955 
III. It was possible à priori that in this acid wherein we assume 
the presence of a floating double bond the reduction velocity should 
be different from that of an acid with an ordinary double bond. I 
have, therefore reduced glutaconic acid and also aconitie acid *) 
(where we meet with -a similar mobile hydrogen atom) according 
to the methods of Paar, Skita, and WirLsrärrer, measured the velo- 
cities of these reductions, and compared them with those of some 
other unsaturated acids. And indeed some peculiar differences now 
came to light. 
Whether the velocity measured is a true reaction velocity or some 
kind of diffusion velocity does not matter here. For in the latter 
case some differences in diffusion power would have been noticed 
between these two kinds of double bonds. The reduction experiments 
executed, however, elucidated this question somewhat; hence, I 
will again refer to the mechanism of these reduction processes 
more fully. 
IV. According to Tuorpr’s investigations *) glutaconic acid yields 
an hydroxyanhydride : 
CH — C=O 
Ue EN: 
CH O (6. hydroxy « pyrone) 
CH = C — OH 
which on hydration under different conditions — hence, also with 
strong alkali or in the presence of casein — reproduces the same 
acid (mp. 137°). The labile acid whose formation precedes that of 
the normal acid is, therefore, very unstable. 
The question now arose, however, whether the existence of 
this labile acid (of the labile acids, respectively) might not be 
demonstrated by a physico-chemical process. For this it was necessary 
to observe as accurately as possible the course of the hydration of 
the hydroxyanhydride under various conditions and at different tem- 
peratures. These experiments will also be communicated fully in 
what follows. 
Preparation of glutaconic acid by different methods. 
A. The object was attained most rapidly by the process of Conrap 
1) BLAND and THoRPE: Soc. 101 1490 (1912). 
2) Soc. 101. 863 (1912). 
