April 26, 1894] 



NATURE 



611 



lated to lead to confusion, and to minimise, if not to destroy, 

 whatever advantas^es may be derived from their use. 



The time has therefore come when this question must be dis- 

 cussed and. if possible, settled, and, in the hope of leading to 

 such a settlement, I propose to point out what seem to me to 

 be the 'essential characteristics of such regions, considering 

 them to be established for the purpose of facilitating the study 

 of geographical distribution as one phase of the problem of 

 evolution. In order to come at once to the question at issue, I 

 will first summarise the statements or assumptions which seem 

 to me to imply a misconception of the nature and uses of zoo- 

 logical regions. These fall under two heads : — 



(i) It is asserted that the same regions will not answer to 

 show the distribution of all groups of land animals. Some of 

 the classes, or orders, or sonietimes even the families, require 

 us to establish different sets of regions — regions which may 

 differ both as to their number and their limits — in order to repre- 

 sent and study the distribution of such groups. 



(2) As a guide to what constitutes a region, it is laid down 

 that areas which have few peculiarities in the higher groups — 

 such as families, even though of continental extent, rich and 

 varied in genera and species, and having a large number of 

 peculiar types, are not of regional status. The criterion of a 

 region is said to be the exclusive possession of peculiar groups 

 of higher rank than genera ; and this without any regard to 

 proportionate area, or to the poverty and monotony of the 

 fauna as a whole. 



Now the first of these assumptions — that the same set of 

 regions will not serve for the study of the distribution of all 

 animals— raises the whole question of the nature and practical 

 utility of zoological regions, and is a proposition which the 

 chief purpose of this article is to disprove : it must therefore 

 be considered in some detail. 



In the first place, it implies that the students of any particular 

 group — reptiles, beetles, butterflies, land shells, &c. — should 

 each mark out the globe into regions exhibiting the chief features 

 of the distribution of its families, genera, and species, and that 

 any other division, arrived at by the study of other groups, will 

 be of little or no use to them. But if this is true, it mu?t be 

 carried further ; for not only do the various classes and orders 

 of animals differ considerably in their distribution, but many of 

 the tribes and families. To take the case of the mammalia, 

 which, for distributional study, has always been treated as a 

 whole, how different is the distribution of the Edentata from 

 that of the Ungulata. In the former group South America is so 

 rich that it is of more importance than all the rest of the globe, 

 while in the latter it is so poor that even when joined with 

 North America it would hardly equal either of the other con- 

 tinental regions in importance. But if we constructed a set of 

 regions to correspond with the distribution of each of these 

 orders, we should not bring out the facts more clearly than can 

 be done by means of the regions most usually adopted for the 

 whole class, while we should lose the ndvantage of easy com- 

 parison with each other, and with the remaining orders of the 

 class, as well as with other classes of animals. '^\i\. comparative 

 distribution is the one essential feature of our study, without 

 facilities for which the bare facts are uninstructive and of hardly 

 any scientific value. 



This point has been very clearly brought out in the case of 

 birds, in a work specially devoted to the geographical distribu- 

 tion of one family — the plovers. These birds are, a> a whole, 

 cosmopolitan, so much so that Mr. Seebohm tells us " they have 

 not even a remote connection" with the usually adopted zoo- 

 logical regions ; and he adds : " These birds only recognise 

 three regions — Arctic, Temperate, and Tropical." Again, 

 after describing the distribution of the chief genera during the 

 breeding season, he says : " The inevitable conclusion is that 

 the Charadriadas do pay considerable attention to the climatic 

 or isothermal regions, but appear practically to ignore the 

 Sclaterian regions." 



These very positive statements would lead a reader to con- 

 clude that here, at all events, the regions established by Dr. 

 Sclater for birds as a whole are of no use. Yet we find that in 

 the great work above referred to — the "Geographical Distribu- 

 tion of the Charadriada;" — Mr. Seebohm rarely uses these 

 three climatal regions, but throughout the book gives the distri- 

 bution of the species of each genus in terms of the six Sclaterian 

 regions. And if we consider the habits of these birds, so many 

 of which get their food on sea-shores and tidal estuaries, while 

 all of them have great powers of flight, and many of them 



NO. 1278, VOL. 49] 



migra'e along the coasts of all the continents, it is really sur- 

 prising to find so many of the genera and species which are 

 nevertheless strictly limited to certain of the Sclaterian regions. 

