THE TAXONOMIC POSITION OF CORYNEBACTERIUM ACNES H. C. DOUGLAS AND SHIRLEY E. GUNTER Department of Bacteriology, University of Washington, Seattle 5, Washington Received for publication February 20, 1946 The organism known as the "acne bacillus" in medical literature was first observed by Unna (1896) in histological sections of acne comedones, but Sabouraud (1897) was the first to cultivate this organism successfully in pure culture from the contents of acne pustules. Subsequent studies on the etiology of acne vulgaris by Gilchrist (1900, 1903), Hall6 and Civatte (1907), Hartwell and Streeter (1909), Fleming (1909), Suldmerson and Thompson (1909), and Molesworth (1910) indicated that the acne bacilli were morphologically similar to the corynebacteria but differed markedly from these organisms in showing a strong preference for anaerobic conditions. Gilchrist (1900) named the organism Bacillus acnes, whereas Bergey et al. (1923) placed it in the genus Corynebacterium because of its morphological relationship to the members of this group. Our interest in these organisms became aroused following the observation of Weiser and Gunter (1942) of this laboratory that samples of human blood plasma destined for a civilian plasma bank contained anaerobic diphtheroids as one of the principal contaminants. These workers were able to show that such organisms occurred on normal skin and probably gained entry into the blood plasma as the result of ineffective skin disinfection preceding venipuncture. The anaerobic diphtheroids isolated from plasma and skin seemed to fit the general description of Corynebacterium acnes (Bergey et al., 1939), but, as the inclusion of such organisms in the genus Corynebacterium seemed questionable to us, we considered it worth while to make a comparative study of a series of isolates and the available authentic cultures of C. acnes in an attempt to clarify their taxonomic position. It also seemed desirable to obtain quantitative data on the occurrence of these organisms on normal human skin. CULTURES EMPLOYED Cultures of C. acnes, numbers 6919, 6920, 6921, and 6923, were obtained from the American Type Culture Collection. These cultures were originally obtained from the National Collection of Type Cultures, England, and had been isolated from clinical cases of acne vulgaris. In addition, 27 cultures of anaerobic diphtheroids were isolated from the skin of ten normal subjects, 4 cultures from different samples of human blood plasma, and 1 culture from a clinical case of acne vulgaris. CULTURE MEDIUM The medium recommended by the U. S. Office of Civilian Defense for sterility testing of blood plasma supports a fair growth of all strains, but because of acid production the pH soon becomes limiting in this poorly buffered medium and 15 on May 15, 2020 by guest http://jb.asm.org/ Downloaded from
Transcript
THE TAXONOMIC POSITION OF CORYNEBACTERIUM ACNES
H. C. DOUGLAS AND SHIRLEY E. GUNTERDepartment of Bacteriology, University of Washington, Seattle 5, Washington
Received for publication February 20, 1946
The organism known as the "acne bacillus" in medical literature was firstobserved by Unna (1896) in histological sections of acne comedones, butSabouraud (1897) was the first to cultivate this organism successfully in pureculture from the contents of acne pustules. Subsequent studies on the etiologyof acne vulgaris by Gilchrist (1900, 1903), Hall6 and Civatte (1907), Hartwelland Streeter (1909), Fleming (1909), Suldmerson and Thompson (1909), andMolesworth (1910) indicated that the acne bacilli were morphologically similarto the corynebacteria but differed markedly from these organisms in showing astrong preference for anaerobic conditions. Gilchrist (1900) named the organismBacillus acnes, whereas Bergey et al. (1923) placed it in the genus Corynebacteriumbecause of its morphological relationship to the members of this group.Our interest in these organisms became aroused following the observation of
Weiser and Gunter (1942) of this laboratory that samples of human blood plasmadestined for a civilian plasma bank contained anaerobic diphtheroids as oneof the principal contaminants. These workers were able to show that suchorganisms occurred on normal skin and probably gained entry into the bloodplasma as the result of ineffective skin disinfection preceding venipuncture.The anaerobic diphtheroids isolated from plasma and skin seemed to fit the
general description of Corynebacterium acnes (Bergey et al., 1939), but, as theinclusion of such organisms in the genus Corynebacterium seemed questionableto us, we considered it worth while to make a comparative study of a series ofisolates and the available authentic cultures of C. acnes in an attempt to clarifytheir taxonomic position. It also seemed desirable to obtain quantitative dataon the occurrence of these organisms on normal human skin.
