
became a nuisance in the trawl, some time at least in 
the eighteen-eighties. How did the showy Plusia 
moneta become a common British moth? No one 
recorded it here before 1890. Extinction must ensue 
from countless causes. If compelled to specify one 
class of cause as operating rather than another, we 
should regard the appearance of a new and antagonistic 
organism as by far the most formidable and effective 
agency of extinction; but we have only to glance 
at anthropological data to observe that no rule obtains 
as to the length of time which the process of extermina- 
tion will take. Whatever doubts be entertained as to 
the significance of adaptation in delimiting specific 
characters, there can be none that survival is deter- 
mined by selection according to the balance of the 
profit-and-loss account on the workings of the machine. 
Wondering at the Ceylonese endemics, Dr. Willis 
asks rhetorically, “‘ Had one arrived in Ceylon just in 
time to see the disappearance of a considerable flora ? ” 
We may reply, What more likely ? Is the alternative 
interpretation, that he had come in time to attend 
the birth of a new flora, more acceptable? About 
half the endemics of Ceylon, he tells us, occur on the 
tops of single mountains or small groups of mountains. 
Does he really suppose that future ages will witness 
the spread of such species downwards from the 
mountain tops ? ¢ 
In reading the chapter on the origin of species and 
the many passages in which references to mutation 
are made, I see signs that Dr. Willis, though making 
large assumptions in the name of genetical experiment, 
is not sufficiently conversant with the present state 
of genetical science. Both from observation and 
from experiment, the certainty that variation is 
_largely discontinuous has been established. If for 
the moment we abrogate the consideration of inter- 
specific sterility we might declare that forms mistak- 
ably like new species do actually arise suddenly. But 
this is scarcely mutation as contemplated by the 
theory of Age and Area. If we were told categorically 
which ‘‘ wide” species is regarded as the putative 
parent of which endemic, we should be in a position 
to consider how far this interpretation is consistent 
with what we know of variation. From anything so 
precise Dr. Willis shrinks. Here and there we get a 
glimpse of what he would like us to infer. The endemic 
Coleus elongatus, for example, he is inclined to claim 
as the immediate product of C. barbatus, from which 
it differs in some ten respects. The shrubby Veronicas 
are characteristic of New Zealand; if pressed Dr. 
Willis would point to the “ wide” V. elliptica (men- 
tioned above) as their putative parent. Similarly 
the Chilian Ranunculus acautis, or alternatively R. 
crassipes (found in Kerguelen), which both occur in 
NO. 2776, VOL. 111] 


[JANUARY 13, 1923 
New Zealand, might be adduced as the parent of the — 
endemic Ranunculi of those islands. Though un- 
deniable as possibilities, we have to consider what — 
warrant for such guesses can be drawn from the 
observed facts of variation. The answer is quite 
clear that up to the present scarcely anything com- 
parable has been observed. The “rogue ’-peas, the 
“fatuoid ” mutations of oats (Nilsson-Ehle and later 
Marquand), with perhaps a very few more, are all that 
can be quoted as precedents, none certainly in point. 
No one familiar with genetical work would be dis- 
inclined to entertain the supposition that such groups 
of.endemics as the New Zealand Veronicas may not 
improbably be co-derivatives from one or more crosses ; 
so also may the hosts of “ species ” of Crateegus which 
Prof. Sargent has described largely as endemics on 
derelict farms of the Eastern States. But to establish 
these propositions, genetical and doubtless cytological 
work on a vast scale is required, and far too little has 
been yet done to justify the bold assumptions lightly 
made in the doctrine of Age and Area. 
The evidence adduced by de Vries from Cnothera 
which led him to propound the theory of Mutation 
is clearly enough the precedent which Dr. Willis has 
at the back of his mind. From the first the meaning 
of the Cnothera work was ambiguous. The re- 
searches of Renner and of Heribert-Nilsson have now 
shown that those early suspicions were justified, and 
that the ‘ mutations” of Ginothera are not genuine 
illustrations of the origin of species by variation in 
descent from a pure form. Had de Vries grasped 
the implications of Mendelian analysis, he could never 
have so interpreted them with any confidence. The 
few words in which he conveys his benediction on 
this new venture should be read with caution and 
reserve by persons unfamiliar with the history they 
purport to relate. x 
Unconvincing as the main argument of “ Age and 
Area” appears, the reader will find in it some curious 
and interesting discoveries. Of these the most re- 
markable is the uniformity of the statistical distribu- 
tion of species among the genera of various and most 
dissimilar forms of life, both plants and animals. 
The monotypic genera, with one species each, are 
always the most numerous, commonly forming about 
a third of the whole group, the ditypics, with two species 
each, are the next in frequency, genera with higher 
numbers of species becoming successively fewer. Set 
out graphically, according to the number of species 
they contain, the genera exhibit what is here called 
a “hollow curve ” of frequency, and there is no gain- 
saying the fact that these curves, though collected 
from such miscellaneous sources, have a remarkable 
similarity. Another curious feature exhibited by this 
