975
V O L U M E 2 4 , NO. 6, J U N E 1 9 5 2 Table 11.
Effect of Source of Chloroplatinic Acid on Anodic Current Current, pa." Lot A
39.6 Mean 0.9 Av. dev., % Electrode 2 used for these measurements.
Lot B
39.8 2.3
in a buffered solution prepared from equal volumes of 0.2 AT potassium dihydrogen phosphate and 0.2 S sodium hydroxide solutions. .4 second test of reproducibility involved the use of different preparations of chloroplatinic acid. Solutions were prepared from each of two lots of solid reagent grade chloroplatinic acid. Electrode 2 was replatinized several times from each solution and the anodic current measured with the same dextrose solution referred
to in Table 1. The currents obtained are shown in Table 11. Although the precision of the currents obtained with platinizing solution B is not so good as was usually obtained, the average current agrees well with the average obtained for platinizing solution A. In the work reported in Tables I and 11, two different electrodes were used. They were prepared in the same way, starting with 20gage wire. I t is apparent on comparing the average currents for electrode 1 shown in Table I, with those for electrode 2 shown in Table 11, that the current response of the platinized platinum electrode can be duplicated with the usual precision of polarographic current measurement. LITERATURE CITED
I. M . , and Lingane, J. J., "Polarography," p. 440, Xew York, Interscience Publishers, 141. (2) MaoNevin, TV. M., and Sweet, T. R., Quart. J. Studies on Alcohol, 12, 46-51 (1951). (1) Kolthoff,
RECEIVED for review J a n u a r y 11, 1952. Accepted March 15, 1952. From a thesis submitted t o t h e Graduate School of T h e Ohio S t a t e University in partial fulfillment of t h e requirements for t h e degree of daetor of philosophy, March 1952.
Determination of Dihydrostreptomycin and Mannosidostreptomycin with Periodic Acid W. AUBREY \'AIL1 AND CLARK E. BRICKER Department of Chemistry, Princeton University, Princeton, N. J .
D
This work was started in order to improve the chemical method for the determination of dihydrostreptomycin. During this preliminary investigation, an observation was made which has led to a differential spectrophotometric method for the simultaneous determination of mannosidostreptomycin and streptomycin in mixtures. Dihydrostreptomycin when treated for a specified time with periodic acid yields 1.02 moles of formaldehyde, whereas streptomycin produces only 0.20 mole. Prior to the formaldehyde determination with chromotropic acid which forms the basis for the determination of the dihydrostreptomycin, the iodic acid and the excess periodic acid are precipitated with lead acetate. The differential spectrophotometric method for mannosidostreptomycin and streptomycin utilizes the fact that when any of the streptomycins are treated with periodic acid and lead acetate, a characteristic color is formed. Because these colors for the streptomycins are very time-sensitive, optimum conditions for maximum sensitivity and differentiation must be carefully followed. These methods should find application in the manufacture, control, and physiological investigation of the streptomycins. The method for dihydrostreptomycin is believed to be the most reproducible chemical method so far reported. The reaction of lead acetate with the streptomycins which have been oxidized with periodic acid should provide a useful check on other chemical methods for deter.mination of mannosidostreptomycin.
IHYDROSTREPTONYCIS, n hich is made by the catalytic hydrogenation of streptomycin, has been determined by two chemical methods which differ in principle from those reported for streptomycin. Garlock and Grove (8) and Colon, Herpich, Johl, Seuss, and Frediani ( 4 ) reported methods based on the oxidation of dihydrostreptomycin with sodium metaperiodate and periodic acid, respectively. In both cases, the fornialdehyde that was formed during the oxidation was removed from the reaction mixture by distillation and then determined by t h e chromotropic acid procedure ( 1 , 3 ) . More recently, Hiscox ( 9 ) suggested a spectrophotometric method for dihydrostreptomycin based on its characteristic absorption a t 265 mp after acid hydrolysis. Although periodates are fairly specific reagents for oxidizing organic compounds containing adjacent hydroxyl or amino groups, this specificity is lost when the reaction mixture is heated ( 7 ) . Colon et al. ( 4 ) report that streptomycin, which theoretically should yield little, if any formaldehyde with periodic acid oxidation, gives with their specified distillation conditions 0.6 mole of formaldehyde per mole of streptomycin. Dih-drostreptomycin, furthermore, yields 1.6 moles of formaldehyde per mole of this compound instead of the expected 1.0 mole. The accuracy of their method depends therefore on a very careful control of the time of distillation. Garlock and Grove ( 8 ) allow the dihydrostreptomycin to stand overnight n-ith sodium metaperiodate, destroy the excess periodate with sodium thiosulfate. and then distill. As the chromotropic acid method is one of the most rapid and reliable methods for determining formaldehyde, and periodic and iodic acids interfere positively with this procedure, a more convenient method for removing this interference from a reaction mixture T V ~ Ssought. Roberts and Bricker, in extending their method ( 2 ) for the determination of end unsaturation in organice 1
Present address, American Cyanamld Co., Stamford, Conn.