CULTURES EMPLOYED
Cultures of C. acnes, numbers 6919, 6920, 6921, and 6923, were obtained fromthe American Type Culture Collection. These cultures were originally obtainedfrom the National Collection of Type Cultures, England, and had been isolatedfrom clinical cases of acne vulgaris. In addition, 27 cultures of anaerobicdiphtheroids were isolated from the skin of ten normal subjects, 4 cultures fromdifferent samples of human blood plasma, and 1 culture from a clinical case ofacne vulgaris.
CULTURE MEDIUM
The medium recommended by the U. S. Office of Civilian Defense for sterilitytesting of blood plasma supports a fair growth of all strains, but because of acidproduction the pH soon becomes limiting in this poorly buffered medium and
growth ceases. A highly buffered modification of this medium supports a lux-urious growth of all strains and is recommended for isolation and routine culturework. The modified medium has the following composition: bacto peptone,2 per cent; bacto yeast extract, 0.5 per cent; glucose, 1.0 per cent; KH2PO4,2 per cent; B. B. L. sodium thioglycolate, 0.10 per cent; and agar, 0.05 per cent,or 1.5 per cent, depending upon whether a fluid or solid medium is desired. ThepH should be adjusted to about 7.1 before sterilization and will be about 6.8 aftersterilization. Considerable darkening of the medium takes place during auto-claving, but this has no apparent adverse effect upon the growth of the organisms.
ISOLATION PROCEDURE AND OCCURRENCE ON NORMAL SKIN
The isolation of cultures from normal skin and the estimation of the relativenumbers of these organisms on skin were accomplished as follows: An area ofabout 2 square inches of skin on the upper arm was scraped with a sterile Bard-Parker blade moistened in sterile saline, and the skin scrapings were transferredt.o a sterile mortar containing 5 ml of sterile saline and some fine sand. After
TABLE 1The occurrence of anaerobic diphtheroid8 on normal skin
SUBJECT NUMBER OF COLONIES EXAMINED % ANAEROBIC DIPHTROMDS
E. F. 45 40J.F. 28 14V. E. 43 9E. J. 33 33R. W. 48 50G. D. 27 78P. M. 48 10
thorough maceration, dilutions of the resulting suspensions were plated, using themedium described above, and incubated under hydrogen in a McIntosh-Fildesjar for 4 days at 37 C. Plates which showed well-separated colonies were counted,and all of the colonies from the counted plates were streaked and stabbed ondeep butt agar slants of the same composition as the plating medium. After4 days' incubation at 37 C, all of the resulting transfers were examined by gramstaining and inspected for growth in the stab and on the slant. Anaerobicdiphtheroids were readily recognized by their morphology and by the fact thatgrowth occurred only in the stab and never on the slant. The flora of the skinof 7 individuals was examined in this way, and the occurrence of anaerobic diph-theroids, expressed as a percentage of the total skin flora capable of developingunder ahaerobic conditions, is tabulated in table 1. As anaerobic diphtheroidsaccounted for from 9 to 78 per cent of the total skin flora capable of growth underthese conditions, it is obvious that organisms of this type constitute a significantpart of the normal skin flora. The remainder of the flora of the subjects ex-amined by us consisted almost entirely of nonpigmented facultatively anaerobicmicrococci.
The occurrence of anaerobic diphtheroids on normal individuals was first notedby Lovejoy and Hastings (1911), who found these organisms to occur in largentumbers and in almost pure cultures in the sebum expressed from the sebaceousglands around the folds of the nose. We have mnade no isolations from thissouirce, but the examination of gram stains of such material has served to verifythe observations of those wAorkers. In all probability the normal habitat of theseorganisms is the sebaceous glands of the skin, and in areas such as those aroundthe nose where the glands are large and very active in secreting sebaceous materialthe organisms occur in great abundance.
