3 



also to the feet, the shape of the bill, and the structure of the tongue, the modifica- 

 tions of the palatal framework, and to the intestinal tract inclusive of its convolutions. 

 The smell of the Drepanididae is different from anything else I know ; but are we 

 prepared to admit this as a diagnostic character, until we know that it does not depend 

 upon the food, although Acrulocercus is devoid of that sort of scent 1 



Additional Data concerning the Melationship of the thick-billed Hawaiian Birds. 



In my former paper, p. 6, and fig. 18, I described the tongue of Psittacirostra as 

 " rather thick, hard, and homy, tapering out towards the tip, and while differing 

 considerably from the tongue of Loxioides, nevertheless truly Fringilline." Several 

 well-preserved specimens of Psittacirostra, brought home by Mr. Perkins, show a 

 different condition. The tongue is fleshy in its basal three-quarters, while the distal 

 quarter is thin and horny, slightly split in the middle, and with the thin lateral echoes 

 turned up and inwards, forming a very imperfect half-tube, and slightly frayed out 

 distally, i. e. at the anterior free end. It is consequently far less " Fringilline" than 

 the tongue of Loxioides. On the other hand it resembles, or approaches, that of 

 Pseudonestor, which is far less fleshy, more slender, more deeply split in the middle ; 

 the distal third of the horny sheath becomes gradually transparent towards the tip, 

 is very slightly frayed out towards the tip and on the lateral edges, but shows no 

 indication of curling up of the free margins. 



The tongue of Bhodacanthis, of which Mr. Perkins has brought home one single 

 specimen but well preserved, is the most compact of all. Its upper surface is slightly 

 scooped out, while the whole under surface is covered with the usual thick and hard 

 horny sheath, the thin lateral edges of which curl over upon the dorso-lateral sides, 

 and are very slightly frayed out at the distal sixth only, where alone they form slightly 

 sharp edges of the tongue, the tip of which is scarcely split at all. 



The shape and structure of the tongue is primarily referable not so much to the 

 nature of the food itself (if soft insects, as spiders, larvae, or hard seeds) as to the way 

 in which these various kinds of food are to be procured. The long- and slender-billed 

 birds probe flowers, or cracks of bark or lava, for insects, which they then coax 

 and brush out with their slender and flexible tongue; the thick-billed birds break 

 open the pods or worm-eaten trees, and then scoop out or simply nip the insects 

 or seeds. 



When arranged according to the tongue, Oreomyza and Psittacirostra assume a 

 somewhat central position, leading on the one side to Pseudonestor and ending with 

 Loxioides as an extreme ; while on the other side Oreomyza leads to the more complete 

 tubular brush-tongue of Chrysomitridops, Loxops, Himatione, and Palmeria, to the 

 extreme as represented by Vestiaria, Drepanis, Viridonia, and Hemignathus. 



A similar divergent development is traceable in the modifications of the operculum 

 and the nostrils. Himatione, Loxops, and Oreomyza assume a central, more indifferent 



2l2 



