Alice Lee 
317 
Taken as a whole the mean results of A, B, C and G may be held to represent the state of 
things at Gi'eiz. But it seems of vital importance in future to record (i) the period in the 
season at which the flowers were gathered and (ii) the differences in local environment, if any, of 
the different series. 
The coefficients of variation deduced from the mean values are 14"6 for petals and 22'0 
for sepals. These are well within the range of the coefficients of variation determined for 
the vegetable kingdom*; and as usual much higher than the ordinary values in the animal 
kingdom. 
Dr Ludwig's next results are for Ficaria verna at Gera, Trogen and Gais. I find the 
following values : 
Correlation 
■1928 ±-0205 
•0188 ±-0260 
•1954 ±-0203 
■1821 ±-0386 
•1964 ±-0478 
■2237 ± -0203 
Now these results are extremely anomalous. While Gera H is in fair agreement with Gera K 
so far as means and correlations are concerned, there are reniarkable changes in the variabilities, 
especially of the petals. Both, however, differ hopelessly from Gera I in the correlation 
between the numbers of petals and sepals. This value is about ^2 in H and K and zero in I. 
I venture to think, either that there is some error here, or that this result needs investigation 
of a special kind. Usually we find a correlation of ^2 to '25 between petals and sepals, 
but either at some period of the season or with a special environment this correlation can be 
reduced to zero. Further in Trogen L, which has very much the same means as Gera H, the 
correlation has become negative, or the greater the number of petals the fewer the sepals ! 
Gais in the neighbourhood of Trogen, while giving flowers with remarkably high means in 
both characters, still exhibits a correlation of a positive kind, not very far from the value of 
Gera H and K, or in the longer series not widely divergent from the results for Greiz (^224 as 
compared with -245) — indeed within the probable error of the difference. If the zero correlation 
at Gera and the negative correlation at Trogen be verified by further countings, then, I think, 
it will probably be found that the correlation between the numbers of petals and sepals varies 
with the period of the season, and may pass through zero fi'om positive to negative values. 
If this result be confirmed it seems of considerable importance from the standpoint of plant 
economy. 
We have seen that the Gais means differ very sensibly from the Greiz, while the difference 
in the correlation of petals and sepals is not significant. A change of environment or a 
collection at a different part of the season may easily show much change in means or variability. 
So far there does not seem any good reason for supposing the Gais and Greiz Ficaria verna to be 
of different race, " petites espuces " as Dr Ludwig calls them. By this, I understand, that if 
* Phil. Trans. Vol. 197, A, p. 361. 
Ficaria verna Mean S. D. 
TT , fPetals 8^225 Ml 131 
Gera, H (1000) ^^^^^^ .^^^^j 
T /^^^\ [Petals 8^058 •9432] 
Gera, I (675) j^^^^j^ g.^^^ .^^^^^ 
n V /-lo^ /Petals 8^263 •50321 
Gera, K (.12) jg^^^^^^ ^^^^^ 
T T (Petals 8^144 '7663 
Trogen, L (285) jg^p^^^ 3^386 ^5487 
^- nr/.o.s fPetals 8^560 •94991 
Gais, M (184) ^.^^^ 
Gais, N (1000) '''''1 
' ^ ' \Sepals 4^538 •6158J 
