the range of variation is increased as a result of food supply and other instances where the variation is apparently greater on poor or sterile soil.
It would seem that any treatment that would increase the range of variation, in plants that are grown for breeding purposes would be valuable, but it still remains to be definitely proved whether such increases in the range of variation are in any marked degree heritable and whether valuable maximum variates can be more frequently produced in this way than would be found in similar groups of plants under ordinary treatment.
It is only very recently that the idea has developed that we can go farther than possibly change the environment. With the publication of MacDougal’s researches in 1906 describing mutations that were apparently caused by injecting the capsules of plants with certain solutions, such as zine sulphate, magnesium chloride and the like, a possible new method of forcing variations was introduced. MacDougal apparently obtained marked variations as a result of his treatment, which were inherited in succeeding generations.
Tower, by subjecting potato beetles during the formation of the germ cells to extremely hot and dry or hot and humid conditions with changes of atmospheric pressure, was able to cause the development of marked changes or mutations which were found to transmit their characters true through several generations and which segregated as unit characters following hybridization. He concludes from his experiments “that heritable variations are produced as the direct response to external stimuli.”
Gager has produced similar changes in plants by subjecting the developing ovaries of plants to the action of radium rays and a number of similar studies by Hertwig and others indicate that radium emanations have a very active effect on both plants and animals.
While the evidence favoring the value of such external stimuli as the above in producing new heritable characters is apparently definite and positive, the extent to which the method can be used in practical breeding has not been determined, and indeed we must await further evidence before we can finally accept the evidence, or the interpretation of the evidence, presented in these very valuable and suggestive researches. Dr. Humbert carried out experiments in the speaker’s laboratory in which the capsules of a pure line of a wild plant Silene noctiflora were injected with the solutions used by Dr. MacDougal, and although the number of plants handled (about 15,000) was apparently as great or greater than was used in MacDougal’s experiments, no mutations were found in the treated plants which were not also found in the untreated or check plants.
Some observations and experiments are recorded in literature which indicate that mutilations or severe injury may induce the development of mutations. Most noteworthy among such observations are those of Blaringham, who by mutilating corn plants in various ways, such as splitting or twisting the stalks, apparently produced variations which bred true without regression and which he described as mutations. My own observations on the great frequency of striking bud variations on recovering trunks of old citrus trees in Florida, following the severe freeze of 1894–5, also furnished evidence in support of this theory.
In general, it is assumed that in hybridization we are dealing merely with characters already present and that new characters which appear are due to the different reactions caused by new associations of unit
