STRUCTURE OF DIATOMS. 199 



able. If the cellules are closed at both their extremities during 

 the life of the organism, each cellule must be full either of proto- 

 plasm or of some other more or less fluid substance, unless, indeed, 

 each contains a gas, or constitutes a perfect vacuum, which is 

 scarcely within the bounds of possibility. If each contains proto- 

 plasm, it is obvious that the remains of this, during the mounting 

 of the specimen, would be recognisable amongst the larger species, 

 either by optical or chemical tests. During the boiling in acid, or 

 burning on mica, the fluid contents would burst the films, and in 

 many cases leave behind the evidence of their former condition. 

 In his experience, such evidence has never been forthcoming, and 

 judging from what is known of cellular structure in organic life 

 generally, there are no examples of truly vacuous cavities, inasmuch 

 as all organic tissues are pervious to dialytic or osmotic action. 



It is no doubt true that the organic silica of the diatom, per- 

 fectly hyaline as it looks, is in reality a " colloid," and hence, as it 

 contains an infinitesimal percentage of water, just as flint itself does, 

 dialytic action may take place through the film under notice. But 

 even then the perviousness to moisture of the diatom, if it really 

 keeps the chamberlets full of fluid during the vitality of the 

 organism, would not suf^ce to settle the question ; for, if any fluid 

 whatever remained in the cellules, should the specimens have been 

 but recently taken from their element, it would burst the film on 

 the application of heat, and inevitably burst the walls, whilst traces 

 of the disruption would occasionally be visible under the micro- 

 scope. Again, if the chamberlets contained gas of any kind, and 

 in spite of the effects of the boiling in acids, this gas were too 

 minute in quantity to burst the walls, we should certainly be able 

 to detect gas-bubbles in some of the chamberlets. But, as is well 

 known, the bubbles so common in mounted specimens are not 

 due to the cellules having originally contained gaseous material, 

 but to the accidental admission of air during mounting. 



The only remaining alternative is that the cellules cannot be 

 considered closed cavities, and hence that the alleged presence 

 of an external investing and closing film is illusory. 



VOL. IV. 



