204 BOTANICAL GAZETTE [September 



vide a roof for the widening chamber^ the form of the cellj necessary 

 under this mode of growth^ would seem almost to preclude so unequal 

 a division (4, fig, i). Certainly if the pit were open at the time the 

 first division took place, the cells cut off could only close it by later 

 growth. As a matter of fact, in Leitgeb's plates no such stages as 

 we have sketched are shown; yet if the pit is formed as he says, they 

 should be found. In spite of diligent search we have seen nothing 

 that could possibly be interpreted so. 



Moreover, it is easy to see that Leitgeb was misled by his desire 

 to homologize the formation of air chambers with the formation of 

 pits for the sex-organs. So, though he admits seeing the primary 

 formation of intercellular spaces in Marchantia, Preissia, and Plagio- 

 chasma, he deliberately rejects the obvious explanation of splitting. 

 He did this in order to apply to Marchantiaceae the idea he had con- 

 ceived for Ricciaceae. So if it can be shown in Ricciaceae that the 

 intercellular space is formed by cleavage, and then splits out to the 

 surface, the immediate reason for such distortion will have been 

 cleared away, though it cannot be justified. Having homologized 

 the air chamber and sex-organ pit in Ricciaceae, he applied the 

 explanation perforce to Marchantiaceae. 



Now in the case of sex-organs, it is easy to understand how the 



^_^ 



enrow 



whole cell line is retarded, while its neighbors surpass it. The initial 

 at first grows more rapidly than do the adjacent cells of the thallus, 

 the protuberance and finally the protrusion of the initial being the 

 visible evidence of its more rapid growth. Later the adjacent cells 

 outgrow the sex-organ and it becomes "sunk" in the thallus. 



But in the matter of air chambers is involved the retardation or 

 the cessation of growth in a very limited part of the walls belonging 

 to four contiguous cells, while all the rest of these four walls and the 

 remaining sixteen grow throughout, and all decidedly outgrow the 



portion of the four! While such a thing 



it is wholly improbable and needs to be supported by most convincing 

 evidence. Such evidence Leitgeb nowhere presents, either in text 

 or figure. We propose to show, on the contrary, that the assumption 

 of such peculiar lagging is unnecessary, because all the observed 

 structures can be produced by common splitting, and the earlier stages 



i 



