606 



ON THl ANATOMY OF THE KINaFISniKS. [MaV 19, 



are which have it not. The expansor secundariorum is another 

 muscle which is sometimes absent and sometimes present. 



In marked contrast to the muscular anatomy (excepting the leg- 

 muscles, of which the formula seems to be always AX—), and to 

 the external characters, is the syrinx. I have examined this organ 

 in Alcedo, Dacelo, Cittura, Geryle, Halcyon, and Sawopatis, and 

 find it to be most uniform in structure. In all it ia of the typical 

 tracheo-bronchial form, vnthout a complete coalescence of the last 

 rings of the trachea, except sometimes in front. The intrinsic 

 muscles (a single pair) are well developed and fan out considerably 

 at their insertion onto the first, or apparently sometimes the first 

 and second bronchial semirings. In Dacelo cervina it is quite 

 plain that there are two pairs of intrinsic muscles. The most 

 anterior of these is the more slender ; the wider muscle arises 

 from the trachea just where the extrinsic muscles are given off'; it 

 covers OA'er the insertion of the first muscle and is pyramidal in 

 form, the first muscle being an elongated strip arising in common 

 with the extrinsic muscle. 



The Kingfishers being a group which shows so much diversity 

 in structure, the following tabular statement may be of use :■ — 



' -)- in one of two specimens of S. vagans. 



The above table not only displays the variation in structure of 

 the family but shows the impossibility of a subdivision of the 

 family, at least without further facts— for it is unnecessary to 

 point out specially the lacunsB in the above table. 



Halcyon is perhaps to be regarded as the simplest form, while 

 Dacelo and Sauropatis are at the opposite extreme. The necessary 

 separation of Sauropatis and Halcyon is the classiflcatory fact upon 

 which I would lay the greatest stress. It may be that the black- 

 billed species will turn out to be Sauropatis, and the red-billed 

 the true Halcyon, 



I would also point out the somewhat disappointing fact that no 

 particular results seem to be obtainable from a comparison of the 

 quintocubital with the aquintocubital genera. 



