Tort] SITNNOTT—FOLIAR BUNDLE } 261 
of Clepsydropsis; and this in turn assumes a tetrarch condition, the 
striking similarity of which to the bundle of Asterochlaena has been 
noted by Bertranp. It is noteworthy that the typical modern 
fern character of endarchy is present only in the stem. In the 
leaf trace in the cortex, the first formed elements of the wood seem 
to be clearly connected with neither face of the bundle. 
Conditions in Lyginodendron are very similar. The trace at 
its base is monarch, and though true endarchy apparently never 
occurs in this genus, the protoxylem group, which lies near the 
outer edge of the bundle on the abaxial side of an immersed island 
of parenchyma, is always continuous and in seriation with the 
narrow arc of centrifugal wood. This relation of the protoxylem 
to the metaxylem persists throughout the course of the leaf trace 
in the cortex, but is apparently lost in the petiolar bundle, where 
the earliest tracheids show no definite attachment to either the 
inner or the outer wood. 
The single protoxylem group soon divides into two (fig. 6), as 
in Calamopitys, and the trace in a short time also separates into 
two equal parts, which shortly become diarch. These four clusters 
of protoxylem are all near the outside of the metaxylem, but are 
seriated with the centrifugal wood if at all. In the petiole the 
two strands become united by their abaxial ends into a tetrarch 
arch, which becomes gradually wider and more flattened and soon 
contains eight to ten protoxylem groups, all of which are clearly 
mesarch. This wide bundle often becomes divided into several 
distinct strands, which in the upper part of the rachis come together 
into a tetrarch and finally triarch and mesarch bundle, such as is 
figured by Scort in his description (6). It is noteworthy that in 
the petiole and rachis the abaxial elements are always smaller than 
the adaxial ones, a condition which is often found in the leaves of 
living ferns. 
The structure of the trace in Heterangium is very similar to 
that in Lyginodendron except that it often remains undivided. 
The mesarch protoxylem here as well as in the stele is always in 
Seriation with the centrifugal wood. 
It is clear that in these three genera we have to deal with plants 
showing a strikingly close resemblance in the lower and more 
