SPECIATION AND FEATHER ORNAMENTATION IN BIRDS

The hypothesis that sexual selection promotes speciation has rarely been tested. We identified 70 evolutionarily independent events of feather ornaments in birds. For each focal species we noted the number of ornamented and nonornamented species belonging to its genus and its number of subspecies, as well as its mating system and the extent of its geographic range. For purposes of comparison, we randomly chose a second, nonornamented species for which we obtained information on the number of subspecies, and in cases in which the nonornamented species was in the same genus, we chose a third, nonornamented species in a related genus and obtained the same information. We then noted the number of species in each genus and the difference in numbers of species, or species richness, between paired genera. For the genera of the focal ornamented species, we regressed number of ornamented species on number of nonornamented species and found a positive relationship. As number of species per genus rose, number of ornamented species per genus rose more rapidly, indicating that more speciose genera have a higher proportion of ornamented species than less speciose genera. We then took the deviations from this regression, the residual number of species, and regressed them on the differences in species richness between the paired genera. This relationship was positive indicating that ornamented genera with more than the expected number of ornamented species were more speciose with respect to their paired genera than were genera with fewer than the expected number of ornamented species. Finally, we compared the deviations from this regression, the residual number of ornamented species, with species' mating system and found a greater residual number of ornamented species among species whose mating system is associated with greater skew in male mating success and thus more intense sexual selection. Ornamented species had more subspecies than nonornamented species, even when controlling for geographic range, suggesting an association between subspeciation and ornaments.     

A Model of Genome Size Dynamics During Speciation

A model developed for the evolving size of the repetitive part of the eukaryote genome during speciation was subjected to analytical and computer treatment. The basic assumption of the model was that two classes of repetitive DNA contribute mainly to macroevolutionary changes in genome size: arrays of tandem repeats (ATR) changing through unequal crossover and mobile genetic elements (MGE) changing presumably through an integration mechanism of the Tn- and Is-kind operating in bacteria. Within the framework of this model, the macroevolution of the MGE size is formally equivalent to that of the ATR in the particular case when shifts of chromatids have only one repeat out of register. This allowed us to consider genome size as a large set of various ATRs. The results obtained are as follows. If the duplication and deletion of repeats have unequal fixation probabilities during each speciation act, the predicted species distributions of genome size significantly deviate from the real ones; if they have equal fixation probabilities, there is a conformance between calculated and real distributions. In the latter case, the model reproduces the salient features of real distributions upon acceptance of 1) upper selective boundary nonspecifically limiting increase in genome size within the evolving taxonomic group and 2) non-neutrality of variability in genome size with respect to speciation.     

Linear age-dependent population growth with seasonal harvesting

A population growth is modelled by the Von Foerster PDE with accompanying Lotka-Volterra integral equation describing the birth rate; the age specific death and fertility rates are assumed to depend only on age and not time. A harvesting policy where a fraction of the population of age greater than a given age is harvested for a fraction of a given season. This introduces a time dependence, but this difficulty is circumvented by devising approximate timeindependent models whose birthrates bracket the true birthrate-the standard renewal equation theory applies to the approximate models so quantitative results can be obtained.The author wishes to thank Professors J. J. Levin and R. K. Miller for some useful remarks. This research was partially supported by NIH Grant GMO 7661-02.