No. 387.] REVIEWS OF RECENT LITERATURE. 261 
he claims to be more generalized than those of the other genera. 
As one of these generalized characters he regards the small number 
of rectrices, assuming that the forms with more numerous tail-feathers 
have increased them in the course of development. Thus he remarks 
that “ Dendragapus has developed from Canace by ... a marked 
increase in the number of rectrices.”’ I think there can be no doubt 
that this is a fundamental error. The whole development of the tail 
of the birds shows that the generalized condition is that of numerous 
rectrices, and that the specialization, so far as number goes, con- 
sists in their reduction. I doubt very much that a pair of rectrices 
once /os¢ can ever be redeveloped. If a reduction has already taken 
place and for some cause or another the tail is to answer a purpose 
temporarily suspended during the course of the evolution of a form, 
other feathers which have not undergone the retrograde movement 
are taken into employ and developed so as to serve the same pur- 
pose. Thus in many cases the upper tail-coverts have had to func- 
tion as rectrices, with the result that very often it is difficult to tell 
them apart structurally from the true rectrices, while in one case, at 
least, the under tail-coverts have so completely assumed the rôle of 
the tail that they were generally regarded as true rectrices until it 
was discovered that the so-called tail-feathers were turned completely 
underside up! 
I believe it is equally erroneous to assume that a structure so 
highly specialized as the “top-knot” of the California Partridge 
(Lophortyx) could be changed into the simple structure of the Moun- 
tain Partridge (Oreortyx), and while admitting the possibility of its 
disappearance and the development of the other feathers on the 
crown into a loose crest like that of the Scaled Partridge (Callipepla), 
I see no necessity for such an assumption, especially as the latter 
seems to show'no apterium on the top of the head to account for 
the lost special tract. The more natural explanation seems to be to 
regard all these forms as having originated independently from more 
generalized types. 
But apart from this last objection, which is one of detail, it will 
not do to turn Dr. Clark’s trees upside down ; they will not give us 
the true ascent of the genera, for the reason that he has assumed 
them to descend directly from the most generalized living form. He 
has not taken into account the fact that the genus among the extant 
‘forms, which in its totality shows the greatest amount of generalized 
characters, may have acquired one or more highly specialized peculiar- 
ities, thus showing that it is somewhat outside the direct line of ascent. 
