§ 26. A theory of general application has been put forward and supported with great ability by Oscar Loew[37] which explains the action of poisons by presuming that living has a different composition to dead albumin; the albumin of the chemist is a dead body of a definite composition and has a stable character; living albumin, such as circulates in the blood or forms the protoplasm of the tissues, is not “stable” but “labile”; Loew says:--“If the old idea is accepted that living albumin is chemically the same substance as that which is dead, numerous toxic phenomena are inexplicable. It is impossible, for instance, to explain how it is that diamide N₂H₄ and hydroxylamine NH₂OH are toxic, even with great dilution, on all living animals; whilst neither of those substances have the smallest action on dead plasma or the ordinary dissolved passive albumin, there must therefore be present in the albumin of the living plasma a grouping of atoms in a “_labile_” condition (_Atomgruppirungen labiler Art_) which are capable of entering into reactions; such, according to our present knowledge, can only be the aldehyde and the ketone groups. The first mentioned groups are more labile and react in far greater dilution than the latter groups.” [37] _Ein natürliches System der Gift-Wirkungen_, München, 1893. Loew considers that all substances which enter into combination with aldehyde or ketone groups must be poisonous to life generally. For instance, hydroxylamine, diamide and its derivatives, phenylhydrazine, free ammonia, phenol, prussic acid, hydric sulphide, sulphur dioxide and the acid sulphites all enter into combination with aldehyde. So again the formation of imide groups in the aromatic ring increases any poisonous properties the original substance possesses, because the imide group easily enters into combination with aldehyde; thus piperidine (CH₂)₅NH is more poisonous than pyridine (CH)₅N; coniine NH(CH₂)₄CH-CH₂-CH₂CH₃, is more poisonous than collidine N(CH)₄C-CH-(CH₃)₂; pyrrol (CH)₄NH than pyridine (CH)₅N; and amarin,[38] C₆H₅-CH-NH | \ | CH-C₆H₅, | / C₆H₅-C=N than hydrobenzamide C₆H₅-CH=N \ CH-C₆H₅. / C₆H₅-CH=N [38] Th. Weyl (_Lehrbuch der organischen Chemie_) states (p. 385) that amarin is not poisonous, but Baccheti (_Jahr. d. Chemie_, 1855) has shown that 250 mgrms. of the acetate will kill a dog, 80 mgrms. a guinea-pig; and that it is poisonous to fishes, birds, and frogs: hydrobenzamide in the same doses has no effect. If the theory is true, then substances with “labile” amido groups, on the one hand, must increase in toxic activity if a second amido group is introduced; and, on the other, their toxic qualities must be diminished if the amido group is changed into an imido group by the substitution of an atom of hydrogen for an alkyl.