392 Anxai.s of the Carxi:gie Mlskum. 



possess a few in which it is jjresent. In specimens from Ilrie and 

 McKean counties, this spine is found more frequently, but not always. 

 The impressions of the hand, and the double row of tubercles of its 

 inner margin are often very indistinct, and chiefly so in young speci- 

 mens, so that it is sometimes hard to tell whether a particular individual 

 belongs to this variety or not, since also the shape of the rostrum 

 shows transitions to the typical form, and very young individuals of 

 the latter generally have a more elongate rostrum. Nevertheless, old 

 and full-grown examples of this variety are easily recognized. 



This form was first reported from western Pennsylvania by William- 

 son, who says that a few specimens of it have Ijeen taken in Squaw 

 Run, near Aspinwall, Allegheny county. Additional material in the 

 Carnegie Museum collected by Dr. D. A. Atkinson comes from Pine 

 Creek, below Bakerstown Station and from Chartiers Creek, Carnegie, 

 Allegheny county. The writer has collected this variety in Squaw 

 Run, and further in McKean county, in the Allegheny River near 

 Larabee, and in French Creek at Union City, Erie county. He has 

 found it abundantly in Erie and Crawford counties in the drainage of 

 Eake Erie : at Conneautville Station, Crawford county, in a small tribu- 

 tary of Conneaut Creek ; in Temple and Conneaut Creeks near Albion, 

 Erie county ; in Elk Creek near Girard, Erie county ; in 'Walnut 

 Creek at Swanville, Erie county. There are a large number of speci- 

 mens in the Philadelphia Academy collections from a tributary of the 

 Allegheny River near Port Allegheny, McKean county. 



In Allegheny and Crawford counties this variety was found with 

 the typical form, and was comparatively rare. The specimens from 

 Port Allegheny in the Philadelphia Academy were associated with a 

 much smaller number of typical C. bartoui. At Larabee, the writer 

 found only this form in the Allegheny River (associated with C. 

 obscurus), but the typical form was abundant in small streams. In 

 Erie county in French Creek, as well as in the lake drainage, this 

 variety prevails, the typical C. bartoui being very rare there in the 

 larger creeks, but the latter may be more abundant in small streams, 

 which have not been examined. In Erie county it was always associated 

 with either C. propinqniis or C. obsctiiiis. According to Williamson, 

 this form is generally found in that part of the streams where C. bar- 

 toui and propinqiius (correctly obscurus') come together, and this is 

 quite right. However, there are many streams where this variety is 

 not found at all, and the writer has never seen it in the southwestern 



