86 GEORGE LEFEVRE AND WINTERTON C. CURTIS 
have encountered the condition in one species only, Obhquaria 
reflexa Rafinesque, which is represented in fig. 3. Here the dis- 
tended tubes project far down below the lower border of the rest 
of the gill, and when graved appear greatly swollen. The marsupial 
region is permanently differentiated and is easily recognized when 
empty. The number of tubes comprising the marsupium varies 
in different individuals of the species from two to eight and during 
the breeding season each tube is filled by a solid cord of a stiff glu- 
tinous consistency in which the embryos or glochidia are em bedded. 
Ptychogence. This group contains a single genus, Ptycho- 
branchus. The marsupium occupies the entire outer gill which is 
thrown into a series of folds, each water-tube- ending below in a 
swollen bulb-like enlargement, as seen in fig. 1, which is drawn 
from P. phaseolus Hildreth. Simpson (l. c, p. 612) states that the 
number of folds varies from six to twenty. In the specimen 
figured there are seventeen. 
Eschatigence. Simpson has established this group for the genus 
Dromus in which the marsupium occupies the lower border only 
of the entire outer gill, the ventral ends of the water-tubes being 
slightly distended and rounded. The condition, which has not 
been seen by us, would seem to differ from the marsupium of 
Ptychobranchus chiefly in the absence of the folds. 
Diagence. Here the outer gills throughout their entire length 
are utilized as a marsupium, which in external appearance is quite 
similar to that of the Homogense. The Diagense, however, differ 
from other Unionidse in that the embryos and glochidia are con- 
tained in peculiar cylindrical sacs which lie transversely in the gills, 
whereas in every other case the egg-masses are placed vertically 
in the gills. The group contains one genus, Strophitus. We have 
observed the gravid marsupium in S. edentulus Say, although we 
have not had an opportunity of determining whether the unusual 
position of the egg-masses is correlated with some peculiarity 
in the structure of the water-tubes, or not. 
It is probable as Simpson concludes (/. c), that the oldest type 
of marsupium ph^dogenetically is that occurring in the Endobran- 
chise in which the inner gills only are used as brood-chambers. 
It is a slight transition from this condition to that existing in the 
