1916.] 15 



Tlie females of the second sreneration, roxae and anglica, can in 

 most cases, but not, I believe, in all, be distinguished from those of 

 the first by the paler hairing of the face and other slight chard<?ter8. 

 Nevertheless on account of the great differences between the males of 

 the 1st and 2nd generation in each case, it is not advisable that the 

 spring and summer forms should lie mixed in collections, nor that the 

 names eximia and aiiglica should be dropped. 



The remarks made by Edward Saimders in his " Hym. Acul. of 

 the British Islands " under A. rosae are rather unsatisiactory and not 

 altogether correct, and this is also the case with some parts of his 

 description. Thus the critical joints of the antennae in true roeae ? 

 are not like those of irimnieraiw in any of my specimens. 



On April loth, 1914, I was able to pay verv special attention to 

 the three spring species on Dartmoor, as they frequented the same 

 bushes, and again on the 18th, the only days on which the locality was 

 accessible to me until the 22nd, when the rarer species were alreadv 

 past. A. spinigera was numerous and frequently seen in cop., and 

 two pairs of the rarer J., eximia were taken in the same situation. 

 A. trimmerana J was numerous, but no $ was seen, yet no ^ was 

 seen to pair with either of the other species. This year I took a J 

 and several ? of true rosae in company on white TJmhelliferae, no 

 anglica ( = spinigera 2nd brood) occurring in the locality. In manv 

 parts of Devon A. spinigera is abundant and occurs regularlv each 

 spring, quite unmixed with A. eximia, while the latter is much more 

 local and much more uncertain, frequently failing altogether in its 

 known localities. 



Mr. F. Smith's collection contains only three males of true rosae, 

 but whether these are the three taken at Shirley with 12 $ in 1844, as 

 recorded in his '■ Bees of Great Britain" (1855), p. 52, there are no 

 labels to show. His other ^J (J are anglica, not true rosae. The $ $ 

 of his series of rosae (or aiistriaca as he latterly called it) consist of 

 about half-a-dozen true rosae and some sixteen examples of anglica, 

 the latter mostly from Sidmouth. The var. scotica of trimmerana is 

 from Loch Eannoch, and a similar form occurs in Ireland. 



Park Hill House, 



Paignton, South Devon : 

 Xoveiiiher 2iid, 191o. 