DESCRIPTION OF STRAINS
Relation to temperature and oxygen. The optimum temperature for growth isapproximately 37 C, and at this temperature the maximum crop is reached in
a b
FIG. 1. STRAIN 22. NEGATIVE STAINS OF FoUR-DAY-OLI) SLANT CULTURE'S INCUBATED(a) ANAEROBICALLY AND (b) AEROBICALLY. X 970
liquid media in 3 to 4 days. No growth occurs at 45 C, and growth at room
temperature is extremely slow.All strains show a marked pr-eference for anaerobic conditions, although growth
consisting of a few isolated colonies can usuially be obtainedi on slant cuiltuiresexposed to the air if a heavy inoculum is used. Six strains, howNever, have con-
sistently failed to grow aerobically, but on the other hand one strain has beenfound to grow almost as well aerobically as anaerobically. Apparently a widerange of oxygen tolerance occurs wiithin this group, but there appears to be no
correlation between the relation to oxygen and other characteristics.Morphology. All strains, wAhen examined from fluid or anaerobic slant cultures,
present a similar mnorphological picture, the organisms ranging from small,plump rods to ellipsoids which tend to occur in pairs with the cells joined at a
slight angle (figuire la). The size of the individual cells in gram stains from suchcuiltures is approximately 0.4 to 0.5 by 0.8 to 0.9 microns.
Cells from aerobic slant cultures appear to be somewhat longer and more
pleomorphic than those in anaerobic cuiltures, buit the, differences in morphology
under these two conditions are, in most cases, slight. AIi occasional strain,however, will show marked (lifferences in morphology between aerobic and an-aerobic cultuires, cells from aerobic cultures being much longer and swNollen orclub-shaped, an(l sometimes show-ing what seems to be rudimentary branching(figure lb). Such morphology is somewN,hat similar to that found normally in thecorynebacteria andl has previously been reporte(I for C. acnes by Gilchrist (1903)and Siidmerson and Thoiimpson (1909). 'T'his effect of oxyen on the morphologyis very much like that (lesciribed by vTan Niel (1928) for the propionic acid bacteriawhich show marked pleomorphism in aerobic cultures.
All strains are strongly gram-positive. Endospores are not formed, andmotility is absent.
Appearance of growtIl, on solid and liqutid media. Surface colonies on streakedplates after 4 to 5 days of anaerobic incubation are circular, raised, smooth,glistening, and have an entire edge. The size varies from 1.5 to 4.0 mm indiameter. The color of the colonies is quite characteristic, for after 4 to 5 daysof incubation there is a faint but definite tinge of pink, which intensifies uponfurther incubation to a pale salmon pink.
Liquid cultures containing 0.05 per cent agar tend to (levelop at first as small,discrete colonies, especially if started from small inocula, but this phenomenontends to become obscured upon further growvth of the ctulture. The sedimentedlcells of such cultures appear cream-colored at first, but, like colonies on agarplates, become a pale salmon pink upon further incubation.
Biochenical ReactionsUtilization of carbon compoiunds. As the examination of glucose broth cultures
revealed that large amounts of volatile acid are produced, acid production wasused as a criterion of the utilization of various carbon compounds. Volatile acidproduction, which in a highly buffered medium containing 1.0 per cent glucosefrequently equtaled 10 ml of M/i0 aci(l per 10 ml of medium, was determined bydirect titration of the cultures, by titration of steam distillates, or by notation ofthe color change of an indicator incorporate(d into the mediium. As the resultsobtained Nvere fouind to be (lualitatively the same regardless of the method used,the latter procedutre was chosen, except in the case of lactate, because of itssimplicity. The utilization of lactate was determined by vistual comparison ofthe growth in the basal mediuim alone and in the basal mediuim containing lactate,and in several cases by volatile acid determinations.The basal medium usedl for (letermining acid prodluction by the indicator
method containe(d 2.0 per cent peptone, 0.5 per cent yeast extract, 0.5 per cent ofthe carbon compound, 0.1 per cent sodium thioglycolate, 0.05 per cent agar, andAndrade's indicator. The medium was sterilized in 10-ml amounts in tubescontaining small inverted vials for gas traps. Incubation was in air at 37 C for7 days. The results are summarized in table 2. It can be seen that several com-pounds are fermented with the production of acid but no visible gas, and althoughthe organisms can be separated into 4 groups on the basis of the carbon com-pounds fermented, wi-e have been unable to correlate these fermentation groups
with the sources of the cultures, or other characteristics, and therefore attach notaxonomic significance to them.
