62 
Hoar.—A Comparison of the Stem 
As to the constancy of the scalariform perforations of the vessels 
in the Cornaceae, I have already noted that simple perforations do occur in 
a few cases mixed with the scalariform type. In the Araliaceae the vessels 
have nearly always simple end perforations, though some species also show 
a few bars. The angle of the perforation is usually quite oblique to the 
lateral walls, and its circumference is usually elliptical. The Umbelliferae 
have simple perforations at the ends of the vessels, and these are often round 
and more at right angles to the lateral walls than in the case of the 
Araliaceae. 
It has already been pointed out that in the sieve-plates of the sieve- 
tubes, the higher Angiosperms tend to lose their lateral plates, and the end 
plates come to take up a position more at right angles to the lateral 
walls ( 10 ). This seems to hold true for the perforations of the end walls of 
the vessels of the higher Angiosperms. I can make no definite statement 
here without a broader study of the subject, but merely offer this as 
a suggestion. 
Conclusions. 
1. That throughout the Cornaceae the parenchyma, where it occurs, is 
scattered throughout the wholfe annual ring (diffuse), while throughout the 
Araliaceae and Umbelliferae it is grouped about the vessels (vasicentric). 
2. That the vessels of the Cornaceae show in every species examined, 
at least in part, scalariform perforations, while all species of the Araliaceae 
and the Umbelliferae show in part the simple pored condition. Also, that 
the simple pores of the Araliaceae are more elliptical and more oblique than 
in the case of the Umbelliferae. 
3. That the general anatomical features of the Nyssoideae and Davidi- 
oideae do not seem to warrant their being separated from the Cornaceae 
and their being placed with the Myrtiflorae. 
4. That the presence of secretory canals in Mastixia is not necessarily 
of importance in determining the relationship of the genus. 
5. Finally that, using the anatomy as a criterion, the Cornaceae 
should not be placed in the same cohort with the Araliaceae and with the 
Umbelliferae. 
The writer wishes, in closing, to express his gratitude to Professor 
Fernald, to the Director of the Harvard Botanical Garden, and to the 
Director of the Arnold Arboretum of Harvard University for specimens. 
Many of the species studied were secured by Drs. Eames and Sinnott 
in New Zealand by means of a gift of Mr. J. S. Ames. This investigation 
has been carried on in the Phanerogamic Laboratories of Harvard Univer¬ 
sity under the direction of Professor Jeffrey, and to him I am greatly 
indebted for advice and for the photomicrographs accompanying this 
article. 
