agridat (version 1.27)

garber.frames.uniformity: Uniformity trials of soybeans and wheat in artificially constructed frames

Description

Uniformity trials of soybeans and wheat in artificially constructed frames with homogeneous soil, at West Virginia Agricultural Experiment Station, 1929-1930.

Usage

data("garber.frames.uniformity")

Format

A data frame with 180 observations on the following 7 variables.

year

year, 1929 or 1930

crop

crop, soybeans or wheat

row

row ordinate

col

column ordinate = frame/plat number

framerow

row within frame

yield

yield, grams of hay (soy) or grain (wheat)

total

total weight, grams of straw (wheat only)

Details

At West Virginia Agricultural Experiment Station, concrete "frames" (walls) were constructed, each 9.33 feet x 4.66 feet, with 6-inch thick walls, 4 inches tall. Frames were arranged side by side (not end to end) in a single series. An 8-inch layer of topsoil was removed from each frame and set to one side. Then another 8-inch layer of subsoil was removed from each frame and set aside. The subsoil from each frame was then distributed among all frames. The topsoil from each frame was then distributed among all frames. In this way, the soil among all frames was made homogeneous.

Field width: 56 inches * 30 frames = 1680 inches = 140 feet.

Field length: 112 inches * 1 frame = 112 inches = 9.33 feet.

In 1929 the frames had 2 rows of soybeans, 28 inches apart. Harvested as hay. Weight recorded in grams for each frame/row.

In fall of 1929 the frames were planted to winter wheat, 4 rows, 12 inches apart, harvested in 1930. Weight recorded in grams for grain and total (straw).

Garber says the wheat data showed marked evidence of border effect--the outer 2 rows within each frame had higher yield than the inner rows. He used only the inner 2 rows for the anova. The large amount of variation in the data suggests that factors other than soil variability are the reason for the variability.

In the paper, 'plat' refers to 'col' (frame number).

Data provenance: Using pdf of original paper, tables were scanned by iPhone, then manually formatted and checked by K.Wright.

References

None

Examples

if (FALSE) {

  library(agridat)
  data(garber.frames.uniformity)
  dat <- garber.frames.uniformity

  # Aggregate framerows
  libs(dplyr)
  dat <- group_by(dat, crop, row, col)
  dat <- summarize(dat, yield = sum(yield, na.rm=TRUE),
              total = sum(total, na.rm=TRUE))
  dat <- ungroup(dat)

  libs(desplot)

  # Soybeans 1929
  dat29 <- subset(dat, crop=="soybeans")
  desplot(dat29, yield ~ col*row,
          tick=TRUE, aspect=112/1680,
          main="garber.frames.uniformity - 1929 soy hay")

  # Wheat 1930
  dat30 <- subset(dat, crop=="wheat")
  desplot(dat30, yield ~ col*row,
          tick=TRUE, aspect=112/1680,
          main="garber.frames.uniformity - 1930 wheat (grain)")
  
  # correlation
  cor(dat29$yield, dat30$total) # -.186 matches Garber p 101
  plot(dat29$yield, dat30$yield,
       xlab="1929 soybeans yield (g)",
       ylab="1930 wheat yield (g)",
       main="garber.frames.uniformity\n1930 wheat grain vs 1929 soy hay")
}