a case where a child
resembles its grandparent more than it does either of its parents; such
phenomena are due, so to speak, to the rise to the surface of a hidden
stream of germ plasm that had flowed for one or many generations beneath
its accompanying currents. I believe that the law is replacing more and
more the laws of Galton and Pearson, formulated as statistical summaries
of certain phenomena of human inheritance taken _en masse_. According to
Galton's celebrated law of ancestral inheritance, the qualities of any
organism are determined to the extent of a certain fraction by its two
parents taken together as a "mid-parent," that a smaller definite fraction
is contributed by the grandparents taken together as a mid-grandparent,
and so on to earlier generations. But Mendel's Law has far greater
definiteness, it explains more accurately the cases of alternative
inheritance, and it may be shown to hold for blended and mosaic
inheritance as well.

De Vries's new "mutation theory" is clearly not an alternative but a
complementary theory to natural selection, the Weismannian and Mendelian
theories. Like these last, it emphasizes the importance of the congenital
hereditary qualities contained in the germ plasm, though unlike the
Darwinian doctrine it shows that sometimes new forms may arise by sudden
leaps and not necessarily by the slow and gradual accumulation of slight
modifications or fluctuations. The mutants like any other variants must
present themselves before the jury of environmental circumstances, which
passes judgment upon their condition of adaptation, and they, too, must
abide by the verdict that means life or death.

From what has been said of these post-Darwinian discoveries, the
Lamarckian doctrine, which teaches that acquired non-congenital characters
are transmitted, seems to be ruled out. I would not lead you to believe
that the matter is settled. I would say only that the non-transmission of
racial mutilations, negative breeding experiments upon mutilated rats and
mice, the results of further study of supposedly transmitted immunity to
poisons--that all these have led zooelogists to render the verdict of "not
proved." The future may bring to light positive evidence, and cases like
Brown-Sequard's guinea-pigs, and results like those of MacDougal with
plants, and of Tower with beetles, may lead us to alter the opinion
stated. But as it stands now most investigators hold that there are strong
general