3o6 A HISTORY OF BIRDS 



rufous belly and no gorget ; H. tytleri (North-East Asia, Burma), 

 with a chestnut belly and a narrow gorget ; H. gutturalis (East 

 Siberia, Japan, India, Indo-Malay region, North Australia), 

 with a white belly and no gorget. Of birds regarded as 

 distinct species we may select H. lucida (Senegambia), with a 

 red forehead, white under parts and a black gorget ; H. dimidiata 

 (South Africa), with no red on the forehead, a white throat 

 and belly and only a half gorget ; H. nigrita (West Africa), 

 wholly steel-blue with a narrow white bar across the throat, 

 and no tail streamers ; H. atroccerulea (South-East Africa), 

 which differs from the last named in having no white on the 

 throat and long tail streamers; H. smithi (Africa, India), 

 wherein the red forehead has extended backwards to form a 

 red cap, while the rest of the plumage is blue above and white 

 below ; H. cucullata (South Africa), which has a red cap, blue 

 upper parts and buff under parts, streaked with dusky lines. 

 Yet other variants represented by other species might be found. 

 Thus several species have a patch of red oyer the base of the 

 tail, and this varies in extent and hue, while finally, two species 

 having green upper parts are known. 



All these are apparently descended from the Genus Phedina, 

 wherein the upper parts are of a dull brown, and the under 

 parts heavily streaked. In the red-rumped Mosque-swallows 

 the species have nearly all more or less streaked under parts, 

 and from the varying amount of red colour en the heads, and 

 the dinginess of the red on the rump of one or two species, it 

 would seem that these are intermediate and older forms, between 

 Phedina and the more typical species of Hirundo. 



To what shall we ascribe this ringing of the changes on 

 three or four colours ? To natural selection, sexual selection, 

 isolation, or discontinuous variation ? 



It would seem that this last is nearer the truth. We may 

 suppose, in short, that these permutations and combinations be- 

 gin as variations in the germ plasm, and that if slight at first, 

 the variation having started in this, or that, new direction goes 

 on increasing till it attains a maximum. What are the factors 

 governing this variation we know not ; but they have apparently 

 free play until, and unless, checked by natural selection. But 

 this sumptuary law is not called into action unless the particu- 

 lar colour variation tends to throw the subject thereof into 



