186 ; BOTANICAL GAZETTE [MARCH 
Perhaps no two actual frequency distributions would be alike in 
thus yielding to a simple arithmetical correction. Such a correc- 
tion is probably fallacious in its simplicity. It serves our purpose, 
TABLE IV 
MEASUREMENTS OF LENGTH OF COROLLA IN F, POPULATION OF CROSS BETWEEN 
Nicotiana forgetiana AND N. alata grandiflora 
24 ° 45 25 66 12 
25 2 46 28 67 10 
26 2 47 29 68 7 
27 2 48 28 09 7 
28 I 49 28 70 6 
_ 29 3 50 26 71 6 
30 5 51 28 72 5 
31 8 52 23 73 6 
32 10 53 20 74 4 
33 Ir 54 20 75 S 
34 15 55 19 76 2 
35 13 56 22 17 2 
36 18 57 22 78 ° 
37 18 58 22 9 I 
38 19 59 26 50 2 
39 26 60 21 I 2 
28 61 22 2 I 
41 33 62 15 3 ° 
42 24 63 15 4 I 
43 23 64 14 5 ° 
24 65 12 6 Ir 
TABLE V 
COMPARISON BETWEEN THE THEORETICAL FREQUENCY DISTRIBUTION FOR FOUR FACTORS 
THE ACTUAL FREQUENCY DISTRIBUTION RESULTING WHEN CLASSES 
WITH A CERTAIN CONSTANTLY INCREASING RANGE ARE USED 
Class limits 24-27 | 28-32 | 33-38 | 30-45 | 46-53 | s4-62 | 63-72 | 73-83 | 84-95 
Class range...... 4 5 6 7 8 9 10 II 12 
Frequency. 5.2... 6 27 04 |183 |210° |189 04 23 3 
Calculated fre- 
quency for 828 
individuals for 
($+4)8........ 3.2 | 25.9 | 90.6 |181.1 |264.4 |18r.1 | 90.6 | 25.9 | 3-? 
however, if it calls attention to the manifest error of expressing 
a wide range of biological variation by a frequency polygon of 
equal size classes. 
