24 Plowman. — The Comparative Anatomy and 
in the mature axis of the Dicotyledons. If we bear in mind the fact that in 
the floral axis of the Cyperaceae there are no large and numerous incurrent 
leaf- trace bundles, we should not be surprised to find here retained the 
primitive siphonostelic central cylinder. And, conversely, if we find that 
this primitive arrangement of the stelar elements is disturbed only where 
there are large and numerous incurrent leaf-trace bundles, we are certainly 
justified in the statement previously made regarding the dominance of the 
Monocotyledonous leaf in determining characteristic stelar development. 
The significance of these facts will be considered more at length presently. 
Fossil Cyperaceae. 
According to Seward (49) there were probably no Monocotyledons 
earlier than the Cretaceous period. Owing to the soft structure of primitive 
plants of this group, their preservation is rendered most unlikely. At any 
rate, it has been shown conclusively by the investigations of Scott (48) and 
others that the supposed Monocotyledons of remotest antiquity were in 
reality primitive Gymnosperms. Certainly no Cyperaceae have been 
described as from earlier than the Tertiary period. Heer (24) mentions 
a somewhat doubtful Cyperacites Dallensis from the early part of that epoch. 
By the close of the Tertiary the grasses and sedges were fully established. 
Schimper (44) describes for that period eleven Carices, eight species of 
Cyperus , thirty-three Cyperacites, and a number of more or less doubtful 
Cyperaceous forms. Hartz (23) has recently described typical spikes 
of Dulichium as occurring in the interglacial beds of Southern Jutland, 
Denmark. 
Phylogenetic Considerations. 
The primitive ancestral Cyperaceous type is necessarily more or less a 
matter of conjecture. Among the earliest known fossils of this Order we 
find the two widely divergent genera Carex and Cyperus, occurring in about 
the middle of the Tertiary epoch. Apparently we must look to a con- 
siderably more remote period for the common ancestor of these forms. 
That is to say, the primitive Cyperaceous stock probably extended back 
well into the beginnings of the Monocotyledons as a whole, and conse- 
quently we should expect to find in this Order some indication of the 
original characters of the division. Miss Sargant (43) has expressed a 
tentative view that the ancestral Monocotyledons were geophilous in habit. 
However, this habit usually implies a more or less xerophytic type of 
structure ; certainly, at least, a well-developed mechanical system in some 
part of the plant. But we have seen that the geological record of the 
Cyperaceae is blank prior to the appearance of species of Carex and 
Cyperus, which would indicate that their ancestors must have been of a less 
enduring type of structure. 
In view of these facts, we are inclined to adopt the hypothesis sug- 
