Many small, numerically abundant animals of low trophic position exhibit polymorphisms for colour and pattern. This variation is assumed to be adaptive and maintained by frequency-dependent selection, the fitness of phenotypes being negatively correlated with their frequency in the population. Noctua pronuba is an abundant moth species of the western Palaearctic and is now established in north eastern America. When the moth is at rest it is apparently cryptically coloured with the visible surfaces polymorphic in colour and pattern. The limitation of this variation to the exposed surfaces of the moth suggests that the variation may be maintained by visual selection by predators. The forewing polymorphism in N.pronuba can be split into three distinct phenotypes: rufous, ochre and silver, with the variation probably controlled by a single locus, with three alleles in a dominance hierarchy. The expression of these alleles is influenced by sex with females lighter in colour than males. The aim of this work was to establish whether natural selection maintains variation in forewing colour and pattern to give a balanced polymorphism, using the null hypothesis that the variation was neutral to selection and non-adaptive. This has been approached by experiment and the analysis of temporal and geographic variation in forewing phenotype frequencies. Background resting experiments failed to show different phenotypes adopting different backgrounds on which to rest but experimental conditions did affect the behaviour of individuals. Selection acting in opposite directions on males and females has been suggested as a mechanism maintaining the forewing polymorphism. Little evidence has been found to substantiate this claim with phenotype frequencies in light trap samples similar in males and females. No systematic changes in phenotype frequencies were observed in an analysis of temporal variation over a twenty-five year period. Only the ochre allele varied significantly but the variation is minimal suggesting that the polymorphism is temporally stable. Geographically there was remarkably little variation in phenotype frequencies with only samples from Finland, Scotland and N.lreland having significantly differentiated phenotype frequencies. A study of polymorphic allozymes suggested that large amounts of gene flow occur in the species. The consequence of this gene flow will be to unite geographically separate populations into one panmictic unit. High levels of gene flow, in conjunction with the local abundance of the species, mean that the effective population size will be large. Previous authors have considered that as crypsis is an adaptive trait, variation in the colour and pattern of a cryptic species must also be adaptive, and maintained by selection. This is not necessarily true, and there may be a number of colour patterns that are equally cryptic in the same habitat. It is hypothesised that the large population size and magnitude of gene flow in N.pronuba gives the forewing polymorphism inherent stability both temporally and geographically, without the need to invoke balancing selection.