25The prevalence of Haldane's rule suggests that sex chromosomes commonly play a key role

The increased rate of psychotic illness that has been reported among the African-Caribbean population in Britain is not due to an increased prevalence of PBCs.

Linking intraspecific and interspecific divergence is an important challenge in speciation research. X chromosomes are expected to evolve faster than autosomes and disproportionately contribute to reproductive barriers, and comparing genetic variation on X and autosomal markers within and between species can elucidate evolutionary processes that shape genome variation. We performed RADseq on a 16 population transect of two closely related Australian cricket species, Teleogryllus commodus and T. oceanicus, covering allopatry and sympatry. This classic study system for sexual selection provides a rare exception to Haldane's rule, as hybrid females are sterile. We found no evidence of recent introgression, despite the fact that the species coexist in overlapping habitats in the wild and interbreed in the laboratory. Putative X-linked loci showed greater differentiation between species compared with autosomal loci. However, population differentiation within species was unexpectedly lower on X-linked markers than autosomal markers, and relative X-to-autosomal genetic diversity was inflated above neutral expectations. Populations of both species showed genomic signatures of recent population expansions, but these were not strong enough to account for the inflated X/A diversity. Instead, most of the excess polymorphism on the X could better be explained by sex-biased processes that increase the relative effective population size of the X, such as interspecific variation in the strength of sexual selection among males. Taken together, the opposing patterns of diversity and differentiation at X versus autosomal loci implicate a greater role for sex-linked genes in maintaining species boundaries in this system.

Effective recognition of conspecific mating signals reduces the risk of maladaptive hybridisation. Dissecting the signal recognition algorithms that underlie preferences is a useful approach for testing whether closely related taxa evaluate the same or different signal features to achieve mate recognition. Such data provide information about potential constraints and targets of selection during evolutionary divergence. Using a series of mate choice trials, we tested whether closely-related, but genetically and phenotypically divergent, field cricket species (Teleogryllus oceanicus and Teleogryllus commodus) use shared or distinct recognition algorithms when evaluating acoustic male calling songs. These species overlap in sympatry, show premating isolation based on female discrimination of male calling songs, yet are capable of producing hybrid offspring. Unexpectedly, female selectivity for features of male song differed between the two species. We found that the two species use a combination of shared and unique signal filtering mechanisms, and we characterised how information about male carrier frequency, pulse rate and temporal patterning is integrated to achieve song recognition in each species. These results illustrate how comparatively few, simple modifications in key components of signal recognition algorithms can lead to striking interspecific discrimination among closely related taxa, despite apparent signal complexity. The finding that some steps during signal recognition and filtering are shared between the species, while others differ, can help to identify behavioural traits targeted by selection during evolutionary divergence.

SUMMARY Seasonal variation in the rates of stillbirth, and of deaths under the age of one year, were studied for England and Wales in order to examine (a) changes in the seasonal variation over the years and (b) the correlation between seasonal rates and seasonal temperatures. The quarterly rates of stillbirths were studied for the period 1928-78; and of deaths under the age of one year, in six different age groups, for the period 1912-78. A disappearance of seasonal variation in rates ('deseasonality') occurred for stillbirths in about 1950, and for neonatal deaths in about 1965. For deaths at 1-2 months a trend towards deseasonality has been apparent since 1955, but there has been no such trend for deaths at 3-11 months of age. In the period before deseasonality, and for the first quarter of the year, there was a high negative correlation between the neonatal death rate and the mean temperature in England and Wales but this correlation fell as the seasonal variation in rates fell. The findings suggest that seasonal variation in the neonatal death rate was closely related to winter temperatures during the period 1921-60. For deaths at 1-11 months old, there has been and still is a relation between temperature and seasonal variation in rates, but the relation was less close than for the neonatal death rate.A number of British studies have shown a correlation between mortality from respiratory disease in young children and the temperature in the week, or in the two weeks, before death." But although it has long been observed that spells of cold weather are associated with an increase in the number of deaths during childhood, there appear to have been no formal reports of a correlation between seasonal death rates (from all causes) in infants and the corresponding seasonal temperature. Such a correlation might reasonably be expected whenever there is a marked seasonal variation in rates. Official statistics show that there was a marked seasonal variation in neonatal deaths during the interwar years which no longer occurs, and it seems likely that this change is an example of the phenomenon which Sakamoto-Momiyama4 has called 'deseasonality'. These considerations led us to make the present study, in which our aims were (a) to trace, for England and Wales, the changing pattern of seasonal variation in death rates for a series of age groups under one year covering as long a period as possible (which was the period 1911-78); (b) to examine correlations between seasonal death rates and mean seasonal temperature, and how such correlations might be affected by the phenomenon of deseasonality; and (c) to examine the correlation between death rate and the extent of seasonal variation.Method