374 BOTANICAL GAZETTE [NOVEMBER 
not centripetal, as one would interpret from the treatise cited. It has 
been shown that the transition from endarch to exarch is carried 
through in the metaxylem, both centripetal and centrifugal wood 
occurring long before secondary wood is developed. It is very hard 
to tell where protoxylem ends and metaxylem begins, and where 
metaxylem ends and secondary wood begins. The separation of 
metaxylem from secondary wood by means of the thin-walled par- 
enchymatous cells is not a safe guide in the region of transition, and 
only becomes well marked above this region, where the separation 
of secondary from primary wood is well marked by the thin-walled 
parenchymatous cells, as noted by many authors. The metaxylem 
above this point gets to be relatively very bulky, while the secondary 
wood is represented merely by a few elements. ‘These were the few 
pitted cells which presented to Von Mout (1) a situation without 4 
counterpart, now known to be the herald of the secondary W 
which has gradually crept up into the petiole, a transformation 
begun in its early ancestry, according to Scott, the “new wood” 
driving out “the old,” the former being the only wood present in the 
higher gymnosperms and angiosperms. 
MATTE (6) argues in very much the same way as does METTENIUS; 
an argument which would hold good if the protoxylem and a large 
part of the metaxylem were left out of account. MATTE says that 
the bundles of the cotyledons have centripetal wood throughout, 
centrifugal wood only below the upper region of the petioles, am 
centripetal and centrifugal wood equally well developed at the bases 
of the petioles. In the present investigation it has been pointed out 
that there is no centripetal wood at first, and that it gradually increases; 
while the centrifugal wood diminishes in bulk to the upper extremiti® 
where it is less than the centripetal but does not disappeat entirely. 
Matte further says that what has been said of the cotyledonary 
traces applies equally well to the foliar traces, except that there * ” 
trace of centrifugal xylem in the youngest leaves. It can be show" tia 
aS soon as there are enough xylem elements to show the direction 
development the centrifugal wood is present, but gradually disappea® 
and the centripetal wood increases in the same ratio, until in the upp 
extremities there is only centripetal wood. This also agrees a : 
BERTRAND and RENAULT, except that their statement that centripe — 
