BRITISH SPECIES OF THE GENUS MNIUM. 21 
It may be pointed out that it is impossible to find the explanation of those 
facts in the ordinary hypotheses of the Mendelian theory—such as the 
prineiple of segregation of the zygotes, the dominant-regressive principle, 
the hypothesis of the hereditary factors, and the presence-absence theory 
with reference to these factors,—because all the leaves in Table XIII. belong 
to one stem. Here the base of the explanation is found in a quite different 
principle—gradation. This calls for reflection *. 
Kemark V.: We have seen (§ 12) that the gradation of a given character 
is a character in itself. It would be interesting to observe the transmission 
of a gradation curve in hybridization. We see, for instance, in Table IX. 
(p. 16) that in JM. rostratum the gradation curve of the breadth of the leaves 
at their base has its summit in interval 10 ; the summit of the corresponding 
curve of M. spinosum is found in interval 3. What would be the form of 
the curve in the hybrids between the two species? 
Hybridization of Mosses seems to be impracticable ; but it is certainly 
possible to find among the Phanerogams two species different from each 
other by the form of the gradation eurve of a given character and suitable 
for experiment. 
PARE- LE 
INDIVIDUAL VARIATION. 
The Use of Figures for the Description of Species and the Identification 
of Specimens. 
$13. INDIVIDUAL VARIATION.—Let us suppose that we have measured 
a given character of the longest leaf of a certain number of specimens 
(fertile stems) belonging to a given species of Mniwn—for instance, the 
length of the longest leaf of eight stems of M. spinosum. In this example 
the figures were (in mm.) : 
655 668 798 758 764 764 748 — 810 
As all the measured leaves are exactly comparable (see § 12, Remark II.), 
the influence of gradation is eliminated. Here the variation is governed by 
the laws of chance ; it is individual variation. What answer will be given 
when we are asked which value is characteristic for the species ? 
Here we meet the first difficulty mentioned in $ 1. 
$14. MEAN VALUE.—According to the classic method, we might calculate 
the mean value (in the above example, 7°38 mm.) and take this as a cha- 
racteristic one. But the significance of a mean value, and especially its 
* See on the general importance of gradation, $ 12, p. 17, second paragraph. 
