242 
C. C. FORSAITH 
American species of Nothofagus, and are shown in figure 11, a trans- 
verse section. Wood parenchyma occurs only in a terminal position, 
and is frequently difTerentiated from the libriform fibers by the 
presence of crystals of calcium oxalate. A tangential section through 
the end of the annual ring, seen in figure 12, shows very clearly paren- 
chyma cells, libriform fibers, and rays. An examination of the 
parenchyma shows, beside the simple pitting, the dilTerence between 
ordinary parenchyma and crystal cells; the latter have been derived 
from the former by further subdivision. 
Since the anatomical characters of Elaeocarpus correspond so 
closely to those found in Aristotelia, especially in respect to the dis- 
tribution of wood parenchyma, it is evident that these features are 
the result of similar conditions of reduction. 
Terminal parenchyma in the Angiosperms is apparently the result 
of reduction from ancestral forms in which either vasicentric or diffuse 
parenchyma was present. This has already been shown to be the case 
in the Salicales by Miss Holden, where those representatives which 
normally present only terminal parenchyma retain the vasicentric 
condition in traumatic and primitive regions, thus illustrating the 
principles of retention and reversion, respectively. Conditions very 
similar to those are to be seen in the Fagaceae, in which group, paren- 
chyma occurs normally as cells scattered throughout the wood. In- 
vestigation of the aberrant South American representative, Nothofagus, 
shows only terminal parenchyma in the stem. This genus thus shows 
similarity, mutatis mutandis, to the Salicales investigated by Miss 
Holden (3). 
From a consideration of the evolutionary tendencies in the two 
above mentioned families, it is evident that those which possess ter- 
minal parenchyma have suffered a reduction from more primitive 
ancestors which showed, in the one case, vasicentric, and, in the other, 
diffuse parenchyma, and consequently represent more advanced 
species. 
Applying the above mentioned principles of Elaeocarpus and 
Aristotelia, two species which show terminal parenchyma, it is evident 
that they have been derived by reduction from more primitive an- 
cestors; and, since the two examples cited show a reversion in recessive 
regions to the type of parenchyma distribution normally present in the 
less advanced members of the order, Elaeocarpus and Aristotelia may 
be considered as having suffered a reduction from ancestors, which 
