John R. G. Turner

Quaternary Refugia in Tropical America: Evidence from Race Formation in Heliconius Butterflies

The hypothesis of Haffer, Turner, and others, that patterns of race and species formation in the tropical forests of South America are the result of the isolation of populations in forest refugia during widespread climatic changes in the geologically recent past, is supported by the distribution of races in the butterfly genus Heliconius: the location of the refuges for these butterflies shows an excellent accord with the refuges deduced by Haffer in his studies of forest birds. The strict parallel variation through most of South America of the various races of H. melpomene, H. erato and of ten similarly-patterned species shows the result of selection for Müllerian mimicry; as the patterns must be subject to strong stabilizing selection, and as the low vagility of the butterflies normally produces isolation by distance even in a continuous population, it is suggested that the extreme divergence of pattern that some (but not all) Heliconius underwent in the forest refugia results from selection pressure in favour of mimicking the most abundant or distasteful local species, which would vary from refuge to refuge, rather than from geographical isolation per se.

Gene frequencies in the domestic cats of york: Evidence of selection

RECENT work on the population genetics of the domestic cat has shown that not only do the frequencies of the various coat colours differ between Europe and the Far East, but also these frequencies differ between Europe and North America; as the North American population was presumably derived from that of Europe about two centuries ago, this indicates rapid evolution (Searle, 1964; Todd, 1966). The present survey gives an additional estimate of frequencies in the cat population of England, the country from which one assumes a large proportion of North American cats, as of North American people, are derived; and makes a comparison between the frequencies of a relatively small town and the large London conurbation, in the same way that certain small French towns have been compared with Paris (Dreux, 1967a, b, c, l968c). The data has also been used to discuss the problems of randomness of mating, randomness of sampling, and selection—natural or otherwise. York is a comparatively small city, based on a medieval walled town, with a human population of about 108,600. It has a small amount of light industry, and is still used as an agricultural market centre. Air pollution from large industrial cities further west and from local domestic fires is fairly heavy. The city is much less urban in character than London, which lies 200 miles to the south. The cats, scored at the R.S.P.C.A. home for animals in York, consisted of adults and kittens brought in by their owners, and stray animals brought in by R.S.P.C.A. inspectors and members of the public, both from York and from surrounding rural areas. We scored the following characters:

Mean Fitness and the Equilibria in Multilocus Polymorphisms

A precise theorem is given for the increase in fitness due to natural selection on diploids subject to random mating, non-overlapping generations and not more than two loci; the method of extension to more loci is given by Kojima & Kelleher, and a precise theorem is given here for any number of loci when there is no recombination. The increase is equal to the haploid (or genic) variance in fitness, multiplied by a factor which is equal to two in the absence of dominance, but which otherwise is a function of gene frequency and dominance. The theorem is compared with that of Kimura, which is more general but harder to apply, and to those of Kojima & Kelleher and Fisher, which are respectively restricted to slow selection and absence of epistasis. The new theorem is used to predict the equilibria in populations polymorphic for two loci, and to deal especially with the quasi-stable equilibrium, for which the critical value of recombination is formulated, and the through point, at which a stable and unstable equilibrium meet and annihilate each other. The effect of this in space is to produce a stepped cline, in which gene frequencies and gametic excess change suddenly over a short distance; in time, the through point brings a new slant to Wright’s multiple peak theory of evolution, as populations can move precipitately from peak to peak without the help of random processes. Mean fitness is related only indirectly to population density. By distinguishing carefully between mean absolute fitness (which is the rate of population growth) and mean relative fitness (which is more useful than the absolute parameter for predicting genetical equilibria) we can show the effects of various types of density control on the genetical composition of the population; density dependent selection may appear to be gene-frequency dependent. The fundamental law of evolution is probably a thermodynamic law of increasing matter energy, which is related only tenuously to the law of increasing genetical fitness.

Sexual behaviour: female swift moth is not the aggressive partner

Abstract Observations on three naturally-occurring courtships of Hepialus show that sexual behaviour in these moths is similar to that of butterflies, with the male pursuing the female in flight and mating with her after she alights. Although the female appears to solicit courtship by flying past a hovering male, and although the male becomes flaccid immediately after copulation, it is not true that the female flies directly at the male and knocks him out of the air, as is widely held by moth-collectors.