262 M.A. de Candolle on Germination 
In these two instances, especially in the latter, the numbers 
become much more equal on deducting the degrees of tempe- 
rature below the minimum. Probably this correction becomes 
more requisite as the minimum becomes higher. 
When in these three calculations the numbers of the be- 
ginning and the end, which are often in non-accordance with the 
others, are abstracted, germination takes place, in widely different 
species, under the influence of tolerably similar conditions of 
time and temperature; for the numbers are comprised between 
fourteen and thirty-four when the minima are deducted. They 
are slightly less in the case of the species which requires the most 
initial heat, but in very unimportant proportion. 
Definitively, the method of the sums of temperature applies 
with moderate accuracy to the facts of germination. What is 
essentially required to be known in the case of each species, in 
regard to this function, is the requisite minimum. The rest 
differs but little in the various plants; and it is easy to foretell the 
effects of an increase of temperature when once germination is 
possible, without having recourse to calculations or direct obser- 
vations in the case of each species. The same probably docs 
not apply to the other functions, nor to the assemblage of func- 
tions, from germination to maturation. This would form a point 
to be decided by experiment. Unfortunately Lam unacquainted 
with any means of causing a phanerogamous plant to undergo 
regular development at a certain temperature, without light. It 
would be requisite at least to be able to furnish a species with 
light which is uniform and of the same kind for several weeks. 
With the progress of knowledge, this will be possible sooner or 
later; but until then our calculations upon the sums of heat in 
botanical geography, in, agriculture, and in horticulture will be 
contaminated with hypotheses and manifold causes of imexacti- 
tude*. 
8. Variable temperatures. 
I have not yet experimented upon germination at variable 
temperatures. I even endeavoured to maintain more constant 
temperatures than M. Burckhardt had done, so as to eliminate 
as much as possible the errors which might arise from variations. 
tion a more advanced phase of development, that at which the cotyledons 
Beene exposed. 
* Tf the other functions agree with germination, the numbers calculated 
from the extreme limit of the species must be distrusted. We see, in fact, 
that near the poimt at which vegetation is arrested, much more time is 
required to compensate for the loss of heat. The numbers calculated near 
the limits would only serve for comparison with each other, and the num- 
bers deduced from the centre of a habitat must not be confidently applied 
to express the necessary conditions at the limits. 
ewes «< 
