


TOWER, DARWINISM. 435 
to germinate in the summer of 1907, and in the first plot the 
census made in August showed 6,009 plants, and in the second 
plot, 49. At Llyano Grande the seeds germinated in 1906 because 
this was the more favorable location, with 8,719 in the favorable 
plot and 832 in the normal plot. Similar tests were made in desert 
areas on the central plateau of Mexico near San Luis Potosi, 
Aguascalientes and Torreon and in semi-desert areas near Cuernavaca 
in the Balsas valley during the years 1906—1909. In all, eight 
pairs of plots were planted and the combined result showed 67,411 
germinating in the favorable plot and 7,856 in the normal plot. 
These experiments, involving a large number of seeds of different 
plants show the uniform result that survival, in so far as it depends 
upon germination and the beginning of growth was not due to any 
adaptive capacities of particular variants, but was dependent upon 
the chance position that the seed occupied for germination. About 
a dozen other plants were tested between 1904 and 1910 in the 
effort. to find a plant in which elimination was on the basis of 
adaptive constitution rather than chance position. I did not find 
One and have not yet succeeded in finding one that gives this 
result, much as [ wish it for experimental purposes. 
It may be objected that these tests were made with plants of 
relatively low variability in their natural habitats but this was defi. 
nitely a part of my plan. The plan of experiment used eliminated 
any possible effects that might come from changed environmental 
conditions, which, hypothetically, at least, might introduce new and 
not easily calibrated factors into the experiment. It will be evident, 
I think, that this is the logical procedure and that it would serve 
to detect in a presumably uniform and stable population, the existence 
of those minute and obscure differences in adaptive capacity upon 
which changes that result in the production of new species by 
means of Natural Selection are supposed to be based The whole 
series of tests was, therefore, carefully planned to eliminate the 
uncertainties and error that might come by using new environmental 
complexes, and, moreover, changed environmental conditions are 
another problem. With plants, all of the tests made give the uniform 
“ result that under native conditions, the elimination incident to 
germination and the beginning of growth is uniformly a function 
