376 
rounded angles, retaining a breadth of 3 inches 3 lines at this end, which is devoid of 
the pair of notches characterizing, as a rule, the Anserine sternum’. 
Cnemiorms follows the rule of keelless, or rudimentally keeled, breast-bones of 
flightless fowl in the integrity of the bony shield. The length of each coracoid groove 
is 1 inch 6 lines, the greatest depth 14 line. From near the lateral end of the outer 
wall the pectoral ridges extend backward, slightly converging, but cease to be traceable 
after a course of 2 inches. ‘The costal process (d) is quadrate, relatively thicker and 
more produced than in Cereopsis or Tachyeres. The outer surface, defined by a low 
curyed ridge, is so smooth as to have suggested the remark at p. 242. The inner 
surface of the base of both right and left of these processes shows a large reticulate 
pneumatic vacuity. 
The costal border indicates the same degree of longitudinal curve, convex outward, of 
the coextensive part of the breast-bone as in Cereopsis; but is relatively more extensive, 
and is trayersed obliquely from within outward and backward by seyen articular promi- 
nences for the sternal ribs. The five anterior of these are ridges expanded at the ends 
into articular tubercles; the sixth and seventh are represented by the inner tubercle 
only, A smaller tubercle (ib. fig, 3,41), in advance of the broad ridges, may afford 
attachment to the heamapophysis of the second free rib. The breadth of this surface is 
shown in fig. 3. Cereopsis has but five articular prominences on each costal border. 
Tachyeres has seven, as in Cnemiornis. The outer surface of the sternum near the 
costal border is feebly concave transversely, before swelling into the convexity producing 
the hollow cavity of the anterior half of that bone next the thoracic abdominal cavity. 
It would seem that a comparison with the view of tracing affinity within the limits 
of the Lamellirostaal group could not profitably be made between the almost keel- 
less breast-bone of Cnemiornis and the deeply keeled ones in all existing members of 
such group; for even the sternum of the flightless Steamer-Duck has “the great 
development of the keel” which the experienced ornithologist Eyton adds to his 
osteological characters of the family Anatide*. However, there is a greater convex 
curye of the free border of the sternum in Cereopsis* than in Anser cygqnoides* or in 
Tachyeres; and, in asmall degree, this approximates 7 achyeres and Cereopsis to Ciconia. 
Limb-Bones. 
The coracoid (Pl. CIII. figs, 4-7) accompanying the collection of Cnemiornis bones 
now described, is of the left side, and wants only the terminal expansion fitting to the 
[ Eyton, ‘ Monograph on the Anatidme,’ 4to, 1838, pl. 1. figs. 7-11. Clangula (fig. 4) and Fuligula (fig. 5) 
agree with the Goosander (Mergus serrator) in the conversion of these 
‘achyeres adhere to the anserine type, 
* Eyton, in his classical Monograph (4to, 1838, p, 5), follows Vigors in making « Anatide ” (which suggests 
rather the tribe or subfamily of Ducks) the equivalent of Cuvier’s well-conccived term * Lamellirostres.”’ 
* Id, Supplement to ‘ Osteologia Avium ’ (4to, 1869), pl. ii, Cereopsis. “ Tb. pl. 3 
notches into foramina. Cercopsis and 
