﻿CLASSIFICATION. 



xi 



b"". Naribus parvis, oblongatis 16. CarcineuUs. 



b". Rectricibus 10 . . 17. Tanysiptera. 



b'. Culmine simo vel sulcato. 



c". Culmine recto, integro 18. Cittura. 



d". Culmine versus apicem valde decurvato et sulcato 19. Melidora. 



Although it is very difficult to determine some of the above genera, I firmly believe that 

 they are all well established, the principal difference being found in the bill, as will be 

 seen by a reference to the plate. As, however, these differences are those of degree, I have 

 sought for secondary characters in my diagnosis of the genera, in order to aid the researches 

 of the student. First of all we have the most Alcedinine-looking birds of the subfamily in 

 the small section where the bill is longer than the tail ; and it is here we must first look for 

 the chain of affinity. This is speedily found in the genera Ispidina and Ceyx on the one 

 hand, and in the genera Corythornis and Alcyone on the other. I have already spoken of 

 the connecting links between Ceyx and Alcyone; and it is not hard to find another direct 

 affinity between Corythornis and Ispidina, not only in the genera, but in the species ; for it 

 would not be difficult to believe that Corythornis galerita and Ispidina leacogastra were one 

 and the same species at no very distant period of time. Secondary facts tend to confirm this 

 suggestion, especially since both species occur in the same limited district, being confined in 

 their range to the countries and islands bordering on the bights of the west coast of 

 Africa*. In ordinary specimens of C. galerita the difference is sufficiently striking; but 

 sometimes the old birds get very white on the belly ; and this circumstance, added to the 

 fact that C. galerita has a shorter crest than the other two species of Corythornis, renders it 

 by no means impossible to mistake an adult example for Ispidina leucogastra. Again, the 

 last-named species exhibits a slight keel on the bill, and is partly piscivorous in habit ; so 

 that it is by no means certain that Prince Bonaparte and Mr. G. R. Gray are so very far 

 wrong in placing it in Corythornis. Indeed the species may be considered either an aberrant 

 Corythornis or an aberrant Ispidina. Let no one, however, suppose that, on account of this 

 apparent connexion between the two genera, I consider them identical. Not in the least ; for 

 the true Ispidina}, such as I. picta and /. natalensis, are entirely different in form and in 

 habit. The Ispidince nest in holes of trees (never in banks), and, again, are purely insecti- 

 vorous. Some, such as /. madagascariensis, are only found in the thick forests, while others 

 frequent the banks of streams, but never feed on fish. Two species, which I have included 

 in the genus Ispidina in my ' Monograph,' are so decidedly aberrant that, on account of their 

 flat bills and longer tarsi, I propose to separate them under a distinct genus, which may be 

 called Myioceyx. The characters are given in the Plate of Genera (figs. G, 6 a, 6 b). Leaving 

 Ispidina madagascariensis as the last of the Ispidina 3 , we have to consider now the next step 

 in our approach to Melidora, the curious genus Ceycopsis from Celebes. This appears to be 

 a direct link between Ispidina and Ceyx, not only in form but also in combining the peculiar 

 coloration of the two genera. The absence of true Ceyx from Celebes has always been 

 noticed by Mr. Wallace and ourselves as a peculiar feature in the avifauna of that island ; 

 * This wo say advisedly ; for we do not believe that C. f/ahrita really occurs in North-east Africa. 



b 2 



