

19 



tubular and brush-like portions of such tongues are formed entirely by the elonga- 

 tion, enlargement, and splitting of the ventral half of the horny sheath, while the 

 dorsal half or covering does not partake of this formation, but tapers out and 

 gradually disappears where the body or fleshy portion of the tongue ends. Indica- 

 tions of an elongation, with a frayed-out margin, of the ventral sheath occur in the 

 tongues of many Fringillidae, e. g. in Loxioides. Morphologically, we can derive a 

 brush-tongue from a Fringilline, but not from either a Sylviine, Laniinae, or Turdine 

 tongue. 



The smooth, not serrated, edges of the bill are a feature of the Coerebidse, while 

 those Meliphagidae which like them possess no serrated edges also differ in the formation 

 of their whole bill from others of their family. It is therefore not advisable to 

 compare the Drepanididae with the smooth-billed Meliphagidae. The general shape of 

 the bill differs so greatly in the various Drepanididae, and is subject to such alterations 

 in the numerous Fringilli formes, that no valid conclusions can be drawn from it. We 

 know for certain that the bill is a most adaptive organ, and the arguments concerning 

 the tongue apply still more forcibly to the bill. 



The nostrils, with their strongly developed opercula, seem to be decidedly Cinnyri- 

 morphous, and in the special description of the various Drepanididae this feature has 

 been laid stress upon in order to differentiate them from Loxioides and Psittacirostra, 

 which are Fringillidas. But here again the Drepanididae are intermediate, their 

 nostrils possessing both the upper and the lower flap, although the upper one is 

 never so complete as in most Meliphagidae. The completeness of the operculum is 

 correlated to the length and shape of the bill and to the use of the latter : hence 

 the variability. 



The condition of the primary quills of the Dicaeidae strongly favours their Fringilli- 

 form affinity, not merely because of the obsolete nature of the tenth quill, but also 

 on account of the entire absence of an eleventh quill. An eleventh quill does not 

 seem to occur in the Fringilliformes, although some of them have the tenth quill 

 not more reduced than many Alaudidae and Ploceidae. In many species of the latter 

 two families, and even in some Icteridae, the tenth primary is distinctly functional, and 

 comparatively of the same size as it is in many Meliphagidae and Nectariniidae ; but 

 these latter two groups have an extra quill, the eleventh. Hence it is not so much the 

 mere size of the tenth quill as its non-association with an eleventh quill that gives 

 it its taxonomic value in the Drepanididae. 



The rest of the pterylosis, the feather-tracts, affords us no help, because the Meli- 

 phagidae and the Fringilliformes seem to differ less from each other in these respects 

 than do the Cinnyrimorphae among themselves, notably Arachnothera and Promerops. 

 However, the fluffy nature of the feathers of the back, flanks, and axillaries in the 

 Drepanididae reminds us of the Meliphagidae and not of the Fringilliformes. The pattern 

 of colour affords no clue at all, because the red of Vestiaria and of Himatione, although 

 remarkable for its absence in all the Meliphaginae, is a favourite colour of the Myzo- 

 melinae, and the Fringilliformes, like the Psittaci, exhibit all conceivable colours. 



i 2 



