On the Genera Bambusicola and Arboricola. 387 



have examined some which do not in the least differ from 

 examples from Turkestan. This form has the underparts 

 white down to the vent, except the flanks^ which are brown ; 

 while the ujiper parts are as in the Baikal dark-bellied form, 

 but the head and nape are much lighter, in some old speci- 

 mens these parts being light whity-brown. 



In conclusion, it appears to me that all the above-mentioned 

 Dippers may be considered as slightly modified forms of the 

 same species, or may with equal justice be treated as very 

 closely allied species ; but I am inclined to favour the former 

 view. 



The measurements of the vai'ious forms do not differ very 

 much, as wnll be seen from the table on page 386. 



XXXIII, — On the Gallinaceous Genera Bambusicola and 

 Arboricola. By W. R. Ogilvie Grant. 



(Plate IX.) 



This paper contains a short review of all the known species 

 of the above-named genera, and gives a brief synonymy of 

 them, showing the various combinations of names^ generic 

 and specific, which have been used by different authors. 

 It also gives references to all the good figures. Keys to the 

 species and their geographical distribution are added, and 

 numerous notes. 



The number of tail-feathers in the various groups of 

 Gallinse is a very useful generic character when taken in 

 conjunction with others : it varies enormously in the dificrent 

 genera. For instance, in Ewcalfactoria the minimum number 

 (8) of tail-feathers is found; in Coturnix there are 10-12, in 

 Rhizothera 12, in Francolinus, as well as in Bambusicola and 

 Arboricola, 14, in Perdix 16-18, in Tetraophasis 18, in 

 Tetraogallus 20-22, and so on. These numbers are quite 

 constant, as I have ascertained by examining a large number 

 of specimens of nearly every species. The number of tail- 

 feathers can be accurately ascertained only by lifting up the 

 upper tail-coverts, so as to expose the bases of the rectrices. 

 When this is done there is usually no difficulty in counting the 



2c2 



