OcTOBER 2, 1913] 
NATURE 
149° 
both lungs and gills are suppressed, and the creature | of a jackal, and others again are traceable to some 
s g§ PP J g 
breathes in a most pseudo-primitive fashion. This 
arrangement, more or less advanced, occurs in 
many Urodeles, both American and European, belong- 
ing to several sub-families, but not in every species 
of the various genera. It is therefore a case of 
apparently recent Isotely. 
There is no prejudice in the making of a new 
organ except in so far that every organism is con- 
servative, clinging to what it or its ancestors have 
learnt or acquired, which it therefore seeks to re- 
capitulate. Thus in the vertebrata the customary 
place for respiratory organs is the pharyngeal region. 
Every organism, of course, has an enormous back 
history; it may have had to use every part in every 
conceivable way, and it may thereby have been trained 
to such an extent as to yield almost at once, like a 
bridle-wise horse to some new stimulus, and thus 
initiate an organ straight to the point. 
Considering that organs put to the same use are 
so very often the result of analogous adaptation, 
homoplasts with or without affinity of descent, are we 
not justified in accusing morphology of having made 
rather too much of the organs as units, as if they 
were concrete instead of inducted abstract notions? 
An organ which changes its function may become a 
unit so different as to require a new definition. And 
two originally different organs may come to resemble 
each other so much in function and structure that they 
acquire the same definition as one new unit. ‘To avoid 
this dilemma the morphologist has, of course, intro- 
duced the differential of descent, whether homologous 
or analogous, into his diagnoses of organs. 
The same principles must apply to the classification 
of the animals. To group the various representative 
owners of cases of isotely together under one name, 
simply because they have lost those characters which 
distinguished their ancestors, would be subversive of 
phyletic research. It is of the utmost significance that 
such ‘‘convergences”' (rather ‘‘mergers,’’ to use an 
administrative term) do take place, but that is another 
question. If it could be shown that elephants in a 
restricted sense have been evolved independently from 
two stems of family rank, the convergent terminals 
must not be named Elephantinz, nor can the repre- 
sentatives of successive stages or horizons of a mono- 
phyletic family be designated and lumped together 
as subfamilies. And yet something like this prac- 
tice has been adopted from Cope by experienced 
zoologists with a complete disregard of history, which 
is an inalienable and important element in our science. 
This procedure is no sounder than would be the 
sorting of our Cartwrights, Smiths, and Bakers of 
sorts into as many natural families. It would be 
subversive of classification, the aim of which is the 
sorting of a chaos into order. We must not upset the 
well-defined relative meaning of the classificatory terms 
which have become well established conceptions; but 
what such an assembly as the terminal elephants 
should be called is a new question, the urgency of 
which will soon become acute. It applies at least to 
assemblies of specific, generic, and family rank, for 
each of which grades a new term, implying the 
principle of convergence, will have to be invented. 
In some cases geographical terms may be an addi- 
tional criterion. Such terms will be not only most 
convenient, but they will at once act as a warning not 
to use the component species for certain purposes. 
There is, for instance, the case of Typhlops braminus, 
mentioned at the beginning of this address. Another 
case is the dog species, called Canis familiaris, about 
which it is now the opinion of the best authorities 
that the American dogs of sorts are the descendants 
of the coyote, while some Indian dogs are descendants 
NO. 2292, VOL. 92] 
wolt. The ‘dog,’ a definable conception, has been 
invented many times, and in different countries and 
out of different material. It is an association of 
converged heterogeneous units. We have but a 
smile for those who class whales with fishes, or the 
blind-worm with the snakes; not to confound the 
amphibian Ceecilians with reptilian Amphisbzenas re- 
quires some training; but what are we to do with 
creatures who have lost or assimilated all those 
differential characters which we have got used to 
rely upon? 
In a homogeneous crowd of people we are attracted 
by their little differences, taking their really important 
agreements for granted; in a compound crowd we at 
once sort the people according to their really unim- 
portant resemblances. That is human nature. 
The terms ‘‘convergence’’ and “parallelism” are 
convenient if taken with a generous pinch of salt. 
Some authors hold that these terms are but imperfect 
similes, because two originally different organs can 
never converge into one identical point, still less can 
their owners whose acquired resemblance depresses 
the balance of all their other characters. For instance, 
no lizard can become a snake, in spite of ever so many 
additional snake-like acquisitions, each of which finds 
a parallel, an analogy in the snakes. Some zoologists 
therefore prefer contrasting only parallelism and 
divergence. A few examples may illustrate the justi- 
fication of the three terms. If out of ten very similar 
black-haired people only two become white by the 
usual process, whilst the others retain their colour, 
then these two diverge from the rest; but they do 
not, by the acquisition of the same new feature, 
become more alike each other than they were before. 
Only with reference to the rest do they seem to liken 
as they pass from black through grey to white, our 
mental process being biased by the more and more 
emphasised difference from the majority. 
To ANa bx Gx DH E 
9 
8 
7 
6 
5 
4 
3 
2 Ax Bx 
tae be Go DLE FE 
Supposing A and B both acquire the character x and 
this continues through the next ten generations, while 
in the descendants of C the same character is in- 
vented in the tenth generation, and whilst the 
descendants of D, E, F still remain unaltered. Then 
we should be strongly inclined, not only to ‘key 
together CX with A—~ and B x but tak» this case 
10 10 10 
for one of convergence, although it is really one of 
parallelism. If it did not sound so contradictory it 
might be called parallel divergence. The inventors 
diverge from the majority in the same direction : 
Isotely. : 
Third case.—Ten people, contemporaries, are alike 
but for the black or red hair. Black A turns white 
and red E turns white, not through exactly identical 
stages, since E will pass through a reddish-grey tinge. 
But the result is that A and E become actually more 
like each other than they were before. They con- 
verge, although they have gone in for exactly the 
same divergence with reference to the majority. 
In all three cases the variations begin by divergence 
from the majority, but we can well imagine that all 
the members of a homogeneous lot change ortho- 
