STUDIES OF THE MACROCIILRES. 



327 



and offers the proportional amount of articulatory surface to the 

 glenoid cavity. Its blade is narrow, rather long, of an equal 

 width throughout, and compressed from above downwards. Pos- 

 teriorly its extremity is obliquely truncate from within outwards, 

 while the end itself is slightly curved in the same direction. 



Either coracoid is characterized by a very extensively expanded 

 sternal extremity of a quadrilateral outline, and of no great thick- 

 ness in the antero-posterior direction. The shaft of the bone 

 a.bove this dilated end is rather slender, subcylindrical, being 

 compressed from before backwards, and is evidently hollow. Its 

 summit is not conspicuously enlarged, though it is rather more 

 tuberous than we find it in such a group, for instance, as the 

 Passeres. The head is directed in the articulated skeleton upwards, 

 forwards, and inwards. Its scapular process is not very wide, for 

 the scapula projects over it a little, both mesially and to its outer 

 side ; while in the former direction it stands between its superior 

 articulating edge and the corresponding head of the clavicle, i. e. 

 the scapula does. Air seems to gain access to the shafts of the 

 coracoids, and perhaps to some extent to the extremities of these 

 bones ; but, so far as I have been able to discover, neither the os 

 farcula nor the scapulae possess any pneumaticity. 



Neither of these Trogons possess, upon either side, the little 

 ossicle at the shoulder-joint known as the os Immero-sca^ulare^ 

 though it is just possible that it may in every case have been 

 removed by accident during the preparation of the specimens. 



Of the Pelvis and the Coccygeal Vertehroe. — No marked differ- 

 ences distinguish the pelves of these two species of Trogons. 

 There is some general resemblance between the pelvis of T. mexi- 

 canus and the bone as we find it in certain Caprimulgine birds, 

 though when we come to the details in such a comparison the 

 divergence is sufficiently marked. 



Viewing the pelvis of Trogon mexicanus from above, we observe 

 that the preacetabular area is considerably more extensive than 

 is the postacetabular (PI. XIX. fig. IJ^). The outline of this 

 upper surface is somewhat quadrilateral, its average width being 

 about equal to its average length. In this specimen there are no 

 existing vacuities among the diapophyses of the sacral vertebra). 

 One or two extremely minute ones are found in these positions iu 

 the specimen of T. ^uella among the ultimate vertebrae. 



Marked lateral extension characterizes the transverse pro- 

 cesses of the sacral vertebrae, more especially those three which 