ANALYTICAL CHEMISTRY
976
compound6. found that the excess periodic acid and iodic acid formed could be quantitatively precipitated with lead acetate. The I esulting precipitate could be centrifuged quickly and the formaldehyde could be determined hy rhromotropic acid, with no interference, on the clear centrifugate. This procedure has been extended to the assay of dihydrostreptomycin and to the detcimination of this compound in the presenre of rtreptoniycin. I
c 0
o.6b
0
Figure 1.
50
100 150 200 250 Standing T i m e , Minuter
300
350
400
Effect of Reaction Time with Periodic Acid on Yield of Formaldehyde
\Vhen lead acetate is used to remove the periodate and iodate ions in the dihydrostreptomycin determination, the supernatant liquid, on etanding, develops a yellow color. Under similar conditions, streptomycin gives nearly the same yellow color, whereas mannosidostreptoniycin produces a rose or pink color. Because of this decided difference in color, the possibility of developing a method for the determination of mannosidostreptomycin in the prwence of streptomycin \I as investigated.
aldehyde. If an appreciable amount of streptomycin, which can be readily determined by the procedure of Eisenman and Bricker (6),is present in the sample, a correction for the formaldehyde produced from this compound must be made. For samples t h a t contain less than 2% of streptomycin, the percentage of dihydrostreptomycin can I)e calculated directly from the formaldehyde produced or from n calibration curve obtained with a sample of kn0n.n purity of this compound. Development of Method for Dihydrostreptomycin. The only operat,ion in the recommended procedure that required thorough investigation was the time of standing of the periodic acid with the sample prior to the addition of the lead acetate, Purified samples of dihydrostreptomycin and streptomycin (both of these samples by eshaustive biological assays showed 98.0% of theoretical purity and 1.40% water) were examined by the recommended procedure, except that the reaction time, a t room temperature, of the periodic acid with the sample was varied. The results of this study are shown in Figure 1, where the moles of formaldehyde produced are plotted against standing time. It is apparent that the amount of formaldehyde produced from either of the streptomycins with periodic acid oxidation increases with the time of standing, hut that the difference in the amount of formaldehyde produced by these compounds remains essentially constant regardless of the length of Ptanding. -4lthough a 60-minute reaction time was chosen for the recommended procedure, this time is not critical ( f l minute) as long as whatever time is used is maintained a t a constant value for all calibrations and analyses. Shorter reaction times would cause less interference of streptomycin with the dihydrostreptomycin assay but would also decrease the sensitivity of the method. streptomycin gradually produces formaldehyde on standing with periodic acid, probably because it is hydrolyzed into coniponents xhich contain a primary carbinol group adjacent to a 1.0
02
REAGENTS A h I ) PROCEDURE
Reagents. Periodic acid is prepared daily by dissolving 1.00 gram of HJOs in 10.0 ml. water. Lead wetate is re ared by dissolving 7.5 grams of lead acetate trihydrate in 50 my. ofwater. Chromotropic acid is prepared by dissolving 1.00 gram of 1,sdihydr0xynaphthalene-3~6disulfonicacid in 5.0 ml. of water. The solution b filtered if not clear and is not used if over 3 days old. Procedure for Determination of Dihydrostreptomycin. Weigh :tccurately a eample of dihydrostreptomycin sulfate or chloride or a sample containing salts of both streptomycin and dihydrostre tomycin and dilute accurately to a volume so that the resuyting solution contains between 4 and 6 mg. of the dihydrostreptomycin d t (or of the mixture of streptomycin salts) per nil. Pipet 2.00 ml. of this solution into a 10.0-ml. volumetric fl:tsk, add 1.0 ml. of periodic acid solution, and allow the solution to stand 60 f 1 mmutes at room temperature. Then add 2.0 nil. of the lead acetate solution and dilute the resulting mixture to 10.0 nil. Centrifuge the contents of the volumetric flask a t approximately 1500 r.p.m. for 3 minutes. Accurately pipet 1.0 nil. of the clear supernatant centrifugate to a test tube. Add 0.50 ml. of chromotropic acid solution and 6.0 ml. of concentrated sulfuric acid and then heat the test tube and its contents in a 1)oilingwater bath for 30 += 3 minutes. Cool the solution, transfer to a 25-mI. volumetric flask, and dilute with water to volume. After the resulting solution has reached room temperature, measure the optical density a t 570 mp against a blank solution n-hirh has been carried through the entire procedure. M'ith this procedure dihydrostreptomycin sulfate or chloride produces 1.02 moles of formaldehyde per mole of compountl, whereas either streptomycin salt gives on1.v 0.20 mole of form-
3 0.E c 01
n
-0
-
LI
0" 0.4
0.2
0.c 4 0
440 Wavelength,
480
520
mv.
