46 



THE AMERICAN MONTHLY 



[March, 



but this is a mistake. It is useful for 

 all that is within the working distance 

 of the objective. It requires a more 

 careful collar-adjustment of the glass, 

 and an attention to the manipulation 

 of the light which is troublesome ; 

 but it gives us facts which we can 

 get in no other way. It is almost 

 impossible to get the field free from 

 glare, but this is chiefly due to the 

 reflection of light from objects not in 

 focus, and does not prevent our get- 

 ting sharp delineation of the surface 

 which is properly focussed upon, and 

 the discomfort and inconvenience can 

 well be borne for the sake of the in- 

 crease of knowledge of objects which 

 we get. It has been a rule with me, 

 therefore, to consider my examination 

 of any diatom incomplete until the 

 diatom has been subjected to careful 

 and repeated observations with the 

 vertical illuminator. It is a matter of 

 regret that it is not always easy to 

 obtain for this purpose dry mounts of 

 thoroughly well-cleaned material 

 burnt upon the cover-glass. 



Triceratium favus. — In studying 

 the diatom-valve, the natural course 

 is to begin with the coarser and more 

 strongly marked forms, and proceed 

 toward the finer and more delicate. 

 Microscopists are generally agreed 

 that the hexagonal marking of Tri- 

 ceratium favus^ of Coscinodiscus 

 radiatus, and of other species of simi- 

 lar appearance, is caused by an ar- 

 rangement of cells of true honeycomb 

 form. That these areolae are closed 

 at one end has been proved in the 

 case of Triceratium by the well- 

 known appearance of smaller dots 

 within the hexagons ; but whether 

 both ends are closed has been a vexed 

 question, which it is almost impossi- 

 ble to decide from the study of trans- 

 parent specimens. Here the exami- 

 nation of an opaque mount quickly 

 removes all doubt. Take Moller's 

 beautiful slide of Cuxhaven diatoms, 

 mounted opaque for the Lieberkuhn, 

 and use upon it one of the quarter- 

 inch objectives mentioned above, or 

 Spencer's two-thirds of forty-seven 



degrees with a half-inch solid ocular. 

 If we find a valve of T. favus with 

 the convex side uppermost, we shall 

 see a white surface, regularly studded 

 with circular bosses, or dark spots, 

 looking as if they were a bubble-like 

 film of the silex lifted a little above 

 the general surface. Each of these 

 is in the centre of a hexagon distinctly 

 traced about it, and when the surface 

 is accurately in focus we find that 

 there is no elevation of the dark spot, 

 but that the appearance is due to the 

 fact that the silex is extremely thin 

 over it, being, therefore, nearly black 

 and transparent, like very thin clear 

 ice in a pond, whilst it grows thicker 

 toward the walls of the areola, assum- 

 ing the dead-white appearance which 

 gives all the thicker diatoms on the 

 slide the look of biscuit porcelain. 

 The three horns, and the general con- 

 vexity of the valve, show that we are 

 looking at its outer surface. Let us 

 find now, as we easily may, a valve 

 with its concave or inner surface 

 turned upward. If we use the binoc- 

 ular instrument, the depth of the 

 hollow shell and the projection of the 

 sides of the valve toward us are very 

 striking. With the higher ocular we 

 notice that the walls of the valve are 

 quite thick, and that they have cross- 

 lines indicating a cellular structure. 

 On focussing vipon the interior surface 

 of the shell we find the same black- 

 line tracing of hexagons which was 

 seen on the convex surface, but the 

 area of each hexagon is flat and white. 

 Careful examination with the highest 

 ocular the objective will bear shows 

 that the white surface is dotted. 



We must now seek the assistance 

 of a more powerful objective and 

 the vertical illuminator, using a dry 

 mount of any gathering in w^hich the 

 Triceratium is found. The conclu- 

 sions we have drawn from the appear- 

 ance of the convex surface of the 

 valve are now corroborated, and we 

 find the concave surface strikingly 

 even, the hexagonal marking being 

 only a dark outline in the general 

 plane ; but we also find that the dots 



