(FeCl3)(FeCl3) (1932), Schunk (FeSO4) (1924) and Levine and co-workers (ferric citrate) (1932). The addition of bismuth car-bonate to the culture medium was recommended by Pacheco

HYDROGEN SULPHIDE STUDIES

I. DETECTION OF HYDROGEN SULPHIDE IN CULTURES

CHARLES A. HUNTER AND H. GILBERT CRECELIUS

Department of Bacteriology and State Health Laboratory, Universityof South Dakota

Received for publication July 23, 1937

A considerable number of investigators have proposed methodsfor the detection of hydrogen sulphide in cultures. Media madefrom different brands of peptone as well as the addition of sul-phur compounds, such as cystine, sodium sulphite, sodium sul-phate, taurin, thiourea, etc., have been tried. Burnet andWissenbach (1915), Kligler (1917), Jordan and Victorson (1917),Thompson (1921), Tilley (1923), Mulsow and Paine (1925),Bailey and Lacey (1927), and others added a lead salt to themedium to serve as an indicator for the presence of hydrogensulphide. Filter papers impregnated with lead acetate or leadcarbonate have been used by Tanner (1917), Meyer (1920),Wilmet (1927), Zobell and Feltham (1934) and others. The useof iron salts for the detection of hydrogen sulphide has beenrecommended by Wilson (FeCl3) (1923), Beckwith and Moser(FeCl3) (1932), Schunk (FeSO4) (1924) and Levine and co-workers (ferric citrate) (1932). The addition of bismuth car-bonate to the culture medium was recommended by Pachecoand Mello (1932). In 1934 Zobell and Feltham compared lead,iron and bismuth as indicators for detecting hydrogen sulphideformation and concluded that iron was more sensitive and lesstoxic than either lead or bismuth. They recommend, however,the use of lead acetate papers when testing new and unknownpure cultures of bacteria.

In considering the detection of hydrogen sulphide produced bycultures two main problems present themselves. First, the

185

on February 27, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

186 CHARLES A. HUNTER AND H. GILBERT CRECELIUS

selection of a sensitive indicator of low toxicity and distinctcolor change; second, the selection of a medium suitable for thegrowth of the organisms, and containing an available source ofsulphur. An indicator may be of very low toxicity but whenadded to a medium it may combine with some of the ingredientsproducing a precipitate, thereby increasing toxicity and decreas-ing sensitivity.

All of the heavy metals will react with hydrogen sulphide butonly a few are desirable for bacteriological work. Those com-monly used as indicators are lead and iron salts. Theoretically,iron is more sensitive to hydrogen sulphide in an alkaline men-struum than in an acid one. As most of the culture media areslightly acid or neutral, it follows that the sensitivity of iron isdecreased and minute quantities of hydrogen sulphide win notbe detected. Therefore, an indicator which will be sensitive inthe acid range is desirable.

Theoretically, bismuth is a metal which is sensitive in an acidmenstruum. The difficulty in the use of bismuth is that mostbismuth compounds are insoluble and therefore not applicableto bacteriological work. There are, however, several solublebismuth compounds which are of low toxicity and which are goodindicators for the detection of hydrogen sulphide.

PREPARATION OF BISMUTH INDICATOR

A soluble bismuth compound, suitable for bacteriological workand commonly known as "bismuth liquor," is prepared as follows:

Place 3 grams bismuth citrate, Merck U. S. P. VIII, in a glass stop-pered bottle and add about 10 cc. of distilled water. Mix well and addapproximately 1 cc. of NH4OH sp. gr. 0.90. The bismuth citratedissolves quickly forming a clear, colorless solution. If it does notclear up readily, gentle heat will increase the solubility. Add distilledwater to bring the volume up to 100 cc., one-half (0.5) cc. of the bismuthliquor is added to each 100 cc. of medium.The bismuth liquor prepared according to the above directions will

remain perfectly clear for several weeks and in that condition can beused with satisfactory results. Should the solution become turbidor opalescent the solution should be discarded and a fresh solutionprepared.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


In place of the bismuth liquor 16 mgm. of Bismuth and AmmoniumCitrate, Merck U. S. P. IX, may be added to each 100 cc. of medium.

COMPARISON OF THE SENSITIVITY OF ION AND BISMUTH SALTS

TO H2SSeveral experiments were performed to determine the relative

sensitivity of ferric ammonium citrate and bismuth liquor. Aseries of test tubes were prepared containing 5 cc. of variousbuffers ranging from pH 6.0 to 9.0 in steps of 0.5 pH. To eachtube of the buffer series was added 2.5 mgm. of ferric ammoniumcitrate and to each tube of a second series 0.125 cc. of bismuthliquor. The tubes in each series were then titrated, to a definitebrownish-black color, with a hydrogen sulphide solution con-taining a known number of millimols of hydrogen sulphide percubic centimeter.

