42 The Wilson Bulletin — Xo. (j2. 



a. m. only four Purple Martins were seen flying about town, 

 where they had previously been common. Tlie contrast in 

 numbers was greater than had been anticipated. At noon on 

 the return to Macatawa the Martin flight was found to be still 

 under full headway as shown by the count of 92 birds between 

 11 :o9 and 13 :02 and of 237 between 12 :12 and 12 :i:>:.. The 

 area over which the Martins were abundant was found to 

 extend scarcely one-fourth of a mile inland. 



The wind changed but little during the day and following 

 night, and the next morning many new arrivals among the 

 Warblers and Flycatchers were seen. They formed the first 

 W^arbler "wave" of the fall migration. 



These observations are readily explicable on the assumption 

 that the direction of the flight of the birds corresponded with 

 that of the wind, while they seem to give no support to the 

 "beam-wind" theory. It seems probable that a series of care- 

 ful observations by persons living on the shores of Lake 

 ^Michigan, and on those of other large inland bodies of water 

 with north and south shore lines, might lead to a more satis- 

 factory knowledge of the relations existing between the 

 direction of the wind and that of tlie flight of migrating birds 

 in general. 



The temperature records made at Grand Rapids and other 

 points in Southern Michigan during the 14th, loth and Kith 

 oi August, 1905, show almost no change of temperature, and 

 so apparently eliminate that factor from the problem of deter- 

 mining wdiat agencies w-ere most involved in initiating the 

 particular migration flight just described. It is not often that 

 the change to a favoring wind is unaccompanied by changes 

 of temperature, but here was an opportunity to observe the 

 results following a change in the wind without the complica- 

 tion of accompanying temperature changes. 



The barometer was rising at the time of the flight, and con- 

 tinued to do so during the day and following night. During 

 the spring, extensive migration movements more commonly 

 take place at the time of a falling barometer. In the autumn 

 they usually occur with a rising barometer rather than with 

 a falling one. Hence neither a rising nor a falling barometer 



