1gtt} CHARLES—ANATOMY OF MARATTIA 93 
origin must have shifted upward and sidewise, so as to come above 
a leaf trace instead of along the margin of the gap left by it, as has 
taken place in Alsophila. The later connections must have made 
the same change in relative position of internal and external 
systems. In view of the wide distribution of medullary strands 
in remote groups of the pteridophytes, however, it is not certain 
that they are homologous even in these closely related families. 
A comparison of the young stem of Marattia with that of Psaro- 
nius shows some points of resemblance more sharply than appear in 
the mature vascular system. The band-shaped bundles, the close 
relation of the medullary system to the leaf traces, and the simple 
leaf traces are alike in the young Marattia stem and in Psaronius. 
The absence of roots on the medullary system of Psaronius may be 
due to the relatively small gaps in the peripheral cylinder. The 
slight relation of the peripheral bundles to the leaf traces may be 
due, as RuDOLPH (21) has suggested, to the ranked or simple spiral 
arrangement of the leaves. It may be that the tendency to free 
origin and ending of medullary strands in Marattia indicates a 
reduction from the high state of development of such strands in 
Psaronius, as such similar free endings have been interpreted in the 
petioles of Ceratozamia (DorETY 8) and Botrychium (CHRYSLER 5). 
PROTOXYLEM 
The variability of protoxylem noted by GWyNNE-VAUGHAN (11) 
and TANSLEY (28) in Filicales generally reaches a high expression 
in Marattia alata. The first xylem elements to develop appear at 
the cotyledonary node. They may be the slender spiral tracheids 
common in ferns, similar small vessels with reticulate thickenings, 
as in Dicksonia apiifolia, or stretched scalariform vessels of the 
same caliber as the other elements of the xylem strand; or there 
may be no distinction between the first vessels and those developing 
later asin Angiopteris (FARMER and Hitt 9) and Danaea (BREBNER 
3). In the last case the change to the primary root is brought 
about by a reduction in the xylem at opposite sides of the 
bundle, development of protoxylem at the ends of the xylem band, 
and disappearance of phloem at the ends and increase at the sides 
of the xylem. In two primary roots there were three poles instead 
