ORTMANN: FAMILIES AND GENERA OF NAJADES. 237 
that it is of prime significance. In all my investigations I have never 
come across an exception or variation inthe marsupium. It is true that 
in species which have normally all four gills marsupial, sometimes only 
two gills are found charged. But all authors, who record such cases, 
only mention the fact that the gills were charged, without saying any- 
thing about the structure of the gills. I have also met with such 
cases: but invariably a closer investigation revealed the fact that 
the other gills which were not charged also possessed marsupial 
structure, and consequently were capable of being charged with eggs. 
On the other hand, in those cases, where the outer gills alone serve 
as marsupium, it was not the simple fact that they alone were filled with 
eggs in the breeding season, which was ascertained. It was the in- 
vestigation of the structure of the gills, which induced me to judge 
the character of the marsupium. Lefevre and Curtis (1910, p. 83) 
are inclined to regard my observations in Pleurobema coccineum as 
due to accidental conditions. But this is surely not so. I have seen 
now a great number of individuals of this species with the outer 
gills alone charged, and I have seen and examined many more, females 
in the sterile condition, which invariably had marsupial structure 
only in the outer gills, while the inner gills were different, and not 
built to receive eggs. Not a single exception was observed. 
Further it is quite evident that the arrangement of four gills serving 
as marsupia is found in a number of groups, the species of which are 
undoubtedly closely allied. This is clear in the peculiar Quadrula 
piicata-group, in the metanevra-group and others. Then again, a 
marsupium formed by the outer gills alone is characteristic of other 
natural groups. To me the most interesting case was that of Pleuro- 
bema coccineum. Here I discovered first that this supposed Quadrula 
differs from the Quadrula-type; I also discovered that this species 
intergrades with Q. obliqua and Q. pyramidata. This being the case, 
I concluded that the latter also should have a marsupium like coc- 
cinea. And this proved to be true! 
Nevertheless the character of the marsupium should not be too 
implicitly relied upon. There is no question that the condition in 
which the four gills serve as marsupia is more primitive than the stage 
where only the outer gills are marsupial.8 But it seems to me that 
8 The functional and morphological progress from the four-gill-marsupium to the 
two-gill-marsupium has been correctly understood and expressed by Haas (1910e, 
p- 19). 
