The focus of this paper is caries in populations. Caries is assessed as a phenomenon and as it occurs in populations. It is observed that in recent literature the notion of causation of caries is restricted to the biological process on the tooth surface. This may be sufficient to explain caries as a phenomenon, but it is argued that a biological causation is insufficient in order to understand caries and its variation in populations. The understanding of the determinants of a population's oral health has been seriously impeded by the absence of a theoretical framework. A framework would be useful to the extent that it would bridge relevant categories for portraying complex causal patterns of caries in populations. An approach to a framework has therefore been outlined, and the idea of such a framework has been confronted with three critical comments. The framework is as yet pragmatic, incomplete, uncertain and fragmented. Yet, even such an approach may invite greater precision in epidemiological concepts about causes of caries than presently prevails. The framework directs attention to health effects of collective phenomena that cannot be reduced to individual attributes. The image proposed in this paper is intended to spur discussion about important aspects of the epidemiology of caries in populations.