l896.j METALLIC COLOURS OF BIKDS. 289 



■well-developed lamina ; therefore in this case metallic barbules can 

 occur only towards the middle of the barb in contour-feathers, for 

 this is the only place where the lamina is well-developed. In 

 other words, in contour-feathers the metallic band cannot be, or is 

 not primitively, terminal. Where it seems to be terminal, as in 

 the feather described above, this is accomplished by a total or partial 

 suppression of the terminal barbules, the naked barb persisting 

 and forming the delicate fringe already noticed. This does not, 

 however, occur in quill-feathers to the same extent, because these as 

 a rule are characterized throughout by having barbules with very 

 well-developed laminfe. Thus it is by no means uncommon to 

 find terminal bands of metallic colour in the tail-quills of Humming- 

 birds, or we may find the whole surface metallic : there never seems 

 to be a longitudinal edging of metallic colour. 



Conviction as to the truth of the statement here made, that 

 metallic bands on the contour-feathers of Humming-birds are not 

 primitively terminal, is best attained by considering a series of 

 cases. 



In Eustephamis fialeritus, female, the breast is covered with greyish 

 feathers which have near their centre a spot of brownish pigment 

 which shows a faint metallic-green sheen. In the breast-feathers 

 of E.fernandensis, female, undoubtedly a more specialized species, 

 the spots are not much larger but the pigmentation is darker, and the 

 metallic colour is much more brilliant. This kind of metallic colour- 

 ing is very frequent, especially among the females of many species or 

 genera (cf. the species of the genus OreotrocJdlus). If the barb of 

 such a feather is examined microscopically, it is seen that all the bar- 

 bules with well-developed laminae contain pigment and are metallic. 

 Towards the base of the barb the barbules are colourless and downy, 

 towards its apex they are colourless and short, the lamina is rudi- 

 mentary, the filamentous portion is somewhat expanded and bears 

 only rudimentary cilia (PI. XI. fig. 11). Between this type and 

 that of the gorget-feathers of Basilimui leucoiis (PI. XI. fig. 12) 

 there are all stages in the suppression of these apical barbules. Thus 

 in Etistephanus fernandensis, female, itself the feathers of the poste- 

 rior region of the back are brilliant green, but are quite distinctly 

 tipped with white. In the specialized crest which this female is 

 almost peculiar in possessing this white edging is much reduced, 

 but the colourless apical barbules are still quite visible with a lens. 



In general, we may say that while the feathers of the specially 

 brilliant patches have barbs with naked tips, the ordinary metallic 

 contour-features have barbs which bear at their tips a series of 

 non-metallic barbules. Thus the green feathers on the back even 

 of Basilinna leucotis have a distinct border of pale rufous colour. 

 This is of some interest from the point of view of the development 

 of the metallic colours of Humming-birds. In PhaetJiomis eurynome, 

 one of the so-called 'hermit' forms without any brilliancy of 

 Colour, the feathers of the back are greyish black edged with a 

 broad band of yellowish colour ; the dark region exhibits a very 

 faint greenish sheen. If we begin with a type like this, the evolu- 



Phoo. Zool. See— 1896, No. XIX. 19 



