Low birthweight (LBW) is highly associated with death during infancy, and countries with the highest LBW rates also have the highest infant mortality rates. We compared temporal trends in LBW with both overall and birthweight-specific infant mortality in United States, Canada, Argentina, Chile, and Uruguay over two time periods, using cohort and cross-sectional analysis of national population-based vital statistics for 1985-89 and 1995-98. Infant mortality diminished substantially (RR = 0.60-0.80 for the later vs. earlier periods) and to a similar degree in all birthweight categories in all five study countries, despite an increase in LBW in the US and Uruguay, minimal changes in Canada and Argentina, and a decrease in Chile. The strength of the (positive) association between LBW and overall infant mortality diminished over the two time periods (from r(s) = +0.80 to +0.25 and RR per SD increase in LBW rate from 2.13 [2.09, 2.17] to 1.76 [1.74, 1.79]). The proportion of infant deaths occurring among LBW infants was negatively correlated with overall infant mortality in both time periods (r(s) = -0.30 and -0.60, RR = 0.68 [0.67, 0.68] and 0.47 [0.46, 0.47]). Developed and less developed countries in the Americas have succeeded in reducing infant mortality in all birthweight groups despite inconsistent changes in LBW rates, and none has achieved this success primarily by reducing LBW. Although our results are not necessarily generalisable to the least developed countries in South Asia and sub-Saharan Africa, it is likely that all countries can substantially reduce their infant mortality rates by improving the care of infants at normal and low birthweights.