258 mr. a. h. garrod on the [May 5, 



their structure. A more extended series of observations may indi- 

 cate reasons for modifying the stress laid on some of the points ; but 

 it cannot, if my dissections are correct, change the facts themselves. 

 As an instance of the direction in which alterations may be shown 

 to be required, the case of the relation of Ptilonopus and Treron may 

 be taken ; for, notwithstanding their general similarity in structure, 

 these two genera present important points of difference. In the 

 possession of a gall-bladder by the former, it resembles Carpophaga ; 

 and it is quite possible that the similarity of the two genera under 

 consideration is the result of similarity in habit only, and that whilst 

 the latter (Treron) sprang from the Treronine stem direct, the 

 former (Ptilonopus) may have descended from the Phapine stock. 

 This appearance of the gall-bladder in these genera is very diffi- 

 cult to explain, except on the supposition that in those in which it 

 is absent it has been lost very lately, or that Carpophaga and Ptilo- 

 nopus are very intimately related forms. The uncertainty of its ap- 

 pearance in closely related genera of Mammalia reduces its value as 

 a character of classi6cational importance among them ; but its very 

 general absence in the Columbse makes it almost impossible not to 

 lav considerable stress, in classification, on its presence. In the 

 accompanying Table Ptilonopus is placed in the same division with 

 Treron because of their similarity in the points there laid stress on ; 

 but it requires a very inconsiderable amount more evidence to 

 necessitate its being removed from there into close proximity with 

 Carpophaga and Lopholeemus. The fact that two forms in many 

 respects so similar should differ so much in others is of itself a 

 strong argument in favour of their separation ; for the probability 

 that forces which tend to produce marked external resemblances 

 should simultaneously develop internal differences is, to say the 

 least, extremely small. 



Suborder COLUMB2E. Schizorhhaal birds with a characteristic 

 pterylosis, humerus-head, and sternum ; with the fourth gluteal 

 muscle, which in other schizorhinal birds covers the femur- 

 head, undeveloped ; with the oil-gland nude, when present, and 

 with the gall-bladder generally absent. 



Family ColtjmbidjE. Columbse in which the intestinal caeca 

 never exceed half an inch in length.) 



Subfamily Columbine. Columbidse possessing the ambiens 

 muscle, intestinal caeca, an oil-gland, no gall-bladder, and 

 12 rectrices. 



Genus Columba. 

 „ Turtur. 

 „ Macropygia. 

 ,, Ectopistes. 



Subfamily Phapine. Columbidae possessing the ambiens 

 muscle and no intestinal caeca. 



