INSTITUTE OF SCIENCE, PHILADELPHIA. 55 



A fragment almost certainly of this species was found in the Caloosa- 

 hatchie marl. 



Marginella semen Lea. 

 Marginella semeti Lea, Contr. GeoL, p. 17S, pi. 6, fig. 190, 1833. 

 Marginella /arvala jun. Conrad, Foss. Tert. U. S., p. 45 (in error). 

 Marginella ovata Emmons, Rep. N. C. Geol, Sur., p. 261, fig. 136 (not of Lea). 

 Marginella sp. ? Conrad, Am. Journ. Conch, vi. plate 3, fig. (13?), 1870; figure in upper 



left-hand corner not numbered. 



Range, Claiborne Sands Eocene, Miocene of North Carolina. 



The Marginella semen of Lea was by Conrad referred to the young of M. 

 larvata Conrad {M. ovata Lea.), but typical specimens now before me show 

 this to have been an error. M. semen is a species which stands on the line 

 dividing Persiciila and Marginella. As Conrad observed in 1870, the subdivi- 

 sions of the Marginellidce in use for recent species are not applicable to the 

 fossils, which illustrate every degree of intergradation. M. semen shows the 

 tip of its spire, which is uncovered at all ages. It would therefore belong to 

 the section Cryptospira H. and A. Adams. Li M. larvata it is covered in the 

 adult as it is in the M. gravida, which seems to be the descendant of M. 

 semen. 



An interesting feature ol M. semen is that some specimens retain the color- 

 marks, which are dark, narrow spiral bands, seven in the adult. I do not find 

 this anywhere referred to. 



In later beds we find the following species also banded with rows of square 

 spots corresponding to the continuous bands in M. semen. In the Caloosa- 

 hatchie form the size is greater and the spire hidden under a vortex of shelly 

 matter, while the number of rows of dots is greater than that of the bands in 

 M. semen. In the recent fauna we find in the Antillean species most nearly 

 allied to M. semen and M. gravida that the size has increased still more, the 

 number of spiral color series is still greater, and the spire is more deeply hid- 

 den. The present Floridian Persieida {P. catenata) is not as nearly related to 

 these fossil species as are P. chrysomelijia, P. obesa Redfield, and P. interrnpta 

 Lamarck. There is no reference in Conrad's text to his figure of 1870, above 

 cited. It may have been intended for the following species, which it resembles. 



Marg-inella gravida n. s. 

 Plate 5, figure 3. 

 Shell stout, short, rounded, with the aperture as long as the shell ; color 

 usually whitish, but when wet sometimes showing traces of eight or nine 

 spiral series of squarish spots ; aperture narrow, arcuate, as long as the shell ; 

 outer lip internally crenulated with about a dozen lirse; inner with two rather 

 strong oblique plaits in front, the anterior plait being at the twisted edge of 

 the pillar, and four or five smaller plaits or ridges, not oblique and growing 

 fainter posteriorly ; body-callus moderate, its posterior extension forming a 



