We could now have a look at the arithmetic means and standard deviations for yield per genotype ( G) and nitrogen level ( N).įinally, we can decide to fit a linear model with yield as the response variable. This type of split-plot design is the most common form, but there are many other forms of „the“ split-plot design depending on the design according to which the main plot and sub plot factors are randomized. Varieties, corresponding to the so-called sub-plot factor, were also randomized according to a randomized complete block design, taking main plots as block. It is important to recognize that nitrogen, the so-called main plot factor, was randomized according to a randomized complete block design. Separately for each main plot, the varieties were randomly allocated to the four sub-plots.
Every main plot was split into four sub plots to accommodate the four varieties. For every block separately, the six fertilizer treatments were randomly allocated to main plots. Each block was divided into six main plots. The field was divided into three blocks (replicates).
Split-plot designs are designs for factorial experiments, which involve two independent randomization steps. Desplot( data = dat, form = rep ~ col + row | rep, # fill color per rep, headers per rep text = G, cex = 1, shorten = "no", # show genotype names per plot col = N, # color of genotype names for each N-level out1 = mainplot, out1.gpar = list( col = "black"), # lines between mainplots out2 = row, out2.gpar = list( col = "darkgrey"), # lines between rows main = "Field layout", show.key = TRUE, key.cex = 0.7) # formatting