Uniformity trials with multiple crops, 9 years on the same land

Uniformity trials with multiple crops, 9 years on the same land at Kansas Agricultural Experiment Station

Usage

data("metzger.multi.uniformity")

Format

A data frame with 468 observations on the following 4 variables.

row: row ordinate
col: column ordinate
crop: year and crop, see Details
yield: yield, see Details

Details

Individual plots are 1/20 acre. This is almost certainly 66 feet long (1 chain) by 33 feet wide. 33 * 66 * 20 = 43560 sq feet = 1 acre. Long, narrow plots are common in field trials.

Field width: 3 series * 33 feet

Field length: 12 plots * 66 feet.

1925. Corn. Bu/ac.

1926 Oats. Grain bu/ac, Straw tons/ac. Straw data look poorly measured.

1927 Wheat. Grain bu/ac, Straw tons/ac.

1928 Wheat. Grain bu/ac, Straw tons/ac.

1929 Alfalfa. Pounds/ac.

1930 Alfalfa. Pounds/ac.

1931 Alfalfa. Pounds/ac.

1932 Grain bu/ac, Stover tons/ac.

1934 Stover tons/ac.

Director's report 1920-1922. Cover has a fuzzy photo of the soil fertility experiment.

Director's report 1934-1936. Project 17: The great majority of the plots were quite erratic in their yield performance. A heavy yield one year was in many cases followed by a light yield the following year. The variability among the plots was highest in years of low rainfall. Hence it appears that under the conditions represented by these tests a uniform cropping trial is of decidedly limited value as a means of indicating relatively permanent differences in the productivity of a group of experimental plots. Source: Eighth Biennial Report of the Director 1934-1936, Agricultural Experiment Station, p. 29. https://www.ksre.k-state.edu/historicpublications/pubs/BR08.pdf

SB228 1922 "Relationship of Crop Yields to Quantity of Irrigation Water in Southwestern Kansas". Describes experiments having side-by-side series of duplicate plots, 33 feet by 66 feet = 2178 square feet = 1/20 acre.

In the source documents from Rothamsted, column 1 is series E, col 2 = series F, col 3 = series G.

Data acquisition: PDFs from Rothamsted archive were OCRd. Text copied and formatted by K.Wright.

This data was made available with special help from the staff at Rothamsted Research Library.

Source

Rothamsted Research Library, Box STATS17 WG Cochran, Folder 5.

References

Metzger, W. H. (1935). The relation of varying rainfall to soil heterogeneity as measured by crop production. Journal of the American Society of Agronomy, 27, 274-278. https://doi.org/10.2134/agronj1935.00021962002700040006x https://archive.org/details/dli.ernet.232701/page/273/

Examples

if (FALSE) { # \dontrun{
library(agridat)
data(metzger.multi.uniformity)
dat <- metzger.multi.uniformity

libs(lattice)
# qq plot for each crop
qqmath( ~ yield | crop, data=dat,
        main="metzger.multi.uniformity - QQ plot",
        xlab="Theoretical Quantiles", ylab="Sample Quantiles",
        scales=list(x=list(relation="free"), y=list(relation="free")) )

# Convert yield values to percent of mean within each crop
dat <- dat 
  group_by(crop) 
  mutate(pom = yield / mean(yield) * 100) 
  ungroup()

libs(desplot)
desplot(dat, pom ~ col*row | crop,
        flip=TRUE, aspect=(12*66)/(3*33),
        main="metzger.multi.uniformity - percent of mean within crop")

# reshape to wide format for splom
libs(tidyr)
dat_wide <- select(dat, row, col, crop, yield)
dat_wide <- pivot_wider(dat_wide, names_from = crop, values_from = yield)
splom(dat_wide[ , 3:15], main="metzger.multi.uniformity - scatterplot matrix")
splom(dat[, 2:14])
dat 

} # }