60 SCIENCE PROGRESS. 



his material permitted, a statistical tabulation of results. 

 Here again the variation is in colour of plumage, which may 

 take various dark metallic shades not easily described in 

 words. Concomitant variation in size and proportions are 

 also shown to occur. The facts as stated by Chapman are, 

 broadly speaking, that Quiscalus cenetcs has a breeding area 

 from the Rio Grande Valley northwards to British America, 

 and north-eastwards to New Brunswick. In the area south- 

 east of this Q. quiscula occurs, ranging from Florida to 

 Massachusetts. In Pennsylvania and Massachusetts (pro- 

 bably also in a belt of country extending south-west from 

 Massachusetts through Pennsylvania to the north of the 

 Mississippi) the two forms meet and intergrade. In the 

 rest of its range Q. ceneus is remarkably constant. The 

 complexity arises from the fact that Q. quiscula occurs in a 

 variety of forms which can be regarded as three phases of 

 coloration. Phase No. i is found in S. Florida associated 

 with Phase No. 2, and many intermediate forms. At the 

 north of its distribution Phase No. 1 is absent, No. 3 being 

 there chiefly found, associated with No. 2 and all intermedi- 

 ates. In the central part of its range all three phases and 

 intermediates occur together. The transition from ceneus to 

 quiscula occurs exclusively through phase No. 3 of the 

 latter. 



From the particulars given it seems likely that though 

 the transition is oradual in the sense that all intermediates 

 occur, being indeed frequent in the area of intergradation, 

 yet here again the connecting links are not a series of nor- 

 mals following each other in succession along lines passing 

 from the area of one race to that of the other. Another 

 example well known to ornithologists is that of the Rollers, 

 Coracias indicus and affinis (see Dresser (3) and Sharpe 

 (5)). Here again the coloration of the two races or species 

 is quite distinct. C. indicus inhabits the Indian Peninsular 

 generally, extending westwards to Asia Minor, while C. 

 affinis belongs to British Burmah and Indo-China. It 

 spreads, however, westwards to the neighbourhood of Cal- 

 cutta and onward to Sikkim and Nepal. This part of its 

 distribution is common to the two forms, and intergrading 



