THE BIRDS OF AUSTRALIA. 
have here a different evolutionary product, though the source is at present 
unknown to me. For this group I propose 
Ceycoides, gen. nov., 
the type and only species being G&ycoides madagascariensis (Linne.) 
In coloration this group recalls the true Ceyx, but among the four-toed 
birds that genus is represented by Ceycopsis, which suggests once more 
convergence in colour-schemes. 
There are, apparently, then, m Africa three small groups of four-toed 
Kingfishers of small size, which are of distinct origin when we examine them 
from the point of view of the colour student. There are two groups if we 
ignore colour and utilise “ structural ” details alone. Once more we find 
these groups at discord, as the structural features would class together the 
two diversely coloured groups, separating the two similarly coloured groups. 
When we deal with the three-toed species we find still more discord if the 
same plan be followed. Thus, if we utilise the broad red bill we have blue 
and red Kingfishers classed together, and if we use the blue dorsal coloration 
we get red- and black-billed forms together. But we can trace the variation 
in coloration in the broad red-billed group, and see it paralleled in the narrow 
black-billed group. Further, we can see the red coloration almost unchanged 
in the four-toed series, and we have the typically coloured four-toed bird 
among the three-toed association. 
The conclusion would be that the loss of a toe could be achieved without 
change of coloration and that the red and blue three-toed groups had a perfectly 
distinct origin. Again, the immature of the red-biUed species have black 
bills, so that the latter are nearest to the original form. Such conclusions 
would indicate that the development of these three-toed forms has proceeded 
rapidly along two divergent lines, so that we might have from an unknown 
ancestor the group Ceycopsis and then Ceyx* 
Along the other more prolific stream we may see many tributaries. If 
Alcedo be regarded as ancestral (a dubious expedient) we can easily see the 
evolution of Ceyxalcyon^ and simultaneously the development of Cyanoceyx and 
Therosa. In these two groups we have colour alteration exactly paralleled 
in the pairs C. lepida and G. cajeli and T. soUtaria and T. meehi. As noted 
above, the extreme colour variation in Gyanoceyx may be due to the 
confusion of diverse elements in this group. An outlying and difficult group 
is Argyroceyx, especially on account of the under-coloration. 
It may be recalled that Sharpe realised the difficulties of these small 
* Since the preceding was written a study of the forms grouped under " Halcyon ” has suggested that 
Halcyon coromanda, now known as Entomothera coromanda, is a large relative of these small red Ceycea. It 
agrees in structural features to a great extent, while the coloration is very similar, and it differs altogether from 
all the other species classed in " Halcyon," which is otherwise polyphyletic. 
98 
