OF THE DIATOME^. 35 



crowned bj^ several bristles or setae. Lastly, in Eiicainpia (II. 48) the 

 junction-surfaces are so excavated, that when the frustules are concatenated a 

 filament is formed, perforated by oval foramina. 



In not a few genera, as above mentioned, the attachment is at opposed 

 angles, and a zigzag chain produced ; but in Istlimia (X. 183) it is peculiar 

 in being indiscriminately made at any part of the adjacent frustule, and 

 thereby produces an iiTcgularly branched filament. 



The above examples will suffice to illustrate the characters and varieties 

 of concatenation in the form of filaments, whether straight or spiral ; but it 

 is necessary to add that the width of a filament " equals the length of the 

 frustule or valve measured along the suture or junction-line, and that the 

 breadth of the valve denotes the thickness of the filament " (Smith, Synojos. 

 ii. 6). In those instances in which frustules are connected together by a 

 process or small isthmus acting as a sort of hinge, the concatenation cannot 

 be ascribed to incomplete division only, for the existence of such a process is 

 the result of a special formation which essentially coiTesponds with the 

 pedicle or stipes of fixed species. 



Numerous Diatomese grow attached to foreign bodies by a stalk of variable 

 length, and which, although generally simple, is sometimes comjDound by 

 dividing and subdividing in a ramose manner. Even among the recognized 

 free Diatomeae, such as Navicula, Pinnularia, Nitzscliia, &c., specimens are 

 not unfrequently seen adherent by one extremity, about v>^hich they turn 

 or bend themselves as on a hinge ; however, in these instances such union 

 is but temporary, and no connecting medium exists. In Synedra (X. 184), 

 on the other hand, a bond of union occurs in the form of a little gelatinous 

 conical nodule, resembling very nearly the hinge-like isthmus which binds to- 

 gether the frustules of many genera in a sort of zigzag chain. By the process 

 of self- division it also comes to pass that groups of Synedrce occiu' attached 

 together by the same point, in a fan-like or a stellate form, as in S. radians, 

 S. affinis, &c. In other species detachment after fission is too speedy to 

 allow of this sort of combination, except of some two or four indi\dduals. 

 The fan-like collection of frustules is said to be flabellate, fan-like, or 

 radiating ; and when the component members are curved, they and the bun- 

 dles they form are described as arcuate. The nodule of attachment occurs 

 in various degrees of development, and attains in this same genus Synedra to 

 the dimensions of a pedicle — ex. in Synedra sujperba, and even to branch, 

 as in Synedra fulgens and Synedra pidcJielJa. 



^Tien a stipe branches, it does so normally in a dichotomous manner by 

 the very circumstance of the process of self- division, each new individual 

 produced by that act developing its o"v\ti secondary pedicle, or pedicel. 



This regular dichotomy is instanced in the genera Dorypliora (XIV. 21), 

 Cocconema (XIII. 10), and Gomphonema (XIII. 11). In Licmopliora (XIII. 

 20), and in one species of Rhiindoijliora, viz. Rh. Dalmatica, an irregular 

 branching — essentially dichotomous, however — is met with, and is thus ex- 

 plained by Prof. Smith : — " In Rhipidophora paradoxa and Rh. elongata, 

 self- division is immediately followed by the separation of the half-new frus- 

 tules and a dichotomy in the filamentous stipes, while in the present genus 

 the frustules remain for a time coherent, and continue dividing and mul- 

 tiplying on the summit of the pedicel, which becomes elongated and incras- 

 sated at each successive repetition of the process .... A branching, or rather 

 longitudinal rupture, of the pedicle takes place at irregular intei^als ; and the 

 entire organism presents us with more or less complete flabeUa (fan-like 

 clusters) on the summit of the branches, and imperfect flabeUa or single 

 frustules irregularly scattered throughout the entire length of the pedicel." 



B 2 



