VARIATION IN DIFFERENT PORTIONS OF PLANT. 



31 



main-stem whorls are distinctly less variable, both absolutely and rela- 

 tively. That these changes are significant is shown in table 9, which may 

 be called a "difference table." It gives for any designated series the 

 difference (and its probable error) between the value of a particular 

 constant for the whorls on all branches of that series as given in table 

 8, and the value of the same constant for the main-stem whorls of the 

 same series. The absolute differences, expressed as per cent of the 

 constant for the "all-branch" series, are tabulated as "relative differ- 

 ences." The sign of a difference is taken positive when the main-stem 

 constant is the greater. 



Table 9. — Difference table, comparing main-stem whorls with whorls on all branches. 



Totals for series. 



Series. 



I 



II 



III 



IV] 

 V^ 

 VI 



Difterences. 



Absolute difference 



Relative difference..per cent. 



Absolute difference 



Relative diff erence..per cent. 



Absolute difference 



Relative difference. .per cent. 



Absolute difference 



Relative difference.-per cent, 



Absolute difference 



Relative difference, .per cent. 



Absolute difference 



Relative difference.. per cent. 



Between means 

 (unit=lleaf). 



0.885±0.061 



10.5 

 1.232± .060 



14.5 



1.254± .063 



14.8 



.742± .076 



8.5 

 .713± .089 



8 9 



.348± .075 



4.0 



Between standard 



deviations 



(unit=l leaf). 



-0.165±0.044 



— 13.4 



— .447± .043 



— 35.6 



— .550± .045 



— 44.5 



— .339± .039 



— 32.5 



— .146± .060 



— 10.2 



— .103± .053 



— 7.6 



Between coeffi- 

 cients of vari- 

 ation. 



-3.153±0.490 



— 21.7 



— 6.432± .459 



— 43.9 



— 7.545± .496 



— 51.7 



— 4.498± .419 



— 37.8 



— 3.125± .783 



— 17.5 

 -1.709± .606 



— 11.4 



There can be no doubt that these differences are significant in compar- 

 ison with their probable errors. The only cases in which the differences 

 are not more than three times the probable errors are for the standard 

 deviations of Series V, and the standard deviations and the coefficients 

 of variation of Series VI. Even in these cases the differences are more 

 than twice the probable errors. There is a very substantial differentia- 

 tion between main-stem whorls and whorls borne on branches. Main- 

 stem whorls have on the average roughly 1 leaf more than do whorls on 

 branches, and are from 10 to 50 per cent less variable, both absolutely 

 and relatively. The very low variability of main-stem whorls both 

 absolutely and as compared with whorls on other portions of the plant 

 may be brought out by a totally different sort of consideration. We 

 say with regard to any organ or character that it varies very little 

 if the chance of finding any other than the typical condition is small. 

 If the chance of the occurrence of deviations from the typical con- 

 dition is large, we say that there is great variation. Now, in the case 



