136 BULLETIN 100, UNITED STATES NATIONAL MUSEUM 



PLEUROSIGMA NORMANII Ralfs 



(Peragallo, Pleuro., pi. 4, fig. 6; Cleve and Grunow, Arct. Diat., pi. 3, fig. 67.) 



A specimen found by Mr. E. Leonard, Liverpool, England, in Phil- 

 ippine Islands material supplied by me, agrees more closely with 

 Peragallo's figure than with that in Cleve and Grunow's Arctic Diat- 

 oms. The union of this species with P. affine Grunow is questionable. 

 Peragallo makes it a variety of the latter, while Cleve (Nav. Diat., 

 vol. 1, p. 40) puts P. affine (1880) in P. normanii (1861). 



PLEUROSIGMA OBESUM, new species 



Plate 29, fig. 9 



Valve broad, sigmoid, tapering to the rather blunt rounded ends, 

 the curve of one side from middle to end being strongly convex, that 

 of the opposite side barely concave; rhaphe strongly oblique to the 

 long axis at the middle of the valve, but running perfectly straight 

 for half the distance to the ends, then curving strongly as it ap- 

 proaches the convex side from which it barely remains separated; 

 terminal and middle ends of rhaphe beaded, the former not quite 

 reaching the ends of the valve; central area expanded into a small 

 transverse oval; markings oblique to the long axis, but crossing each 

 other at right angles, so that no transverse or longitudinal lines are 

 present. 



Length, 0.312-0.351 ; width, 0.063-0.067; 11.2 lines in 0.01 mm. 



There is a rather close resemblance between this species and P. 

 heros Cleve (Nav. Diat., vol. 1, p. 44, pi. 4, fig. 20), but the markings 

 of the latter are finer, its shape is relatively less obese, and its rhaphe 

 is parallel with the long axis at the middle of the valve instead of 

 strongly oblique to it. 



In connection with the markings of this species it is in place to 

 say that the generally accepted custom of dividing this genus into 

 two groups, first, those with quincunx markings — that is, one set of 

 lines transverse and the other two oblique — and, second, those with 

 rectangular lines, one set transverse and the other longitudinal, is 

 not correct. As stated in the discussion of the genus, one other 

 pattern is more or less prevalent, namely, species having two sets of 

 rectangular lines, not transverse and longitudinal, but both equally 

 oblique to the long axis of the valve. In the truly quincunx forms, 

 like P. angulatum W. Smith, P. robustum W. Smith and the next 

 species P. obtusum, new species, the lines radiating from any given 

 point are six in number, and the six angles between them are theoret- 

 ically 60°; but in the above diatom there are but two sets of lines; 

 that is, four radiating from any given point, and the angles between 

 them are all 90°. The most conspicuous example of this arrange- 

 ment is P. japonicum Castracane in his Report of the Challenger 

 Expedition, plate 29, figure 14, which occurs abundantly in the Phil- 



