1908] THIESSEN—DIOON EDULE 375 
tal wood is primary and centrifugal wood secondary does not hold 
true, a mistake apparently shared by Marre. It has been shown 
that protoxylem and metaxylem may have both centripetal and cen- 
trifugal elements. This is very well seen in cross-sections at a low 
level of a young strand, where only a few xylem elements are devel- 
oped. In such a section the protoxylem lies against the inner edge 
of the bundle, and the procambium can be seen to occupy considerably 
more than half of the bundle; the boundary of the procambium and 
the protophloem is distinctly recognizable, the protophloem forming 
only about one-third of the bundle. In the upper extremities of the 
petiole, also, there may be seen at an early stage a considerable amount 
of procambium outside the protoxylem; as these usually fail to de- 
velop xylem, thin-walled cells occur between the primary and second- 
ary wood in the upper extremities (fig. 21). 
The girdling habit was first noticed and described by KARsTEN 
(2) in Zamia muricata,®without, however, giving the definite number 
of traces, LEstTrBouDors (3) adds nothing new except that the 
traces branch and anastomose. METTENIUS (4) next misinter- 
preted the situation, as described in the historical introduction 
(P. 357). 
In the embryo and young seedling,”at least, the leaf traces pursue 
definite and well-defined courses and constitute a definite system. 
Four traces are invariably found to leave the vascular cylinder for 
fach leaf. A few anastomoses occur here and there, but these are 
always reducible to four strands. On two occasions only five strands 
"ere found in one of the cotyledons. In the older seedlings anas- 
tomoses are more abundant, but so far as observed these can be 
Teduced to the system found in the young seedling. 
When Karsten (2) described the girdling habit, he suggested a 
Cause for it in saying that the bundles are formed very early in the 
Young leaf, and that the originally narrow curves are later crowded 
far out by subsequent growth and the appearance of new organs. 
"TTENIUS (4) also gives a reason for girdling as follows: “In the 
developmental stage the traces of the youngest leaves lie in the region 
of the vegetative point, and at first ascend in an almost perpendicular 
ction, but during the further growth assume gradually an almost 
Zontal position, and with subsequent growth are lengthened and 