Figure 2. Transmittance Curves of Streptomycin Sulfate Solutions after Periodic Acid-Lead Acetate Treatment Open circles, streptomycin sulfate; closed circles, mannosidostreptomycin sulfate; open squares, dihydrostreptomycin sulfate; closed squares, dihydmmannosidostreptomycin sulfate
-
977
V O L U M E 2 4 , N O . 6, J U N E 1 9 5 2 Table I.
Results of Dihydrostreptomycin Analyses Biological Turbidimetric Method, % ' Purity Laboratory Laboratory A Ba
Lot S o .
Chemical .\lethod, % Purityb
a Data neceawrry to calculate per cent error for these analyses were not available. b Separatesamples were weighed for each analysis.
carlioii atom containing another hydroxyl group. In order to eliminate or minimize such hydrolysis, sodium metaperiodate instead of periodic acid was used for the oxidation. This modihcation made little difference in the yield of formaldehyde from either of the two streptomyeins.
Procedure for Determination of Mannosidostreptomycin. Prepare a solution of the sample containing mannosidostreptomycin and qtreptomycin, so that 1 ml. contains the equivalent of 5 2~ 1.0 ing. of the sulfate salts of the two streptomycins. Pipet 1.00 ml. of this solution into a 10-ml. glass-stoppered volumetric flask and place the flask in a water bath mamtained a t 25" C. Add 1 ml. of periodic acid solution and, after standing for 30 i- 0.5 minutes, add 2 ml. of lead acetate solution. Dilute the resulting niisture t o 10.0 ml., mix thoroughly, and filter through fine filter paper into a test tube suspended in the 25" C. temperature hath. hleasure the optical density of a portion of the clear filtrate a t 415 and 485 mp against a water blank 180 i- 1 minutes after adding the lead acetate. The amounts of streptomycin and mannosidostreptomycin can then be calculated by solving the following simultaneous (quations: 0.147 4 0.093 B = 111 0 131 il 0.194 B = Dz
+ +
= mg of streptom cin in the sample H = mg. of mannosidSlostreptomycin in the sample lli = measured optical density of unknown a t 415 nip ll2 = measured optical density of unknown a t 485 mp The constants for these equations were calculated from t h p optical density produced by 5.0-mg. samples of purified streptomycin and mannosidostreptomycin sulfates which had been put through this procedure. As A and B are given in milligrams the constants are actually the optical densities that 1 mg. of each rompound 11 o d d give and are thus one fifth of the values shoa II i n T:ihl? I1
11
lirie .l
DISCUSSIOh AND RESULTS
0 -
.
0
1
1
I
0.01
I
I
30
0
60
1
I
120
90
Minules
Standlng
Time
with
Periodic
Acid
Figure 3. Effect of Reaction Time with Periodic Acid on Development of Colors 0 Mannosidostreptomrcin Q
Streptomtcin
13cc:iuse of the gradual increase in formaldehyde Erodured in this procedure and even though lead acetate precipitates periodate quantitatively, centrifugates that have stood overnight before being analyzed for formaldehyde gave high results. I t is necessary, therefore, to analyze the centrifugates for formaldehyde within an hour after the addition of the lead acetate in order to get reliable results. Results of Dihydrostreptomycin Analyses. Analyses of several lots of dihydrostreptomycin sulfate by this proceduic and by two independent laboratories using a turbidimetric bioassay procedure and test organism Klepsiella pneumoniae (culture S o . P.C.I. 602) are reported in Table I. Because the biological aways are divergent in most cases, it is difficult to make any accurate comparisons of the t n o methods. However, it seems that the chemical method gives plausible results and that the reproducibility of the method is good. .4s six or eight samples c m he analyzed concurrently by the reconimended procedure, the man-hours required per sample are IOW.
