36 THE HABITAT 
check, the order is 10:2. It has already been shown, however, that these 
precautions are not necessary for ordinary purposes. In computing the 
holard and echard, there is no need to show the figuring, if the process is 
checked and then proved. Notés upon sky conditions aid in explaining the 
daily decrease in water-content. The amount of rain and the:period during 
which it falls are of great importance in understanding the fluctuations of 
the holard. Under community it is highly desirable to have a list of all the 
species, but it is impossible to include this in the table, and a glance at the 
formation list will show them. The form indicated above serves for a day- 
station series, a daily series in one station for any number of check series 
in one spot, and for scattered readings. In many cases the echard will not 
be determined, but on account of its primary importance, there should be a 
space for it, especially since it may be desirable to determine it at some 
later time. 
56. The permanent record. This should be kept by formation, or if the 
latter exhibits well-defined associations, the formational record may be 
divided accordingly. This may seem an unnecessary expenditure of time, 
but a slight experience in finding the water-content values of a particular 
habitat, when scattered through a chronological field record, will be con- 
vincing. The form of permanent record is the same as for the field, except 
that the column for the formation and that for the society are often un- 
necessary. 
57. Sums and means. From the great difficulty of determining the abso- 
lute water-content, and of obtaining a standard of comparison between soils 
on account of the varying ratio between bulk and weight, water-content sums 
are impracticable. For the same reasons, means of actual water-content 
are. practically impossible, and the mean water-content must be expressed in 
per cents. Daily readings are not essential to a satisfactory mean. In fact, 
a single reading at each extreme enables one to approximate the real mean 
very closely; thus, the average of 26 readings in the prairie formation is 
18 per cent. The extremes are 5 per cent and 28 per cent, and their 
average 16.5 per cent. A few readings properly scattered through moist 
and dry periods wili give a reliable mean, as will also a series of daily read- 
ings from one heavy rain through a long dry period. The one difficulty with 
the last method is that such periods can not well be determined beforehand. 
Means permit ready comparison between habitats, but in connecting the 
modifications of a species with water-content as a cause, the extremes are 
significant as indicating the range of conditions. Furthermore, the ex- 
tremes, i. e., 5 per cent and 28 per cent, make it possible.to approximate 
