CVl 
THE VOYAGE OF H.M.S. CHALLENGEE. 
shall simply make a third of it, and if the localities are not so far removed, it will not 
make any difference to our procedure so long as the difference of habitat is, as supposed, 
associated with the difference in character. On this manner of viewing species, the 
different races of mankind would be regarded as species ; and as very different species, 
much better defined than many species amongst the Invertebrates, I certainly would regard 
them ; but to take a specific instance from the Sponges, in the Atlantic there is a group 
of similar forms of Lithistid which, without knowledge of any other similar group, we 
can with certainty call a species, and its differences from all other forms of Lithistid 
are sufficiently great to justify us in distinguishing it generically, it thus acquires the 
name Siphonidium ramosum ; in the Challenger collection is a similar group of Lithistids, 
brought home from the Pacific, in outward form and general structure this is not 
distinguishable from the Atlantic group, one of its spicules, however, is different, being 
an oxystrongyle, while the corresponding spicule in Siphonidium ramosum is an 
oxytylote ; had these Sponges been obtained from the same locality, they would have 
been placed in the same species (they differ far less than a Negro and a European), 
but coming from such distinct geographical regions I have placed the Pacific form in 
a new species — Siphonidium capitatum. With more hesitation I have distinguished the 
form of Pachymatisma, which occurs off the Norwegian coast, from that which is found 
in the English Channel, simply on a difference in the relative size of the spicules and 
in the thickness of the cortex. 
The outcome of the preceding remarks would appear to be that assemblages of similar 
individuals, occurring in different distributional areas, may be regarded as species, if 
they present any constant difference of structure, however trivial. 
W e now pass to groups of genera, and the distinction between these will naturally 
be more marked than between species ; generally there will be no difficulty in finding 
some clear and definite character by which one genus is readily separated from its 
nearest allies, but here again transitional forms cannot, as Nageli, quoted by Polejaeff, 
maintains, be regarded as destructive of the genera which they unite, for transitional 
forms on the evolutional hypothesis are as much to be expected between genera as 
between species. Thus from an origin a let two divergent lines ah ac be drawn to 
represent two groups of species evolved along different lines of descent and of the value 
of genera, then if the whole of the twm series should be known to us it is plain that 
however divergent the two stems may be they must pass into each other at the origin 
a. Unless therefore nature exercises a selective destructive power specially at such 
points as a, we may frequently expect to find what are really good genera passing into 
each other as they receive fresh accessions 'of species. As to what constitutes a good 
genus, I see no escape from the admission, that as in the case of species it is a matter 
resting chiefly on the judgment of the describer, but there is one condition which should 
be satisfied as far as possible, — it should not, except in special cases, be founded on a 
