By continual bisection a diatom is propagated through many generations, but at some stage or other, owing to an unknown cause, propagation by conjugation takes place. When two frustules are near to each other, two little swellings arise in one, which meet two little swellings in the other opposite to it. These soon unite and elongate, the septum or division between them is absorbed so that they form two tubes in which the endochrome of the two frustules becomes mixed, and a spore is formed in each of the two connecting tubes, which increase in size and change in form till they resemble in every respect the parent, except in being much larger. As these young diatoms swell, they split the two parent frustules, become free, and lay the foundation of a twin series of generations. In the Fragillaria only a single spore is formed.[33] In Surirella and Epithemia the manner of conjugation is somewhat different. In the former the valves of two free adjacent frustules separate from each other at the suture or line of junction and the two endochromes are discharged; they coalesce and form a single mass, which becomes enclosed in a gelatinous envelope, and in time this mass shapes itself into a frustule resembling that of its parent, but larger. In Epithemia, however, the endochrome of each of the conjugating frustules divides at the time of its discharge into two halves; each half of the one coalesces with each half of the other, and two frustules are formed which become invested with a gelatinous envelope and gradually assume the form and markings of the parent frustules, but grow to a much larger size, for the spore masses have the power of self-increase up to the time that their envelopes are consolidated. This double conjugation seems to be the ordinary type of the process among the diatoms.[34] But these plants multiply also by gonidia. It is thought probable that as long as the vegetative processes are in full activity diatoms multiply by self-bisection, but when a deficiency of warmth, of moisture, or of some other condition, gives a check to these, that they increase by gonidia, some of which becoming encysted, possess a greater power of resisting unfavourable circumstances, and thus the species is maintained in a dormant state till a change enables them to germinate. It is even thought they may be the origin of distinct species.