212 
THE CONDOR 
Vol. XVI 
reveals such an utter disregard of practically all of the various barriers serving 
as checks to most other animals, that one is rather at a loss to distinguish the 
factor or factors that finally limit its dispersal. Temperature and humidity, 
usually of prime importance, seem here, at first glance at least, to be without 
their customary potency. Rock Wrens are found from the floor of the hottest 
desert to the summit of the coldest mountain top ; while as strong a contrast 
is afforded by the aridity of these same deserts compared with the various 
islands on which the wrens are also abundant. The infinite variety of environ- 
ment which throughout western North America has produced such marked 
results upon many of the animals of the region, is without visible effect upon 
this unyielding organism, although over much of the country the birds are evi- 
dently non-migratory, a habit of life generally favorable to the production of 
variations. However, there is probable significance in the fact that the eastern 
limit of the Rock Wren in North America approximately corresponds with the 
eastern boundaries of the arid division of the Austral Zones, the main habitat 
of the species; so that it seems safe to say that it is the increasing humidity 
eastwards that finally acts as a check in this direction. The assumption is borne 
out by the fact that there is no marked change in the topography of the coun- 
try at this point. The species covers a part of the Great Plains region but does 
not extend over the whole of it. 
In local distribution, however, it seems apparent that features of environ- 
ment other than the variations in temperature or humidity encountered de- 
termine the boundaries. The relatively great difference in humidity between 
the Colorado Desert and the Pacific slope of southern California, for example, 
obviously is of no effect. The feature essential to the presence of the Rock 
Wren is open, unforested country. Furthermore, open plains, uniformly grass- 
covered, will not answer. There must be areas of bare rock, the steep walls 
of gulches or creeks, precipitous cliffs, or other similar surroundings. These 
provided, and there evidently are not in the part of North America occupied 
by this bird, variations of temperature or humidity, from valley to mountain, 
or from desert to ocean, sufficient either to check its distribution or to obvi- 
ously modify its appearance. 
Though the species is found over so much of western North America, it is 
useless to look for it amid forested country, and it is not frequently found in 
even moderately dense chaparral. Its northward dispersal along the Pacific 
coast is evidently stopped by the forests of the region. 
The islands off the coast of California and Lower California offer, in their 
barren and generally unforested condition, surroundings evidently highly fa- 
vorable to the Rock Wren, and the species has found its way to every one. 
Here, if anywhere, it would seem that variations from the general type should 
appear, for the birds are isolated on each of the islands, while the species is 
flourishing on all of them. This isolation in most cases, however, has had so 
far hardly any perceptible effect, and while there is apparently a slight general 
tendency of island birds toward the development of at least one feature, there 
are specimens at hand from each of the California islands which are not to be 
distinguished with certainty from the mainland form. 
Some years ago an insular form was described by Grinnell (1898, p. 238), 
Salpinctes ohsoletus pulverius, from San Nicolas Island. Its habitat was re- 
garded as confined to San Nicolas Island, but in a later publication the de- 
scriber (Grinnell, 1902, p. 68) extended its range to San Clemente Island also. 
This race was founded upon characters of structure and coloration, and 
