naturalists who believe in the immutability of species often
maintain that, even in the most variable forms, the characters which
they consider of specific value are unchangeable. To give an example
from a conscientious writer,[753] who, relying on the labours of M.
Naudin and {359} referring to the species of Cucurbita, says, "au
milieu de toutes les variations du fruit, les tiges, les feuilles, les
calices, les corolles, les etamines restent invariables dans chacune
d'elles." Yet M. Naudin in describing _Cucurbita pepo_ (p. 30) says,
"Ici, d'ailleurs, ce ne sont pas seulement les fruits qui varient,
c'est aussi le feuillage et tout le port de la plante. Neanmoins, je
crois qu'on la distinguera toujours facilement des deux autres especes,
si l'on veut ne pas perdre de vue les caracteres differentiels que je
m'efforce de faire ressortir. Ces caracteres sont quelquefois peu
marques: il arrive meme que plusieurs d'entre eux s'effacent presque
entierement, mais il en reste toujours quelques-uns qui remettent
l'observateur sur la voie." Now let it be noted what a difference, with
regard to the immutability of the so-called specific characters, this
paragraph produces on the mind, from that above quoted from M. Godron.

I will add another remark: naturalists continually assert that no
important organ varies; but in saying this they unconsciously argue in
a vicious circle; for if an organ, let it be what it may, is highly
variable, it is regarded as unimportant, and under a systematic point
of view this is quite correct. But as long as constancy is thus taken
as the criterion of importance, it will indeed be long before an
important organ can be shown to be inconstant. The enlarged form of the
stigmas, and their sessile position on the summit of the ovary, must be
considered as important characters, and were used by Gasparini to
separate certain pumpkins as a _distinct genus_; but Naudin says (p.
20) these parts have no constancy, and in the flowers of the Turban
varieties of _C. maxima_ they sometimes resume their ordinary
structure. Again, in _C. maxima_, the carpels (p. 19) which form the
Turban project even as much as two-thirds of their length out of the
receptacle, and this latter part is thus reduced to a sort of platform;
but this remarkable structure occurs only in certain varieties, and