70 VEGETATION OF A DESERT MOUNTAIN RANGE. 
intervals from 4,000 to 9,000 feet. All of the instruments in this series 
were located on the summits of ridges, so as to give comparable readings 
from similar topographic situations. In addition to the six instruments 
in this series there were also thermometers in the bottoms of cafions at 
5,000, 6,000, and 7,600 feet; a thermometer was exposed on the top 
of the fire tower of the Forest Service on Mount Lemmon, the actual 
elevation of the instrument being 9,225 feet, and thermometers were 
buried in the topmost layer of soil at 6,000 and 8,000 feet. 
Alcoholic minimum thermometers were used in the earlier years of 
these observations, but were replaced by mercuric Six’s thermometers 
in 1913 and. 1914. Various types of thermometer have been used at 
the station at 7,600 feet, and as many as three instruments have been 
exposed simultaneously at that place. All thermometers have been 
calibrated before use and have been verified in place from time to time 
by comparison with a portable thermometer of known error. The 
readings of the thermometers have been taken at irregular intervals, 
as opportunity afforded, and most of the figures secured are for periods 
of several weeks, or for the several months which elapse between the 
last visit in the autumn and the first in the spring. Only at the 7,600- 
and 9,000-foot stations has it been possible to expose the thermometers 
in such a manner as to secure reliable maxima; at all other stations 
the only data secured have been the absolute minima for the intervals 
between visits. The placing of the thermometers in small boxes, with 
numerous perforations, has made possible the securing of good minima, 
but no record has been made of the maxima secured under such con- 
ditions of exposure. The conspicuousness of adequate instrument 
shelters would have invited human interference with the thermometers 
which would have been productive of errors. 
A few records of temperature from the same locality for a number 
of consecutive days have been secured by Professor J. G. Brown and by 
Dr. H. A. Spoehr, as well as by the writer. A large number of single 
observations of minima and of current temperatures have been made 
by the writer at various localities, and it has been possible to use these 
in connection with data from the regular stations in determining the 
normal gradient of temperature decrease and in ascertaining the verti- 
cal shortening of the frostless season. 
LENGTH OF FROSTLESS SEASON. 
It has been impossible, for the most part, to make direct observations 
of the dates of last vernal and first autumnal occurrence of a tempera- 
ture of 32° F. at the several stations on the Santa Catalinas. The 
dates at which visits were made to the mountain were occasionally 
such as to establish the dates exactly for one of the stations, and in 
several cases visits were made at such frequent intervals as to place 
the date within a week or two. In the majority of cases, however, 
