agridat (version 1.16)

kalamkar.wheat.uniformity: Uniformity trial of wheat

Description

Uniformity trial of wheat at Rothamsted, UK in 1931.

Usage

data("kalamkar.wheat.uniformity")

Arguments

Format

A data frame with 1280 observations on the following 4 variables.

row

row

col

column

yield

yield, grams/half-meter

ears

ears per half-meter

Details

Kalamkar's paper published in 1932. Estimated crop year 1931.

Plot 18 of the Four Course Rotation Experiment, Great Hoos, at Rothamsted, UK was used. Sown with Yeoman II wheat.

Field width = 16 segments * 0.5 meters = 8 meters.

Field length: 80 rows * 6 inches apart = 40 feet.

The grain yield and number of ears for each half-meter length were recorded. This is quite a small field, only 1/40 acre in size.

Edge rows have higher yields. Row-end units have higher yields than interior units. These border effects are significant. Variation between rows is greater than variation within rows. Negative correlation between rows may indicate competition effects.

For ears, Kalamkar discarded 4 rows from each side and 3 half-meter lengths at each end.

Kalamkar suggested using four parallel half-meter rows as a sampling unit.

Note, the Rothamsted report for 1931, page 57, says: During the year three workers (F. R. Immer, S. H. Justensen and R. J. Kalamkar) have taken up the question of the most efficient use of land in experiments in which an edge row must be discarded...

References

None.

Examples

Run this code
# NOT RUN {
  
  data(kalamkar.wheat.uniformity)
  dat <- kalamkar.wheat.uniformity
  plot(yield ~ ears, dat, main="kalamkar.wheat.uniformity")
  
  # totals match Kalamkar
  sum(dat$yield)
  sum(dat$ears)
  
  require(desplot)
  desplot(ears ~ col*row, dat,
          flip=TRUE, aspect=(80*0.5)/(16*1.64042), # true aspect
          main="kalamkar.wheat.uniformity - ears")
  desplot(yield ~ col*row, dat,
          flip=TRUE, aspect=(80*0.5)/(16*1.64042), # true aspect
          main="kalamkar.wheat.uniformity - yield")

  # Show the negative correlation between rows
  ## require(asreml)
  ## require(lucid)
  ## dat <- dat <!-- %>% mutate(rowf=factor(row), colf=factor(col)) %>% -->
  ##   arrange(rowf,colf)
  ## m1=asreml(yield ~ 1, data=dat, rcov= ~ ar1(rowf):ar1(colf))
  ## vc(m1)
  ##  ##     effect component std.error z.ratio constr
  ##  ## R!variance  81.54      3.527      23        P
  ##  ## R!rowf.cor  -0.09465   0.0277     -3.4      U
  ##  ## R!colf.cor   0.2977    0.02628    11        U
  
# }

Run the code above in your browser using DataCamp Workspace