OF THE FOSSIL PLANTS OF THE COAL-MEASURES. 
499 
the smallest Sigillarian fragment could be discovered in it, but in vain. I found Lepi- 
dodendra in abundance of various sizes, and Stigmarice in sufficient numbers; but no 
one specimen existed indicating the possibility that these trees might have been 
Sigillarice. 
ZJlodendron. 
In my second memoir (Phil. Trans., 1872) I described the only specimen of Uloden- 
dron hitherto met with in which the organization was preserved. I have received 
a second specimen from Mr. D’Arcy W. Thompson, of Edinburgh, which, though less 
perfect, is confirmatory of the truly Lepidodendroid character of these stems, and their 
conformity to the type of Lepidodendron Harcourtii. In that memoir, in speaking of 
the bilateral series of scars which characterise the genus, I observed : “It seems 
probable that these scars sustained objects which were chiefly developed from the 
epidermal layer, and whose bases rested upon the outer bark ; they certainly were not 
roots or branches, and I incline to the belief that they were organs of fructification” 
( loc. cit., p. 210). My young friend Mr. Thompson has recently collected a fine series 
of these Ulodendra from the Edinburgh coalfield, some of which conclusively settle 
this question. I am indebted to him for a fine specimen, in which the branch bifurcates 
in the usual Lepidodendroid manner. The large characteristic scars are seen, as usual, 
on both the inner and outer surfaces of these branches. The inner series is curious, 
since one of them is located at the exact point of bifurcation of the branch—a position 
from which no serial root could possibly be given off. The scars average about 0‘9 in 
diameter. Mr. Thompson has also obtained two branches, each of which exhibits the 
usual rows of scars, but in each specimen one scar supports an actual cone in situ. The 
actual area of attachment of each cone to its branch is a very circumscribed one, being 
wholly different from that of a secondary to a primary branch. On the other hand, 
the diameter of the entire base of the cone is very considerable—corresponding, in fact, 
with that of the peripheral margin of each characteristic Ulodendroid scar. It is now 
easy to understand these scars. We constantly find the greater part of their surfaces 
covered with the modified bases of leaves. The cone obviously originated as a small 
lateral bud, but its development into a branch was arrested, leading to the formation 
of a strobilus. As this expanded, though the actual attachment of its central axis to 
its parent branch was a very circumscribed one, the entire base of the cone became 
broad, and as a result of its active growth it pressed down the leaves of the branch 
over an area corresponding to the diameter of that base. The very large scars seen on 
many specimens bear no relation to the size of the cone, but have obviously increased 
in size through the continued growth of the stem after the deciduous strobili had 
fallen. 
Spores. 
In my last memoir (Phil. Trans., Part I., 1878, Plate 23, figs. 59-64) I figured a 
series of remarkable macrospores, which occur abundantly in some portions of the 
