ill-health (e.g. malaria), omission of these weighings does not have a large effect on the results. Some significant sex differences were found, but do not appear to be importantly confounded with group differences. Of the three groups compared, the East Mianmin live at moderate altitude, the Imnai live at low altitude, and the West Mianmin are a migrant group from moderate to low altitude. In principle the analysis could have shown the groups to be differentiated by altitude of current residence (suggesting the influence of environmental factors), by altitude of ancestral residence (suggesting the influence of genetic and/or cultural factors) or by some interaction of causal factors. Findings on most parameters suggest interaction. Neither moderate nor low-altitude conditions appear consistently more favourable to rapid growth; it is suggested that the advantage of a more favourable disease environment at moderate altitude may be at least partially offset by a nutritional disadvantage. The effect of environmental conditions may also be non-uniform, especially for young infants, since infants of the migrant group, the West Mianmin, are born heaviest but grow more slowly in the early months than infants of either non-migrant group.