ROCKY MOUNTAIN HERBARIUM STUDIES. III. THE ELAEAGNACEAE—A MONO‐GENERIC FAMILY

Abstract

In studying some of my collections from the Southwest, I came upon some sheets of the well-known but comparatively rare plant, hitherto known as Shepherdia rotundifolia Parry, Am. Nat. 9: 350. 1875. My first impulse was to place it in the genus Elaeagnus, but to my surprise I found it in She pherdia. This led to critical examination of its characters. Its aspect and its fruit are typically Elaeagnus. Its floral structure is neither Shepherdia nor Elaeagnus, for while its stamens are normally 8, occasionally 2 or 4 have aborted. As in Shepherdia it may be said to be dioecious, but as in Elaeagnus the pistillate flowers approach the monoclinous type i.e., there are rudimentary stamens (staminodia). In endeavoring to compare the three genera into which the family has been divided-Elaeagnus (the type genus), Ilippophae, and Shepherdia-the characters relied upon to separate them break down at one or more points. These characters pertain to length of calyx tube, number of sepals, number of stamens, perfect or imperfect flowers. Elaeagnus, for example, is said to be perfect or polygamous. Four is the prevailing number in the family, and it is not difficult to see that in the normal variation, in a given genus, this number might be either doubled or halved. There is one character, however, that has not been stressed-viz., the fruit. This, it seems to me, is more fundamentally significant than the characters referred to above, and a study of the fruit characters leads to the conclusion that the family may be best understood if considered as mono-generic. There are only four indigenous American species, and these divide into two fairly distinct sections with characters as shown in the following key. Were the material at hand, it seems certain that the thirty or forty species found on the other continents could be assigned to these two sections as readily as they are now placed in the two genera, Elaeagnus and Ilippophae.