Osmundacece. 
257 
regarded from without) forms the right-hand portion of the eighth leaf- 
trace above, while the right-hand strand forms the left-hand portion of 
the thirteenth leaf-trace above. When one of these oval strands 
joins a similar strand from another node it closes a leaf-gap, and 
these strands therefore correspond with the “compensation strands” 
of dicotyledons. Phylogenetically, of course, they are the caufine 
parts of the original cylinder. 
In some Osmundacece (e.g., O. regalis ) the leaf-gaps (which 
only affect the xylem of the cylinder) are extremely long, being 
continued upwards for eighteen internodes above the exit of the 
trace 1 (Fig. 83). It is this great length',and the consequent overlapping 
of many gaps, which gives rise to the long, independent course of the 
“compensation-strands.” In the species with shorter gaps the 
strands fuse with their neighbours so much the sooner. 
Fig. 84. Todea barbara. Diagram of course of xylem strands in the stele. 
Leaf-gaps black, protoxylems dotted. From Seward and Ford. 
The single protoxylem in the concavity of each leaf-trace is 
continued downwards into the stem as the protoxylem already 
mentioned in the concavity of each U-shaped strand of the stem- 
cylinder. Traced downwards the two arms of the U approach and 
fuse, so as to enclose an island of parenchyma containing a proto¬ 
xylem group, enclosed in a solid mass of xylem. The xylem-strand 
thus becomes mesarch and circular or oval in section. At a lower 
1 According to Lachmann (’89) Plate V., Fig. 1), the phyllotaxy is 
sometimes pj, and in such a case of course the leaf-gap 
cannot exteniffor more than thirteen internodes. 
