148 Proceeclimjs of the Royal Physical Society. 



Another pony — one of the Celtic variety — proved sterile with 

 a zebra, a Kiang, an Arab, and three other ponies, but had off- 

 spring to a crossbred pony presenting the more striking traits 

 of the Celtic type. There is no record of equine mules having 

 perfect germ-cells, but one of my zebra hybrids, after return- 

 ing from a prolonged visit to England, had a small number 

 of ripe sperms. Before leaving for the south, and within six 

 weeks after his return, mature germ-cells were conspicuous 

 by their absence. 



Both horses and mares are sometimes quite sterile until 

 some decided change is made in their surroundings, such as 

 a change of food, a change from an outdoor to an indoor life, 

 or a change from one district to another. 



If, in obedience to the Mendelian law, a considerable 

 number of the offspring inherit the characters of a new 

 variety which happens to be pre-eminently adapted to its 

 environment, or if because of its exclusiveuess a new and 

 highly favoured variety produces a number of offspring 

 identical with itself, or if the members of a variety only 

 prove fertile in certain areas or when some change takes 

 place in the immediate surroundings, or only consort with 

 members of a similar variety, it will be evident that there 

 are many possible ways in which the foundations of new 

 species may be laid. Here it may be asked. Can these and 

 other like forms of isolation be regarded as forming part of 

 the doctrine of natural selection ? I think they would all be 

 included in natural selection by Mr Wallace, and I see no 

 reason why any of them need have been excluded by Mr 

 Darwin. What sort of subsidiary factors Mr Darwin had in 

 his mind when he spoke of natural selection being the main 

 factor it is difficult to imagine. It, I believe, might make 

 for clearness if natural selection were divided into environ- 

 mental and germinal selection. 



The differences between the germ-cells of a variety and 

 the parent species would afford scope for germinal selection 

 — differences in fertility and vitality, in the ability to 

 respond to various kinds of stimuli, and to transmit the 

 distinctive characters of the variety to which they 

 happened to belong. 



