Dr. E. J. Mills on Aniline Derivatives. 23 



but it is difficult to purify. The only substance I was able to 

 separate, in two large operations, was volatile in a current of 

 steam, white, and crystalline ; it was soluble in aqueous hydric 

 chloride, and precipitated therefrom by ammonia, but not by 

 platinic chloride after repeated crystallization from spirit. When 

 melted, it gave no definite fusion-point. 



Properties. — Having contrived a very delicate apparatus for 

 the determination of fusion-point, 1 have been enabled to make 

 some interesting comparisons among these derivatives ; but the 

 actual numbers (though otherwise completely corrected) have not 

 yet been converted into degrees of the air-thermometer. All 

 the determinations are made upon substances repeatedly purified 

 by fractional methods ; and in no case is the probable error of 

 the fusion-point greater than 0°*01. The result of fifty-five 

 observations with monochloraniline from acetanilide was 69°'69 ; 

 0°*03 higher than that of forty-eight observations with the deri- 

 vative of the direct acetic process. Dichloraniline could not be 

 obtained in sufficient quantity for systematic examination by 

 either of the methods ($) or (7). 



Trichloraniline was very easily made by the direct acetic pro- 

 cess, and purified by distillation in a current of steam. The 

 anilide process yields little, if any. For the organic analysis of 

 this substance I am indebted to the kind offices of my friend 

 Mr. Valentin. 



0*3392 grm. substance gave 0*7486 grm. argentic chloride. 



0*2311 grm. of the same substance gave 0*3103 grm. carbonic 

 dioxide. 



0*2758 grm. of the same substance furnished 0*3690 grm. 

 carbonic dioxide and 0'0600 grm. water. 



Found. • Calculated. 



Argentic chloride . . 220*7 219*1 



Carbonic dioxide . . . 134*3, 133*8 134*4 



Water 21*7 18*3 



These numbers agree very well with theory. The fusion- 

 point, as determined by ninety-six observations with two mer- 

 curial thermometers, was 77°*05 ; and it occurs so sharply, that 

 the probable error of the result is only 0°*0014. Lesimple 

 (Ann. Chem.und P harm. vol. cxxxvii. pp. 126 & 127) describes a 

 trichloraniline which he obtained by reducing nitrotrichlorobenzol. 

 He states that it has a very unpleasant and persistent smell, and 

 that it melts at 96°*5. On these two points my derivative differs 

 from his : it has rather a faint odour, but not very unpleasant, 

 and melts 19° lower. In all the other reactions mentioned by 

 Lesimple the two bodies exhibit a complete agreement. 



Monobromaniline prepared from acetanilide melted at 61°*80 



