738 PROFESSOR W. C. WILLIAMSON AND DR. D. H. SCOTT ON THE 
diameter of from 2 to 24 millims. The largest root has a diameter of 7 millims. 
This possesses a considerable zone of secondary tissue, but even without this the 
diameter would have been about 5 millims. We also find, however, many rootlets 
of quite small dimensions, several examples of which were figured in Memoir XIII. 
It requires a good deal of care, in the case of these very minute specimens, to determine 
with certainty to what plant they belong. We believe, however, that the structure of 
the cortex may be safely relied on. Where the inner cortex consists of lax tissue, with 
very numerous secretory sacs, while, at the same time, the well-marked “ epidermal 
layer” is present, we feel justified in referring the rootlet to Lyginodendron. The 
organs with which confusion is possible are : 
1. The finest rootlets of Calamites, formerly known as Astromyelon ;* and 
2. Those of Rhizonium, a root of unknown affinity. 
From the Calamitean rootlets, those of Lyginodendron van usually be distinguished 
by the absence of the thick-walled superficial layer, which is so characteristic of the 
former. The inner cortex of the rootlets of Lyginodendron has also less definite 
lacunze than that of the rootlets of Calamites, a distinction which, however, becomes 
of doubtful value in the case of the very finest rootlets. 
The rootlets of Rhizonium differ from those of Lyginodendron in the entire structure 
of the cortex, which is formed throughout of uniform closely-fitting cells, with few 
secretory sacs. Mhizontumt is also characterized by its extraordinarily abundant 
branching, much exceeding that even of the roots of Lyginodendron. In some even 
of the small rootlets, the identification is quite satisfactory. Thus triarch rootlets, 
with the characteristic structure of the Lyginodendron root, are found, as to the true 
nature of which there can be no doubt.t 
When Memoir XIII. was published, the fact that these finer branches were 
rootlets was already recognized as probable, though Kaloxylon was at that time 
regarded as an independent plant. The doubt suggested in that memoir as to the 
centripetal development of the xylem disappears as soon as we realize that the 
rootlets of different sizes are not developmental stages of the same organ, but 
represent branches of different orders or simply of different dimensions. 
The typical roots of Lyginodendron, before the commencement of secondary growth, 
often bear a striking resemblance to the smaller roots of recent Marattiaceze. On 
the other hand, they have little in common with those of any Cycadez at present 
investigated. 
* See “ Further Observations,” &c., Part II. The Roots of Calamites. 
+ See Witiiamson, ‘ Organization,” Part XV., Plates 3 and 4, figs. 16-21. 
t As in C.N. 1907, where triarch rootlets of Lyginodendron and of Rhizoniwm lie side by side. The 
former measures about ‘6 millim. in diameter. The still smaller rootlets, some of which are diarch, are 
more doubtful, but some of them, such as those figured in Memoir XITI., most probably belong to our 
plant. 
