‘So far, so good. What Lamarck next asks us to accept, as his “second law,” seems not only to lack the support of experimental proof, but to be inconsistent with what has just preceded it. The new character which is _ex hypothesi_, as was the old character (length, breadth, weight of a part) which it has replaced--a response to environment, a particular moulding or manipulation by incident forces of the potential congenital quality of the race--is, according to Lamarck, all of a sudden raised to extraordinary powers. The new or freshly acquired character is declared by Lamarck and his adherents to be capable of transmission by generation; that is to say, it alters the potential character of the species. It is no longer a merely responsive or reactive character, determined quantitatively by quantitative conditions of the environment, but becomes fixed and incorporated in the potential of the race, so as to persist when other quantitative external conditions are substituted for those which originally determined it. In opposition to Lamarck, one must urge, in the first place, that this thing has never been shown experimentally to occur; and in the second place, that there is no ground for holding its occurrence to be probable, but, on the contrary, strong reason for holding it to be improbable. Since the old character (length, breadth, weight) had not become fixed and congenital after many thousands of successive generations of individuals had developed it in response to environment, but gave place to a new character when new conditions operated on an individual (Lamarck’s first law), why should we suppose that the new character is likely to become fixed after a much shorter time of responsive existence, or to escape the operation of the first law? Clearly there is no reason (so far as Lamarck’s statement goes) for any such supposition, and the two so-called laws of Lamarck are at variance with one another.’ In its most condensed form my argument has been stated thus by Professor Poulton: Lamarck’s ‘first law assumes that a past history of indefinite duration is powerless to create a bias by which the present can be controlled; while the second assumes that the brief history of the present can readily raise a bias to control the future.’[19] An important light is thrown on some facts which seem at first sight to favour the Lamarckian hypothesis by the consideration that, though an ‘acquired’ character is not transmitted to offspring as the consequence of the action of external agencies determining the ‘acquirement,’ yet the tendency to react exhibited by the parent _is_ transmitted, and if the tendency is exceptionally great a false suggestion of a Lamarckian inheritance can readily result. This inheritance of ‘variation in tendencies to react’ has a wide application, and has led me to coin the word ‘educability’ as mentioned in the section of this address on Psychology.