98 THE HABITAT 
comparison of corresponding level or factor curves, both with respect to 
position and character. 
142. The amplitude of all the curves described above is determined by _ 
the unit values of the factors concerned, while the length is dependent upon 
the number of stations, points, or times. The value assigned the latter 
upon the plotting paper is purely arbitrary, but it is most convenient to fix 
this at the centimeter square. The unit value for temperature is 1° Centi- 
grade per square, each subdivision of the latter representing 0.2, and the 
range being 22 degrees. For water-content curves, each square represents 
a value of 2 per cent, the smaller square being 0.4 per cent, and the range 
2-48 per cent. The unit value for humidity is taken as 5 per cent, making 
each small square I per cent, and giving room on the sheet for the entire 
range from I~Ioo per cent. Owing to the anemometer used, curves of 
wind velocity have been based upon the number of feet per minute. One 
hundred feet is taken as the unit value, and the range is from o-2200 feet. 
The unit value for the curve of light intensity is .o05. Each small square 
is .0o1, which permits a range from .ooI to .oI on one sheet. Consequently, 
when it is desired to plot the curve of a series of habitats with a range in 
intensity greater than this it is necessary.to use a double sheet. This is the 
usual device when the range of curves is too great, except where the excess 
is slight. In this case the curve is left open at the top, and the value which 
the crest attains is indicated. All curves in combination are labeled at the 
beginning or left to indicate the level, station, or point, and at the end or 
right to show the time, or day, if this is not the basis of the curve or series. 
The discussion that precedes deals exclusively with curves representing 
factors determined in the field. It applies with equal force to results 
obtained by instruments in control houses. In these, however, all factors 
except those directly experimented with, usually water-content and light, 
are practically equalized, and the curves based upon them are used chiefly 
to show how nearly equal they have become. The important curves aré 
those of the water-content series, both holard and chresard, and of the 
shade tents. Where several houses are differentiated with respect to tem- 
perature or humidity, curve series of both these factors are necessary. 
Factor Means and Sums 
143. It has been shown elsewhere that the daily mean of temperature can 
be closely approximated from the maximum and minimum of both day and 
night. Maximum-minimum instruments for the other factors are lacking, 
however, and for light, humidity, and wind these values can only be ob- 
