540 BRITIS 



in a greater or less degree of all the species. Neither the border of the 

 hind-wings nor the antennfe offer fixed characters, and in the latter a 

 difference of one-third in length will sometimes occur in the males of 

 the same species from a single locality. 



One may ask what is left on which a differentiation can be made. 

 Bateson discovers (ante, pp. 420-421) distinct differences between the 

 genitalia of A. filipenthdae, A. trifolii, A. riciac, A. e.ndans and A. pur- 

 puralis, yet he fails to distinguish those of A. lonicerae and A. trifolii, 

 and A. filipendulae and A. hippocrepidis, Stephs. A distinction in the 

 genitalia of A. Jilipendulae and A. lonicerae has not been sufficient to 

 prevent Fletcher from obtaining hybrids between these insects, nor 

 Standfuss from successfully crossing A. filipendulae with A. trifolii. 

 We have left for consideration, .^then, the habits of the insects, their 

 relation to their environment, their specialisation to their food-plants, 

 and similar factors which have caused them to take on those characters 

 which, in the mass, make us consider them as species. We are not in 

 a position to satisfactorily deal with the effect of environment on the 

 fixity of specific characters, yet a few considerations may be useful. 

 The Anthrocerid, which is best known to us as having a very specialised 

 environment, is A. e.ndans. It is an alpine and arctic species, 

 apparently incapable of carrying on its existence apart from a long and 

 severe winter, and an alpine flora. It haunts the marshes at the 

 sea-level along the shores of the Arctic Ocean, in Eussia and Finmark. 

 As we come south it gradually leaves the lowlying land, following the 

 gradually increasing altitude to which the arctic (and alpine) flora is 

 restricted, and reaches up the mountains of Scandinavia and Fin- 

 land as far as plant-life can exist. In southern Scandinavia and 

 Scotland it is rarely found below 1,500 ft. -2,000 ft. above sea-level, 

 and as we go south, the altitude at which an alpine flora is found con- 

 tinually increases, until in the Pyrenees, the Alps of central Europe, 

 and the huge mountain chains of central Asia, it never occurs at a lower 

 elevation than from 5,000 ft.-6,000 ft. A. e.ndans is distributed over 

 almost the whole of these ranges, in districts separated by thousands of 

 miles by lowlying land in which the species is never found. We have 

 only to reach a certain height to make sure of finding this insect, 

 slightly changed it may be, according to the district, but still undoubted 

 A. exulans, never coming below these comparatively barren regions of 

 the high alps, but going as high as vegetation can be found to 

 support the larva. Here, then, we have one of the most specialised 

 of the Anthrocerids, so far as its environment is concerned, and 

 here we have (except within certain very narrow limits) one of 

 the most constant, its constancy probably due to its isolation 

 through an almost incalculable period, for when an insect occurs 

 practically unchanged in the mountains of Scandinavia, Scotland and 

 France, the Pyrenees, the Alps, and the Altai, confined to a certain 

 elevation, and absent from all the intermediate grounds, which are of 

 immense extent, we must conclude either (1) that the species existed 

 in its present form at a time when the whole of the intermediate areas 

 were suited to its existence, and subsequent climatic changes have 

 driven it into still suitable places, or (2) that in common Avith many 

 other species, it was able to exist in the lower lauds under different 

 climatic conditions, and that the competition for existence being less 

 severe, it retreated into those districts which it now occupies, where 



