1897] ORIGIN OF SPECIES AMONG PLANTS 175 
obstruction to self-fertilisation produced by the rostellum is 
common in orchids, and generally occurs in all the species of 
any particular genus. Yet it is obliterated in Phajus blumei, 
Chysis aurea, species of Chrysoglossum, Arundinia speciosa, and Eria 
Jlavescens, &c., so that these species set plenty of good seed by self- 
fertilisation, whereas 40,000 blossoms of Dendrobium speciosum set 
one pod. I have already had occasion to allude to the rudimentary 
organs of Mercurialis, Erodiwm, &c., which are retained in all the 
species alike. 
Is Dr Wallace, therefore, justified in making the above asser- 
tion at all, or in demanding that either useful or useless characters 
should be limited? Why should either one or the other be 
obliterated when a new variation arises? The fact that a genus, 
which is the result of sufticient variation in a species (unless it be 
monotypic), does retain both useful and useless characters in some, 
many, or all of its species, shows that there are no grounds for his 
statement. Natural selection may demand it, but nature utterly 
refuses to be obedient to that theory. 
Too great stress is laid upon a necessary fixity, as a proof of 
specific characters, by many writers. This is purely a relative 
matter. Cultivation has been suggested as a test of a species; but 
this is the very best means of inducing a wild plant to vary, as all 
cultivators know. The fixation of any variation is a matter of 
time. About five years may, perhaps, be regarded as the average 
period under cultivation in “fixing” races: but nothing is known 
about wild varieties. In either case the rule is that the environ- 
ment must be constant. 
Indefinite Variations, non-existent.—This is the second 
hypothetical source of new variations according to Darwinians. 
With regard to all the offspring varying approximately alike 
and not “indiscriminately ” (Romanes) or “ indefinitely ” (Darwin) 
when subjected to changed conditions of life, I wish to emphasise 
the fact most strongly that experiments show conclusively that 
if seedlings are subjected to a markedly different environment, 
when they grow up to maturity, the rule is, that all that do 
change, change in precisely the same way. They do not vary 
indefinitely among themselves; so that there is no material here— 
any more than with “individual differences ”—for natural selection 
to act upon. Thus, in cultivating the wild parsnip or carrot, all 
the seedlings that change, do so by beginning to assume the same 
new characters—viz., an increased size with a greater fleshiness in 
the root, larger dimensions of the leaves, reduction of hair, &c., 
with a corresponding alteration in the anatomical structures. 
So, too, if the seeds of an amphibious plant as Ranwneulus 
heterophyllus be sown in a garden border, all grow absolutely alike 
