1888.] PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 221 



This disappears again in the dorsal series, where they are closely in- 

 terlocked with each other, and the neural tube once more becomes con- 

 tinuous. For the rest we find that the "keteroccelous" plan of articu- 

 lation prevails among these vertebras thus far described ; that the 

 centra are much compressed laterally in the dorsal region, where also 

 the transverse processes are unusually wide and some of their spiculi- 

 form interlacements more than commonly broad. With the exception 

 of the atlas they are all pneumatic. 



The pair of free ribs that are attached to the sixteenth vertebra are 

 long and pointed, with free extremities. They do not, however, bear 

 epipleural appendages. 



Nothing peculiar marks the ribs of the dorsal series nor the hsema- 

 pophyses that connect them with the sternum. The epipleural append- 

 ages are large and all are closely, though freely, articulated with the 

 posterior borders of their ribs. 



The first pair of sacral ribs are like the dorsal ones, except they have 

 no epipleural appendages. The last two sacral pair, however, anchylos 

 with the pelvis, and their lnemapophyses do not reach the sternum. 



Of the sternum (Figs. 3 and 4). — Mergus has an interesting form of 

 this bone, and it differs in a number of points from the sterna of its sup- 

 posed nearest allies among the Ducks. The body is of au oblong out- 

 line and moderately well concaved above. Eight over the anterior 

 border in the median line there is a single semi-globular pit, but there 

 appears to be no pneumatic foramina of any size at its bottom. 



The costal processes are large, prominent, and quadrate plates. They 

 extend behind the first h asm apophyseal facet. These latter articulations 

 are six in number, and the lateral borders behind them are sharp, curv- 

 ing at first outward, before they extend backward, to the xiphoidil 

 margin. 



Upon the convex, pectoral aspect of the bone we are to notice the 

 principal muscular lines. These extend directly backward, one on either 

 side, from the lip of bone that overarches the outer end of the coracoidal 

 groove, to pass along the inner side of the vacuities behind, where they 

 become very faintly marked. 



A transverse straight line limits the xiphoidal extremity, and en- 

 grafted upon this in its middle we find a distinct convex prolongation 

 of no great size, its base being rather less than one-third of the border 

 upon which it occurs. 



Just over this latter, in the apertures of the postero-external angles 

 of the bones, we find on either side a large, oval fenestra. 



A sternum of this shape, differing as it does in this particular from 

 the notched style of the bone among most of the Geese and Ducks (for it 

 is the same as we find it in Glaucionetta), forms an exception to the 

 character laid down by Huxley for his Chenomorpluv, which includes the 

 subfamily to which Mergus belongs. (Fig. 3.) 