Five-millipiam samples of the sulfate salts of streptomycin, maiino6ido.ticptoiiiycin, dihydiostreptomycin, and dihydroinannosidost reytoiiir cin were run through the procedure except that the entire transmittance curve of each of the r e d t i n g solutions n as measured (Figure 2). Bccause streptomycin has an absorption maximum a t 415 nip, \\ hereits manno4lostreptoinycin shone its mavinium absorption a t 485 m+, a method for the simultaneous determination of mannosidostrcptoniycin in the presence of streptomycin was suggested. In the development of this method, the follo~ingvariables nere studied: time of standing a t 25' C. Nith periodic acid, amount and concentration of periodic acid, pH of initial and final solutions, time of waiting after addition of lead acetate solution, amount of lead acetate, and amount of the streptomycins. The temperature a t which the entire procedure is rarried out is certainly another variable, and although not studied, a temperature of 25' C. was arbitrarily maintained for all experiments. The effect of the reaction time of periodic acid TFith each of the streptomycin compounds is shoq-n in Figure 3. It is obvious that a more intense color is produced for the shorter reaction times, but after considering the sensitivity, the rate of change of sensitivity with time, and the difference in color given by the compounds, a reaction time of 30 minutes wm chosen. The amount and concentration of periodic acid used in thiF determination are not critical. The amount of periodic acid nas varied by adding 1 ml. of various solutions containing betneen 20 and 130 mg. of this arid. The readings a t 415 and 485 nip for the two streptomycins nere lowest for the smallest quantity of periodic acid but remained essentially constant when 80 and 120 nig. of periodic acid were used. The concentration of periodic acid was varied by adding various volumes of different periodic solutions, so that a total of 100 mg. of the acid M a p a h a y s present. llthough this variable did not shor5 an appreciable effect until the most dilute acid solutions were used, the most suitable concentration seemed to be 100 mg. per ml. A periodic acid solution containing 100 mg. per nil. has a pH of 1.1. If the p H of this solution was adjusted to 0.5 \\ith sulfuric acid before it was used in the recommended procedure, no color was obtained from either streptomycin or mannosidostreptomycin. Likewise, if sodium acetate was added in various amounts so that the periodic acid solution had pH's of 1.7 to 3.8, no colors Tvere formed.
978
ANALYTICAL CHEMISTRY
The effect of the p H of the solution after the addition of lead acetate was investigated and found to be less critical. In the recommended procedure, the pH of the filtered solution is 3.8 and is apparently well buffered. In the pH range of 3.6 to 4.0, no detectable variation was observed, but a t lower or higher pH's less color was developed for each of the streptomycins. One of the most important variables in this procedure is the time necessary to develop the color after the addition of lead acetate. A plot of the optical density a t 415 and 485 mp for streptomycin and mannosidostreptomycin a t various times after adding lead acetate is shown in Figure 4. The optical density of streptomycin a t both wave lengths becomes essentially constant after 100 minutes. However, the optical density a t 485 mp for mannosidostreptomycin does not reach its maximum value until about 180 minutes, whereas the reading a t 415 mp does not reach a limiting value even after 4 hours' standing.
0.2
However, when only 150 mg. of lead acetate was added, no color developed for either streptomycin. One of the serious limitations of this method is due to the fact that neither streptomycin nor mannosidostreptomycin gives linear calibration curves over the range of 1 to 5 mg. However, in the 3- to 5-mg. range, the optical densities are almost linear with concentration. In addition, as long as a total of 4 to 6 mg. of both streptomycins is present, the readings a t 415 and 485 nip appear to be linear and additive, The optical densities a t t h e two wave lengths when plotted against per cent streptomycin, where the v eight of streptomycin plus the weight of mannosidostreptomycin equals 5.0 mg. in all cases, are shown in Figure 5. In this procedure, it is necessary to know accurately the optical density produced a t the two wave lengths by 5.0 mg. of pure streptomycin sulfate and pure mannosidostreptomycin sulfate. The average values of six determinations on streptomycin sulfate with two different solutions and of nine determinations on three different solutions of mannosidostreptomycin sulfate are given in Table 11. It is obvious from these values that the colors for both of these compounds a t the two wave lengths are reproducible to uithin only about 2%.
