22 On the Structure of Diatoms. 



sharply-defined elongated hexagons.* Hartnack's hexagons as 

 figured are too much elongated ; although sometimes such an ap- 

 pearance is presented when the illuminating pencil is at right angles 

 with the median line, the transverse lines being less distinctly per- 

 ceptible. When the light is so arranged as to show every side with 

 equal perfection, the form of the markings is nearer that of regular 

 hexagons. 



The Amici prism is found to work excellently on the SurireUa, 

 and when it is used with the ^V objective and the blue cell the 

 slightly elongated hexagons are easily exhibited on an average 

 frustule. 



Aulacodiscus Kittonii. — This splendid disk is traced with easy 

 angular figures evidently elevations, and the spaces between the 

 Hnes are undoubtedly depressions. Some of the markings are cir- 

 cular, others square, some pentagonal, some hexagonal, and others 

 heptagonaL Broken specimens of Brightivellia Johnsonii with Hke 

 surface markings show the line of fracture running through the 

 areolae. 



Navicula rhomhoides. — Individual frustules of this species vary 

 considerably in degree of difficulty of resolution. Some of the 

 smallest valves when mounted in balsam tax the powers of excellent 

 instruments. The writer has found all specimens, whether mounted 

 dry or in balsam, to yield readily transverse striae with oblique illu- 

 mination direct from the lamp. Under the same conditions an 

 average valve exhibits well-defined longitudinal striae. With the 

 ammoni-sulphate cell it is instantly and clearly shown covered in 

 every part with squares, like Pleurosigma Balticum. 



Navicida crassinervis. — The specimens of this variety, in my 

 possession, are more difficult than Frustulia Saxonica, and even 

 rival A. 2)eUucida under lamp illumination ; but any clean frustule 

 is satisfactorily resolved. 



Using monochromatic light with plain mirror and Wenham's 

 paraboloid, longitudinal hnes are discovered. After careful mani- 

 pulation both sets of lines are seen at the same time, and an appear- 

 ance of beading results. 



Navicula cuspidata. — Both sets of Hnes are easy, but the 

 longitudinal are much closer together than the transverse. Conse- 

 quently the light interhnear spaces are elongated and no semblance 

 of beading is to be seen. In diatoms where the intersecting striae 

 are of nearly equal fineness the little square spaces, when not well 

 defined, seem circular ; and if the illumination by transmitted hght 

 is intense, they present a raised appearance due to refraction. 



JMr. Charles Stodder called my attention to this diatom with the 

 view of ascertaining with the ^V whether or not the two sets of 

 lines lie in difierent focal planes. My observations, many times 

 * ' The Microscope,' Carpenter, page 182, 



