(purified through the methoxy compound) in 500 cc. of water containing 4.3 g. of sodium bicarbonate was treated with 20 cc. of 30% hydrogen peroxide (superoxol). T h r solution rapidly darkcried and, after standing for thirtyfour hours, a crop of clull red large leaves had separated, 3.2 g. (43% based on 2-hydroxy-l,4-naphthoquinoneutilized). One crystallization from tlioxane gave piire isoiiaphthazarin. Acidification ol the aqueous filtrate yielded 2.1 g. of orange precipitate which consisted of unchanged 2liyciroxy-l,~-uaphtlioc~uinoneplus a small ainouut of isoiiaphthazarin. On . rai occasions, u s e of less pure 2liydroxy-l,-l-naphtiioquinolie as starting material led t o iiiuch lower yields.
Z-Hydroxy-d-( 6-cyclogeranylj -1,4-naphthoquinone.Isonaphthazarin (2.9 g.) was reduced as described by Fieser and Gates' and the leuco compound heated in the dark under nitrogeri for forty-eight hours a t 85-iO' with 1.0 g. of $-cyclogeranini2 ( i n . p. +lo), 0.ii g. of anhydrous oxalic acid and 20 cc. of dioxane. The processing of the reaction misture included the following steps: extraction of the unchanged lcucoisoriaphthdzarin with aqueous hydrosulfite, rec1uctio:i with ~oiicetitraied aqueous hydrosulfite, anti extraction from ether- petroleum ether with Claiscn's alkali. 'Ihc crude phenolic portion thus obtained was chrotiiatographed af I el' air oxidation on freshly ignited ruagncsium sulfate. On develop men^ with peiroieuin ether, a weakly adsorbed bright yellow baiid readily passed into the liltrate. Similar filtratrs from systematic readsorptions of the column eluate were coinb i n d and on co~icentration to dryness under reduced pressure afforded 99 nig. of solid residue which aftcr three crystallizations from ether--petroleurn ether gave :;ti nip. OF golden yellow rectangular plates, 111. p. 135135.5'. I t is quite soluble in the ordinary organic solvents, fairly soluble in warm petroleum ether, much less soluhle cold, and dissolves in dilute alcoholic alkali to give the heautiFill scarlet characteristic of alkali salts ot 2-hydroxy-i ,-Ii~aphthoquiiioiies. 1t dissolves i n concentrated sulfuric acid to give a deep orange-red solution. l'wo further crystallizatioiih to ohtaiii a -ample for analysis t i i d ncit alter the melting l:itii111.
A n d . Calcti. for C.t,,FI&: C , 77,:iX; H, 7,1*5. Fouiitl: C, 77.50; H, 7.2:i. e-Cyclogeranolapachone (1).-- A solution of I I nig. of 2hydroxy-3-(~-cycioge~dIiyl)-l ,~-naplithoquino~ie in ice-cold concentrated sulfuric acid ([i.X cc.) was allowrtl to stand several minutes, then diluted with ice water. The prcscipitated dark orange-brown material was taken into ether, washed with water, bicarbonate arid brine, and coiicentrated t o dryness. The, residue was taken into benzene hexane and chromatographed on frvshly ignited inagnrsiuiri sulfate. Development with . i O ~ o henzen-hexane left a broad salmon-pink hand iu the middle of the column which was sectioned out arid t.lirtLd with ether. After evayorar i o n of tiir ether, i l i e solid residue was crystallized twice from piire a c c ~ o n eto give l j . 5 mg. ol orange-red prisinatic hlatlcs, 111. p. 2:12 233.3 '. A inisetl melting point with (3-yeranolapachone prepared according to Fieser and Gate,' \howecl 110 depreszioii
DEPARTXENT OF CHEMISTRT B R Y N h!!A\VR COLLEGE
RECEIVEDMAY 29, 1942
BRVNWIAWR, P.EWSYI.VAXJA
Riboflavin Estimation in Fruits and Vegetables BY G. MACKINNEY AND J. M. SUCIHARA
As part of a collaboratiw project,' i t was recently necessary to make a series of chemical determinations of thiamin and riboflavin in certain fruits and vegetables, and the Comer-Straub procedure2 was followed. Unfortunately, a t the beginning, S ~ p e r s o r b the , ~ specific adsorbent for riboflavin was unavailable. An empirical method was, therefore, evolved, and we hoped, later, to correlate results into the series by concurrent assays on additional samples on arrival of the adsorbent. This comparison may iiow be made and, subject to certain provisos, we believe the modification accurately reflects differences in riboflavin content within a series. With respect to absolute values, it is in accord with microbiological assay by means of Luctobacillus casel:. It has, further, certain advantages : increased light stability, 110 adsorbent is needed and the riboflavin in the aqueous buffer exhibits approximately twice the fluorescence found in pyridine-acetic solution, with consequent decrease in the percentage reading crrcr. The Cornier-Straub procedure is followed in detail in extraction and prcparation of the sample, except that, in thc case of fruits, 10 ml. of pectinol (1 g. in 25 ml.) is added per .iO-ml. of sample, in addition to the clarase. The whole is then iricuhated a t 45' for two hours. The pectinol is absolutely nccrssary for prunes, apricots, dates, etc., to produce a satisfactory solution. A 10--20 ml. aliquot is theu heated t o tioilirig with 5 ml. of 2yoacetic, as in ( l ) , inark t o voluine, 50 mi., with buffer, and a 15-ml. aliquot ~ i ithree minutes with 1 nil. of rrcated for a ~ i ~ i n i m u of po~assimnpcrmanganat e , and decolorized with 3 ml. of ;3(70hydrogcn peroxide. The solution is then filtered and wmparecl with Ixiffercd standards a t PH 6.0 in a Coleman iluorop1iotoriit.ter. The R? filter for !lie exciting light (Hg arc) cuts out coinpietcl>-above 4900 A. and for the fluorescent light, below 3100 .k. Thc cut-out is sharp, and, for the latter filter. tlie transmission rises from zero a t 5100 t o over 90% a.t 5400
