PREC1PITANTS FOR FURFURAL 43 



results were very much higher when phloroglucinol was used as 

 a precipitant than when barbituric acid was employed, the latter 

 giving normal values. This substantiates the statements of other 

 workers who found that hydroxymethylfurfural was not precip- 

 itated by barbituric acid. 



The barbituric acid method possesses, therefore, certain ad- 

 vantages over the phloroglucinol method, in that the reaction is 

 more specific and a definite condensation product is formed. The 

 precipitate, however, is sufficiently soluble to render a solubility 

 correction necessary. Then again, a large excess of the reagent 

 appears to be necessary, indicating that possibly an occlusion of 

 the precipitant leads to a compensation of errors. 



The possibility of obtaining better results by using some de- 

 rivative of barbituric acid will be discussed in the experimental 

 part of this paper. 



EXPERIMENTAL. 



Barbituric acid ordinarily is prepared by the condensation of 

 area with the sodium salt of malonic ester. The corresponding 

 thio derivative was prepared by Michael 10 17 and by Gabriel 

 and Colmann 18 by condensing thiourea with sodium malonic 

 ester, and the imino derivative was prepared by Michael 17 and 

 by Traube 19 from guanidine and malonic ester. These two de- 

 rivatives are analogus in many respects to barbituric acid. It 

 remained to be determined whether they would react in a simi- 

 lar manner with furfural, and possibly give a more complete 

 precipitation. 



The barbituric acid used in this work was a Kahlbaum prepara- 

 tion, which we purified further by recrystallization from water. 

 Analysis showed it to contain 21.80 per cent nitrogen; theory 

 21.87 per cent. 



Our first preparation of thiobarbituric acid was made accord- 

 ing to the method of Gabriel and Colmann. Two and three- 

 tenths gms. of sodium was dissolved in 50 cc. absolute alcohol, 

 and 16 gms. malonic ester added, then 7.6 gms. dry thiourea, 

 previously dissolved in absolute alcohol. The mixture was 

 heated on a water bath under a reflux condenser for ten hours. 

 The white pasty mass which resulted was then treated with 

 80 cc. water and 7.6 cc. hydrochloric acid and gently warmed 

 until it had dissolved. Upon standing, thiobarbituric acid crystal- 



