68 Gregg. — Anomalous thickening in 
bium makes its appearance as above described, but it never 
becomes continuous with the normal cambium. The orienta- 
tion of its products is here also reversed as compared with the 
typical arrangement. The activity of this cambial layer 
results in the first instance in the production of two isolated 
bands of secondary tissue lying on either side of the 
normal vascular mass ; subsequently a new series of divi- 
sions begins at the outer limit of the pericambium cor- 
responding to those producing the outermost secondary 
ring in the first type ; the products of these divisions re- 
semble those in type i ; the outer cambial layer, however, 
never forms a complete ring round the whole vascular cylinder, 
but on the other hand becomes continuous on either side with 
the first anomalous cambial layer (see Fig. 4). If we take only 
diarch roots into consideration, the differences between the 
two types may be summed up as follows : — In Type 1 we get 
two inner rings of secondary tissue surrounded collectively by 
a single outer ring; in Type 2 we get a single internal ring 
flanked on two sides by two anomalous outer rings. As 
follows from the mode of development described, the outer 
rings in Type 2 have the opposite arrangement of the xylem 
and phloem to that in the inner rings of Type 1 1 . It is a 
point of some interest that all the cambial layers observed 
appear to continue their activity simultaneously. Judging 
from the relative thickness of the cell-walls, there is no reason 
to suppose that the inner cambial layers pass over into per- 
manent tissue when the outer ones begin their activity ; how 
long this simultaneous growth of all the layers may continue 
could not of course be determined in the material available. 
As regards the nature of the xylem and phloem-elements 
derived from the inner and outer abnormal cambium, no 
important differences were found as compared with the 
normal secondary tissue. 
It is important to mention that the peculiarities of structure 
described in these roots have no relation to any abnormality 
1 The structure of the rings in question may be compared to that of concentric 
bundles. 
