270 NORTH AMERICAN MUSTELID.E. 



line in general is therefore wedge-shaped. The rostral part of 

 the skull (all that in advance of the zygomatic arches) is about 

 one-third, or rather less, of the total length. The sides of the 

 rostrum are approximately parallel in old skulls, owing in a 

 measure to the swollen tract of rooting of the canines, in young 

 specimens somewhat tapering; the nasal extremity is abruptly 

 narrower than the rest; the obliquity of bevelling of the nasal 

 aperture is about 45o. In old skulls, the nasal and maxillary 

 sutures are obliterated ; in young ones, the nasal bones are 

 seen to be narrow, with approximately parallel edges for the 

 anterior half of their length, where they begin to narrow, and 

 extend as slender acute processes very far back — to opposite 

 the middle of the orbits. Their suture with the superior max- 

 illaries is very brief; for the intermaxillaries reach far up, and 

 for nearly all the rest of their extent they are received betwixt 

 long, pointed processes of the frontal. Similarly, the superior 

 maxillary runs up in a recess of the frontal to a point opposite 

 the ends of the nasals. This deep wedging of lateral processes 

 of the frontal between processes of the nasals and maxillaries 

 forms a complete letter W, better marked than in any other 

 North American genera of the family, though they all, excepting 

 Lutra and Unhydris, show an approach to the same character. 

 The anteorbital foramen is large and rather triangular than cir- 

 cular. The orbits are much better defined than in Mustelhue 

 and ^[epliltinw — not that supraorbital processes are stronger 

 than usual, but because the zygoma sends up a spur to mark 

 the orbital brim below — much as in the others. The approxi- 

 mation of these two (zygomatic and supraorbital) processes 

 completes about two- thirds of a circle. The point of greatest 

 constriction of the skull is a little back of the supraorbital pro- 

 cesses, at a point about midway in the whole length of the 

 skull ; except in some very old skulls, the constriction is little, 

 if any, greater than that of the interorbital space. The top of 

 the skull is marked with an average sagittal crest, whence 

 forks curve outward to the supraorbital processes. In young 

 specimens, there is little or no trace of these ridges. The occip- 

 ital crests appear more flaring than they really are, owing to 

 the general breadth of the skull behind ; they are in fact only 

 moderately developed in the oldest specimens, excepting at 

 their lateral extremities. From a moderate median emargina- 

 tion, the crests proceed on either hand with a moderate con- 

 vexity, which suddenly increases at the bend around to the 



