of the Drsinulica" and DiatomcT. 18 



weak solution of carbonate of ammonia), the contents retract on 

 both sides from tlic annular rim, and constitute two completely 

 sei)ariite cell-like structures (halves of a primordial utricle), 

 each of a very long ellipsoidal form, and each lyinjr close 

 against one of the primary sides {faces of halves) of the cell 

 (fig. 31). When the annular rim has grown inwards to about 

 the sixth part of the shortest diameter of the cell, its develop- 

 ment is arrested. In natural conditions, this stage is suc- 

 ceeded by the rctractitm of the primordial utricle from it. 

 Each of these halves of the cell-contents becomes clothed, on 

 the side turned away from the ])rimary side of the cell, with a 

 new membrane (tigs. 32, 33), which soon exhibits the first indi- 

 cations of the peculiar thickening ribs and nodules of one of 

 the primary sides of our Pinnularia. The cell has now completed 

 its division. Seen from one of the secondary sides, it contains 

 two new individuals, equal to the mother-cell in length and 

 breadth, but only possessing one-third of its thickness. The 

 externally-situated primary side of each of them is the old pri- 

 mary side of the mother-cell, to which we must imagine the 

 newly-formed membrane of the daughter-cell closely adherent 

 at all points. Perhaps the narrow secondary sides of the new 

 cells may be in the same condition. But the contiguous pri- 

 mary sides of the daughter-cells are totally new structures, 

 which, developed ra})idly, in a short time become similar to the 

 old primary sides in every part. The two daughter-cells are at 

 first held together by the broad, middle piece of the secondary 

 sides of the mother-cell, bearing the above-mentioned annular 

 rim inside. The contents of the intermediate space consist of a 

 transparent fluid destitute of any solid structures, doubtless pure 

 water. The two daughter-cells are finally set free by the gradual 

 ' weathering' of the zone-membrane which holds them together. 

 The division of Surirella hifrons takes place exactly in the same 

 way. An essentially similar kind of vegetative multiplication 

 is widely diffused, if not general, in the Diatomejc. The well- 

 known phenomenon of the formation of a tubular membrane, 

 often impregnated with silex, and elegantly dotted or areolated, 

 connecting the two segments of Isthmia, Melosira, &c., depends 

 upon the same process*. 



An analogous case is met with in the formation of the spores 

 of Pellia epijjhyUu. The mother-cell here produces six ridges of 



* [The author speaks of the amnilar Tiur as if it were a product of the 

 wall of the ])arent cell. It is rather a neiv product, formed in the interval 

 between the two valves which have separated at the suture (somewhat as 

 happens m (Edngonium). According; to the degree to which the valves 

 separate, the intermediate piece is a slender ' hoop/ as in Xavicula, or a 

 long tubular piece, as in Melosira, &c. — A. H.] 



