3 1 8 Gwynne- Vaughan. — On Polystely 
growth of the vascular tissues on either side to meet each 
other across it. This type is by far the most prevalent ; the 
other, which probably corresponds to those plants in the 
mature structure of which the perfect stele predominates, was 
present in two cases only out of fourteen examined. 
In the second type of seedling the central cylinder is, as 
before, maintained intact until the third or fourth leaf, but 
the gap formed by the departure of the fourth or fifth trace 
becomes an actual gap or break in the cylinder itself. The 
endodermis, instead of passing straight across, falls in around 
the margins of the gap, and passes right round the periphery 
of the medulla, just within the xylems of the vascular ring 
(Fig. 8); so that when this gap becomes closed up by the 
approximation of its margins, the endodermal layers first 
meet and fuse (Fig. 9), then the pericycles, and finally the 
ring of vascular tissue again becomes continuous. But now 
that portion of the endodermis which has been invaginated 
into the medulla, in consequence of the process just described, 
has become cut off and separated from the external endo- 
dermis, and forms a distinct internal endodermis, which often 
also extends into the internode below the gap at which it 
was actually invaginated (cf. Fig. 10). When the next leaf- 
trace departs, the internal endodermis is once more put into 
continuity with the external endodermis around the corners 
of the gap. Further up the stem they become separate 
again in exactly the same way as before. The seedling has 
therefore attained at this point exactly the same structure 
as that termed (for the sake of convenience) imperfect 
gamostely in the mature plant. 
According to Van Tieghem, almost identical processes are 
gone through in the seedling of Botrychium Lunaria % resulting 
in an exactly similar structure, which, however, he here speaks 
of as ‘ gamodesmy.’ If it be held necessary to retain Van 
Tieghems terminology throughout, I see no alternative to 
calling the imperfect steles ‘ gamodesmic groups,’ both in the 
1 Journ. de Bot, Vol. iv, No. 23, p. 405. 
