4 BULLETIN 339, U. 8. DEPARTMENT OF AGRICULTURE. 
The spikes in the tops of the stakes were leveled with the weir crests" 
by means of long carpenter’s levels, great care bemg exercised to 
keep them level at all times. It was found that excess velocity of 
approach, due to the filling up of the weir pools, was the most fre- — 
quent source of error in weir measurements, and special care was 
taken throughout the entire investigation to maintain weir pools of — 
sufficient size and depth to insure a slow velocity of approach. The 
formula used in computing the discharge over the weirs was Q=3.367 
L H?, where Q=discharge in cubic feet per second, L=length of crest _ 
in feet, and H=head or depth of water on the crest in feet. a 
Usually the head was measured every hour, but where the flow 5 | 
remained rather constant measurements sometimes were taken at 
two-hour intervals, and where much fluctuation occurred the head 
was measured oftener than every hour. Plate I, figures 1 and 2, 
should furnish a clear idea of the type of weirs used. The volume of — 
water diverted by the large projects was determined by daily gage = 
readings and current-meter determinations. 
DETERMINATION OF AREAS AND YIELDS. 
With the exception of the large projects, all areas included in this — 
investigation have been determined by transit and chain surveys. 
The transit notes were plotted on a scale of 100 feet to the inch, and _ 
the acreage was determined from the map with a polar planimeter. 3! 
It is believed that the areas of all of the tracts in this vari ation eX- 3 
periment have been determined within 0.01 acre. | 
The crops from the three plots into which each tract has been di- _ 
vided have invariably been cut, stacked, and‘thrashed separately, 
and weighed wherever possible. In cases where grain or alfalfa fields 
were located at long distances from a set of wagon scales and weigh—@ 4 
ing has been impracticable, the yields from the different plots have © 
been determined by the automatic weighers attached to the thrash- 
ing machines and by the measurement of the hay in the stack. 
While the determination of the yields by the latter method has not 
been absolutely accurate, the same method has been used with all | 
plots into which a tract was divided, thus insuring comparable | 
results. 
CLASSIFICATION OF SOIL. 
The soils on the different tracts have been classified with respect — 
to their mechanical composition only, chemical analyses having beous 
made in but few cases. Representative samples of the first, second, © 
third, and fourth foot of soil from each tract ees upon — 
have been secured and prepared and are now filed away for further | 
reference. The soils varied from the finest of adobe clays to the — 
coarsest gravels. The extremes of soil which have been included © 
