56 BOTANICAL GAZETTE [JANUARY 
not only in the leaf trace, but also in the wood of the axis imme- 
diately around the leaf trace. Fig. 10 represents an outgoing 
foliar bundle; subtending it, there is a mass of ray parenchyma 
forming a true multiseriate ray. Often in the case of numerous 
small bundles going into the petiole, this mass of tissue extends 
all the way from one bundle to the next. 
These photographs are all of Aesculus Hippocastanum, but the 
conditions of Aesculus glabra are essentially the same, except that 
under the leaf trace there is seldom as much parenchyma as here 
shown. 
Fig. 11 is a tangential view of the wood of a root, showing an 
outgoing rootlet and the tissue immediately under it. The condi- 
tions are very much as in the branch, each rootlet trace being sub- 
tended by numerous multiseriate rays. Usually, however, this 
condition is not so pronounced in the root as in the branch. 
Fig. 12 is a tangential section of a floral axis, showing the con- 
ditions immediately below an outgoing flower stalk or peduncle. 
The rays are characteristically biseriate, and extend sometimes a 
long distance below the trace. Often above the trace they are 
broader than below, but they never extend as far. The woody 
tissue is as a whole poorly developed, except at the end of each 
flower stalk, where it becomes much thicker. Just below the end, 
there are 5 or 6 traces going out simultaneously, each of which has a 
small number of multiseriate rays below it. 
Three of the recognized primitive localities have thus been shown 
to have well marked multiseriate rays, and wounded wood was 
examined for the same structures. None were found, either because 
the injury was not sufficiently severe, or because the degeneracy of 
Aesculus has gone too far to be recalled traumatically. 
Aesculus then presents a condition just the reverse of that found 
in Quercus. The former has uniseriate rays normally, with multi- 
seriate rays persisting in primitive localities; the latter has com- 
pound rays normally, with uniseriate rays in primitive localities. 
Accordingly, multiseriate rays are primitive for the Sapindales, 
and Aesculus, instead of being the most primitive of the Sapindales, 
on the basis of ray structure, is really advanced, its simplicity being 
due to degeneracy. 
