820 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 
both sides of the Pacific, that the ethnological evidence afone might 
not be sufficient, but the addition of definite botanical data may 
yield conclusive proof. The earlier idea of an Asiatic origin of Amer- 
ican civilizations having been given up, the tendency has been to 
believe that agriculture and other arts of civilization have developed 
quite independently on the two sides of the Pacific. But even if we 
were willing to believe in closely parallel developments in customs 
and arts, this could not explain the prehistoric distribution of the 
same cultivated plants over the Tropics of both hemispheres. 
ORIGIN OF CULTIVATED VARIETIES OF THE COCONUT PALM. 
One of the most convincing arguments for the Malayan origin of 
the coco palm was the existence of many and very diverse varicties 
in the Kast Indies. It was a favorite line of reasoning with De 
Candolle that the native home of a species was the region in which 
it had been longest in domestication. The length of the period of 
domestication was inferred from the number of varieties as a meas- 
ure of the time that selection had been at work. Such calculations 
were applied to the sugar cane, the taro, and numerous other plants. 
In dealing with the coco palm this plan seemed to be particularly 
effective, for nearly all of the recorded varieties are in the East 
Indies. The coco palm in America is not as uniform as commonly 
supposed, though the varietal diversities do not approach those of 
the Malay region. 
Careful consideration of the evolutionary argument will lead, how- 
ever, to a conclusion directly opposite to that reached by De Candolle, 
for the greatest and most definite variations of a cultivated plant are 
much more likely to occur and be preserved outside its natural range, 
where intermixture with the wild type of the species is prevented. 
There are many reasons for believing that the abrupt and striking 
“‘sports”’ that appear among our cultivated plants are not, in reality, 
‘aused by selection, but are induced by new conditions and by the 
state of inbreeding that generally accompanies domestication. 
The normal or wild type of a species is generally prepotent over 
the varieties which have arisen in domestication, so that the ‘‘im- 
proved” breed rapidly ‘‘deteriorates’’ when allowed to become 
crossed with the wild stock. Darwin and many later experimenters 
have proved, also, that when diverse breeds are crossed the offspring 
are very often not intermediate between the breeds, but tend to 
revert to the ancestral form. The breeders of high-grade varieties 
look upon such mongrels as degenerate, but from the standpoint of the 
evolutionist they may be said to be recovering from the injurious 
results of inbreeding. It was noticed, for example, that in parts of 
Guatemala, where the wild tropical papaw (Carica) is common, the 
cultivated trees also have very small fruit. The tendency to such 
