THE ENTEROPNEUSTA. 667 
In fact from these figures one may affirm with some confidence that laccadivensis and 
muscula (though they cannot be separated from one another by this criterion) are quite 
separate from maldivensis, gracilis, and parva, whilst of the last three gracilis is very 
probably distinct from the other two. By this criterion alone it is not possible to separate 
maldivensis from parva any more than laccadivensis from muscula. No doubt this could be 
done in the case of these forms if the material were sufficient to treat another ratio (e.g. 
the length ratio) in the same way. 
The total length ratio was worked out for laccadivensis and maldivensis, in which forms 
alone the amount of material rendered the attempt feasible. As there are no definite segments 
by which to reckon, the absolute value obtaimed for o will depend upon the number of groups 
into which the material is broken up. In this case the mean in the two varieties is 23'1 and 
17-9 respectively. The series have been broken up into groups separated by an interval of 
two units, which is approximately 10°/, of the mean in each case. The treatment has been 
uniform in each case and the results are as follows :— 
TABLE 9. 

var. laccadivensis var. maldivensis 


M. = 23-1 17-9 
PE.M.=| + 129 + -260 
c = +1°798 | +1°589 | 
P.E.c= + -091 | + -184 | 
Nee 49 | 17 | 


Hence the mean is widely separate in the two forms—separated by so many times 
the probable error as to render it absolutely certain that no random selection of 17 specimens 
from a large quantity of lJaccadivensis could ever have a mean length ratio with the remotest 
approximation to the above 17 specimens of maldivensis. 
Leaving now the question of variation as a criterion of species and variety we may 
pass on to consider one other point in connection with it. A great feature in the group 
of the segmented Chordata is the process of cephalization which becomes more and more 
pronounced as we pass upwards from Amphiozus to Mammals. Accompanying this process, 
and doubtless intimately bound up with it, is the phenomenon of unequal variability in the 
different regions of the animal’s body from the meristic pomt of view. With increased 
cephalization we find associated a decreased variability at the cephalic as compared with the 
caudal end. The nearer we approach to the caudal extremity the greater appears to become 
the meristic variability’. From this point of view the relative variability of the different regions 
of the Enteropneust’s body becomes of interest, especially as most writers see in the collar 
cord the commencement of the cephalization process. From the data given in Table 11, p. 671, 
1 For the present this statement is based upon a number dorso-lumbar and sacral regions a small amount, whilst the 
of unpublished data from different Vertebrata, chiefly fishes, caudal region exhibits most of all. Probably some such 
collected by the writer. The phenomenon is perhaps most phenomenon is to be found also among Invertebrates, but 
familiar in the case of the Mammals where, speaking gener- absence of data permits only of conjecture. 
ally, the cervical region exhibits no meristic variation, the 
