176 
Ethel N. Thomas. 
In transverse section these dividing septa appear as round or 
oval pitted areas upon the more or less oval cells (PL III., figs. 2 and 4). 
In longitudinal section the cells are somewhat elongated, and 
frequently present the appearance of a flask with laterally directed 
neck (PI. III., fig. 5). It is only when the arms are presented to the 
observer end on, that their septa are seen as perforated discs, and 
as the walls of the chambers are usually but one cell thick, it is 
clear that they are found only on the cells of the radial wall in a 
tangential section (PI. III., fig. 6) and on the tangential cells as seen 
in a radial section. The tangential and radial cells in these sections 
respectively, shew the arms in profile (PI. III., fig. 6). 
The connecting arms frequently expand at the cross wall, 
recalling the similar expansions seen in sieve-tubes and “ trumpet 
hyphae.” 
Karsten 1 has described similar tissue in several Mangroves, 
notably in the roots of Bruguiera gymnorhiza. He figures disc-like 
connecting walls, which look almost precisely like those just 
described, but he does not mention any perforations. On the other 
hand the cells of Bruguiera possess thickening bands not found 
in Acrostichum aureum. Experiments on mangroves and on water 
plants have shown that such tissues are for the purpose of aeration, 
and we cannot doubt that here too the same function is performed. 
The “air tissue” is found most abundantly on the proximal 
portion of the large roots. At the distal end it is almost absent, 
and the lateral roots form short sharp spines similar to those found 
on the aerial roots of Dioscorea prehensilis . 2 
The very close similarity of response met with in the mangroves, 
whose members are drawn from very widely separated orders of 
flowering plants, has always been a point of considerable interest, 
but this parallelism of development is brought out with much 
greater force, when we find the same response in the one member 
of this oecological association derived from the Vascular Cryptogams. 
Under normal conditions, roots are very constant in their 
general anatomical features, and this is doubtless correlated with 
the constancy of the medium in which they grow, and of the 
functions which they have to perform. When one or both of these 
conditions is unusual, as in the present case, a very definite modifi¬ 
cation is met with, and this modification can with greater certainty 
1 G. Karsten. Ueber die Mangrove-Vegetation in Malayischen 
Archipel. Bibliotheca Botanica, Heft 22, 1891. 
2 Hill and Freeman. The Root-structure of Dioscorea prehensilis. 
Annals of Botany, 1903. 
