Maech 23, 1900.] 



SCIENCE. 



453 



be a subspecies of E. iowce, and would re- 

 ceive a trinominal name. 



But if, in the course of time, invasions of 

 one habitat by individuals from the other 

 should cease, then the intergrading forms 

 would, through interbreeding with the ex- 

 treme forms, be gradually absorbed by them 

 and finally disappear altogether. In the 

 creek would then be found a form differing 

 clearly and constantly from the lake form 

 and without any connecting forms. Under 

 these circumstances the form in the creek, 

 as well as that in the lake, must rank as a 

 distinct species. 



This is the present condition, so far as 

 our investigations have enabled us to de- 

 termine. There is no difficulty in distin- 

 guishing individuals taken in the lake from 

 those found in the creek, and neither form 

 seems to invade the habitat of the other. 

 Large collections were made, not only of the 

 fishes inhabiting the lake, but also of those 

 in the creek. The latter was carefully 

 seined twice, from its source to its mouth, 

 and not a single example of E. xowcb or any 

 form showing intergradation was seen. 

 Similarly careful investigations were made 

 in the lake without discovering any individ- 

 uals of the creek form or any showing in- 

 tergradation. Whether further collecting 

 will discover connecting forms cannot, of 

 course, be stated. The small size of the 

 creek and of the lake, and their close geo- 

 graphic relation, render it almost certain 

 that individuals of the. one form would oc- 

 casionally invade the habitat of the other, 

 and vice versa. While the environment 

 of the creek is markedly different from 

 that of the lake, it is improbable that a 

 change from one to the other would prove 

 disastrous to the individuals concerned. 

 Some of such individuals would, it seems, 

 be able to survive, and some would prob- 

 ably interbreed with individuals of the 

 other form whose habitat they had invaded. 

 This was, quite likely, the condition in the 



beginning, and the creek form, so long as it 

 remained connected with the parent species 

 by the intergrading forms resulting from 

 such interbreedings, would be a subspecies 

 of the parent species. But, as already 

 stated, no such connecting forms have yet 

 been found and the form inhabiting the 

 creek is a distinct species. 



There is one other possible condition 

 worth considering. Let us suppose that, 

 after the creek colony had become well es- 

 tablished, and for many generations had not 

 intermingled in any way with the parent 

 species in the lake, the habits of one or the 

 other, or both, should change somewhat and 

 that they should again begin to invade each 

 other's habitat and to interbreed. The 

 result of this interbreeding would be the ap- 

 pearance of individuals possessing morpho- 

 logical characters more or less intermediate 

 between the lake and the creek forms. In 

 other words, individuals would be found 

 showing that the two forms intergrade and 

 placing them again in the relation of species 

 and subspecies. If we could hnow%h\s, to 

 have been their history, however, we should 

 certainly not place them in the relation of 

 species and subspecies. We should regard 

 them as two distinct species, and the indi- 

 viduals which seem to show intergradation 

 we would call hydrids, which they really 

 are. But we can rarely, if ever, hnoiu that 

 such has been the history. So long as inter- 

 gradations are found connecting the two 

 forms, the one last discovered must be re- 

 garded as a subspecies of the other. In the 

 present case, however, no intergradations 

 seem to exist, and the relation is that of two 

 distinct species. While the occurrence in 

 nature of hybrids is doubtless very unusual, 

 such a condition as the one supposed is cer- 

 tainly not improbable. 



And this suggests a further consideration 

 of subspecies. An examination of the de- 

 scriptions of forms which have been pub- 

 lished as subspecies does not show that the 



