Bases for the Prediction of Corn Yields

When it became evident that the Experiment Station plots at Ames formed too narrow a base for the development of correlations and regressions of preharvest measurements and yield, plots were established in the fields of cooperating Story County farmers to obtain moderately intensive records from fields specifically chosen to give a wide variation in soil fertility and yield.The present report is based upon these studies and is an attempt to deter- mine the preharvest measurements most likely to be of value in yield predic- tion, together with practical considerations of methods of collecting such data under field conditions.It is not assumed that the equations developed from these data will be of universal application.The principles and methods used, however, are thought to form bases for the prediction of corn yields.Methods Twenty-five areas of 8 x 8 hills were selected in farmers' fields near Ames in 1941 and 20 areas in 1942.These areas were chosen to represent poor, average and good yields and were not necessarily representative of the fields of which they were a part.Each experimental area was divided into 16 plots of four hills each.Half of these plots were used for all those measurements which would not affect their value as a record of final yield.The remainder were used for early harvests, early tests of shucked against unshucked ear measurements, etc.All of the fields in 1941 and all but one in 1942 were planted with hybrid seed.Eleven varieties were included in 1941 of which Ia-939, planted on eight fields, was the most popular; 1942 varieties were comparable.Soils varied from Dickinson sand, a terrace soil of low yielding capacity, through Wabash, Webster and Clarion sandy loams to silt and clay loams.The plots thus included a cross section of central Iowa varieties and soils.Measurements were made in four groups: measurements of stand, leaf area, height, basal stalk diameter, and date of 50% silking in July or early August; ear counts and ear growth measurements in August; ear size with and without shucks and early harvest yields in September, and final ear sizes and harvests in October.Actual dates of measurements were based upon stage of development, calculated from date of 50% silking, and varied be- tween fields and between years by as much as two weeks.Stalk and ear diameters were taken with calipers.Stalk diameters were taken as maxi- mum, or maximum and minimum diameters of the somewhat flattened stalks, in the first internode which was not blocked by brace roots.Ear diameters were taken at the center of the ear, either over the shueks (unshucked) or over the grain (shucked).Stalk height was measured after silking to the ligule of the uppermost leaf and recorded as maximum ligule height.Leaf measurements included number of active leaves per stalk and area (length x max.width) of the leaf arising at the same node as the upper ear (ear leaf).Ear length was measured with a steel tape as length of cob covered with