aqtiilittiim, with some Notes on Marsilict quadrifolia , &c. 577 
are a watery fluid. The terminal walls of the sieve tubes are generally 
very much inclined (PL LI V, Fig. 26), so much so that in longitudinal section 
they are often indistinguishable from the lateral walls. Both in the terminal 
(PI. LIV, Fig. 13) and in the lateral (PI. LIV, Fig. 14) walls large thin areas 
occur ; these are the sieve plates. They are generally more numerous on the 
terminal walls, but no definite rule can be laid down, and it is almost im- 
possible to distinguish with absolute certainty between a lateral wall with 
numerous large plates and a terminal wall with few small ones ; the two 
classes merge into one another, and a separation of the two types is not 
founded on any fundamental difference in their structure. In a transverse 
section terminal plates can often be distinguished by the fact that they may 
run right across the diameter of what would otherwise be a single circular 
sieve tube, dividing it into two elements, each having a half-moon-shaped 
cross section (PI. LIV, Fig. 26). The distinction of sieve areas into sieve fields 
and sieve plates, employed by Hill in describing the Phanerogams, will not 
be of value here, since there is no sharp differentiation of the two types. 
The development of sieve plates on the terminal walls is sometimes very 
extensive indeed, the major part of the wall being thin, and the thickened 
portion consisting only of bars of cellulose running transversely from side to 
side of the wall ; such a wall is very reminiscent of that of a tracheide. 
In Marsitia the general relations do not differ very materially ; the 
stele is a solenostele in which the sieve tubes form two fairly regular con- 
centric rings, or rather two pairs of arcs, for laterally phloem development 
is feeble as in Pteridinm. The sieve tubes have somewhat thicker walls 
than those of Pteridium, but in general appearance they resemble them 
very markedly. The development of sieve plates on the lateral walls of the 
internodes is not so extensive as in Pteridium , and gives more the impression 
of scattered pits. At the nodes, however, the lateral sieve plates seem to 
be developed fully. This is interesting in view of the fact that roots are 
developed at the nodes only in Marsitia quadrifolia , roots, petiole, and 
sporocarp all growing out at the same level round the procumbent rhizome. 
There is, therefore, a special demand for conduction around the stem at the 
nodes, while the internodes are comparatively passive in this respect. 
Russow 1 asserts that terminal walls in the sieve tubes occur, as a rule, only 
at the nodes, the tube running unbroken from node to node, but this state- 
ment has not been found to be correct. Terminal walls are not very 
frequent in the internodes, but they do occur. They are often considerably 
inclined, which may have caused Russow to overlook them ; and they bear 
numerous sieve plates ; in other details the phloem of Marsitia quadrifolia 
does not differ materially from that of Pteridium. 
1 Russow (72), p. 5 . 
