44 
as it were. We have now for several years discussed the specific 
identity (and the reverse) of Tephrosia crepusculaiia and Tephrosia 
bistortata, and have arrived at no very satisfactory results. Everyone 
seems willing to grant that they are what may be called “ doubtful ” 
species, that is, that the characters by which they may be differen¬ 
tiated from each other are not sufficiently decided to leave no doubt, 
in some cases, as to which species a particular individual specimen 
should be referred to. In other words, it is stated, that even specialists 
cannot invariably determine them. 
HYBRIDITY OF ALLIED SPECIES. 
The practical work which Mr. Bacot and Dr. Riding have success¬ 
fully carried out in the direction of hybridising these species, appears 
to me to be of the greatest importance, and the results arrived at 
may possibly be very far-reaching in their character. That these 
species are very closely allied, everyone allows ; that the distinctive 
characters are so ill-defined as to make it a matter of difficulty, except 
for the trained specialist, to discriminate, may also be conceded. They 
are species in the making, as I have just said, and their specialisation 
is not yet completed. It is clear that, if the theory of evolution by 
natural selection be sound, there must be many such cases, and there 
must be, in nature, every gradation between the polymorphic and 
unstable species, in which almost every individual varies from almost 
every other in some slight and unimportant manner, through every 
gradation of varieties (local races), and sub-species, to clearly defined 
species. 
Each species is separable from its nearest allies by certain cha¬ 
racters which will differentiate it from all other species. These we 
call specific characters. Some naturalists, and I disagree entirely 
with them, go so far as to include all species that will pair and pro¬ 
duce offspring as being one species. Under these conditions we should 
have to unite Srnerinthus populi with S. ocellatus, Amphidasys strataria 
with A. betularia, Saturnia carpihi with S. pyri, and a large number 
of species well-defined on many characters in all their stages of 
existence. We should also have to unite Phasianus colchicus with 
P. torquatus, the hare [Lepus europaeus (timidus )] with the rabbit 
(L. cuniculus ), and endless other birds and mammals which are 
abundantly distinct. When the question of hybridity was first 
studied, it was laid down as an axiom that all hybrids were sterile, and 
when it was found that fertilisation between two plants or animals was 
possible, and that fertile progeny resulted, the plants and animals were 
reduced at once to the rank of varieties. As, however, our data on this 
subject accumulates, it appears to be certain that a very large number 
of closely allied, but, in the generally accepted sense of the term, 
perfectly distinct, species, are not only reciprocally fertilo, but their 
hybrids are also fertile inter se. Before, however, we can assert that wo 
really know anything about the subject, a very great number of 
careful experiments on many different species must bo carried out, and 
the results compared. 
VARIATIONS ALWAYS PRESENT IN ORGANIC BEINGS. 
Tho general tendency not only for all organisms to vary, but also, for 
every constituent structure and part of an organism to vary, is so well- 
known, that, in ro-roading any of Darwin’s works, one is struck by tho 
frequency with which ho prefaces his remarks with “ if tho spocios 
