48 
Scott . — On a Palaeozoic Fern , //£<? 
established between the protoxylem of the stem contained in the internal 
rays and that of the leaf-trace enclosed in the peripheral loops. It must, 
however, be remembered that the approach of the protoxylem-rays to the 
surface of the wood is not only connected with the supply of the leaf-trace ; 
continuity with adventitious roots has also to be established. Root-steles, 
for example, are given off from arm i at a level between that of Phot. 4 and 
Phot. 5. 
The condition at a somewhat higher level, in a case not complicated 
by root-insertions, is shown in detail in Fig. 4, PI. III. 1 Here the arm 
already has a squarer end than lower down. The internal ray (which 
is still continuous) is expanded, at the point where it forks, into a little 
triangular body, the first indication of the central strand destined for 
the axillary stele. Beyond this the two branches of the ray run out nearly 
to the surface. 
Still higher up, the arm becomes cornute ; the peripheral loops are 
lengthened and are here more evident as loops ; they now become isolated, 
so far as the transverse section is concerned, by the closing up of the branch 
rays which connected them with the triangular body. This also becomes 
isolated by the closing up of the internal ray behind it, a change which may 
take place at a lower level. (See Phots. 3-5, arm 3 ; for the shutting off of 
the main ray at a lower level see Phot. 6, arm 1.) 
The rest of the process may be followed rapidly in arm 5. At the 
bottom of the series (Phot. 1) this arm has long lobes, each of which is 
obliquely truncated and tipped by a peripheral loop ; the loops are quite 
cut off from the median triangular strand, but the latter is still in con- 
nexion with the internal ray. Above this (Phot. 2 ) the ray is. beginning 
to die out behind the strand, which is lengthening tangentially. At the 
next higher level shown (Phot. 3) the ray has completely closed up for 
a long distance and the median strand appears as an island, surrounded 
by the large-celled wood. The arm is here distinctly constricted about 
midway between the island and the end of the internal ray. 
Here the trace separates from the wood, as seen in Phot. 4 and Phot. 8. 
The outward passage of the trace and the slight changes which it under- 
goes within the limits of the series require no further description. 
From the facts observed it is clear that the peripheral loops are 
constituted by the two branches of the internal ray passing out nearly but 
not quite to the surface of the wood, the parts which enclose their termi- 
nation becoming small-celled. The internal rays contain the protoxylem, 
which thus comes to lie inside the peripheral loops. The point of bifurca- 
tion of the internal ray supplies the median strand for the axillary stele. 
All these parts become isolated as we follow them further up. In other 
words, the ‘ undivided leaf-trace’ receives three branches from the internal 
1 Arm 2 , from the section next above that shown in Phot. 6. 
