SO-CALLED FORMATIVE SUBSTANCES. 179 



to me, however, that the relative size of structures produced from 

 different levels and the relative position of intermediate struc- 

 tures are as truly an expression of polarity as the existence of 

 different structures at the two ends, since they must be the result 

 of characteristically different physiological conditions present at 

 the different levels. 



Though I have not attempted it, it is possible that with the 

 aid of special assumptions these cases could be made to fit into 

 a formative substance hypothesis, but the value of a hypoth- 

 esis is to a certain extent inversely proportional to the number 

 of special assumptions required, and there is always danger that 

 a special assumption may introduce more difficulties than it 

 removes. 



Moreover, if axial polarity is due to a gradation of substances, 

 we must logically conclude that a gradation of substances from 

 the median line to the lateral margins of the body in both direc- 

 tions exists, for the structural differentiation in these directions is 

 just as truly an expression of " polarity " as is the axial differen- 

 tiation. This leads us into numerous further difficulties. This 

 lateral gradation must exist in the developing parts as well as in 

 the old, otherwise the differentiation between lateral and median 

 regions could not arise. How, for example, are we to account 

 for the formation of the other side of the body and of a median 

 region in longitudinal pieces of planarians of less than half the 

 width of the body ? In whatever manner we may proceed to bring 

 these cases under the general hypothesis certain other special as- 

 sumptions are necessary. If we say that the formation of the 

 opposite side along the cut surface is due to the excess of mar- 

 gin-forming substance in the piece, how shall we account for the 

 formation of a median region between the two margins ? If, on 

 other hand, we attribute the formation of a median region to 

 gradation of the formative substances for this region, supposing 

 them to be in excess at the cut surface, how shall we account for 

 the formation of the opposite side ? Furthermore the new mar- 

 ginal region is formed first. Consideration of other experimental 

 data adds further difficulties. 



But perhaps enough has been said to show some of the special 

 difficulties in which a hypothesis of formative substances involves 



