OF THE DIATOME-E. 61 



millions in a single month, — a circumstance which will in some degree explain 

 the sudden, or at least rapid appearance of vast numbers of these organisms 

 in localities where they were, but a short time previously, either unrecognized 

 or only sparingly diffused." 



This multiplication by self- division now described, is generally supposed, 

 after a time, so to speak, to exhaust itself, and thereby to render necessary 

 other plans of propagating species. That some other modes do really exist 

 is suggested by the fact of the considerable variations of size of frustules of 

 the same species obtained at one time from the same locality, and moreover 

 by diversities in the relative distance and in the delicacy of the strise of the 

 surface. One such mode of propagation Mr. Thwaites has demonstrated to 

 consist in the production of sporangial frustules by a process of conjugation 

 analogous to that in the Desmidiese and many other Algae. 



Con jroATio^r. — The method of conjugation, although essentially alike in all 

 cases, exhibits several important modifications in the genera of this family. 

 These were more or less clearly perceived by Mr. Thwaites, who spoke of them 

 as exceptional varieties ; but to Mr. Smith belongs the credit of reducing all 

 of them under four principal forms: viz., 1. That in which two parent frustules 

 produce two sporangia by conjugation, as mEpitliemia,Cocconema,Gomjphonema, 

 Encyonema, and Colletonema. 2. Two parent frustules generate a single spo- 

 rangium, e. g. in Himantidium. 3. '' The valves (vol. ii. p. xii) of a single 

 frustule separate, the contents set free rapidly increase in bulk, and finally 

 become condensed into a single sporangium. This may be seen in Cocconeis, 

 CijcIoteUa, Melosira, Orthosira, and Schizonema. 



" In Melosira nummuloides, M. Borrerii, and M. subflexilis, the second valve 

 of the conjugating frustule is rarely found united to the mucus surrounding 

 the sporangium, the conjugation taking place only in the last frustule of the 

 filament ; but in Melosira varians and Orthosira orichalcea, conjugation taking 

 place throughout the entire filament, both valves are usually found adherent 

 to the sporangium or its surrounding mucus. 



" From a single frustule, as in the last method, two sporangia are produced 

 in the process of conjugation : this takes place in Achnanthes and Bhahdonema.^' 



In describing the process as generally as possible, we cannot do better than 

 follow Mr. Thwaites's account, although it is illustrated by an example taken 

 from the fii^st categorj" of variations. " For the most part," he tells us, 

 *' conjugation in the Diatomeee, as in the Desmidieae, consists in the union 

 of the endochi'ome of two approximated fronds, — this mixed endochrome 

 developing around itself a proper membrane, and thus becoming converted 

 into the sporangium. In a veiy early stage of the process, the conjugated 

 fnistules, as in Eunotia turgida, have their concave surfaces in nearly close 

 apposition (XI. 1), and from each of these surfaces two protuberances arise, 

 which meet two similar ones in the opposite frustule (XI. 3) ; these protu- 

 berances indicate the future channels of communication by which the endo- 

 chrome of the two fnistules becomes united, as well as the spot where is 

 subsequently developed the double sporangium, or rather the two sporangia. 

 A front view of two frustules at the same period shows each of them to have 

 divided longitudinally into two halves (XI. 4), which, though some distance 

 apart, are stiU held together by a very delicate membrane : this, however, 

 soon disappears. 



'' The mixed endochrome occurs, at first, as two irregular masses between 

 the connected frustules ; but these masses shortly become covered, each with 

 a smooth cyUndiical membrane, — the young sporangia, which gradually 

 increase in length (XI. 5, 6), retaining nearly a cjdindi'ical fonn (XI. 7), 

 until they far exceed in dimension the parent frustules, and at length, when 