Similar titrations were performed using a 2 per cent tryptonephosphate agar adjusted to the same H-ion concentration as thebuffers employed in the previous experiment.Some difficulty was encountered in the titration of the ferric

ammonium citrate solutions. The iron compound reacts veryslowly with hydrogen sulphide to form iron sulphide. Therefore,it was found necessary to run several series of these tubes andafter the addition of a known amount of hydrogen sulphide thetubes were allowed to stand five or ten minutes before readingswere made. This indicates that iron is not as sensitive to hydro-gen sulphide as some other metals, for, in the case of bismuth, theformation of the sulphide was instantaneous.

In examining the data in table 1 the superiority of bismuth overiron is clearly shown. It should be noted that the amount ofH2S required to produce a brownish-black precipitate in thebuffered solutions using bismuth as an indicator, was only 0.0048millimol. When ferric ammonium citrate was used as an in-dicator the amount of H2S varied from 0.02885 millimol at pH8.0 and above to 0.101 milhimol at pH 6.0. In the bufferedmenstruums bismuth proved to be 6 to 21 times more sensitiveto H2S than iron, depending upon the H-ion concentration. Itis noteworthy that bismuth indicator is more sensitive to hydro-

187


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


gen sulphide in the presence of agar than in the buffered liquidmenstruum. The ratio of sensitiveness of bismuth to iron variedfrom 8 to 1 at pH 9.0 to 24 to 1 at pH 6.0. Why the bismuthwas more sensitive to hydrogen sulphide in an agar menstruumthan in the buffer solutions cannot be stated. Inasmuch as thepresence of hydrogen sulphide is generally determined in an agarmedium this difference in the sensitiveness of bismuth over ironis of considerable importance, especially with organisms that pro-duce such small quantities of hydrogen sulphide as to cause nocolor change with the ordinary indicators. The results of theseexperiments indicate that ferric ammonium citrate is not nearly

TABLE 1Sensitivity of Bi and Fe as shown by the number of millimoles of HVS required to

show the presence of sulphides at various H+ concentrations

BUJFFURD SOLUTIONS NUTRNINT AGAB

pH Sensitivty sensitivityBismuth Iron of Bi over Fe Bismuth Iron of Bi over

6.0 0.00481 0.101 21X 0.00240 0.0577 24X6.5 0.00481 0.0481 lox 0.00240 0.0481 20X7.0 0.00481 0.0481 lox 0.00240 0.0481 20X7.5 0.00481 0.0481 lox o.00240 0.0384 16X8.0 0.00481 0.02885 6X 0.00240 0.02885 12X8.5 0.00481 0.02885 6X 0.00240 0.02885 12X9.0 0.00481 0.02885 6X 0.00240 0.0192 8X

as sensitive to hydrogen sulphide as bismuth. Ferric ammon-ium citrate reacts better in an alkaline menstruum while the re-action has no apparent effect upon the sensitivity of bismuth.The data are also presented graphically in figure 1.

It is believed that the reaction of ferric ammonium citrate withhydrogen sulphide at various hydrogen-ion concentrations shouldform a straight line but due to the difficulty in determining theend point this was not obtained.The data presented here do not agree with the results reported

by Zobell and Feltham. They found that ferrous iron was moresensitive than bismuth salts. Due to several factors it is, how-ever, difficult to compare data. First, these investigators usedbismuth carbonate which is an insoluble form of bismuth, and


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

HYDROGEN SULPHIDE STUDIES 189

therefore the sensitivity to sulphides was markedly reduced.Second, the medium containing the indicator was titrated withsodium sulphide which is a basic compound and, due to hydroly-tic dissociation, would tend to make the medium more alkaline,in which case iron becomes more sensitive to the sulphide.The results presented here not only agree with the theoretical

consideration of these indicators, but definitely show the supe-MIl~sles

0.09L'

0. A OAdawism of Bhat & Io as IndicatorsIbr the Detectd of Hydrogen DAlPhidEatV\lariedrogm Ion ans

0.070 'ric Aionii Qitfate in DBffer Solutions

a- Aesmfins Citrate In Agar0.080-fmatb & Amaaim Cinrat. ia kuffer Solutions

-imth & Ascdin Mtrate In Apr0.080

0.040 _ _ _ _ __ _

o.020 __\__-

0.020

0.000.