Liquefaction of gelatin. Gelatin is liquefied in 10 days at 37 C by all strains inpeptone yeast-extract phosphate glucose gelatin "deeps."
Litmus milk. Standard litmus milk was exhausted by steaming, cooled,inoculated, and layered with sterile vaspar. All strains produced a rennet curdafter about 2 weeks' incubation at 37 C, followed by a slow peptonization of thecasein.
Reduction of nitrates. All strains reduce nitrates to nitrites in peptone yeast-extract phosphate glucose thioglycolate broth containing 0.2 per cent KNO3.
Production of indole. About one-half of the strains produce indole in tryptoseyeast-extract phosphate glucose thioglycolate broth.
Catalase reaction. Strongly positive for all strains.Hemolysis. All strains produce beta-hemolysis on peptone yeast-extract
glucose phosphate agar containing 5.0 per cent by volume of citrated humanblood.
TABLE 2The utilization of carbon compounds
GROUP NUMBER OF GLUCOSE MANNOSE GALACTOSE FRUCTOSE GLYCEROL MANNTOL MALTOSECULTURES
None of the cultures utilized lactate, arabinose, sucrose, lactose, raffinose, salicin, inulin,or starch.+ = acid production but no visible gas.
Although the 37 strains show certain differences in their cultural characteris-tics, we can see no logical basis for their separation into more than one groupand consequently consider them closely related, if not identical. It thus seemsjustifiable to conclude that the strains isolated from the skin, blood plasma,and cases of acne vulgaris belong to a single species that is now known as Coryne-bacterium acnes. The question of the inclusion of this organism in the genusCorynebacterium, however, will be discussed after a consideration of the natureof its catabolic process.
NATURE OF THE FERMENTATION OF GLUCOSE
Nothing has been reported concerning the nature of the fermentation producedby Corynebacterium acnes beyond the routine observation by several workersthat acid but not gas is produced from certain carbohydrates. It is evident,however, from the fact that only a negligible amount of growth takes place inthe absence of sugars that sugar fermentation represents the principal mode ofenergy liberation for this organism. The significance of the nature of the fermen-
tation in the classification of facultative and anaerobic bacteria has been empha-sized by Kluyver and van Niel (1936) and Barker and Haas (1944), althoughthis important point is frequently not appreciated by bacteriologists.The general morphological picture, the relation to oxygen, and the fact that
volatile acids are produced in the fermentation of glucose suggested a closerelationship between Corynebacterium acnes and the propionic acid bacteria.To identify the volatile acids produced by C. acnes analyses were made by meansof the Werkman (1931) partition method using the systems ethyl ether and waterand isopropyl ether and water. The results indicated the presence of propionicand acetic acids in the ratio of 1.7 to 2.2 parts of propionic to 1 part of aceticacid. The presence of propionate in concentrated solutions of the sodium saltsof the volatile acids was verified microchemically by the characteristic appear-ance of the mercurous salt (Klein and Wenzl, 1932), but attempts to demonstrateacetate by means of the mercurous salt or by the formation of sodium uranyl
TABLE 3The fermentation of gluco8e by Corynebacterium acnes, strain A.T.C.C. 6950
AMM-ATOMS ILXI-ECQUIVALENTSmm LU-ATONS OF AVAILALE
* Calculated according to the method of Barker (1936).t (-) = product used; (+) = produced produced.
acetate (Klein, 1932) failed because of interference by the propionate. In orderto demonstrate conclusively the presence of acetate, the acetic and propionicacids were separated by the azeotropic distillation procedure of Schicktanzet al. (1940). After this had been accomplished, acetate was easily detectedby either of the microchemical reactions just mentioned.That fixed acids are not formed in appreciable amounts in the fermentation
of glucose was demonstrated by the fact that practically all of the acid producedcould be accounted for in the acids volatile in steam. In some cases a smallamount of nonvolatile acid is apparently produced, but because of the amountsinvolved its identity has not been determined.
