22 THE NAUTILUS. 



not extend far enough north to allow the localities in Keewatin, 

 Canada, to be plotted. 



While the exhibit as a whole affords what is probably a fairly 

 accurate view of the actual distribution, yet it also shows how 

 lamentably small our knowledge really is of the range of one of our 

 most common species, and how much still remains to be done before 

 any generalization can be other than tentative. Still the chart is 

 of interest as showing what we do know, and serves to indicate the 

 regions to which attention of American collectors should be particu- 

 larly directed. Thus the apparent absence of the species from the 

 entire littoral Atlantic region (with the exception of a single record 

 from Wilmington, N. C.) from the Potomac to Florida, and along 

 the coastal region of the Gulf from Florida to Texas is very striking 

 and worthy of investigation. So, too, the exact range towards the 

 west is important and of considerable scientific interest. No doubt 

 many of the deficiencies of the present map can be supplied from 

 material already accumulated in the many private collections of the 

 country, aud it is hoped that this review ot the subject will induce 

 the publication of all such information for the benefit of those inter- 

 ested in the geographical distribution of the American fauna. For 

 it is only by the accumulation of detailed information, such as is 

 afforded by the publication of carefully authenticated local lists, 

 accompanied by exact localities, that any substantial advance can be 

 expected. 



Taking the chart, such as it is, certain facts can be deduced with 

 a great degree of certainty, while others must still remain more or 

 less tinged with uncertainty until more exact knowledge shall either 

 prove or disprove them. On the north it seems reasonably certain 

 that hicarinatus extends through British America from Anticosti, 

 New Brunswick and Nova Scotia west to at least Lake Winnipeg 

 and Manitoba and north to Keewatin. In all probability it may be 

 found to range west through Saskatchewan and Assiniboia well 

 toward the Rocky Mountains. 



The data already accumulated shows a distribution through the 

 northern United States from Maine to Oregon. The northwestern 

 data, though scant, tend to show that the invasion of the valley of 

 the Columbia was through transfer from the headwaters of the Mis- 

 souri or its tributaries to the Snake, Pend Oreille or other tributaries 

 of the Columbia. 



