22 
G wynne- Vaughan . — On Polystely 
took place, it would appear that the epicotyledonary cylinder, 
having lost the faculty of adequate expansion, and being no 
longer able to enclose these leaf-traces within its slender 
circumference, would have to accommodate them in some 
other manner. 
Now, we may reasonably regard the manner in which the 
epicotyledonary stele of the seedling becomes converted into 
the mature structure of the stem as a reminiscence of the 
phylogenetic progress of the race itself, though, of course, with 
the earlier incidents blurred by subsequent adaptation to the 
surroundings. It therefore appears probable that the narrow 
central cylinder, being unable to expand, became broken up, 
in connexion with its constituent leaf-traces, into so many 
vascular arcs, and that these arranged themselves in a ring of 
suitable diameter around the centre of the stem, fusing with 
each other laterally at intervals, and opening out again. In 
fact, a structure was produced similar to that found in some 
specimens of P . japonica and P. obtusifolia at the present 
time. 
Again, it is evident that even such a ring as this might at 
length reach a limit of expansion, relative to the size of the 
stem, while room could still be found for more vascular tissue 
on the inner side of the cauline portions of the leaf-traces, 
thus converting them into steles, and at the same time without 
breaking up the ring. The various gamostelic species may be 
cited as exhibiting such a structure. 
If, however, we suppose that a continual increase in the 
number or size of the leaf-traces entering the stem necessi- 
tated a still further increase in the amount of vascular tissue 
contained in the latter, there appears to have been no alterna- 
tive but to break up the ring and to scatter the steles 
throughout the ground tissue, as is to be observed in many 
dialystelic species. 
However that may be, one result appears clear— that, in 
the Primulas, the gamostelic condition is more primitive and 
nearer normal monostely than is the dialystelic type, and that 
probably a gamodesmic condition preceded either. 
