OP ARTS AND SCIENCES. 153 



on the reaction which takes place between benzil and hydroxyl- 

 aniine. As is well known, Auwers and v. Me3 T er have obtained 

 two isomeric benzilmonoxinies, and three isomeric benzildioximes. 

 They sought, in the first place,* to explain this isomerism by a 

 different grouping of the atoms in space, but soon gave up this 

 explanation as insufficient,! especially because of the observations 

 of Hantzsch. The latter, together with Werner, J has brought for- 

 ward a new hypothesis to explain this isomerism, which depends 

 upon the assumption that space isomers of trivalent nitrogen can 

 exist. 



The experimental proof, however, upon which everything de- 

 pends, that the three dioximes, for example, have the same chemical 

 constitution, cannot be regarded as absolutely settled. 



From the observations developed in this paper, hydroxylamine 

 (one molecule) will react with benzil, forming in the first place an 

 addition product, 



/OH 

 C 6 H 5 C 



l x NHOH 



C 6 H 5 -C 



OH 

 OH 



which can then lose water (two molecules) in two different ways, 

 giving 



C 6 H 5 -ONOH C 6 H 5 C=N N 



I and q 



C 6 H 5 -CO C 6 H 5 -C( 



OH 

 7 oxime, I. a oxirue, II. 



In a like manner two molecules of hydroxylamine will react on 

 benzil, forming first the addition product, 



P XT p'-'H- 



uw 5 y HN0H 



which can then lose two molecules of water in three different 

 ways : 



* Ber. d. Chern. Ges., XXI. 946, 3510, XXII. 705. 

 t Ibid., XXIII. 2405. 

 t Ibid., XXIII. 11. 



