106 PTERYLOGRAPHY. 



was here interrupted by a true or (in Dacelo gigantea, PI. VI, fig. 7) false gap, behind which it again 

 commenced as a sparsely feathered hinder part enlarged in all directions, and continued in the 

 same form to the caudal pit, at which it was contracted into a stronger rump-band two or three 

 feathers in width. By the side of this in A. coromanda, smyrnensis, capensis, and omnicolor, I 

 find scattered contour-feathers which extend to the lumbar tract and render this indistinct. In 

 A. cottaris the very long rump-band which reaches to the commencement of the pelvis is accom- 

 panied by two rows of contour-feathers at a moderate distance apart ; and the lumbar tract is 

 very distinct, and indeed strong : in A. rudis, maxima, and senegdensis, on the contrary, the 

 rump-band is somewhat broader, stronger, and definitely limited. All these species, like 

 D. gigantea (PI. VI, fig. 7), have a rather strong and broad lumbar tract. In the smaller Kingfishers 

 the number of remiges is twenty-two, in the larger ones twenty-four or twenty-five, of which 

 the third is always the longest, whilst the first and second remain perceptibly shorter, the former 

 being not longer than the tenth. In A. cottaris alone (which is further strikingly distinguished 

 by having the outer branch of the inferior tract very near the main stem), the first primary is as 

 long as the second, third, fourth, and fifth, which exceed all the rest in length. The number of 

 rectrices was twelve in all the species. 



9. AMPHIBOLY. 



The presence of an after-shaft on the contour-feathers is a pterylographic character 

 common to all the members of this family ; as also the occurrence of a circlet of feathers on 

 the tip of the oil-gland, and the number of rectrices, which is always ten. With these characters* 

 however, we seem to have exhausted the points of agreement of the genera belonging to the 

 group, and their essential differences necessitate the following arrangement : 



A. fPith the bands of the inferior tract narrow, and the outer branch distinct and freely 



divergent. 



This group includes the genera Corythaix and Musopliaga, of which the latter is pterylo- 

 graphically represented on Plate VI, figs. 8 and 9. Besides the species figured (M. paulina), I 

 have examined M. violacea and M. variegata (Phasianus africanus LATH.), as also the well-known 

 Corythaix persa. All agree perfectly with the figure here given, possessing a general covering 

 of feathers on the head (except when naked rings are present round the eyes), from which two 

 main tracts originate, one on the nape, the other on the throat. The former, the commencement 

 of the dorsal tract, passes as a simple narrow band to the anterior end of the trunk, where it 

 becomes a little dilated and then breaks off. Beyond a large gap situated between the shoulder- 

 blades, which, however, is less distinct in Corythaix persa and which undoubtedly constitutes 

 the chief pterylographic character of the Musophagidse, commences, at the extremity of the inter- 

 scapulium, the gradually widening hinder part of the dorsal tract, which gradually extends into 

 a sparsely feathered saddle, consisting, however, of rows of feathers running obliquely from before 

 backwards, in which the true main stem is still indicated by two closer central rows. It is only 

 in M. paulina that the femoral tracts remain distinctly separated from this large saddle, and the 

 portion situated on the tail is here only distinguishable as a narrow triserial band ; in all the 



