a 
a 
ci 
ie 
1912] HOLDEN—SAPINDALES 53 
different. These are a few of the uniseriate variety, but the major- 
ity are multiseriate. 
Figs. 4 and 5 represent transverse and tangential sections of the 
wood of Sapindus sp. Like the two genera described, the wood 
is ‘‘diffuse porous’; the parenchyma is vasicentric and abundant. 
The fibers are characterized by delicate cross partitions of cellulose, 
constituting the so-called ‘‘septate fibers.”” Though some of the 
vessels are small, the majority are large, serving at once to separate 
Sapindus from the other members of the Sapindales. They have 
porous perforations and spiral thickenings; the side walls in some 
places have closely crowded pits, but in other places there is a 
decided tendency toward fusing into rows of slitlike bordered pits. 
The rays of Sapindus are multiseriate, much like those of Acer. 
Fig. 6 represents another member of the Sapindales, Staphylea 
trifolia L. Here the wood parenchyma is vasicentric, and usually 
on the radial side of the vessels. The wood elements are not as 
specialized as in the other genera; instead of being libriform or 
septate fibers, they are fiber tracheids, with thinner walls and con- 
spicuously bordered pits. The vessels have both porous and 
scalariform perforations; the pits on the side walls are sometimes 
unfused like Acer and Aesculus, but are more often united to form 
large slit like openings. Staphylea is the only one of the four 
genera examined in which there are no spiral markings on the inner 
walls of the vessels. The rays range from 1 to 1o-cells wide. — 
These broad rays cause a local “dipping in” of the annual ring, like 
that in Quercus. 
Having considered the general characteristics of these four 
members of the Sapindales, the question arises as to which is the 
most primitive and which the most advanced. Disregarding the 
evidence furnished by the ray, Staphylea, with its scalariform: 
perforations and fiber tracheids, seem to be the lowest; but taking 
the ray structure into consideration, Aesculus seems to be the lowest. 
Thus the question narrows down to which is the more primitive 
for the Sapindales, uniseriate or multiseriate rays. 
In this connection it is interesting to note the work of EAMEs 
of this laboratory on the genus Quercus.’ He found the rays of 
5 Eames, A. J., On the origin of the broad ray in Quercus. Bot. Gaz. 49: 161-167. 
pls. 8, 9. 1910. 
