OF THE DIATOME^. 97 



Slimed species of Navicidm under the name of N. varians. So again Dr. 

 Greville, speaking- {A. N. H. 1855, p. 258) of the Grammatopliora (?) Bal- 

 fouriana (Smith), which he erects into a new genus Diatomella, observes, 

 *' There is greater variation in the rehitive length and breadth of the fnistules 

 than would be likely to occiu- in other Diatomaceous grouj)s. In some the 

 length is more than equal to twice the breadth, while others are exactly- 

 square ; and between these two extremes every gradation may be observed ; 

 resembling in this inequality Fmgilaria, Odontidium, Grammatopliora, and 

 other filamentous genera having piano- compressed frustules." But in this 

 very case a difference arises between Dr. Gre\ille and Mr. Smith respecting 

 the value of internal marldngs as a characteristic distinction ; for the latter 

 author remarks, '' The absence of a curve in its septa, relied upon by Dr. 

 Greville, I cannot regard as of sufficient importance to constitute a generic 

 distinction, as this featui'e is scarcely noticeable in some states of Gramma- 

 tophora macihnta, and is imiformly absent in G. strktaj^ (Synojms, vol. ii. 

 p. 44.) 



" The size of the matui^e frustule " (says Prof. Smith, J. M. S. 1855, p. 132) 

 "before self-di\asion commences, is, however, dependent upon the idiosyncrasy 

 of the embryo, or upon the cii'cumstances in which its embryonic growth 

 takes place ; consequently a very conspicuous diversity in their relative 

 magnitudes may be usually noticed in any large aggregation of individuals, 

 or in the same species collected in different localities. 



" It may also be easily conceived that, while a typical outline of its cell 

 must be the characteristic of a certain species, such outline may to some 

 extent be modified by the accidental circumstances which suiTOund the em- 

 bryo during its earlier growth and development. A lanceolate form may 

 become linear, elliptical, or even somewhat oval, by the pressure of siuTound- 

 ing cells ; and acute ends may be transformed into obtuse or rounded ex- 

 tremities. 



" Those who understand the process of self- division will see here a suffi- 

 cient reason for the occurrence of multitudes of frustules deviating from the 

 normal form, or even for the existence of myriads at one spot, all having a 

 form different from the type,- — the single embryo from which they have all 

 sprung by self- division (which process stereotj^es the shape with which it 

 commences) having from some accidental circumstances become modified in 

 its outline. 



" It follows, then, from these considerations, that neither size nor outline 

 is sufficient to enable the observer to determine the species of a Diatoma- 

 ceous frustule. If he has the means of comparing specimens in sufficient 

 numbers and from various localities, he may fix with tolerable certainty 

 upon the magnitude and form which may be regarded as the average and 

 type of the species ; but, without such opportunities, a reliance upon such 

 characters will inevitably lead to the undue multiplication of species and to 

 a confused and erroneous nomenclature." 



In the construction of genera, similar difficulties present themselves. Thus, 

 Mr. Brightwell complains (J. M. S. i. 252) — " It appears as if we could carry 

 our real knowledge Httle beyond that of species ; and when we attempt to 

 define kinds and groups, we are met on every side by forms which set at 

 nought our definitions. With reference to the species of the present genus 

 {Triceratium), looking upon T. favus or T. megastomum as what we con- 

 ceive to be the most perfect plan (if any) on which this group is constructed, 

 we find all the species diverging from it, and caiTying us to analogous forms 

 in other groups, or lost in them. Placing the perfect triangular forni of 



