144 SCIENCE PROGRESS. 



have already considered the addition of two hydroxyl groups 



to fumaric and maleic acids respectively, and the proof 



that the acid (maleic) which yields inactive tartaric acid 



OH . 



HCCOOH 1 . „ r , HCCOOH 



HrrODH must possess the " cis formula t-j/ydqu 



OH 



depended on the assumption of the above principle, for it 

 assumed that both the OH groups attach themselves to the 

 same side of the molecule; or, in other words, that only one 

 pair of tetrahedron-corners is separated to take up the two 

 added groups ; similarly the " trans " formula is proved for 

 fumaric acid (compare the figures for fumaric, maleic, and 

 tartaric acid). 



In considering the effect of adding atoms to two carbons 

 trebly linked, we must remember that these are represented 

 by two tetrahedra placed face to face, and that when one 

 pair of corners is separated to take up two new atoms, two 

 pairs of corners will remain joined, so that the added atoms 

 will be on the same side of the molecule. On addition of 

 bromine to acetylene-dicarboxylic acid we should get, then, 



2 , the "cis" or dibrom-maleic acid. As a fact we 



L0 2 HCBr 



get this acid, but at the same time there is formed twice as 



f^O HfRr 

 much of the other isomer * ,-. ; the formation of the 



BrCcO.H 



latter is, however, not surprising when we consider that 

 it is much more stable than its isomer, which may be sup- 

 posed to be the sole immediate product, and to undergo 

 transformation while in the nascent state. The numerous 

 similar objections to the theory which have been raised, 

 especially by Michael, are similarly answered, for in every 

 case the objection is the same, that instead of the expected 

 product there is formed another which is more stable in the 

 conditions of the experiment. Moreover, these objections 

 refer exclusively to halogen-derivatives, which we have seen 

 to be especially liable to transformation. 



As when atoms are added to the molecule, so when they 

 are abstracted from it, a difference between the isomers is 



