PRESENT POSITION OF PALEOZOIC BOTANY—SCOTT. 399 
Measure age. (See pl. um.) The leaf bases, with typical d/yel- 
oxylon structure, are attached to the stem. The steles are three in 
number, each with a solid axis of primary wood, surrounded by sec- 
ondary wood and phloem. (PI. II.) The leaf-trace bundles, given 
off from the outer surface of the steles, are concentric in the lower 
part of their course, but soon break up into a number of collateral 
strands, with external protoxylem. A large number of these col- 
lateral bundles enter the petioles, which thus have a very Cycadean 
type of structure, chiefly differmg from those of recent Cycads 
in the fact that the wood of the bundles is, as a rule, wholly 
centripetal, while in the living family the foliar bundles are mesarch. 
The triarch adventitious roots, which spring from the stem between 
the leaf bases, also bear a considerable resemblance to those of Cycads. 
The leaf of Medullosa anglica, as shown by the characters of the 
rachis and leaflets in the petrified specimens, was that of an Ale- 
thopteris, probably identical with the species A. lonchitica, which 
is common, in the form of impressions, at similar horizons. There 
is a considerable probability that the 7rigonocarpum originally de- 
scribed by Hooker and Binney in 1854 and referred by Williamson 
to 7. oliveforme (but apparently identical with 7rigonocarpum Park- 
insoni Brongniart ), was the seed of M/edullosa anglica. The petrified 
specimens of the seed are invariably associated with the rachis and 
other organs of the J/edullosa, and there are certain points of struc- 
tural agreement which confirm the probability of the attribution. 
The seed is a large, ovate one, 40-50 mm. long; quite half the 
length, however, is occupied by the micropylar tube, the most re- 
markable feature of the seed. The testa consists of two layers— 
the outer sarcotesta, of delicate, partly lacunar tissue, bounded ex- 
ternally by a sharply differentiated hypoderma and epidermis, and 
the inner, ribbed, sclerotesta, consisting of dense, thick-walled tissue. 
The nucellus appears to have been free, from the chalaza upward, 
and terminates at the apex in a pollen chamber, provided with a 
distinct beak, as in the seed of Cordaites. The vascular system of 
the seed was double, the outer system of bundles traversing the sar- 
cotesta, while the inner formed a complex tracheal network in the 
nucellus. The membrane of the megaspore is evident, but the pro- 
thallus has not yet been found preserved. The structure of the seed 
as a whole presents marked analogies with that of recent Cycadean 
seeds, the differences depending chiefly on the free nucellus, while 
in the Cycads it is adherent to the inner layer of the integument. 
The attribution of the 77igonocarpum just described to Alethopteris 
is further rendered highly probable by Mr. Kidston’s discovery of 
the seed of Neuropteris heterophylla, and I have little doubt that 
it is correct, though the direct proof of actual continuity has not 
