Sa Fae 
ae 
1907] MACDOUGAL—HYBRIDIZATION OF WILD PLANTS 55 
considerations this is not disproved. Synthetic tests, which would con- 
sume a large fractioa of a human life-time, and anatomical examina- 
tions are yet to be resorted to. Meanwhile this oak, which seems to 
be constant in successive generations within the limits of its variability, 
must be regarded as a species in accordance with current taxonomic 
practice, until some positive evidence to the contrary is obtained. 
The principles illustrated by the foregoing facts may be briefly 
summarized as follows: 
It is obvious that the facts of geographical distribution may be 
relied upon to furnish conclisive evidence as to the origin of a species 
or a hereditary quality only under very exceptional conditions, in 
Which other possibilities are excluded, and then only in a circum- 
stantial manner. It is of course a basal and necessary fact that 
species not in contact may not hybridize, but the converse is true 
only when otherwise proven. . 
So far as the plants of suspected hybrid origin from parents sug- 
gested by distributional relations are concerned, the methods of 
investigation available are two, which may separately secure affirma- 
tive evidence of conclusive value, while the third may bring no more 
than confirmations and suggestions. 
Attempts at synthetization, if successful, yield dependable con- 
clusions as to the composition of a hybrid, yet a failure to secure a 
form by synthesis may be due to innate and almost intangible diffi- 
culties in the hybridization of the forms concerned, by the different 
Tesults of reciprocal crossing, and the difference in physiological 
attributes of elementary species included under one name. Further- 
more, the natural form, the ancestry of which is under search, may 
have been a derived hybrid which became fixed in the mth generation 
by a fortuitous combination of dominant characters. To secure a 
Similar result in an experimental test might be beyond the range of 
Probability. 
In an anatomical examination such a combination of dominant 
and recessive characters with fluctuations in meristic features away 
beyond either parent may make the results of but little value until 
confirmed by data derived from other sources, before their full value 
may be known. 
A study of a fixed hybrid by cultural tests of its progeny will reveal 
