OP THE DIATOME^. 99 



the epiderm ; and this combination seems to follow certain and invariable 

 laws, which are subject to no derangement from the external circumstances 

 in which the growth of the embrj^o may take place. The structure of the 

 valve reveals itself in the character of the striation, which may therefore be 

 found a good specific distinction." Thus the striae may be costate or monili- 

 form, parallel or radiate, reach the median line or be absent from a greater 

 or lesser portion of the siu'face, &c. The relative distances and the distinct- 

 ness of the stria3 are also other features to be recorded, allowance being made 

 for the influence of localities and of age, and for the fact of their having 

 originated from the same or from different sporangia. 



Next to striation in importance is locality, which will often aid to discri- 

 minate between closely allied forms, since fresh- and salt-water species 

 cannot exchange habitats. Locality also seems even more restricted by other 

 external conditions of a more limited nature. 



Lastly, the arrangement of the endochrome confers a specific character 

 more certain than habitat. Examples of various arrangement of gonimic 

 substance, and of the large, constant, oil-like globules, have been already 

 given. 



It follows, therefore, that the difficulty of defining species is much en- 

 hanced where examples occur only in a fossil state. Even in the living 

 state, shape and size cannot be implicitly relied on, but gatherings are re- 

 quired from different localities, and everj condition of growth observed, before 

 an average size or a typical outline can be decided on. And although stria- 

 tion is an important guide, it often happens that this feature is so nearly 

 alike in allied species of the simple forms, such as Cocconema, Cymhella, and 

 Navicula, that oiu' determination must be influenced by less important con- 

 siderations, and the habitat, outline, and an^angement of cell-contents all 

 require to be brought imder review before we should feel justified in consti- 

 tuting a species. 



In the construction of genera, the several conditions (viz. form, size, stria- 

 tion, habitat, and disposition of endochrome) employed in the determination 

 of species are also resorted to. Other peculiarities, however, are noted, 

 such as the transverse or longitudinal lines or bands, indicating thickenings 

 of the valves, the presence of a central spot (umbilicus) or of terminal ones, 

 and (as Prof. Smith mentions) ^' the obvious varieties of form or combination 

 to which the ceUules submit in the progress of their formation, exhibiting 

 themselves as hexagonal, circular, or irregular in outline, as distinct from 

 each other, or as more or less confiuent." {Synops. vol. ii. p. xxiv.) 



Kiitzing has extensively used the circumstance of the presence or absence, 

 the number and the position of apparent pores, not only in constituting 

 genera, but also the higher divisions, families and orders. The figure of 

 frustules on a transverse section, or an end view, is another point he has 

 resorted to in framing his classification. He would, indeed, appear to assign a 

 yet higher importance to the central spot or umbilicus than Ehrenberg him- 

 self, siace he has distinguished his tribes Striatce and Vittatm, respectively, 

 into two orders, Stomaticce and Astomaticce, according as this structural pecu- 

 liarity is present or absent. So, again, in the case of the Navicular frustules, 

 he has constituted Surirella with some other genera into a family Surirellece, 

 separated from Navicula, Pinnularia, and other genera, and placed in a dis- 

 tinct order of Striatce, because the former group is destitute of an umbilicus 

 (hence Astomaticce), which the latter possesses (the Stomaticce). Moreover, 

 as the family Naviculeae, along with others, presented an umbilicus on each 

 valve of their frustule, the term Distomaticce was applied to distinguish them 



h2 



