108 NORTHERN PAL^ARCTIC BULLFlXCnES. 



grading iuseusibly from a somewhat grayish " burnt carmine" down to 

 a just perceptible red wash over the deep cinereous gray. On the hack 

 the burnt carmine tint is deeper in Xo. 2 than in Xo. 1, and from these 

 gradually fading until Xo. 7, which shows the last trace of red on the 

 upper parts. 



It is impossible for me to draw a line anywhere in this series, but, 

 judging from Mr. Seebohm's original description (Ibis, 1882, p. 371), 

 where he particularly enlarges on the red color of the back, I presume 

 that he would refer the lirst six ones to P. rosacea. If we now look at 

 the localities given in the above table, it will be seen that among the 

 six first numbers are specimens from Kiusiu, Hondo, and Yezo, conse- 

 quently, from the three principal islands and from both sides of " Blak- 

 iston's Line ;" it is also clear that true P. griseiventris occurs both south 

 and north of that line. 



It may be said, however, that the above series proves but little geo- 

 graphically, since all the specimens are probably winter birds ; that 

 there would be nothing surprising in finding F. griseiventris migrating 

 south to Hondo during the cold season; and that even the appearance 

 of a true and typical P. rosacea from Hakodadi at that time of the year 

 is of little importance. 



I find, however, in the manuscript notes which Captain Blakiston 

 kindly placed at my disposal, several remarks which are of some con- 

 sequence in the present connection, for it is evident that Xo. of the 

 above table (Blakiston's Xo. 1U57) is by no means the only specimen 

 from Yezo with red on the back. Here are Captain Blakiston's remarks : 

 " Xo. 10.J7 [the one in U. IS. Xational Museum just alluded toj, S , Feb- 

 ruary, Hakodate, Jiush on bacl-j represented in Hakodate Museum by 

 Xo. 772, Hakodate, October, which has slight flush on back, and Xo. 

 1952, Hakodate, May, good deal flush on back." Judging from the 

 Avording, the latter specimen must be something like Xo. 2 of my 

 table above, and having been obtained at Hakodadi in May, it goes a 

 long way to prove that locality has nothing to do with the presence or 

 absence of red on the back or its greater or lesser intensity generally.* 



I have carefully gone over my whole series in order to ascertain 

 whether there might not be any other characters possibly distinctive of 



* Since the above was submitted for publication I learn from an article in the 

 Ibis, 18d7, p. 101, that Mr. Seebohm, whose type specimens of F. romceu came from 

 Yokohama, now considers this bird peculiar to Yezo and the opposite portion of the 

 Siberian mainland. Unfortunately I have onlj' one very gray F. (irisciventris from 

 Hakodadi to disprove this, and Blakiston's manuscript notes do not assist me in 

 this case, except that he mentions a Sapporo specimen without "Hush" on the back. 

 That Blakiston's notes prove nothing in this respect is due to the fact, however, that 

 he only made notes in regard to the red siiecinieus from Yezo, and not to the common 

 gray ones, which seem to be much more common. He who for twenty years had col- 

 lected in Yezo, was under the same impression as I, that F. rosacea was intended for 

 a supposed southern race. Nothing could prove more conclusively how worthless is 

 the claiui of F. rosacea to be regarded otherwise than a phase of the gray bird. 



