770 The American Naturalist. [September, 
much of the increase in number is to be sae by the 
division of hair follices and how much by new formation. 
So, too, in the replacement of molted hairs it is doubtful 
whether a new papilla is formed, or 
whether the old papilla retains its 
functions. 
When we analyze the phenomena 
of hair formation we find that the 
epidermis takes the initiative so far 
as cell multiplication is concerned. 
With feathers, on the other hand, the 
increase in cells begins in the dermis, 
the result being a slight elevation of 
the surface of the skin. Next th 
deeper cells of the epidermis form 
themselves into a double layer (fig. 
5) and the whole is strikingly sugges- 
tive of a scale, in that one edge of 
the elevation projects more than the y 
other. This outgrowth increases in 
extent until there results a cylindrical process protruding from 
the skin (fig. 6) with a very slight insinking, the rudiments of 
the future feather follicle, 
at its base. In this out» 
growth, which is to give 
T rise to the future down- 
a n SET ithe development feather, both dermis and 
epidermis; ep, splivichideac d, epidermis may be recog- 
deri TEs 
The epidermis has increased in thickness by cell multipli- 
cation, while the epitrichium retains its primitiye condition. 
Fic. 6. Pier esr section of a later oo of the down ecm after Davies, 
bv, blood vessel ; i ermis; ep, epitri ium: >p pulp ra) illa. The 
basal layer of iee Soli ermal cells have pict a cylindrical ch 