6.0 6.5 7.0 7.5 8.0 8.5 9.0se

FIG. 1

riority of bismuth over iron in detecting hydrogen sulphide irres-pective of the H-ion concentration.

PREPARATION OF A MEDIUM INCORPORATING BISMUTH

The second phase of work was to produce a medium incorporat-ing a soluble bismuth compound as an indicator. This reallydeveloped into a problem of considerable magnitude. Solublebismuth compounds are very active and react with many sub-


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

190 CHARLES A. HUN'TER AND H. GILBERT CRECELIUS

stances producing insoluble compounds under certain conditions.Experiments have been performed with over 650 media, somebeing of no value and others showing definite promise.

In the preparation of a medium containing a soluble bismuthcompound one of the difficulties encountered was the formationof an insoluble precipitate. This difficulty was finally eliminatedwhen it was discovered that Bacto tryptone and 0.5 per cent agarwould not precipitate in the presence of bismuth, K2HPO4, glu-cose and sodium sulphite but that the use of any other peptonewould produce a precipitate (Feldman and Hunter, 1935).After using a medium made from these ingredients for severalmonths with good success the supply of Bacto tryptone, Bis-muth citrate and glucose was exhausted. In obtaining new lotsof these reagents difficulty with- a precipitate developed as wellas the reduction of the bismuth. In the past the reduction ofbismuth by the glucose had never been encountered. Numerousexperiments were performed to eliminate the precipitate in themedium and the reduction of bismuth. The reduction of bis-muth was prevented by substituting mannitol, a non-reducingcarbohydrate, but the elimination of the precipitate was moredifficult.During the experiments on the preparation of a medium which

would not precipitate with bismuth it was discovered that Bacto-peptone iron could be used. The medium was prepared by add-ing 12 grams of the dehydrated peptone iron agar (one-thirdusual quantity) to 1000 cc. of distilled water, boiling until dis-solved and then adding 10 cc. of 20 per cent Na2SO8 solution,25 cc. bismuth liquor and 5 grams of mannitol. This mediumcontained such a small quantity of ferric ammonium citrate thatit did not react with H2S.

Further experiments brought out the fact that not all lots ofBacto-peptone iron could be used, as some would produce aprecipitate. Many media were made, varying the ingredients,using different lots of tryptone, agar and other chemicals. Theresults of these experiments showed that the tryptone was thefactor causing the precipitate. Some lots of tryptone did con-tain or did not contain a small quantity of material that reacted


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

HYDROGEN SJLPHIDE STUDIES

with the bismuth causing the precipitate. Just what the factorsare that cause this trouble has not as yet been determined.When using the bismuth sulphite medium made with Bacto-

peptone iron the criticism has been raised that ferric ammoniumcitrate may be responsible for the blackening. It has been defi-nitely proven by numerous experiments that the blackening isdue to bismuth sulphide and not iron sulphide. With organismsthat produce only a small quantity of hydrogen sulphide a brown-ing develops rather than the deep brownish-black precipitate.Another experiment was performed to prove that the iron wasnot responsible for any blackening and also that the bismuth wasnot reduced. In this case the medium was prepared omittingthe bismuth liquor, inoculated with various organisms, incubatedfor 24 hours at 370C. and then a small quantity of bismuth liquorwas added. The results of this experiment check with thoseobtained when bismuth was present before inoculation.

In trying to overcome the development of a precipitate a seriesof experiments were performed varying the amount of each in-gredient. The outcome of these experiments revealed that nearlyas good results could be obtained by reducing the content ofthe bismuth to one-fifth (1/5) the original quantity. Thecolor change of the medium would be brownish with small quan-tities of hydrogen sulphide and a deep brownish-black whenlarger amounts of sulphide were formed. Accordingly, thefollowing medium was prepared using only 0.5 per cent of bismuthliquor with excellent results.

Medium ITryptone .................................... 7.0 gramsK2HPO4....................................................0.3 gramAgar ........... ......................... 5.0 gramsWater.................................... 1000 cc.

Dissolve and filter, then add:Na2SO3 (20 per cent) .................................... 10 cc.Bismuth liquor (3 per cent) ................................... 5 cc.Mannitol.................................... 5 gramsTube and sterilize at 15 pounds for 15 minutes.

