170 BOTANICAL GAZETTE [SEPTEMBER 
agreement in reproductive structures and processes easily out- 
weighed in importance the most conspicuous differences in general 
morphology. In 1794 MoENCHHAUSEN set up for the box elder 
the new genus Negundo, at the same time recognizing the close 
relationship to Acer in the name Negundo aceroides. Extreme 
opposition to the Linnaean rating of the box elder was developed 
by KarsTENn, who in 1880 gave the name Negundo Negundo. Each 
of these three names is held in favor by a considerable group of 
botanists at the present time, though KarsTEN has a much smaller 
following in this matter than either LINNAEUS or MOENCHHAUSEN. 
To all three of these men, as well as to most of their followers, 
the minute anatomy of plants was of course a sealed book, whose 
importance was wholly ignored because unknown and unsuspected. 
It remained for DE Bary, VAN TIEGHEM, and a host of more 
recent investigators, to discover and demonstrate the fundamental 
significance and value of anatomical features in the study of 
phyletic relationships. 
The following study of the comparative anatomy of the box elder 
was undertaken a few years ago, in the hope that some additional 
light might be thrown upon the problem involved in the title of 
this paper. 
General morphology 
As is well known, the box elder is usually a rather low-growing, 
irregularly and diffusely branching tree, hardly more than a shrub 
in many parts of its range, attaining to its maximum size of 50-70 
feet in height, with a trunk diameter of 2—4 feet, only in the lower 
Ohio Valley. In a number of characters the box elders show a 
remarkable range of variety. 
Referring to fig. 1, it appears that the compound leaves are 
3-0-foliolate, with the leaflets variously toothed, lobed, and 
divided. The range in leaf types is apparently not so great in the 
eastern part of North America as it is in the central part. Most of 
the descriptive works published in America follow the early writers 
in stating that the leaves are 3—5-foliolate. As the result of a careful 
statistical study of 1250 box elder trees in southeastern Wisconsin, 
it was found that the leaves were predominantly of the 3-foliolate 
type on a little more than 5 per cent of the trees; 5-foliolate leaves 
