42 IOWA ACADEMY OF SCIENCE 



molecular weight of the condensation product. Furthermore, 

 this method calls for solubility corrections. Krober assumes that 

 two molecules of water are split out in the reaction between 

 furfural and phloroglucinal. Goodwin and Tollens 10 claim that 

 only one molecule of water is liberated at ordinary temperature, 

 but if the reaction is carried out at a temperature of 80° three 

 molecules are liberated. A slight variation in the conditions 

 may, therefore, affect the result considerably. Krober noted the 

 fact also that when the phloroglucide is allowed to stand in the 

 air for a time, it takes up moisture which cannot be expelled by 

 subsequent drying. From this brief survey of the literature, 

 it is obvious that the phloroglucinol method in common use is 

 not altogether satisfactory. 



Other reagents also have been tried with varying success. 

 Kerp and Unger 11 used semioxamizine as a precipitant for fur- 

 fural, but obtained results that were too low. Conrad and Rein- 

 bach 12 found that furfural and barbituric acid condensed in the 

 presence of dilute hydrochloric acid. Subsequently, Unger and 

 Jager 13 applied this reaction to the quantitative determination 

 of furfural. They found that six to eight times as much barbi- 

 turic acid as the theory required was needed to give the cal- 

 culated value for furfural. The condensation product had the 

 advantage of being only very slightly soluble in hydrochloric 

 acid (1.22 mgm. per 100 cc). They claim that barbituric acid 

 does not precipitate the furfural derivatives of hexose origin and 

 that these merely tend to color the solution yellow. The reaction 

 is a very simple one, consisting in the condensation of one mole- 

 cule of furfural and one molecule of barbituric acid, through the 

 aldehyde group of the former and the methylene group of the 

 latter, with the splitting out of one molecule of water. The pro- 

 duct was found to contain 13.75 per cent nitrogen, which is 

 in close agreement with the calculated value of 13.63 per cent. 

 When prepared from the furfural distillate from natural sources, 

 the product was found to contain 13.96 per cent nitrogen. 



Fromherz 1 * used barbituric acid as a precipitant for methyl- 

 furfural, and found the condensation product to be not appre- 

 ciably soluble. Fallada, Stein and Ravinka 15 found that barbi- 

 turic acid and phloroglucinol gave very nearly the same results 

 when pure xylose and arabinose were distilled and precipitated. 

 On the other hand, when sucrose was added to the pentose, the 



