150 
SEXUAL SELECTION: BIRDS. 
pi*' 
smaller and less perfect than on the other f 6 ^ 1 ^- 
with the upper parts of the external black ring®^,, 
ficient, as in the case just mentioned. The imp® ^ 
tion here seems to he connected with the fact tin 1 
spots on this feather shew less tendency than ngU1 ’ ^ 
become confluent into stripes ; on the contrary, th e y ^ 
often broken up into smaller spots, so that two o r 
rows run down to each ocellus. _ $ 
be 
We have now seen that a perfect series can ^ 
V 
lowed, from two almost simple spots, at first '^.Ji 
distinct from each other, to one of the wonderful s 
and-socket ornaments. Mr. Gould, who kindly 
some of these feathers, fully agrees with me in tk e ^ 
pleteness of the gradation. It is obvious that the ^ ^ 
in development exhibited by the feathers on th e 
bird do not at all necessarily shew us the steps " A 
have been passed through by the extinct progeu 1 ' t | 1( 
the species ; but they probably give us the clu e 
actual stops, and they at least prove to denn'^jil 
tion that a gradation is possible. Bearing 111 ||f 
how carefully the male Argus pheasant display!: 
plumes before the female, as well as the man) n’ 
rendering it probable that female birds prefer tin ^ 
attractive, males, no one who admits the ag e ' 1 ' jtl' 
sexual selection, will deny that a simple dark sp'-’ ^ 
some fulvous shading might be converted, 
the approximation and modification of the a ‘ 
spots, together with some slight increase of c f ^ 
into one of the so-called elliptic ornaments. 
latter ornaments have been shewn to many 
and all have admitted that they are extremely 1^ fv 
some thinking them even more beautiful th 1 _ pr 
ball-and-socket ocelli. As the secondary P^ u ^d 
came lengthened through sexual selection, ‘ l ^ 
the elliptic ornaments increased in diameter* 