a E
2 0.1
Table 11. Optical Densities of Streptomycin Sulfate and Mannosidostreptomycin Sulfate"
t
Wave Length,
Streptomycin hlannosidostreptomycin Sulfateb Sulfate 415 0 . 7 3 3 2= 0.008 0 . 4 6 4 i 0.008 486 0 . 6 5 7 =t0 . 0 0 8 0.968 2= 0.018 a All spectrophotometric measurements made on Beokman Model B spectrometer using 1-cm. Corex cells. Showed 98.0% of theoretical purity from exhaustive biological assays a n d 1.40% water. This material using paper chromatographic technique of Winsten and Eigen ( 1 2 )gave only one zone and contained 1.0% water. Mp
z I
2 0.C 01
0
-;0.2 4 -
n
0
a €0.1
Yt
I
I
60
120
,~
ac
Standing
Time
Minuter with Lead
I I80
I
240
Acetate
Figure 4. Effect of Standing Time with Lead Acetate on Development of Colors 0 Mannosidostreptomycin
0
Streptomycin
The steady increase in the optical density at 415 mp for mannosidostreptomycin is probably due to the hydrolysis of this compound into the same intermediate that produces the yellow color from streptomycin. This is substantiated by the fact that solutions of streptomycin and mannosidostreptomycin which were put through the recommended procedure and allowed to stand overnight appeared to have the same yellow color. Because of the instability of the pink color from mannosidostreptomycin, it is necessary to measure the optical density a t the two wave lengths a t a very definitely specified time. The difference between the two compounds appears to be greatest after 180 minutes of standing .iyith lead acetate, and consequently this standing time is recommended. The amount of lead acetate used in the procedure is not too critical, as long as a larger amount of lead ion is present than is needed to precipitate the iodate and periodate. Changing the amount of lead acetate from 280 to 340 mg. in the recommended procedure did not alter the readings at either wave length.
'U 40 80 0.4 0
20
60
%
100
Streptomycin
Figure 5. Calibration Curves for Mixtures of Streptomycin and Mannosidostreptomycin
Various synthetic mixtures of the sulfate salts of streptomycin and mannosidostreptomycin xere analyzed by the recommended procedure. The results of these analyses, shown in Table 111, clearly indicate that the accuracy of any one determination is not better than about 10%. When several analyses are averaged, the results are usually coneiderably better, as is shown from the average values given in this table. Although this procedure has been applied only to mixtures of streptomycin and mannosidostreptomycin of clinical purity (greater than 85% purity) it may be possible to use it on less pure samples if suitable blanks are carried through the procedure and
V O L U M E 2 4 , N O . 6, J U N E 1 9 5 2
919
Table 111. Analyses of Mixtures of Streptomycin and Mannosidostreptomycin Sulfates Streptomycin, llg. Taken Found Average 1 0 1 06 0.9; 0 88 0 97
Jlannosidostreptomycin, hIg. Taken Found Average 4 0 3.94 4 03 4 12 4 03
2.0
2.22 1.93 1.94
2.03
3.0
2.78 3.07 3.06
2.97
3.0
3.31 2.94 2.88
3.04
2 . CI
1.69 2.06 2.12
1.96
4.0
4.12 3.98 4.22
4.11
1 .@
0.88
0.89
1.02
0.78
manganese, bismuth, cadmium, and strontium ions produced no color change after a sample of streptomycin was treated with periodic acid. It is possible that streptomycin after a short periodic acid treatment may provide a specific color reaction for lead. ACKNOWLEDGMENT
The authors nish to thank the Heyden Chemical Corp. for supplying the samples of' the streptomycins used in this investigation and also for the biological assays that are reported. The assistance of X. James Sage, x h o ran some confirmatory experiments on the determination of dihydrostreptomycin, is also gratefully acknowledged. LITERATURE CITED
Bricker, C. E., and Johnson. H. R., ISD. ENG.CHmI., A s z ~ . used in place of the wat'er blank for reading the optical densities. The very exacting conditions that niust be f o l l o ~ e d the , lengt,h of time required for the determination, the comparatively low wnsitivit,y, and the necessity of solving simultaneous equations are all disadvantages of the method. On the other hand, as contrasted to other chemical methods for determining mannopidost,reptoniycin (6, 10, 11 ), sugars and carbohydrate material do not interfere. Furt,hermore, this method provides an entirely new method for the simultaneous estimation of tiyo streptomycins in the presence of each other. Lead ion appears to be the only cation that gives a yellow color with streptomycin and t'he rose color with mannosidostreptoniycin after periodate treat'nient. Although a study over t,he entire pH range was not made, ferric iron, aluminum, calcium, magnesium, mercuric, mercurous, thallic, silver, beryllium, zinc,
ED.,17, 400 (1945).
