THE ENTEROPNEUSTA. 669 
usually correspond fairly closely (cf. Willey, 799, Pl. XXVI. fig. 1, and Spengel, 703, 
Pl. XXIV. fig. 8). With regard to the second part of the question it may be easily demon- 
strated that in a given stretch of body length the branchiae are more numerous than are 
the epidermal annulations or the hepatic caeca in a stretch of the same length in the 
post-branchial region. The number of tongue bars averages about 5°6 per mm. of length 
(see Table 14, p. 675). On the other hand the number of epidermal annulations in the 
post-hepatic region is about 16 per mm. Therefore if we look upon these repeated parts 
as evidences of a rudimentary segmentation’, we must regard the post-branchial segments 
(or at least those of the hepatic caudal region) as being on the average 3°5 times as long 
as the branchial segments. In other words the branchial segments are three times as 
numerous as the hepatic-caudal ones per unit of length. Consequently if we calculate the 
coefficients of variation on the basis of “segments” instead of collar length units we must 
multiply the coefficient of variation for hepatic-caudal region by 3°5 in order to compare 
it directly with that for the branchial region. This brings it up to the value 41:26 as 
compared with 29°66 for the branchial region. If therefore we regard the Enteropneusts as 
exhibiting a rudimentary form of segmentation the variability in the different body regions 
falls into line with that of the Holochorda*. Much phylogenetic stress however must not be 
laid upon this point since it is not improbable that the phenomenon of cephalization may 
be shewn to be widespread among the Invertebrates also. What is of more importance is 
the influence, small though it be, which the occurrence in the Enteropneusta of a form of 
segmentation, similar in kind though less marked, may exert upon our conception of the 
manner in which the merism of the Vertebrata may have arisen. And here a few remarks 
of a more speculative nature may not be irrelevant in connection with 
METAMERISM IN THE ENTEROPNEUSTA. 
Willey 99, p. 303) has recently made a brilliant suggestion with regard to the origin 
of gill clefts. The discovery of the peculiar dermal pits of Spengelia has led him to regard 
gill-slits as having originally arisen as perforations in the interannular grooves formed for 
the aeration of the gonads, a function which they eventually gave up for that of respiration 
on the establishment of an elaborate vascular system. In this way Willey establishes a 
connection between the segmentation of the gills and of the epidermal annulations, the 
connecting link being the gonads. It has seemed to me possible to extend this conception 
of Willey’s and to see in the gonads and their arrangement the prime factor in the 
segmentation of the chordata. We may take as our starting-point a small Triploblastic 
(and perhaps also Tricoelomate) creature with three more or less definite body regions, of 
not widely unequal length, and of which the hindermost contains the gonads. The repro- 
ductive elements would be small and fertilization external in the sea. Under these conditions 
it is obvious that, other things being equal, those with the largest gonads would have the 
best chance of leaving offspring to perpetuate and amplify this feature. Increased size of 
1 Spengel (’03, p. 276 seq.) disagrees with Willey’s view 
of the significance of the epidermal annulations and sees in 
them only the effects of muscular contraction on the dis- 
position of the epidermal glands. He points out that there 
exists a certain amount of irregularity. Such irregularity 
however seems to me in no way to militate against the 
view of the significance of these structures taken by Willey. 
Goer. 
° It is interesting to notice that in Ptychodera the differ- 
ence between variability in the different regions of the body 
is not nearly so marked as in Amphiorus where we may 
regard the cephalization process as having extended further. 
Ihave found the following values for C.V. in different regions 
of the latter genus. Preatrioporal segments, C.V.=1-792; 
postatrioporal preanal, C.V.=2°655; postanal, C.V.=6-185. 
86 
