8292 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 
or black. On the inside the husk may contain coarse or fine fibers 
or may consist of a reddish edible pulp. The meat may be thick or 
thin, hard or creamy, and rich or poor in oil. One sort has an outer 
layer of the meat red instead of white. Finally, the varieties differ 
in size, number, and color of the leaves, some being light green and 
some dark, or yellowish, grayish, or reddish. Most kinds have the 
pinne separate, but in a few they remain united as in the young 
seedlings. 
No wild palms have been found with any such individual or varietal 
diversification as the coconut varieties; indeed, most of the species 
of palms are less easy to distinguish from their nearest relatives 
than are these varieties. Professor De Vries has described many 
such variations of the evening primrose as new species, on the ground 
that they answer all the requirements of species as treated by sys- 
tematists, but if this be true it would seem to be better to modify 
our taxonomic criteria than to set ourselves the unnecessary task 
of naming as new species all the cultivated varieties of coconuts and 
other plants. 
The objection to the classification of these mutative variations 
as species Is not that the amount of difference may not be as great 
as between some wild species, but that the evolutionary status of 
the mutation is not the same as among the wild species. The nat- 
ural species is a group of freely interbreeding diverse individuals, 
while mutative varieties generally represent the progeny of a single 
individual variation, induced and maintained by inbreeding. 
To give Latin names to fifty or a hundred Malayan mutations of 
the domesticated coco palms might appear to place them in the 
same taxonomic rank as the South American species of Cocos, but 
it would amount, after all, only to an assertion of the identity of 
conditions essentially unlike. 
The Malayan varieties of the coconut are said to propagate true 
from seed with great uniformity, and this fact may have strength- 
ened the idea of diverse ancestry in the wild state, but it is now 
known that the so-called mutations or abrupt variations of inbred 
domesticated plants, such as coffee, are often as constant in their 
characters as the parental type, or even more so. In this respect 
they often differ strikingly from new varieties, originated by hybrid- 
ization, in which it is generally thought necessary to ‘fix the char- 
acters’? by selection. 
The coconut palm appears to be the only cultivated tree whose 
varieties are in the nature of mutations that come true from seed, 
if exception be made of the varieties of mangoes and citrous fruits 
that can be propagated from seeds because of the peculiar habit of 
polyembryony, the formation of false embryos from nucellar buds. 
The mutative varieties of the coffee shrub form a parallel with those 
