April 29, 1898.] 



SCIENCE. 



607 



species and domesticated varieties in respect of 

 cross-sterility : the other, the fact that natural 

 selection cannot possibly give rise to polytypic 

 as distinguished from mouotypic evolution." 

 The author adds : ' ' Now it is my belief that the 

 theory of physiological selection fully meets 

 these diflBculties. " 



To the present writer, these objections to the 

 sufficiency of natural selection seem not to carry 

 weight, but the matter can only be suitably 

 tested by observation. In the first place, it re- 

 mains to be shown that polytypic evolution, in 

 the sense intended by Romanes, occurs com- 

 monly in nature. Cases are cited in which 

 different species of plants appear to be origina- 

 ting in the same area, but it is by no means 

 proven that they did not originate in different 

 localities and have since intermingled. And 

 even if they have been confined from the begin- 

 ning to an apparently uniform environment, it 

 will be very hard to prove that the uniformity 

 is real, so that there is no opportunity for nat- 

 ural selection to act differently on the different 

 parts of the original species. It seems to me 

 that, given diverse conditions of life within a 

 limited area, natural selection may operate not 

 only to bring about the division of a species into 

 two or more, but also sterility between them. 

 Let us take the case of such grasshoppers as are 

 found on the sandhills of New Mexico. Some 

 are green and live on the herbage ; others are 

 light brown, the color of the ground on which 

 they commonly rest. Each kind will gain some 

 advantage from its color, combined with its 

 habits. Suppose that these are still mutu- 

 ally fertile varieties of one species, the greenish 

 and brownish specimens will cross, and the 

 formation of two races will be hindered.* But 

 there will be variations in fertility, and those 

 which are most fertile together, and at the same 

 time of identical colors, will have an advantage. 

 Hence the correlation of, say, greenness with 

 certain variations of the reproductive organs, 

 will be seized upon by natural selection and per- 

 petuated. Even in ordinary monotypic evolu- 

 tion those individuals which are specially fertile 

 inter se, or infertile with those unlike them, and 

 at the same time possess beneficial characters, will 



* Unless there arises a green and brown diohroism 

 ■within .speoiiic limits, as in one species known. 



have an advantage, and the peculiarites of the 

 reproductive organs will be increased through 

 natural selection. Another case is that in which 

 two varieties, mutually fertile, have originated 

 in different localities, but afterwards occupy the 

 same territory. "When they meet (as has been 

 actually observed in some cases) hybrids, or 

 mongrels, will be formed all along the line, and 

 it appears as if the peculiarities of each form 

 will eventually be lost. But any small percent- 

 age of either variety which has in common some 

 variation of the reproductive organs, leading to 

 physiological isolation, will be left alone to 

 represent the original variety ; and as that va- 

 riety is fitter to survive than the mongrel it 

 will supplant it, and we shall have a species 

 which is Infertile with its allies. That it will 

 be infertile with all, or nearly all, its allies is 

 probable because the variations of the repro- 

 ductive organs are so numerous and so diverse 

 that it is not likely that exactly the same sort 

 of variation will be selected in any two cases. 



Thus, while natural selection might employ 

 physiological (more properly, sexual) isolation 

 in the formation of species, it is not quite appa- 

 rent to me that such isolation would of itself 

 cause divergence. For while sexual isolation 

 would be homogamy as regards that variation 

 — which is by itself the reverse of beneficial — it 

 would be apogamy as regards other variations, 

 since it is not shown that it is correlated with 

 any particular modifications of other kinds. So 

 the isolated individuals, having no advantage, 

 but some disadvantages from their isolation, 

 would tend to be eliminated rather than to in- 

 crease, owing to the operation of natural selec- 

 tion. 



As telling against the necessity for natural 

 selection in the formation of species, Mr. 

 Gulick's statements regarding the land-shells 

 (Achatinellidis) of the Sandwich Islands are 

 quoted. In these islands it appears that 

 almost every valley has a distinct variety 

 or species, and yet the conditions of exist- 

 ence are ajjparently the same in at least 

 many of the localities.* Mr. Gulick considers 



* I have never been to the Sandwich Islands, hut 

 can testify that in Jamaica, which also has a remark- 

 ably varied molluscan fauna, there are many different 

 conditions of soil, moisture, vegetation, etc., consti- 



