The relation between variability and abundance shown by the measurements of the eggs of British nesting birds

For some years it has appeared to the author of importance to ascertain by collocation of direct observations whether, in fact, Darwin’s generaliza­tion that “widely distributed, much-diffused and common species vary most”, is well established; and, if so, whether we can form an idea of the quantitative relationship between variability and abundance. In a paper published nine years ago in collaboration with Mr E. B. Ford (1928), I examined the variability, shown by Mr Ford’s observations, of thirty-five species of night-flying moths, mostly British, from which examination it appeared that the difficulties in the way of comparing variabilities could be overcome at the expense of some elaboration of the calculations. When a satisfactory basis of comparison had been obtained it appeared that in both sexes the group of species classified in entomological handbooks as “abundant” or “very common” were, in fact, on the average, some 60% more variable than the group of species classified as less than common, while the group of species classified as common held, in both sexes, an intermediate position. Within each of these three groups there was a wide range of variability and I should give an altogether false impression if I suggested that variability among these thirty-five species was closely associated with abundance, or that, if influenced by abundance, it is not also much influenced by other causes. Nevertheless, the apparent influence of abundance was, in this research, statistically significant, and encouraged me to think that other cases might be found in which the reality of the phenomenon could be tested and its numerical importance gauged. For many years the Rev. F. C. R. Jourdain has made and compiled measurements of considerable numbers of the eggs of British nesting birds. At the suggestion of Mr Julian Huxley, I approached Mr Jourdain, proposing at the same time that, in order to ensure the preservation of the large number of individual measurements he had made, I could have them copied in my heritable variance of the species would be nearly proportional to the logarithm of the population. For example, if one appeared in each genera­tion, this would be enough in a population of a million to maintain, on the average, 30·4 factors segregating. While in a population of a billion (a million million) it would be enough to maintain 58·0. The additional segregating factors in the more numerous species would, indeed, add little to its observable variance because of the very great inequality of the generatios with which, in such species, factors can be maintained. They would, on the other hand, be available to increase the observable variance when­ever, by change of circumstances, the mutant gene happened to become slightly advantageous. When this occurs the gene ratio will shift slowly towards equality, and the contribution of the factor in question to the observable variability of the species will steadily increase, until the rival genes are equal in numbers. After this it will decrease again as the new gene replaces the old. Such a course of events must often take thousands of generations. It appears that only if the more numerous species, through its capacity for keeping a larger number of neutral mutations in store for future use, or because of the larger number of individuals exposed to the chance of very rare advantageous mutations, manages to maintain a larger supply of new genes, each gradually spreading throughout the whole population, that its observable variance will be larger than that of a species of smaller population. Hence, although I believe many unknown factors may largely affect the variation of different species, any consistent difference in variability, observable between more and less populous species, must be ascribed to that fraction of the specific variance contributed by those particular factors by means of which evolutionary progress is now taking place. It is for this reason that I believe the attempt to ascertain whether such a difference in variability can be measured is worth the very con­siderable computational labour which the attempt involves. It would be too much to claim that we have here a means of measuring the rate of evolutionary progress. What it does supply is the order of magnitude of the contribution to the observable variance, of those genetic factors by means of which evolutionary progress is now proceeding.