 Owing, no doubt, to the peculiarities of habit just referred to, 

 about half the genera are cosmopolitan, being found in all the 

 i six regions during some part of the year ; but, even of these, 

 \ certain groups of species are often confined to one or two 

 j regions. 



When we turn to the non-cosmopolitan genera, however, we 

 find some very instructive facts, which well serve to illustrate 

 my main conten'ion as to the sufficiency of one set of regions. 

 The following table gives the distribution of these genera, taken 

 from Mr. Seebohm's volume: — 



\ .(^dicnemus (Stone Curlews) [ All regions except the Nearctic. 

 j Lobivanellus (Wattled Lap- 

 wings) All regions except the Nearctic 



and Neotropical. 



Vanellus (Lapwings) All regions except the Nearctic 



i and Australian. 



j Cursorius (Coursers) Palsearctic, Oriental, and Ethi- 



1 opian regions. 



I Glareola (Pratincoles) All regions except the Nearctic 



I and Neotropical, 

 ! Ibidorhynchus (Ibis-billed 



Oyster-catcher) Palaearctic only. 



I Phalaropus (Phalaropes) ... Palsearctic and Nearctic, mi- 



j grating or straggling into most 



I of the other regions. 



Limosa (Godwits) All regions except the Ethi- 

 opian. 

 ■ Ereunetes (Snipe-billed Sand- 



j pipers) Paliearctic and Nearctic. 



I Phegornis (Short-winged 



Sandpipers) Australian and Neotropical 



Rhynchasa (Painted Snipes)... The four Tropical regions. 



Now we have here to notice two points : 



(i) That in most of these genera not only are they absent 

 from one or more of the S-lateriaii regions and present in 

 others, but in the regions where they do occur they are usually 

 widely dispersed, thus showing that their range is defined and 

 limited by the very same barriers which so well mark out the 

 general range of land birds. 



(2) We also find that the use of the old-established and 

 widely accepted six regions of Dr. Sclater, enables us very 

 clearly and concisely to describe or to tabulate the comparative 

 distribution of the genera and species of this great family of 

 wading birds, which have been thought to be such erratic 

 wanderers that the author of a work devoted to them declares 

 that — "the zoological regions of Sclater have nothing what- 

 ever to do" with them.. 



Now this case of the plovers is perhaps as strong as any that 

 can be brought to prove that different groups require different 

 setsof regions ; and it at once brings us to the question at issue, 

 which is, whether anything would be gained by establishing a 

 set of Charadriine regions. The climatic regions — which Mr. 

 Seebohm suggests as more natural in this case — would not 

 bring out such fact? as the absence of /Edicnemus from the 

 Neaictic and of Limosa from the Ethiopian regions ; the limita- 

 tion of Glareola to the eastern hemisphere, and of Phegornis to 

 the Australian and Neotropical regions, unless the Sclaterian 

 regions were also used as sub-regions — thus introducing com- 

 plication in place of simplicity, and gaining, so far as 1 can see, 

 no advantage whatever. 



But further, if the plovers are to have their own regions and 

 sub-regions, there are probably 50 or 100 of the orders and 

 families of the animal kingdom which would equally require to 

 be so treated ; and as in all these cases the new regions must 

 have separate names, it is quite clear that by far the larger part 

 of them would remain for ever unknown, except to their 

 inventors. 



A little consideration will, I think, convince us that this plan, 

 of practically unlimited distinct setsof regions, would be a posi- 

 tive hindrance to any intelligent study of the distribution of 

 animals, a study which derives its chief interest and importance 

 from its relation to ihe theory of organic evolution, and which 

 must therefore include the comparative distribution of the 

 various classes, orders, and rninor groups. But how will it be 

 possible to make the necessary comparison if the distribution of 

 the groups to be compared is given in terms of as many distinct 