Neutral volatile products were shown to be absent by applying the usualmicrochemical tests to distillates obtained from neutralized cultures.Although no visible gas is formed in ordinary cultures, CO2 formation could
be demonstrated either by culturing in Eldredge tubes, or by collecting the gasesover mercury from 500 ml of culture contained in an all-glass fermentation vessel.
The gas produced in the latter procedure was shown to be pure C02 by the factthat it was completely absorbed by 5 per cent KOH.Although the qualitative analyses indicated that Corynebacterium acnes
produces a propionic acid fermentation of glucose, it was desirable to verify thisconclusion by means of quantitative experiments. This was accomplished byanalyses of cultures grown in large Eldredge tubes containing 200 ml of 0.18 NBa(OH)2 in one compartment for the absorption of C02, and 250 ml of culturemedium containing 1.0 per cent glucose in the second compartment. Becauseof the interference of sodium thioglycolate in the determination of reducingsugars, it was necessary to omit this substance from the medium, otherwise themedium was as previously described. Incubation was at 37 C under an atmos-phere of oxygen-free nitrogen for a period of 10 days, after which time thefermented medium and an uninoculated control were analyzed. The resultsof such an experiment, using C. acnes A.T.C.C. no. 6920, are presented in table 3and demonstrate clearly that this organism produces a propionic acid fermenta-tion, as practically all of the carbon and available hydrogen in the glucosefermented can be accounted for in carbon dioxide, propionic acid, and aceticacid. Less extensive work with the remainder of the cultures indicates that allstrains produce the same type of fermentation of glucose.
DISCUSSION
C. acnes and a second anaerobic diphtheroid, C. lymphophilum, were placedin the genus Corynebacterium by Bergey et al. (1923) purely on the basis ofmorphological considerations. C. lymphophilum was isolated by Torrey (1916)from abnormal lymph glands, and although the original cultures are no longeravailable for study, it seems highly probable from Torrey's description of thisorganism that it is identical with C. acne8.On a morphological basis C. acneg might be placed in either the genus Coryne-
bacterium or the genus Propionibacterium, for these two groups have certainmorphological features in common. Two facts, however, indicate that C. acnesis much closer to the propionic acid bacteria than to the corynebacteria. First,the nature of the catabolic process clearly relates C. acnes to the propionic acidbacteria. It is true that Tasman and Brandwijk (1938) found propionic acidto occur as a product of sugar dissimilation by C. diphtheriae. In the Tomcsikstrain studied by these workers propionic acid accounted for 30.3 to 49.9 percent of the glucose fermented, and in the Bandoeng strain, 5.9 to 6.6 per cent.The fermentations studied by Tasman and Brandwijk were conducted underhighly aerobic conditions, however, and may not be representative of the cata-bolic process of these organisms under strictly anaerobic conditions. Fujitaand Kodama (1934) and Friedemann (1940), on the other hand, did not reportthe presence of propionic acid in the anaerobic fermentation of glucose by C.diphtheriae, but indicated that the principal dissimilation products producedwere lactic, acetic, formic, and succinic acids, and ethanol. Although it issomewhat difficult to reconcile the conflicting results of these workers, it seems
reasonable to conclude that the typical mode of sugar fermentation by C.
diphtheriae is not a propionic acid fermentation, although propionic acid may beformed under certain conditions or by certain strains.
Secondly, the effect of oxygen in either completely or strongly inhibitinggrowth of C. acnee markedly distinguishes this organism from the typicallyaerobic corynebacteria and relates it to the propionic acid bacteria, which alsoshow a decided preference for anaerobic conditions.
