70. Four plots were available for an experiment to compare clover accumulation for four different sowing rates ( Performance of Overdrilled Red Clover with Different Sowing Rates and Initial Grazing Managements, New Zeal. J. Experiment. Agric., 1984:

71—81). Since the four plots had been grazed differently prior to the experiment and it was thought that this might affect clover accumulation, a randomized block experiment was used with all four sowing rates tried on a section of each plot. Use the given data to test the null hypothesis of no difference in true mean clover accumulation (kg DM/ha) for the different sowing rates.

a. Test to see if the different sowing rates make a difference in true mean clover accumulation.

b. Make appropriate plots to go with your analysis in (a): Make a plot like the one in Figure 11.8, make a normal plot of the residuals, and plot the residuals against the predicted values. Explain why, based on the plots, the assumptions do not appear to be satis ed for this data set.

c. Repeat part

(a) replacing the observations with their natural logarithms.

d. Repeat the plots of

(b) for the analysis in (c). Do the logged observations appear to satisfy the assumptions better?

e. Summarize your conclusions for this experiment.

Does mean clover accumulation increase with increasing sowing rate?

Sowing Rate (kg/ha)

Plot 3.6 6.6 10.2 13.5 1 1155 2255 3505 4632 2 123 406 564 416 3 68 416 662 379 4 62 75 362 564