DR. L. F. DE BEAUFORT, BIRDS FROM DUTCH NEW GUINEA. 391 



Kaiser Wilhelmsland. Consequently Japen (Jobi) stands somewhat more apart and has — although 

 small — affinities to the West in the same way as Kaiser Wilhelmsland has affinities to the 

 East. Four species inhabiting the Papuan isles in the Pacific are also found în German N. 

 Guinea and so it is of spécial interest that we found Munia spectabilis, hitherto only known from 

 New Britain (New Pommerania), at Humboldt Bay. From this place and the rest of the northcoast 

 to Geelvink Bay 5 species hâve been described, that hâve not been found elsewhere up to now, 

 while the number of species that this part has in common with Jobi and German N. Guinea 

 has been told above. 



YVhen I spoke hère of „species" I mostly meant „subspecies" or geographical repré- 

 sentative forms, which are of still more value to our discussion, as.they are not likely to be 

 found in other parts, where they are replaced by other subspecies. I am obliged to make 

 a little remark on the notion ^subspecies" in modem ornithology, as I am afraid that, 

 as it is considered now, systematic ornithology is on a wrong and dangerous path. To begin 

 with, what is a subspecies, considered from a physiological point of view ? It is not easy to 

 find an answer to this question in ornithological literature and there is rnuch diversity of 

 opinion, although thèse hâve never been clearly exposed. However it seems to me that 

 most ornithologists consider subspecies as varieties, caused by the différent external circum- 

 stances in différent localities from a species with a wide range; otherwise I cannot understand 

 why the rule is so severely stuck to that two subspecies of one species never can occur on 

 one and the same place. 



Following this rule forms are often considered as species, which would doubtlessly be 

 called subspecies, if not one or the other unlucky form overlapped the area of his neighbour, 

 e. g. the species of the genus Myristicivora (cf. Nov. Zool. VIII p. 116). On the other hand 

 there are many species that so gradually change from one extremity of their area of distri- 

 bution to the other, that when spécimens of both extremities are compared, they differ 

 considerably, but thèse différences are smoothed away when a large séries from the whole 

 range is examined. It is not impossible that we hâve hère examples of the origin of species 

 by isolation, and that the gradually changing species are the first step, while the overlapping 

 subspecies are the last in that direction. As long as experiments are not carried on, and I 

 don't think this ever will be possible with birds (butterflies would be better material for 

 studying the subspecies question), we can only make hypothèses, but we must be careful not 

 to grow dogmatic and consider every geographically separated form as subspecies and ail 

 varieties occurring together as distinct species, however little they may differ inter se. To 

 prove that I am not fighting against windmills and that there is real danger that arguments of 

 geographical distribution are getting too much power against morphological ones in systematic 

 ornithology I will cite one of our most eminent ornithologists in one of his latest papers. 

 Speaking about Casuarius casuarius bistriatus from Dutch New Guinea, which is very closely allied 

 to C. c. beccarii, he says that the author of this subspecies „ought not to hâve confused it even 

 for a moment with C. c. beccarii as I specially state in my monograph that C. c. beccarii is 

 confined to Wokan Island in the Aru Islands." When such a weight is laid upon geographical 

 distribution, we not only run the risk ofneglecting the natural affinities of the différent species, but 

 what is still worse, such species will be absolutely worthless material for the study of zoogeography. 

 In the systematic list of the birds collected I hâve generally cited the Catalogue of 

 Nova Guinea. V. Zoologie. 50 



