The tendency of animals suckled in different litters to grow at different rates is already known to be a serious problem for experimental design in developmental biology. However, the contribution made by avoidable technical practices, such as failure to standardise litter size, to this litter effect is not clear from the literature. The present study sought to minimise the litter effect by carefully contriving five litters of rats to be similar in many respects likely to affect growth rate. These litters were raised simultaneously by mutually similar maternal animals in the same room. Their growth rates were not, in fact, equal: significant effects of litter were resolved on three growth rate dependent variables by the thirtieth postnatal day. The extent of this effect (as great as the between animal variance, despite the attempts to minimise it), and its implications for the interpretation of publications in which it is ignored are briefly discussed. Solutions are offered for the problem litter effects cause for experimental design.