802 rROCEEDINGS OF THE NATIONAL MUSEUM. vol.x.xiv. 
however, are, at least to a considerable degree, migratory, but among 
man}'^ of the others, while there exists a greater or less individual 
inclination to wander during the winter, there is no regular nor well- 
delined movement. 
Altitude, unaccompanied l)y other influences, appears to have little 
or no potency as a modifying agency, for the present material shows 
that, notwithstanding some statements to the contrary, there is no 
appreciable difference either in size or color between mountain birds 
and those from adjacent lowlands. Neither are the boundaries of the 
currently accepted faunal areas, except in very restricted application^ 
of any value in circumscribing the ranges of the horned larks. In fact 
these birds, in many details of geographical distribution, seem to be 
quite anomalous. Why, for instance, is the form inhabiting the San 
Joaquin Valley, California, not the same as the one in the Sacramento 
Valley, as the similar conditions of climate and topograph}^ would lead 
us to expect'^ 
Quite interesting, however, is a comparison of the distribution of 
the American horned larks with that of the song sparrows, a group 
of approximately equal dispersion and similar plasticity. There are of 
each of these groups twenty-one recognizable races in Mexico and North 
America.^ Of these, four song sparrows and five horned larks are, 
during the breeding season, largely or entirely confined to Mexico; 
twelve of each are similarly peculiar to the United States; while five 
song sparrows and four horned larks are wholly or mainly British 
American and Alaskan."^ It should be borne in mind that in Mexico 
and the United States the song sparrows are not quite so widely dispersed 
as the horned larks; that in British America and Alaska the horned larks 
appear to be entirely absent from the area in which the greatest differ- 
entiation occurs among the northern song sparrows — the Pacific coast 
south of Bering Sea. Although corresponding thus fairly well in gen- 
eral distribution, there is of course considerable divergence in minor 
details. Particuhirly in the arid regions there seems to be a marked 
tendency among the horned larks toward strong differentiation within 
very limited geographical areas, so that the degree of difference is no 
criterion at all for extent of range. 
One peculiarity of widely dispersed types, the reduplication of forms 
in far se})arated localities, seems to be carried to the extreme among 
the horned larks — the American forms at least. The original case was 
mentioned )jy Dr. Dwight,^ but the present investigation has revealed 
at least five others in which the resultant bird is more like some dis- 
tant race than any of those whose range is contiguous or nearest. 
^Inthiy estimate no account is taken of Mclospiza c. juddi, as it is not a tenable 
subspecies. 
-Since tlic above was written two additional races of song sparrows liave been 
descril>ed, l)Oth from tlie Pacific coast of tlie United States. 
^'Auk, VII, April, 1890, pp. 139, 144. 
