The plots were scored by each of us and the combined score used as a measure of performance. The data
for the revised blocking were


Treatment
 1
 2
 3
 4
 5


 New Block
1 2

4 13
 10,17


3,3
5


5 8


We sweep as usual alternatively by blocks and by treatments. The figures hereafter are multiplied by 10 to
avoid decimals and reduce spatial confusion.


108 76


 -2 +22
 -8+62
-12-12 -13
 +8 -58
 +18


-20 +17
 -25 +45
-5 -5 +2
+16 -37
+16


-23 +17
 -28 +42
-4 -4 +6
+14 -36
+15


+14
+14

 +4
-6 -26


 +5
 -7

 +21
+1-19


 +4
 -11

 +21
+2 -18


Treat Mean
 +11
 +23
 -14
 -15
 -5

Block Mean

Treat Mean
 +1
 +4
 -3
 0
 -1

Block Mean

Treat Mean
 -1
 +1
 -1
 0
 0

Block Mean


 Residuals
 -13 +11
 -31 +39
+2 +2 -4
+23 -43
+23


 Residuals
 -21 +16
 -29+41
-2-2 +5
+16 -37
+17


 Residuals
 -22 +18
 -29+41
-3-3 +7
+14 -36
+15


The sum of squares of the residuals is 6010 (60.1 in terms of the original data). For comparison the Error
SS for the analysis using the original blocks was 58.3. The new blocks do not appear to be an improved
description of the pattern of variation between the plots.


Block Mean

Residuals


Residuals


Residuals


 +3
 -9

 +19
-1-21


 +4
 -11

 +21
+2 -18


 +5
 -12

 +21
+2 -18