Introduction of complementary foods (weaning) before 4 to 6 months of age and unmodified cow's milk before age 12 months are associated with several health risks. To develop effective interventions to discourage these practices, evidence of their determinants is needed. This systematic review identified documents from seven electronic databases (database inception 2008) and reference lists, and by contacting authors. Seventy-eight studies in developed countries, published between 1976 and 2008, quantifying the association between either feeding practice and its potential determinants were included. Study quality was systematically assessed in terms of representativeness, sample size, method of outcome ascertainment, and approach to statistical analysis. The distribution of evidence for each determinant was visualized in a harvest plot showing the strength and direction of associations found and the quality of relevant studies. The strength of evidence for each determinant was summarized as strong, moderate, limited, or inconclusive, using an algorithm based on the consistency of the results of studies of the highest available quality. Strong evidence denoted that the determinant was examined in three or more high-quality studies and >/=75% of results were consistent. Strong evidence was found for six determinants of early weaning (ie, young maternal age, low maternal education, low socioeconomic status, absence or short duration of breastfeeding, maternal smoking, and lack of information or advice from health care providers) and for two determinants of early introduction of unmodified cow's milk (ie, low maternal education and low socioeconomic status). Of these determinants, improving advice given by health care providers appears the most tractable area for intervention in the short term.