OF THE DIATOME^. 33 



Diatoms, referred by Mr. Brightwell {J. M. S. 1856, p. 106) to the geniis CJice- 

 toceros, have highly developed spines on the valves, besides the two pairs of 

 very long filiform smooth or spinous horns spiinging from the fi'ustules 

 themselves, or froin the interposed cingulum. Cerataulus is another genus 

 provided with a pair of long horn-like processes. 



Great variety of outline may prevail in a genus, so considerable indeed 

 that an accurate definition is mth difiiculty laid doT^Ti, the characteristics 

 shading off through several species, untU at length the similarity to an as- 

 sumed typical form is much diminished, whilst on the other hand an ap- 

 proach is made towards the features of another genus. The like latitude of 

 form prevails also with species, and gives rise to very numerous and fre- 

 quently perplexing varieties. On this topic Prof. Smith remarks — '^ While 

 a typical outline of its frustule is the general characteristic of a species, this 

 outline may be modified by the accidental circumstances which smTound the 

 embryo during its growth and the development of its sihcious epiderm ; then, 

 any such aberration of form becomes stereotyped by the process of self-divi- 

 sion of the frustule, generating multitudes of others slightly deviating from 

 the normal form." It must not be forgotten that the figui'e is greatly modified 

 or entu-ely changed by the position of the valves, whether seen in face or on 

 one side ; for each frustule generally presents four planes or sides, and, unless 

 regard be paid to this circumstance, one genus may be mistaken for another, 

 or even each view be presumed a distinct genus. Thus in the genera Navicula, 

 Pinnularia (XII. 5, 6, 15, 18), and in many others, the frustules are on one 

 aspect boat-shaped, but on the other oblong with truncated ends, or prismatic. 

 In the genus Triceratium (XI. 43, 44), the difference of figure is very re- 

 markable according to the side viewed (as presently illustrated). It is there- 

 fore necessary to examine a specimen on eveiy aspect it presents : this can 

 generally be effected by the accidental rolling over of frustules under inspec- 

 tion, or can otherwise be brought about by a very slight sliding movement 

 of the thin glass cover upon the slide under the microscope. 



Mr. Brightwell thus describes and explains the transitions of form produced 

 by a change in position of the frustules of the genus Triceratiuin (J. M. S. 

 i. 248) : — " The normal view of the frustule may be represented by a vertical 

 section of a triang-ular prism. If the frustule be placed upon one of its flat sides, 

 we look down upon its ridge and obtain a front view of its two other sloping 

 sides. If it be placed upon one of its ridges, we have a front view of one of 

 its flat sides, generally broader than long, and of its smooth or transparent 

 suture or connecting membrane. If the frustule be progressing towards 

 self- division, it is then often considerably longer than broad, and when nearly 

 matured for separation presents the appearance of a double frustule." 



It would be in vain to attempt to describe all the numerous forms assumed 

 by the members of this extensive family ; the representations in the plates 

 of this volume will convey the clearest notions of their diverse outline and 

 markings (see Plates 4 to 17). Great difference unfortunately has existed re- 

 specting the sides which should be esteemed primary and afford specific cha- 

 racteristics, and those which should be held as only secondary; and the 

 nomenclatui-e of the surfaces has been equally a matter of dispute and imcer- 

 tainty. Ehrenberg employed the terms dorsum, venter, and lateral surfaces 

 or sides, but so loosely that they do not always Indicate homologous portions. 

 Thus he has often called a convex surface the dorsum, simply from its convexity, 

 and a concave one the venter, on account of its concavity. Kutzing attempted 

 a more certain and scientific phraseology by calling those sides which have 

 no central opening (umbilicus), but through which self-di\ision occurs, the 

 primary sides, and the other two the secondary sides, further distinguishing 



D 



