18S4-S5.] Edinburgli Xaturalists Field Ciub. 239 



of Triceratium, where, in the one view, the organism is triangular, 

 but in the other quachilaterah 



Seeing, then, that the configuration of the siliceous cell-wall is 

 so prominent, it is evident that it will be of the greatest import- 

 ance in the elaboration of any system of classification ; hence it 

 becomes necessary to designate its various component parts by a 

 well-defined phraseology. This has been recognised since the time 

 of Ehrenberg, who introduced the terms dorsum and venter to 

 signify a convex and a concave surface respectively ; but it is 

 obvious that these designations could readily be applied to non- 

 homologous parts, hence Kiitzing, Eabenhorst, and others adopted 

 an improved but somewhat misleading nomenclature, which may 

 be most clearly explained in the following manner : — 



T7, , . /T> ir 1 Ttr "^ (. Lateral surfaces (Ehrenbero,-). 

 1* ront view (Kalfs and W. i i ^i • ■^ /t^-. • s ' 



o -^1 N . 1 , f ) rrimar}^ side Kutzmo-). 



Smithj — I.e., hoop next V = < „ / • 1 /o i i i.\ 



1 i i becondary side (liabenhorst). 



^^^"^■^'^^' Zonal 



view. 



0-, . /T) ir 1 -nr ^ T Dorsum and venter (Ehrenberg). 



bide view (Kalis and \ V. / \ r, i -t /t- -. • n 



o -^1 \ ^ • c e{ I becondary side (Kutzine;). 



bmith) — I.e., surface of > = ■-, -n ■ • n /t> 1 1 .\ 



^ . \ i i rrimary side (Kabenhorst). 



valve next observer, \ I ^^ •' . ^ J 



' J \ Valval view. 



There can be no doubt that the terms having reference to the con- 

 necting zone and valves are those which are most natural, and 

 which will ultimately gain the acceptance of all. 



In habit, as in form, great variations occur in this interesting 

 group. In very many instances the individual frustules are met 

 with in a free condition, but in others stipitate and concatenate 

 or ramose forms are to be found. It is hardly necessary to note 

 that the form of any chain, whether straight or curved, will depend 

 on the figure of the component frustules ; but it may be men- 

 tioned that both the degree and nature of the attachment of adjoin- 

 ing frustules vary within wide limits. Thus in some cases the 

 union is so slight that one frustule can glide over the other 

 {e.g., Bacillaria) ; while in others the junction surfaces are dentated 

 or provided with excrescences and pits, in order to effect a more 

 complete connection. The nature of the attachment, moreover, may 

 be temporary or permanent. In the former case the free forms be- 

 come adherent by one extremity, but no special connecting medium 

 exists ; while in the other there may be a nodule or pedicel for 

 fixation. This stipes, again, may branch either in a dichotomous 

 manner, as in Gomphonema, or irregularly, as in Licmophora ; but 

 its length is a matter of secondary imj^rortauce, as it is found to 

 vary in any species according to the inherent vigour of the species 

 or the nature of the external influences to which it is subjected. 



