The second generalisation is that "the variation of the degree of animal fertility in response to the direct action of the environment shall bear an inverse proportion to the variation of the survival capacity under that environment." [54] Here Mr. Pell and I part company. I have already (Chapter III) disputed the causal connection between birth-rate and death-rate which Mr. Pell here asserts. His generalisation is made by assuming that birth-rates and death-rates rise and fall together: that conditions which produce a high death-rate will also produce a high birth-rate and that conditions which cause a low death-rate will also cause a low birth-rate; that the increase or decline of a population is due to the direct action of the environment; and finally that "the _actual_ degree of fertility is decided by the direct action of the environment." [55] On that last rock Mr. Pell's barque sinks. The mistake here is analogous to the old Darwinian fallacy, abandoned by Huxley and by Romanes, that natural selection is a creative cause of new species. Even if the hypothesis of evolution--and it is merely a hypothesis--be accepted, the only view warranted by reason is that variation of species and their actual degree of fertility may be produced, not by the direct action of environment, but by the _reaction_ of species to their environment--a very different story. There is no statistical evidence to prove a uniform correspondence between birth-rates and death-rates, and it is improbable that there should be a physical law of nature whose operations cannot be demonstrated by mathematical proof. Moreover, we know that the same conditions which cause a high birth-rate may cause a low death-rate. In the case of the first settlers in a new country the death-rate is low because the diseases of civilisation are absent and the settlers are usually young, whereas the birth-rate is high. If fifty young married couples settle on the virgin soil of a new country it is probable that for many years an enormous birth-rate, of over 100, will coexist with a low death-rate.