in Myzodendron punc hi latum , Banks et Sol . 1 8 1 
stoma Itself is elevated, the guard-cells are not flush with the 
surface, but are depressed and covered by a very thick cuti- 
cularised layer. A look at Fig. i will show that it is difficult 
to say whether the outer and inner thickening ridges belong 
to the guard-cells or not The latter lie with their long axes 
parallel to that of the stem, and consequently are seen In 
section, only In cross sections of the stem. The formation of 
these swellings on the stem, foliage-leaves being absent, in- 
creases the assimilative surface of the stem— much In the same 
way as the grooved surface of the stem does in Equisetum — 
and Is one of the chief distinguishing features of the sub-genus 
Gymnophyton , as contrasted with Eumyzodendron, 
The internal structure of the stem is to a great degree ex- 
plicable by a knowledge of the habit of the plant. I would 
refer readers to the Flora Antarctica, reminding them that it 
was not until nearly twenty years later than the time of pub- 
lication of that work, that the nature of sieve-tubes was at all 
understood. M. punctulatum stands out horizontally from the 
stem of the beech, resulting In a one-sided development of 
its wood (Fig. 3). The pith and the greater part of the pri- 
mary medullary rays consist of woody prosenchymatous fibres, 
containing starch. There is in the stem of Myzodendron the 
same lignification of the pith as occurs in the pith of the 
adventitious roots of Hedera Helix . Secondary thickening, 
by the activity of a normal cambium ring, the cells of which 
are of the typical form, occurs as in Dicotyledons generally. 
The activity is, however, confined chiefly to the formation of 
secondary xylem, in which secondary medullary rays, very 
small and inconspicuous, occur. Annual rings too are evident, 
though, possibly owing to the absence of foliage-leaves, and of 
much secondary phloem, they are not well marked. In the 
secondary phloem there are no sclerenchyma fibres. These 
form an Incomplete irregular zone, lying half-way between the 
vascular tissue and the epidermis (Fig. 3 s). The sieve-tubes 
are very narrow, have simple oblique terminal sieve-plates, 
facing the radial plane, and are peculiar in being nucleated. 
Treatment with Feh ling’s solution shows that the sieve-elements 
