31 



experience and his results on samples New. 2 and 3 are very close to the theoretical 

 The following re.-ults were obtained on samples Nos. 4 and 5: 



Analyses of vanilla extract, 1907. 



& Omitted from averages. 



e 



that 



Also reported 0.0 14 iH-rn-nt rnum:irin in this samph-. 



"MM! \ H "I TMK ANALYSTS. 



.1 /.. A. /'/.,-.- it \v<>uid in- much more accurate to detennine the specific gravity of th 



-olution. and pipette oif a known volume for ana!y.-i-. Ii \va- found in this work tha 

 solun -landing in the lM-aki-n.ii the balance pan for five minutes lost 10 mg. 



Thi- mean- th.it under ordinary condition-^ the third decimal place in the result is 

 meaningleM. 



In thi- lal"!-- use a new te-t for coumarin which was originated by Doctor 



Q. It depend- mi the fact that a drop of alcoholic potash solution containing 



mi- of poia--iiini h)dr".\id to a liter will, when placed on crystals of coumarin, 



.i lemon yellou color. Thi- color i< very intense, but disappears rapidly. 



Neither vanillin nor acetanilid i- ai'feeted by the rea-jent. The test is applied 



direi il\ on thecry-tal- by ii-in^ a glass rod for a dropper, and the results have always 



much more -ati-i'a<tory than those obtained by any other test, because of its 



h vanilUo and coumarin give a yellow solution on standing for 



time, but the cry-tul- of \aiiillin remain uncofored until dissolved, while the 



coumarin inten.-elv colored. 



The color te- 1 iiiilhl are n i a- -at i-factory as they might be. In this work 



un. I the polari/inu' inuro-rope of the greatest use. There is little danger of 

 niiifu.-ini: cither of the thr. e cry-tal- which are likely to occur in a vanilla extract. 

 The di-Tepancie- in th i.--ults for vanillin are mute usual where the crystals are 

 \\eL r hed directly fi"in an ether .-olutioii. They seldom come out well, usually being 

 colored, often Dearly black. This can he remedied by another extraction, not with 

 ether. 



.!. /'. NV. On te-tiiiij this residue for acetanilid, (RHsert's tests) no reaction for same 

 could be obtained A- only 25 grams extract are taken, the residues actually obtained 

 in and 0.0026 irrain for the duplicates. Using 4 mg of pure acetanilid, 

 no r- ,!d be obtained by llitsi-rt's tests as given in Bui. 107. Using chlorin 



\\ater d'.S. I' i in -lead of a solution of chlorinated lime (I : 200) a good reaction was 

 obtained with I mu' acetanilid; 2 mg gave fair test. The chlorin water is mixed with the 

 a etanilid; a pink color form- in a few ,-econds, changing gradually to a purple and 

 finally ( a blue. 



./. r. oiwn: Vanillin: The residue of the ether extraction for vanillin almost inva- 

 riably contain- a lar_re amount of resin and other impurities. It has always been 

 our cu-tom to extract the vanillin with petroleum ether and deduct the residue from 

 the weight of impure vanillin, it will be noted that the difference in results is consid- 

 erable. In 18 determinations the impurity with the vanillin has varied from 2 to 30 

 mg. the average beinir In. 4 mg. 



We have al-o found that an easier drying residue of vanillin has been obtained by 

 extraction from the "2 per cent ammonia solution with chloroform. 



/e For the separation of coumarin and acetanilid we have been unable to 

 obtain petroleum ether with a boiling point 30^40 0. We have used gasoline, 86 B. 

 On fractionating this naphtha, fractions boiling at 35-40, 40-45, 45-50, 50-60, etc., 



