Please read the first three pages of this article about Split Plot Experiments.

The data is provided below - the values represent a measure of corrosion resistance.

Furnace Run	Temperature (C)	Coating 1	Coating 2	Coating 3	Coating 4
1	360	67	73	83	89
2	370	65	91	87	86
3	380	155	127	147	212
4	380	108	100	90	153
5	370	140	142	121	150
6	360	33	8	46	54

Create a dataframe with one row per observation with four variables, wholeplot, temp, coating, and resist - make sure the first three variables are factors. The classical split plot analysis of variance table may be produced in R using the aov function,

summary(aov( resist ~ temp*coating + Error(wholeplot)))

The sums of squares in the table produced by the above may be extracted from the following simpler analysis. Attempt to do so and be prepared to explain the logical relationships that justify this derivation.

anova(lm( resist ~ coating + wholeplot))
anova(lm( resist ~ coating * temp))