Dosage at the scute locus, and canalisation of anterior and posterior scutellar bristles in Drosophila melanogaster.

A N extensive analysis, of the genetic variation underlying the canalised character scutellar bristle number in Drosophila melanogaster, has led to the development of the following model by RENDEL (1959, 1962). A population of individuals has an observed mean phenotypic value which is a continuous function of a single developmental variable. This variable has been termed “make” to avoid the implication that a particular substance is necessarily involved (RENDEL 1962). A number of scutellar bristles greater or less than four occurs only when the level of make in an individual falls outside a range of values bounded by an upper and a lower threshold. Assuming the variation among individuals in a population to be normally distributed on the underlying scale, the frequency of flies with missing bristles ( 3 or less) , or with additional bristles (5 or more), can be related to the mean and variance of the distribution of make by use of the probit transformation. Control of development to give the normal four bristles on the scutellum operates over a range of approximately six standard deviations on this underlying scale ( RENDEL, SHELDON and FINLAY 1965). ROBERTSON (1965) has questioned the biological validity of using the probit transformation to analyse this character, since the formation of four bristles in their normal sites may not be a sequential or cumulative process. ROBERTSON suggested an alternative model in which the primary phenomenon is the control of development leading to one and only one bristle at each site, and argued that the probit transformation would then be applicable only to the events at a single site. Total scutellar bristle number is then a secondary measure which is a function of the respective siite probabilities, which ROBERTSON assumed to be independent in any given individual. In support of his model, RENDEL (1965) presented convincing evidence for a lack of independence of bristle formation at different sites in populations homozygous for the sex-linked mutant scute. This mutant leads to a marked, average reduction in scutellar bristles. From observations of the respective probabilities of formation of a bristle at anterior and posterior sites on the scutellum, RENDEL calculated the expected distributions of scutellar bristle patterns for a number of populations based on ROBERTSON’S model of independence. In all populations the observed frequencies were shown to depart from expectation in a characteristic manner, which in scute populations of high mean bristle number led to a higher