With most lots of tryptone, medium made according to thisformula will not precipitate. It has, however, been noted that

191


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

192 CHARLE A. XiththI AND It. GILBERT CTEOPTUS

certain lots of tryptone will precipitate. If this is the case it isadvisable to use the formula for medium II. The Difco Lab-oratories have been able to reproduce medium I in a dehydratedform. This product has been thoroughly tested using severalhundred cultures and the results are comparable to those of theoriginal bismuth medium.

Medium IIThe formula is the same as that for medium I except that 5.0

per cent of skimmed milk is added. No precipitate will form withany of the various lots of tryptone that have been tried. Themedium is an opalescent color and the formation of hydrogensulphide can easily be detected as the contrast in color is quitedistinct.

Medium IIIThe medium is the same as No. II except that the agar content

is increased to 1.5 per cent. The medium is slanted and boththe slant and butt inoculated. The presence of H2S on the sur-face of the slant will be noted by a brownish coloration withmetallic sheen while in the butt of the tube the color will be adeeper brown or black. This medium has been used to prove thatthe brownish-black precipitate in media I and II was not reducedbismuth.In comparing the bismuth medium with peptone iron it was

found, as should be expected from the theoretical considerationof the indicators, that the bismuth medium would detect hydro-gen sulphide formation in a larger percentage of cultures thanwould peptone iron. Organisms which are normally classifiedas non-hydrogen-sulphide-formers gave a definite reaction in thebismuth medium. In one series of experiments in which 80stock cultures were inoculated into peptone iron and into bis-muth medium it was found that after 48 hours 19 of the cultures.or 23.7 per cent were positive in peptone iron and 46 or 57.5 percent were positive in bismuth sulphite medium.The comparison of results of some of the organisms in peptone

iron and bismuth sulphite medium can be seen in table 2. Figure


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


2 shows the appearance of the different media with a few or-ganisms. One thing to be noted is that Salmonella paratyphi,contrary to general opinion, does produce hydrogen sulphide ascan be demonstrated by the use of bismuth indicator. This isnot surprising when considering that bismuth is many times more

TABLE 2H2S reactions in various bismuth and iron media

ORGANISMS

Esch. coli......................Esch. communior "R".Esch. communior...............Esch. coli Soule "S" 1.........Esch. coli Soule "R" 2.........Aer. aerogenes.................E. typhosa 6...................E. typhosa 1...................E. typhosa 2...................E. typhosa 3...................E. typhosa 4...................S. paratyphi 90................S. paratyphi 89................S. paratyphi 88................S. paratyphi 87................S. paratyphi 10................S. schottmzilleri 21.............S. schottmulleri 20.............S. schottmfilleri 93.............S. schottmidleri 92.............S. schottmulleri 91.............Shig. paradysenteriae Flexner.Shig. paradysenteriae Flexner.Shig. dysenteriae...............

BISMUTH-SULPHITE MEDIUM

6-hour

++

+++

++++++

++

24-hour

++

+++

++++++

++

++++++

++

48-hour

++

++++

+++++

++++

PEPTONE IRON (DIFCO)

6- 24- 48huhour hour 48-hour

_ + +++- + +++- + ±+++- + +++_ - +

+ +++I++++

sensitive to hydrogen sulphide than are iron salts. The mediahave been criticized on account of being too sensitive and thedifferential value of hydrogen sulphide production is no longervalid in some cases. It is true that more organisms will probablyindicate hydrogen sulphide production than by other methodsbut this should not eliminate its routine use. What the bac-

193

++++I++++


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


teriologist desires to know is whether an organism does or doesnot produce hydrogen sulphide regardless of past knowledge ofthe organisms. The differential value of the test may be de-creased for the time being but future experience and investigationmay prove it to be of distinct value. Until such time arrives it issuggested that hereafter the presence or absence of hydrogen

I

A_

-Z

I._

'1t.

I

F..;4 *4&tll' -[ ia II M 1Wi -t*4

I.

1 1G. 2

sulphide be recorded according to medium or indicator used. Itis not sufficient to record that an organism does or does not pro-duce hydrogen sulphide.

SUMMARY AND CONCLUSION

1. Experiments to determine the sensitivity of ferric ammoniumcitrate and bismuth liquor for detecting hydrogen sulphide defi-nitely show the superiority of bismuth over iron irrespective of

B1,11111111- 111101,111, Modillill .-


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


the pH. The most sensitive range of iron is in an alkaline men-struum while with bismuth the H-ion concentration has no effect.

2. After experimenting with more than 650 media, incorporat-ing a soluble form of bismuth, 0.5 per cent bismuth liquor or 16mgm. bismuth and ammonium citrate, Merck, U. S. P. IX,several formulae were developed which give good results. Thevariable factor is the tryptone, not all lots being of uniform com-position. In case precipitation occurs with certain lots of tryp-tone the addition of 5 per cent of skim milk overcomes this diffi-culty.

