Feb., 1922] LA RUE AND BARTLETT PESTALOZZIA GUEPINI 89 
Part 2. Appendage length 
Strain 
No. of 
Genera- 
Total 
Range of 
Variation 
Range of 
Generation 
IVIeans 
Mean 
Standard 
Deviation 
Coefficient 
Variation 
I ... 
6 
22.0 
4.8 
10.9 
4.54 
41.6 
0 
e; 
0 
IQ.O 
3.6 
1 1. 6 
4.60 
39.7 
1 . . 
6 
26.0 
6.6 
18.9 
4.42 
23.3 
A . . 
18.0 
1.7 
19.7 
3.42 
17.4 
c: 
8 
2^.0 
5.4 
21.3 
3.96 
18. 1 
6 
4. 
23.0 
2.7 
21.6 
3.69 
17. 1 
7 
6 
17.0 
4.2 
16.4 
3.47 
21.2 
8 
c 
0 
26.0 
2 =; 
26.^ 
16.8 
Q 
24.0 
7.1 
23.7 
4.56 
19.2 
10 
30.0 
1 1.4 
23.6 
5.86 
24.8 
II 
6 
21.0 
4.6 
25.2 
4.78 
19.0 
12 
5 
24.0 
4.4 
28.7 
4.15 
14.5 
IT. 
6 
26.0 
7.9 
30.0 
4.7^ 
15.8 
14 
7 
23.0 
8.7 
25.1 
5.05 
20.1 
I . ... 
6 
19.0 
S.2 
23.0 
4.0^ 
17.5 
16 
6 
20.0 
2.6 
22.8 
3-93 
17.2 
17 
7 
24.0 
6.6 
22.9 
5.33 
23.3 
18 
7 
23.0 
5.1 
22.0 
4.13 
18.8 
IQ 
6 
27.0 
6.5 
19.4 
4.22 
21.8 
20 
6 
24.0 
Q.5 
20.6 
4.13 
20.1 
21 
6 
30.0 
10.8 
19.4 
6.38 
32.9 
22 
7 
21.0 
5.1 
19.9 
3.95 
19.9 
23 
5 
21.0 
7.0 
19.9 
4.60 
23.1 
24 
7 
19-0 
2.7 
1^-7 
3.45 
18. 5 
25 
5 
16.0 
3-4 
23-5 
4.09 
17.0 
26 
6 
25.0 
3.8 
22.7 
4.36 
19.2 
27 
7 
23.0 
10. 0 
21.5 
4.44 
20.7 
28 
6 
48.0 
13. 1 
24.2 
5-05 
20.9 
29 
8 
29.0 
5.0 
24.0 
3-96 
16.5 
30 
6 
29.0 
3-2 
24.6 
4.96 
21.6 
31 
6 
27.0 
2.8 
22.9 
4.25 
18.6 
32 
6 
17.0 
2.2 
22.6 
3.36 
14.9 
33 
7 
19.0 
3-9 
13.4 
3.10 
23.1 
34 
6 
23.0 
10. 0 
23.3 
5.01 
21.5 
35 
22.0 
6.2 
24.6 
4-93 
20.0 
The least number of generations upon which any strain mean was based 
was four, the greatest, eight. The least accurate comparisons, therefore, 
would be between two strains both of which had been carried only four 
generations. Conversely, the most accurate comparison would be between 
strains which had both been carried eight generations. The standard errors 
of the differences between the means of the several strains were multiplied 
by three, and if an observed difference between two strains exceeded this 
value, the difference in question was considered to demonstrate the probable 
distinctness of the two strains. For spore length 1.8 was considered a 
significant deviation in the case of the most dependable comparisons; 2.7 /x 
in the case of the least dependable. The corresponding criteria for the 
appendage lengths were 3.1 m and 4.5 p.. It must be remembered that a 
criterion of three times the standard error of a difference between means 
secures practical certainty of the distinctness of strains. On the contrary, 
it still remains highly probable that unlike strains remain in association 
