A two-way layout with missing values:
  Balanced Incomplete Block Designs (BIBD)

Starting random-number seed: 123456
(Enter a number on the command line to specify a different seed.)

FIRST DATASET:

Reactions of rats to chemical substances on their skin
Text (Hollander & Wolfe, 2nd ed) Table 7.13, p315
Data Source: Mendenhall (1968):
  (7 treatments, 7 subjects)

Data:
          A        B        C        D        E        F        G   
--------------------------------------------------------------------
Rat#1   10.2      6.9              14.2                              
Rat#2                      9.9     12.9              14.1            
Rat#3            12.1     11.7               8.6                     
Rat#4                              14.3      9.1               7.7   
Rat#5             8.8                                16.3      8.6   
Rat#6   13.1                                 9.2     15.2            
Rat#7   11.3               9.7                                 6.2   
--------------------------------------------------------------------
Avg:    11.5      9.3     10.4     13.8      9.0     15.2      7.5   


Midranks:
          A        B        C        D        E        F        G   
--------------------------------------------------------------------
Rat#1     2        1                 3                              
Rat#2                       1        2                 3            
Rat#3              3        2                 1                     
Rat#4                                3        2                 1   
Rat#5              2                                   3        1   
Rat#6     2                                   1        3            
Rat#7     3                 2                                   1   
--------------------------------------------------------------------
Sum:      7        6        5        8        4        9        3   

Data is a BIBD design with
  k=7, s=3,  n=7, p=3,  la=1,  p*(s-1) = la*(kk-1) = 6

Durbin-Skillings-Mack statistic  D=12
  Large-sample chi-square approx:  P=0.06197  df=6

Carrying out 100,000 permutations to estimate
  the exact P-value for D=12:
Observed permutation score = Sum of squares of
  treatment-group rank sums = 280
    P = 1891/100,000 = 0.01891  95% CI (0.01807, 0.01891, 0.01975)

SECOND DATASET:

Logarithms of toxic dosages of seven chemicals to kill 95% of aphids
Text (Hollander & Wolfe, 2nd ed) Table 7.12, p311
Moore and Bliss (1942):
  (7 treatments, 7 subjects)

Data:
          A        B        C        D        E        F        G   
--------------------------------------------------------------------
Run#1   0.465    0.343             0.396                             
Run#2   0.602             0.873             0.634                    
Run#3                     0.875    0.325                      0.330  
Run#4   0.423                                        0.987    0.426  
Run#5            0.652    1.142                      0.989           
Run#6            0.536                      0.409             0.309  
Run#7                              0.609    0.417    0.931           
--------------------------------------------------------------------
Avg:    0.497    0.510    0.963    0.443    0.487    0.969    0.355  


Midranks:
          A        B        C        D        E        F        G   
--------------------------------------------------------------------
Run#1     3        1                 2                              
Run#2     1                 3                 2                     
Run#3                       3        1                          2   
Run#4     1                                            3        2   
Run#5              1        3                          2            
Run#6              3                          2                 1   
Run#7                                2        1        3            
--------------------------------------------------------------------
Sum:      5        5        9        5        5        8        5   

Data is a BIBD design with
  k=7, s=3,  n=7, p=3,  la=1,  p*(s-1) = la*(kk-1) = 6

Durbin-Skillings-Mack statistic  D=7.71429
  Large-sample chi-square approx:  P=0.25979  df=6

Carrying out 100,000 permutations to estimate
  the exact P-value for D=7.71429:
Observed permutation score = Sum of squares of
  treatment-group rank sums = 270
    P = 30429/100,000 = 0.30429  95% CI (0.30144, 0.30429, 0.30714)

Random numbers used:  2,800,000