46 University of Michigan 



until after counting the laterals. The fifth lateral is turned 

 considerably inward, and is almost completely hidden by the 

 distal end of the first marginal. The break to the marginals 

 is a sharp one, and shows as a raised edge, even under low 

 magnification. The well-developed marginals are all tricuspid 

 and point obliquely inward, and the transverse row itself also 

 slopes obliquely backward (i. e., in the direction towards which 

 the cusps point). As the inner cusp of each marginal overlaps, 

 to a certain extent, the outer one of the preceding tooth, it is 

 sometimes difficult to make out more than 2 cusps, which prob- 

 ably accounts for Binney's statement that only a portion of 

 the marginals are tricuspid. The lenses in his time were con- 

 siderably inferior to the modern oil-immersion objective. 

 Amongst the outer reduced teeth, the thirtieth and thirty-first 

 have four cusps each, while the outermost is a mere denticle, 

 and is lacking in some of the transverse rows. 



For comparison, the jaw and radula (fig. 2) of Guppya 

 stcrkii (Dall)^ was also examined. This species, as Vanatta 

 ( 1920) has already pointed out, has a similar dentition to that 

 of G. gundlachi, only the radular ribbon is so minute as not 

 to fill the field of the oil-immersion objective. The central 

 tooth, for instance, is only about 4 microns (.004 mm.) in 



I 5 13-15 

 width. The formula is approximately : C — ; L — ; M . 



3 3 3 



The number of cusps out to the ninth marginal was determined, 

 but their shape on this tooth could not be made out very accu- 

 rately, as the ends of the cusps are smaller than the limit of 

 possible microscopic vision, and so could only be detected as 



1 Dried animals; A. N. S. P. No. 46177; collected at the Clydesdale 

 Brick and Stone Company Farm, Beaver County, near Elwood City, 

 Pa., by J. B. Clark. 



