BIRDS. 177 



(' Zool. Proc., 5 1865, pp. 593617, figs. 18). The scapula ( op. cit., figs. 36) is like that of 

 Caprimulgus, but larger and more delicate ; the coracoid is short ; the meso-coracoid process is 

 small in Cypselus (figs. 1 and 2), large and distinct in CluBtura (fig. 3). The epicoracoid spur is 

 absent, and this part is but little produced. 



The furcula (figs. 1 8) agrees with that of Caprimulgus, being U-shaped, strong, and having 

 even a still more distinct prse-coracoid, the flat top of which abuts against the lower surface of the 

 head of the coracoid (pp. cit., figs. 1 8) ; the inter- clavicle is also well developed, but does not come 

 near the sternal keel in Cypselus, Chcetura, or Collocalia ; it nearly reaches it in Dendrochelidon, 

 where it is nearly as large as in the Swallow (Hirundo rusticd). The structure and development 

 of the furcula answers evidently to that of the Hawk (Nisus), for the thick prae-coracoid wedge 

 does not appear to have an osseous centre of its own, but to derive its bony matter from the 

 clavicle. In Plate XIII, figs. 13, 14, I have shown the Sternum of Cypselus opus, magnified 

 two diameters ; fig. 14 shows it from above, and fig. 13 from the side. Its general form is wedge- 

 shaped, being of great length, narrow in front, and very wide behind ; its lateral margins are 

 concave ; its posterior margin sinuoso-arcuate ; its anterior, narrow end having the oblique 

 costal processes (c. p.) standing out like ears in front of the five costal condyles (c. c.). The 

 rostral process is oblong, and obliquely placed between the shallow coracoid grooves (cr. g.). 

 The keel is immense, gently concave in its front outline, and gently arcuate below ; it reaches 

 to the end of the xiphoid region, which is wholly ossified, and has a ribbed margin. Behind the 

 coracoid grooves there is a deep (primordial) fossa, which receives the air from the contiguous cells 

 (see fig. 14). In both figures the body of the bone is seen to be irregularly pierced with two rows 

 of fenestrae ; another double fenestra partly severs the first third of the keel from the rest (fig. 13). 

 Before I studied Dr. Sclater's paper, I had determined the meaning of these " fenestrae," namely, 

 as being irregular, because feeble, attempts at segmentation ; the upper rows as partly cleaving 

 the body of the Sternum into five parallel bands ; and the one on the keel tending to cut up that 

 part into succeeding segments. In Chcetura zonaris (op. cit., figs. 3 and 4) there is a very large 

 fenestra on each side, at the junction of the costal and xiphoid regions ; behind these there are 

 several small fenestrae. In Collocalia francica (op. cit., figs. 5 and 6) there is a similar deficiency 

 of the bone ; and Dr. Sclater mentions two openings in the keel, and quotes Mr. Gosse (' Birds 

 of Jamaica,' p. 59) as describing a similar structure in his Tachornis phcenicobia. But the unique 

 condition of the Sternum in Dendrochelidon Wallacii is most instructive, for this Bird has 

 true fenestrae as large as the closed notches of jEgotlieles cristatus. The long, oval inner 

 fenestra runs up to the penultimate costal condyle on each side, and a little behind its middle 

 the outer fenestra begins ; it is a long, irregular triangle, and has its smallest side parallel with 

 the broad, concave, posterior margin. The inner fenestra half divides the ento-sternum from the 

 lateral parts ; the intermediate xiphoid bar is narrow, and lies obliquely between the two fenestrae. 

 Here, let it be remarked, is the greatest tendency to a morphological breaking-up of this relatively 

 huge Sternum, in a case where, functionally, 1 most solidity is required. 



Family" TROCHILIN^E." 



The Shoulder-girdle and Sternum of the Humming-birds (see Hunt. Mus., Nos. 1519 1523) 

 are very similar to those of the Swifts ; but the strongly marked characters seen in these Birds are 

 23 



