33 



northern deciduous forest, which normally does not reach south of the 

 Assiuiboine Biver, in an isolated case crops out again. Practically all 

 of that part within Manitoba has been made a forest reserve by the 

 Dominion government. We will thus see that a place that is virtually 

 a transposed portion of the country north of the limit of the prairie 

 region, far beyond the limit of spretus and totally unsuited for the con- 

 tinuation of that species, has been suspected, owing to a lack of definite 

 knowledge, of furnishing breeding grounds. It is the one part of the 

 country in which it may be positively stated that spretus does not breed. 



There is along the escarpment of the Turtle Mountains, extending 

 from about the point where the international line crosses the western 

 limit of the hills along the southern slope to the vicinity of St. Johns, 

 an oftentimes broken and narrow outcropping of a sandy substratum. 

 This area is covered with a very sparse growth of grass and was often 

 referred to by the residents as a probable breeding ground for spretus. 

 In fact, I was informed by one observer that he had seen a species, 

 presumably spretus, breeding there for years before the cultivated 

 lands had been invaded. After determining definitely that the moun- 

 tain proper was in no seuse the sought-for breeding ground, some 

 attention was devoted to this formation. Although doubtless more or 

 less suited for the deposition of eggs by locusts, it was found to be of 

 very limited extent. There will be found a few square yards upon the 

 brow of a hill, and perhaps no more will be found exposed for several 

 rods. Below this the alluvial soil of the low land along the southern 

 slope becomes well marked, and in the western part below this sandy 

 formation we find extensive fields of gumbo. Both of these are, of 

 course, quite impossible places for the permanent harboring of spretus. 

 After a succession of favorable seasons the locust might breed in this 

 narrow strip in numbers to invade the cultivated parts and cause 

 damage. But there were none here this year; the swarms doing dam- 

 age at this time did not originate there, and I am of the opinion that 

 this place never will furnish any considerable number of locusts. 



After I had arrived at the conclusion that it would be necessary to 

 look elsewhere for the origin of the swarms that from time to time have 

 come upon North Dakota and Minnesota than in or about the Turtle 

 Mountains, I received information regarding a swarm flying high on the 

 afternoon of the 17th of August over Whitewater Lake and in an 

 almost due southeasterly direction, far above the mountains, into North 

 Dakota. It was very remarkable that all the spretus along the northern 

 slope on the mountains had joined this swarm in motion. Where a week 

 before this species had been seen everywhere between Boissevain and 

 Deloraine, upon my return none were to be found. It had been expected 

 that swarms would pass over that region at about that time, since they 

 have always appeared by the 15th of August in Minnesota. On that 

 day, for the first time in a fortnight, the wind had changed from a 

 southerly direction and blew toward a point south of southeast. 

 11608— No. 22— -3 