3. The bismuth sulphite medium is far superior to any mediumincorporating lead or iron for the detection of hydrogen sulphidein cultures. The results show that organisms producing a smallquantity of hydrogen sulphide give negative results with ironwhile in the bismuth sulphite medium definite browning occurs.

4. Organisms which are usually stated in the literature to benegative hydrogen sulphide formers, such as Salmonella para-typhi and Shigella dysenteriae, give positive reactions with bis-muth.

5. It is recommended that the indicator and method be statedwhen recording whether an organism does or does not form hydro-gen sulphide. It is not sufficient to state that the organism is anegative or positive hydrogen sulphide producer.

REFERENCESBAILEY, S., AND LACEY, G. R. 1927 A modification of the Kligler lead acetate

medium. Jour. Bact., 13, 183-189.BECKWITH, T. D., AND MOSER, J. R. 1932 The reduction of sulphur containing

compounds in the wood pulp and paper manufacture. Jour. Bact.,24, 43-52.

BURNET, E., AND WISSENBACK, R. 1915 Valeur des renseignements fournis parla culture en g6lose a l'acetate de plomb, pour la differenciation desbacillus typhique, paratyphique A et paratyphique B. Comparaisonavec les r6sultats obtenus par l'agglutination, dans identification de517 echantillons de bacilles typhique et paratyphiques. Compt. rend.de Soc. de Biol., 78, 565-568.

FELDMAN, M., AND HUNTER, C. A. 1935 An improved method for the detectionof hydrogen sulphide produced by microorganisms. Proc. S. Dak.Acad. Science, 15, 41-45.

HUNTER, C. A., FELDMAN, M., AND CRECELIUS, H. G. 1937 Detection of hydro-gen sulphide in cultures. Jour. Bact., 33, 31-32.

195


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


JoRDAN, E. O., AND VICTORSON, R. 1917 Differentiation of the paratyphoid-enteritidis group. Jour. Inf. Dis., 21, 554-555.

KLIGLER, I. J. 1917 A simple medium for the differentiation of the membersof the typhoid-paratyphoid group. Amer. Jour. Pub. Health, 7,1042-1044.

LEVINE, M., VAUGHN, R., EPSTEIN, S. S., AND ANDERSON, D. Q. 1932 Somedifferential reactions in the colon-aerogenes group of bacteria. Proc.Soc. Exper. Biol. and Med., 29, 1022-1024.

MULSoW, F. W., AND PAINE, F. S. 1925 Hydrogen sulphide production by bac-teria. Iowa Acad. Science, 32, 63-70.

MEYERS, JOHN 1920 The production of hydrogen sulphide by bacteria. Jour.Bact., 5, 231-252.

PACKECO, G., AND MELLO, J. T. 1932 Sur un proc6d6 de determination del'hydrog6ne sulfure dans les cultures bact6riennes. Compt. rend. deSoc. de Biol., 110, 131-132.

SCHUNK, I. V. 1924 Methods for class demonstration of hydrogen sulfideformation by bacteria. J. Elisha Mitchell Sci. Soc., 40, 107.

TANNER, F. W. 1917 Studies on the bacterial metabolism of sulfur. I. Forma-tion of hydrogen sulfide from certain sulfur compounds under aerobicconditions. Jour. Bact., 2, 585-593.

THOMPSON, L. S. 1921 The group of hydrogen sulphide producing bacteria.Jour. Med. Res., 42, 383-389.

TILLEY, F. W. 1923 The relation between chemical composition of peptonesand hydrogen sulfide production. Jour. Bact., 8, 287-295.

WILMET, M. M. 1927 Sur la sensibility de quelques r6actifs de l'hydrog6nesulfure gazeux. Compt. rend. de l'Acad. des Sciences, 184, 287.

WILSON, W. J. 1923 Reduction of sulphites by certain bacteria in media con-taining a fermentable carbohydrate and metallic salts. Jour. Hyg.,21, 392-398.

ZOBELL, C. E., AND FELTHAM, C. B. 1934 A comparison of lead and iron asdetectors of hydrogen sulphide produced by bacteria. Jour. Bact.,28, 169-176.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

Documents

(FeCl3)(FeCl3) (1932), Schunk (FeSO4) (1924) and Levine and co-workers (ferric citrate) (1932). The addition of bismuth car-bonate to the culture medium was recommended by Pacheco