Drinking behavior of the isogenic mouse strain C57BL/6J was analyzed into nongenetic components: stochastic fluctuations, responses to fluctuations in the current environment, and persistent differences between individual animals. The latter accounted for the major part of the variance. The variance was neither increased by differences in diet during the postweaning rapid growth period (prior to assay for drinking choice) nor diminished by uniformity of treatment during this period, suggesting that significant differentiation had occurred prior to weaning. The large variance between animals could be explained by assuming that the genetic role in consumption of alcohol by C57BL mice is permissive-a relative insensitivity to the aversive orosensory and pharmacological effects of 10% alcohol-rather than a specific drug-seeking predisposition.Mice of the inbred C57BL strain differ significantly in consumption of alcohol, even under carefully controlled conditions (1, 2). They also differ in other behavioral traits such as fighting capacity, emotionality, and susceptibility to audiogenic seizures despite having essentially the same genome (3). There is thus an interesting contrast between genetic identity and behavioral diversity, a phenomenon that must stem from differential effects of the environment (4).In the present study, this isogenic strain was used to analyze the variance of drinking behavior into three components: stochastic fluctuations (transient variations in consumption of alcohol, uncorrelated between animals), responses to concurrent stimuli (shown by correlated variations in a replicate group), and persistent differences between animals in the same environment (found when the daily consumption of each animal is averaged over several weeks). The latter differences point to an effect of earlier experiences, not necessarily involving alcohol, and are the principal concern of this report. If genetically heterogeneous animals living under apparently identical conditions exhibited as great a variance in drinking behavior as that of the C57BL mice, it would be natural to attribute their variance to genetic diversity and to use the differences in their drinking as guides to selective breeding. With such animals, selective breeding will segregate only the genetic contribution. Previous work in our laboratory has dealt mainly with the immediate effects of diet or drugs on consumption of alcohol by C57BL/6J mice (5-8). For these purposes, a preliminary stratification of animals into groups was made after a period of baseline observation to ensure similar group averages before treatment. These experiments therefore always involved a substantial within-group variance of alcohol consumption. The present study examines the source of this pretreatment variance.In principle, a persistent bias toward or against alcohola substance with both nutrient and pharmacological actions-could be established by experiences at any time after conception. The problem is to identify the epochs in development when the animal's...