590 ON INCREASE IN COMPLEXITY [pt. iii 



animal or by one of its segments are not measured by the time 

 required to dissolve the animal or one of its segments in a solution 

 of potassium cyanide." Nevertheless, the susceptibility method does 

 show up the existence of gradients of something. The very numerous 

 studies of Child; Hyman; and Bellamy on the susceptibiHty of dif- 

 ferent regions of embryos at different stages are quite sufficient to 

 prove that. They will therefore be called in the remainder of this 

 discussion axial or physiological gradients, and their nature will be 

 left undefined. An example not tied up with any theory of the nature 

 of the gradients is the staining gradient of the chick embryo found 

 by McArthur to hold for many acidic and basic vital dyes. It must 

 be admitted that throughout Child's treatment of the subject he 

 confuses growth-rate with organisation- or differentiation-rate though 

 the two are certainly not identical. In fact, it is often difficult to 

 tell from Child's arguments whether he means growth-rate, differ- 

 entiation-rate or metabolic rate, and it is not very helpful to lump 

 them all in one as "level of physiological activity". But it is time 

 to come to the facts obtained by the use of the susceptibiHty method. 

 Child used two variations of the susceptibility method, the direct 

 technique and the indirect technique. In the former case the 

 resistance or susceptibility is determined directly by concentrations 

 of toxic agent which kill the animals within a few hours. For a 

 particular species a concentration must be determined which kills 

 without acclimatisation, but which does not kill so rapidly that no 

 differences between parts of the body are discernible. Child has 

 used all kinds of toxic agents in his experiments, various cyanides, 

 alcohol, ether, chloroform, chloretone, acetone-chloroform mixtures 

 and other narcotics, X-rays, ultra-violet rays, ammonia, soda. He 

 has obtained the best results with those substances which have a 

 narcosis time and a killing time very close together, in other words, 

 with those substances whose effects are not complicated with nar- 

 cosis. The most favourable poisons, he found, were the cyanides. 

 As Hogben has pointed out, it would have been far more favourable 

 to the metabolic gradient theory if Child had confined his attention 

 to the cyanides, which are known to have an inhibitory action upon 

 some, though not upon all, tissue oxidations. For then it would not 

 have been known that substances such as ether and chloroform had 

 precisely the same effects, and the association between axial gradients 

 and oxidation-rate would have been more convincing. 



