234 



correctly points out, and differs from S. purpurascens merely in having the back with a tinge of 

 green and the head purple without any tinge of green, whereas S. purpurascens generally has the 

 head and neck purple with a faint greenish or bronze tinge; but these differences are so slight, 

 and depend much on the light in which the specimen is placed, that I hesitate to acknowledge 

 this species as a good one. A specimen obtained by Mr. Michaeloffsk between Satchany and 

 Mzchet, for the loan of which, together with a series of selected specimens of Starlings, I am 

 indebted to Mr. Pleske, Director of the St. Petersburg Museum, has the entire head and neck 

 purple, the interscapulary region green, the scapulars, wing-coverts, rump, and underparts purple, 

 the abdomen violet-black, and the flanks purple. 



Sturnus caucasicus, Lorenz, Beitr. Orn. Cauc. p. 9, pi. v. fig. 1 (Sharpe, torn. cit. p. 37), is 

 another subspecies very closely allied, if indeed separable from S. purpurascens. Dr. Sharpe 

 certainly states that it has " the wing-coverts dark steel-green, externally glossed with purple," 

 which would be a character to separate it from S. purpurascens ; but this must be a mistake, as 

 in the original description Mr. Lorenz states that the wing-coverts are " violet," and one of his 

 original specimens, for the loan of which I am indebted to Mr. Pleske, has the wing-coverts 

 decidedly purple, exactly similar to S. purpurascens. This specimen differs only from S. polta- 

 ratsMi in having the rump glossed with green, whereas in S. poltaratskii the green does not 

 extend below the lower part of the back, the rump being purple. 



Sturnus purpurascens, Gould (Sharpe, torn. cit. p. 37). — Dr. Sharpe in his description says 

 that this species has " the mantle and back green," but Gould in his original description (P. Z. S. 

 1868, p. 219) expressly states that " the entire back is of a lovely purple," which is the distinctive 

 character claimed by Dr. Sharpe for his S. porphyronotus, which he states (Ibis, 1888, p. 438) 

 differs from S. purpurascens in having the back entirely reddish purple like the rump and upper 

 tail-coverts. 



Sturnus porphyronotus, Sharpe (torn. cit. p. 38, pi. ii.). — As above stated, the distinctive 

 character claimed by Dr. Sharpe for this species is the purple back, which is the character claimed 

 by Gould in 1868 for his S. purpurascens, and it therefore appears very doubtful if Dr. Sharpe's 

 will stand. After a careful comparison of a large series oi S. purpurascens and S. porphyronotus, 

 I can discover no difference except that in S. purpurascens there is in the centre of the inter- 

 scapulary region a slight gloss of steel-green, which is not apparent in S. porphyronotus ; but it 

 requires a good light to detect this difference, and I can scarcely consider it of specific value. 



Sturnus minor, Hume (Sharpe, torn. cit. p. 39), does not occur within the Western Pake- 

 arctic area. 



Sturnus unicolor, Temm. (Sharpe, torn. cit. p. 39), differs from all the other Starlings in 

 being uniform in colour, the head and back uniform and not differing in coloration. Full 

 particulars respecting this Starling are given in the ' Birds of Europe,' iv. p. 415. 



After a careful examination of a large series of specimens in my own collection and in the 

 British Museum, as well as a selected series from the St. Petersburg Museum kindly placed at 

 my disposal by Mr. Pleske, I arrive at the following conclusions : — 



Sturnus unicolor may be placed on one side, as it is quite distinct from any other species of 

 Starling. The remaining Starlings may be divided into two groups, viz., those having the wing- 

 coverts green or steely blue, and those having the wing-coverts purple. The first group contains 



