600 THE DESCENT OF MAN. 



be seen gradually to become more and more softened and 

 shaded into each other, with the upper lighter part toward 

 the left-hand corner rendered still lighter, so as to become 

 almost white and at the same time more contracted. But 

 even, in the most perfect ball-and-socket ocelli a slight dif- 

 ference in the tints, though not in the shading, between 

 the upper and lower parts of the ball can be perceived, as 

 before noticed; and the line of separation is oblique in the 

 same direction as the bright-colored shades of the elliptic 

 ornaments. Thus almost every minute detail in the shape 

 and. coloring of the ball-and-socket ocelli can be shown to 

 follow from gradual changes in the elliptic ornaments; and 

 the development of the latter can be traced by equally 

 small steps from the union of two almost simple spots, the 

 lower one (fig. 58) having some dull fulvous shading on its 

 upper side. 



The extremities of the longer secondary feathers which 

 bear the perfect ball-and-socket ocelli are peculiarly orna- 

 mented (fig. 61). The oblique longitudinal stripes sud- 

 denly cease upward and become confused; and above this 

 limit the whole upper end of the feather (a) is covered 

 with white dots surrounded by little black rings standing 

 on a dark ground. The oblique stripe belonging to the 

 uppermost ocellus (b) is barely represented by a very short 

 irregular black mark with the usual curved, transverse 

 base. As this stripe is thus abruptly cut off we can per- 

 haps understand from what has gone before how it is that 

 the upper thickened part of the ring is here absent; for, as 

 before stated, this thickened part apparently stands in 

 some relation with a broken prolongation from the next 

 higher spot. From the absence of the upper and thickened 

 part of the ring the uppermost ocellus, though perfect in 

 all other respects, appears as if its top had been obliquely 

 sliced off. It would, I think, perplex any one who believes 

 that the plumage of the Argus pheasant was created as we 

 now see it to account for the imperfect condition of the 

 uppermost ocellus. I should add that on the secondary 

 wing-feather farthest from the body all the ocelli are 

 smaller and less perfect than on the other feathers and 

 have the upper part of the ring deficient, as in the case 

 just mentioned. The imperfection here seems to be con- 

 nected with the fact that the spots on this feather show 

 less tendency than usual to become confluent into stripes; 





