J. E. DUERDEN 143 



first ordinary one are diminutive or vestigial, as if undergoing degenera- 

 tion in the same manner as described for the under-coverts. The third 

 row of upper-coverts is nearly complete in some southern birds, but 

 rarely in northern ; along with the succeeding rows it often shows a 

 rapid shortening from the proximal end forwards. 



The region of the elbow, between the plumes of the fore-wing and 

 those of the upper-arm, is also one in which plumage reduction is in 

 progress. In some southern birds no sharp line of separation can be 

 drawn between the two, the feather rows of the fore-wing passing un- 

 interruptedly into those of the upper-arm (Fig. 3) ; but in most northern 

 birds an irregular naked area occurs, and the proximal limits of the 

 different rows are then well defined (Fig. 4). In these cases the second 

 and third rows of upper-coverts are not continued upwards so far as to 

 end in the marginal row, but terminate above it. Thus the rows of 

 upper-coverts may be undergoing reduction at the elbow end of the 

 row, though the chief loss hitherto is from the distal end. 



From the above it is clear that, if the reduction of the upper-coverts 

 continues along the lines indicated, the outer surface of the fore-wing 

 will in time become as naked as is the present under surface, the 

 manner and order of succession of the losses being much the same in 

 both cases. In each a loss in the number of rows takes place, as well as 

 of the individual feathers from each row. A study of the one is supple- 

 mentary to a study of the other, and the results are confirmatory, though 

 at the present time the two surfaces exhibit very different degrees of 

 degeneration. 



In correlation with the number of rows of upper-coverts a marked 

 structural difference is displayed in the width of the fore-wing. In some 

 northern birds only the remiges and first-row coverts extend the entire 

 length of the wing, the other rows being reduced in number and scarcely 

 passing beyond the proximal half, while in some of the best Cape birds 

 two or three rows of coverts may be practically complete, followed by a 

 number of other shorter rows (cf Figs. 3 and 4). Taking the more 

 extreme cases it is found that where the rows are complete and numerous 

 the fore- wing structurally is proportionately broad, while where the 

 number is small it is distinctly narrower. Were the measurements of a 

 large number of wings to be taken and also the number of feathers borne 

 by them, there is no question that a distinct correlation ratio would be 

 disclosed. A marked degree of independence undoubtedly characterises 

 most of the directions along which degeneration is taking place, but it 

 is manifest that certain correlations may also appear. The words of 



