AY ye ae OR Se ere Pe 
A RY ee) 
1909} SINNOTT—MESARCH STRUCTURE IN LYCOPODIUM 143 
etal primary wood. In the higher plants, the centrifugal xylem pre- 
ponderates, and the protoxylem becomes continuous with it,even in the 
cases where centripetal elements are present. The only exceptions to 
this rule occur in foliar bundles. The leaves of living cycads and of the 
Cordiates show true cryptogamic wood in their vascular strands, with 
the protoxylem closely associated with it, though in all cases centrif- 
ugal secondary wood is present. This persistence of a cryptogamic 
structure in the leaf, among the higher plants, furnishes good evidence 
of the conservatism of the foliar bundle. Scattered centripetal 
elements appear in the peduncle of certain cycads and in the coty- 
ledons of Ginkgo, also seats of vestigial characters, and centripetal 
wood is fairly well developed in the leaf of Prepinus (JEFFREY 9) 
and in the stems of many of the Cycadofilices; but in all these cases, 
the bundle is of the higher type, the protoxylem being continuous with 
the centrifugal wood. 
The structure of the vegetative stem, of the floral axis, and of the 
leaf, in certain living Cycadaceae, shows clearly the transition from 
the lower type of wood to the higher. In the stem, the protoxylem joins 
the centrifugal wood and there are no centripetal tracheids. In the 
stalk of the male cone of Stangeria paradoxa, and to a less extent in 
the male and female peduncles of three other species of cycads, as 
shown by Scorr (8), scattering centripetal xylem elements are present. 
They are not connected with the protoxylem, however, but are sepa- 
rated by parenchyma from the rest of the bundle, which otherwise is 
exactly similar to those found in the stem. In the petiole and blade 
of the cycad leaf, however, there is more than this faint suggestion 
of ancestral structures; for here, as above described, true cryptogamic 
Wood occurs, consisting of well-developed centripetal elements 
directly continuous with the protoxylem. In the petiole, and some- 
times in the blade, centrifugal wood is present as well. This is sepa- 
rated by parenchyma from the protoxylem, and appears to be largely, 
if not entirely, laid down by cambial activity. Its resemblance to a 
Sigillarian bundle, as noted by RENAULT, is thus pretty exact. Scotr, 
however, believes that some, at least, of the centrifugal wood 1s 
primary, and consequently that true mesarch development takes 
place, a condition which he compares to that found in the stem bundles 
of Lyginodendron Oldhamium. He presents the mesarch structure 
