54 MEMOIRS OP THE CARNEGIE MUSEUM. 



have been observed subsequently. All four gills are used as marsupia. The 

 eggs form placenta? of lanceolate shape, of a peculiar yellow-brown or pale orange 

 color. Glochidia have been observed in specimens from Holston River in Tennessee. 

 Their shape is subcircular; length and height about 0.19 mm. Thus they resemble 

 those of Q. metanevra (0.18 X 0.19), but they are more nearly circular. 



Breeding season: Gravid females were found on May 22, 1914; May 23, 1914; 

 May 24, 1911; May 25, 1914; July 8, 1913. Glochidia were observed on May 23. 

 The species apparently is bradytictic, as usual in the genus, but the date for glo- 

 chidia is rather early. Wilson & Clark (1914) found gravid specimens in the 

 Cumberland in June and July. 



Remarks: This is an eminently characteristic species. The chief variations 

 observed concern the general shape of the shell and the sculpture. The typical 

 shape is "subcjdindrical," with the height about the same, or nearly the same, as 

 the diameter (see measurements of Nos. 1 and 3, above), but there are rather 

 flat individuals. Sterki (1907o, p. 390) calls attention to a form from the Tus- 

 carawas River, which lacks the tubercles of the posterior ridge. This form is the 

 prevailing one in western Pennsylvania, and is most abundant in the Beaver 

 drainage and in French Creek. I have many specimens in which no trace of the 

 tubercles is seen, and generally also the smaller nodules of the anterior part of 

 the shell near the beaks and the sculpture of the posterior slope are absent in them, 

 so that the shell appears absolutely smooth. Yet there are all sorts of intergrades, 

 connecting this smooth with the normal form (See PL V, figs. 2, 3). Farther down 

 in the Ohio this smooth form is not found. These conditions are to a degree parallel 

 with those observed in Q. metanevra and its var. wardi. 



Very often the degree of obesity of the shell is correlated with the develop- 

 ment of the sculpture, so that smooth specimens are at the same time unusually 

 compressed. But this is not always the case, and smooth individuals may be 

 subcylindrical, as our No. 1 (See measurements) and sculptured individuals may 

 be somewhat compressed (See No. 4). 



This smooth, and sometimes compressed form of the headwaters is so far 

 known only from the Tuscarawas River in Ohio, from Beaver River and French 

 Creek, in Penns^'lvania, and from the ui)per Monongahela (West Fork River) in 

 West Virginia. 



It should be mentioned in this connection that a similar, compressed form is 

 known from the headwaters of the Clinch River, in Virginia. This is the var. 

 strigillata (Wright), but it is not a smooth form. On the contrary it is covered 

 by a multitude of small tubercles. But here also the large tubercles of the posterior 

 ridge are absent. 



