76 Hirsu : AIR CHAMBERS IN THE RICCIACEAE 
almost exactly indicated by the diagrammatic scheme given by 
Barnes and Land in their figure 1. During the formation of the 
filaments there is a marked growth in the thallus in which the pro- 
truding filaments do not share, so that the air spaces broaden per- 
Ficure 5. éiccia Frostii, Later stage, showing similar structures. 
ceptibly. In the older parts of the thallus the air spaces may 
become nearly completely closed (FriGuRE 6, P) through the en- 
largement of the terminal cells of the filaments; but, nevertheless, 
they retain their canalicular form and present an entirely different 
appearance from the irregular but more or less polygonal air spaces 
of Ricctocarpus natans, 
This method of development is clearly in harmony with Leit- 
geb’s account of the origin of intercellular spaces. In his ‘‘ Un- 
tersuchungen tiber die Lebermoose”’ Leitgeb maintains that near 
the growing point on the dorsal surface there appear between four 
cells small pits that in profile view are simply dark points, ‘that 
these impressions arise not by cleavage but in this way, — that the 
free outer walls of the enclosing cells grow up over these places.” 
B, C, and D in Figure 6 are without doubt these pits. He states 
further that there is a possibility that occasionally these depres- 
sions could elongate through a splitting from without in; but he 
rejects this in favor of the view that their increase in size is accom- 
plished by a growth in thickness of the thallus dorsally.* He 
thus believed that the sex organs are visible at the same time as 
the pits that form the very beginning of the intercellular space 
formation, and that, as in the case of the former, the whole cell 
* Leitgeb, Untersuchungen iiber die Lebermoose 42 EJ; 27) 1870, 
