Vol. VIII] VAN DENBURGH AND SLEVIN— GARTER-SNAKES 251 



fectly good subspecific character. A small number of the 

 specimens also show an increased number of body scale-rows. 



Specimens from northwestern Nevada, as those from the 

 Pine Forest Mountains, Virgin Valley, and Quinn River Cross- 

 ing, in Humboldt County, appear to be intermediate between 

 this form and true T. o. vagrans, the coloration being typical 

 of the latter while a tendency toward an increase in the num- 

 ber of preoculars is still present. These are listed with T. o. 

 vagrans. 



In the region of Puget Sound snakes of the vagrans type, a 

 majority of which have two preoculars, are again encountered. 

 We can see no reason for not including them here. It seems 

 best to include here also the snakes from Del Norte County, 

 California, and from Josephine and Coos counties, Oregon, 

 although the number of specimens from these localities is so 

 small as to leave one in doubt as to the usual number of pre- 

 oculars, and the coloration is more like that of T. o. couchii. 



Perhaps nowhere else in the world are snakes so abundant 

 as near Klamath Falls. We counted a hundred and eighty on 

 a small rock about a yard in diameter in Link River, and, at 

 another point on the same river, caught fourteen with one grab 

 with both hands. ^ They feed upon small fish and toads. Most 

 of these snakes are of this subspecies, but a few are Thamno- 

 phis sirtalis infernalis. 



Thamnophis ordinoides couchii (Kennicott) 



Giant Garter-Snake. 



Diagnosis. — Normally with eight supralabials ; twenty-one 

 rows of scales ; no red in coloration ; dorsal line absent or indis- 

 tinct posteriorly, usually distinct on neck; usually some dark 

 markings on gastrosteges, preocular usually single ; infralabials 

 often more than ten. 



Type Locality. — Pitt River, California. 



'In June, 1918, some nine years later, they were not especially abundant here. 



