APR. 4, 1923 CLARK: ORIGIN OF VERTEBRATES 137 
lome; the chaetognaths suggest a relationship with the molluscs, and 
also with the nemerteans; the phoronids suggest a relationship with 
the polyzoans, but have a well developed coelome, and the colonial 
habit is reduced to the budding off of new individuals; and the brachi- 
opods suggest both the polyzoans and the barnacle-like arthropods. 
While by this third readjustment all the four original features are 
recombined in each animal type, the balance between them is imper- 
fect, for the influence of one of these features in each case is greatly 
overshadowed by the influence of the other three. 
A fourth readjustment (fig. 2) would correct this imperfect balance 
and result in the appearance of four animal types all very much alike. 
There are four types which appear to belong here (fig. 3), the tuni- 
cates, the cephalochordates, the balanoglossids and the pterobran- 
chiates. The tunicates seem to be in line with the polyzoans, while 
they also suggest both the brachiopods and the phoronids; the cephalo- 
chordates clearly stand in the cestode-arthropod line, and at the same 
time show indubitable affinities with the echinoderms; the balano- 
glossids, with no trace of asexual reproduction, may be considered in 
line with the flukes and molluscs and between the chaetognaths and 
echinoderms; while the pterobranchiates seem to fall between the 
chaetognaths and the phoronids. 
These four closely related types resulting from this fourth readjust- 
ment are each slightly excentric; but they are so close to each other 
that a fifth readjustment would presumably give a final perfected 
type in which at last all of the four chief features of the original types 
would be reunited in the economically most perfected form. 
The vertebrates appear to occupy this central position (fig. 3). In 
them we are able to recognize the segmentation of the cestodes, anne- 
lids, and cephalochordates, combined with the coelomic structure 
first indicated in the flukes, both enclosed in the undivided body of the 
turbellarians. Unless the limbs can be compared to budded units 
recalling certain highly reduced and specialized units in tunicate or 
polyzoan colonies, the influence of the feature represented by Micro- 
stomum seems to have disappeared. 
In the course of the various readjustments which culminated in the 
formation of the vertebrates numerous secondary features, such as 
visual and other sense organs, appendages of different kinds, diver- 
ticula and other outgrowths from the enteric canal, chitinous and 
calcareous skeletons, etc., all of which exist in the coelenterates and 
in one or other of the four types derived immediately from them, 
became enormously developed and specialized in correlation with the 
