500 Development of Mouth and Gills in Bdellostoma 
condensation of the brain, though I shall be better prepared to discuss 
this subject after a more careful study of the brain development. 
The hyomandibular cavity continues to degenerate and there is only 
a slight indication of it in embryos of later stages, 7. e., those in which 
the branchial pouches and head cartilages are forming. In Fig. 18 
through the posterior auditory region of such an embryo the hyomandi- 
bular is indicated by the slight lateral diverticulum marked Hm just 
at the beginning of the throat where the gut wall arches dorsally 
to map out the velum. Older stages lose even this indication and the 
gut arches up around the anterior end of the velum without any sugges- 
tion of the hyomandibular diverticula. 
Thus in Bdellostoma the hyomandibular cleft early attains an enor- 
mous development, then gradually decreases in size, and disappears be- 
fore the embryo has reached a very advanced stage. 
THE PRESENCE AND FATE oF Two Pairs oF Post-HYOMANDIBULAR 
Ginn, CLEFTS: 
Dean, 99, writes (op. cit., page 270) that the question of the dis- 
appearance of a pair of gill clefts lying post-hyomandibular has been 
strikingly revived by his studies. In a foot-note on the same page he 
also mentions this “ doubtful cleft” which lies close behind the hyo- 
mandibular as corresponding to the thyroidian cleft of Dohrn. Study- 
ing sections of these embryos I have been able to confirm the observations 
made on cleared totals by Dean and I have further demonstrated a second 
cleft following close behind the first, which likewise disappears during 
development. 
The existence of these additional clefts in the embryo is of evident 
interest with regard to the maximum number of gill slits in chordates. 
Since the adult Bdellostoma often possesses as many as thirteen pairs of 
gills we may now include from embryonic data a total number of fifteen 
pairs exclusive of the hyomandibular and mandibular clefts. There seems 
to be no valid argument against this greater number representing a 
primitive feature. The tendency in Bdellostoma is to lose its large num- 
ber of gills as is shown first by the suppression of these two embryonic 
pairs and further by the observations of Miss Worthington, 05, on 
the adult. She has found (p. 633) the number of gills in different 
species ranging from seven to thirteen pairs, and in the case of the 
Californian species no number seems to be either normal or 
fixed. Of more immediate interest is her observation that traces of 
lost gills are often to be found not at the hinder end of the series of 
gills, where one would at first suppose, but at the cephalic end. 
