Affinities of Fossil Plants from the Palaeozoic Rocks. 251 



Medullosa might well be concisely described as a polystelic Heter- 

 angium. 



The course of the leaf-trace bundles was followed very completely in 

 consecutive series of transverse, and in longitudinal, sections. The 

 leaf-traces leave the steles precisely in the same manner as in Heter- 

 angium. On becoming free the trace is a large concentric bundle, 

 surrounded by its own zone of secondary wood and bast. As it 

 passes obliquely upwards through the cortex, the trace loses its 

 secondary tissues, and undergoes repeated division into a number of 

 smaller bundles, each of which has collateral structure. These collateral 

 strands have in all respects the same arrangement of their elements as 

 the well known bundles of Myeloxylon. 



The base of the leaf received a large number of bundles, consisting 

 of the ultimate branches derived from the subdivision of several of the 

 original leaf-traces. This distribution of the bundles is peculiar and 

 unlike that in any known plants of Cycadean affinities. 



In a few cases accessory vascular strands, of concentric structure, 

 recalling- the cortical bundles of a Cycas, were found to the outside of 

 the normal stelar system. 



The stem formed a well marked zone of internal periderm. In one 

 specimen the whole of the outer cortex, with the leaf-bases, had been 

 exfoliated, so that in this case the periderm formed the external surface. 



The leaf -bases and petioles present in all respects, as regards hypo- 

 derma, vascular bundles, and gum-canals, the characters of the Myeloxy- 

 lon Landriotii of Renault, which was evidently not a species, but a type 

 of leaf-stalk common to various Medulloseae. The petioles branched 

 repeatedly, the finest ramifications of the rachis having a diameter of 

 about 1 mm. only, but retaining in essentials the " Myeloxylon " struc- 

 ture. The leaf was thus a highly compound one; the structure of 

 the leaflets associated with the rachis agrees well with that of the 

 Alethopteris leaflets, figured by M. Renault. 



The roots, never previously observed in any species of Medullosa, 

 were of triarch structure, with abundant formation of secondary wood 

 and bast, and an early development of internal periderm, by which the 

 primary cortex was thrown oft'. Developmental stages show that the 

 periderm originated in the pericycle. The roots, which branched 

 freely, were borne on the stem in vertical series, between the bases of 

 the leaves. They were attached to pedicels, through which the 

 vascular tissues of the roots were continuous with those of the stem. 

 The author is indebted to Mr. J. Butterworth and Mr. G. Wild, for 

 specimens which have thrown important light on the connection 

 between root and stem. 



The full paper concludes with a short historical re'sumt*, and a dis- 

 cussion of affinities. 



Medullosa anglica, in the structure of its stem, shows unmistakable 



VOL. LXIV. IT 