Bricker, C. E., and Roberts, K. H., 4 x 1 ~CHEM., . 21,1331 (1949). Bricker, C. E., and Vail, W.-4.. Ibid., 22, 720 (1950). Colon, A , Herpich, G. E., Johl, R. G.. Keuss. J. D., and Frediani, H. A , , J r A m . Pharm. Asaoc., 39, 335 (1950). Eisenman, W,,and Bricker, C. E., ANAI.. C H m r . , 21, I507 (1949).
Emery, W.B., and Walker, =i. D., AnciZyst, 74,455 (1949). Fleury, P., and Boisson, R., C o m p t . rend., 208, 1509 (1939); J . pharm. chim., 30, 145, 307 (1939).
Garlock, E. A , Jr., and Grove, D. C.,J . CEin. Invest., 28, 843 (1949).
Hiscox, D. J.,AKIL. CHEM., 23, 923 (1951). Kowald, J. A., and AlcCormack, R. B., Ibid., 21, 1383 (1949). Perlman, D., J . Bid. Chem.. 179, 1147 (1949). Winsten, IT. A., and Eigen, E., J . A m . C h m . Soc., 70, 3333 (1948).
RECEIVED for review October 2 , 1951. Accepted March 26, 1952.
Improved Ferric Chloride Test for Phenols SAUL SOLOWAY AND SAMUEL H. WILE" The City College, College of the City of New York,New York,N . Y. The ferric chloride test for phenols as described in standard works on organic analysis fails in many cases. This study shows that this qualitative test can be greatly improved if carried out in organic solvents using small amounts ofpyridine as an alkalinizing agent. Some of the best tests are obtained in chloroform solution, a medium which had been described previously as giving negative results. All phenols, with some minor exceptions, give the test using these modified procedures. The chemistry of the test is discussed, showing the important role plaved by the solvent and an added base.
F
ERRIC chloride in solution give? colors with a number of organic derivatives. The common ones are phenols, enols, oximes, h>droxamic acids, and some carboxylic acids. Howc~vei,the appearance of a color change on the addition of ferric chloride depends to a great extent on such common factors as solvent, acidity, and concentrations of reactants. This study shows that a consideration of these factors leads to a greatly improved qualitative test for the phenolic group. The recipe given for applying the ferric chloride test to phenols is the solution or suspension of the unknown compound in water, methanol, or ethanol, to which is added a solution of ferric chloride in the same solvent ( 5 , 8, 9, 1 1 ) . If a marked color change does not result, some investigators have recommended the addition of solid or aqueous sodium bicarbonate. This procedure was parried out in water and methanol with 78 phenols (Table I). 1 Present address, Department of Chemistry, University of Kansas, Lawrence, Kans.
Among other phenols cited in the literature, Table I shows that such common compounds as p-benzylphenol, trans-diethylstilbestrol, meso-hexestrol, ethyl p-hvdroxgbenzoate, etc., gave negative tests with the usual procedure. -4consideration of the chemistry of this color test led to its trial in other solvents. Kesp and Brode ( 1 4 ) noted that color changes were obtained for many phenols in oxygenated or nitrogenated solvents, but not in hydrocarbons or their halogenated derivatives. Preliminary experiments verified thr observations of these authors. Holyever, the addition of relatively small amounts of pyridine to solutions of anhydrous feiric chloride and phenol in such solvents as benzene, toluene, o-xylene, chloroform, chlorobenzene, but>-lbromide, and ethylene dibromide, gave deep purple colors. It is the authors' experience that these solvents are superior to the ones usually used in this color test. With the exception of 2,6-di-fert-butyl-p-cresol, a highly hindered phenol that does not even react with metallic sodium in anhydrous ether