In view of the foregoing considerations we feel that C. acnes shows far strongeraffinities to the genus Propionibacterium than to the genus Corynebacterium,and suggest, therefore, that this organism be transferred to the former genus asPropionibacterium acnes (Gilchrist) Douglas and Gunter comb. nov. Such atransfer would require a modification of the diagnosis of the genus Propioni-bacterium (Bergey et al., 1939) to exclude the phrase "ferments lactic acid."The strains of P. acnes studied here seem to form a homogeneous group that
can be differentiated from other species of Propionibacterium on the followingbases:
(1) Optimum temperature. Previously described species of Propionibacteriumhave an optimum temperature of 30 C, whereas P. acnes has an optimum of 37 C.
(2) Gelatin liquefaction, action on milk, and nitrate reduction. None of thepreviously described species of Propionibacterium liquefy gelatin or digest milk,and only one species, P. pentosaceum, reduces nitrates.
(3) Lactate fermentation. All previously described species of Propionibacteriumferment lactate, but P. acnes does not ferment this substance.
(4) Habitat. To the best of our knowledge, P. acnes has been isolated onlyfrom man, whereas practically all other species of Propionibacterium have beenisolated from dairy prodlucts.
SUMMARY
The organism previously known ag Corynebacterium acnes is transferred to thegenus Propionibacterium, as P. acnes, on the basis of its relationship to oxygenand the nature of its catabolic process.
P. acnes occurs on normal human skin, where it constitutes a significant partof the skin flora.
ACKNOWLEDGMENTS
The authors wish to acknowledge the assistance of Mr. Edward L. Foubert,Jr., in some of the experiments, and the assistance of Dr. Waldemar F. Kirch-heimer, King County Tuberculosis Hospital, in furnishing clinical material.
REFERENCESBARKER, H. A. 1936 On the fermentation of some dibasic C4-acids by Aerobacter aero-
genes. Proc. Koninkl. Akad. Wetensechappen Amsterdam, 39, 3-12.BARKER, H. A., AND HAAs, VICTORIA. 1944 Butyribacterium, a new genus of gram-posi-
tive, non-sporulating anaerobic bacteria of intestinal origin. J. Bact., 47, 301-305.BERGEY, D. H., et at. 1923 Bergey's manual of determinative bacteriology. 1st ed.
BERGEY, D. H., et al. 1939 Bergey's manual of determinative bacteriology. 5th ed.Williams and Wilkins, Baltimore.
FLEMING, ALEXANDER 1909 On the etiology of acne vulgaris and its treatment by vac-cines. Lancet, I, 1035-1038.
FRIEDEMANN, T. E. 1940 Carbohydrate metabolism of Oidium lactis and Bacillus subtilisin complex carbohydrate-rich culture medium. Proc. Soc. Exptl. Biol. Med., 43,148-151.
FuJITA, A., AND KODAMA, T. 1934 Studies on the energy producing functions of patho-genic bacteria. Kitasato Arch. Exptl. Med., 11, 87-131.
GILCHRIST, T. C. 1900 A bacteriological and microscopical study of over 300 vesicularand pustular lesions of the skin, with a research upon the etiology of acne vulgaris.Johns Hopkins Hosp. Rept., 9, 409-430.
GILCHRIST, T. C. 1903 The etiology of acne vulgaris. J. Cutaneous Diseases, 21, 107-120.HALL*, J., AND CIVATrE, A. 1907 Contribution A la bacteriologie des glandes s6bacees.
Ann. dermatol. syphilig., 4 Serie, 8, 184-188.HARTWELL, H. F., AND STREETER, EDWARD C. 1909 Bacillus of acne-B. acnes. Boston
Med. Surg. J., 161, 882.KLEIN, G. 1932 Handbuch der Pflanzenanalyse. Zweiter Band, Spezielle Analyse.
Julius Springer, Vienna.KLEIN, G., AND WENZL, H. 1932 Der mikrochemische Nachweis flilchtiger Fettsiiuren
in der Pflanze. II. Mikrochemie, 11, 73-130.KLUYVER, A. J., AND VAN NIEL, C. B. 1936 Prospects for a natural system of classification
of bacteria. Zentr. Bakt. Parasitenk., II, 94, 369-403.LovEjoy, E. D., AND HASTINGS, J. W. 1911 Isolation and growth of the acne bacillus.
J. Cutaneous Diseases, 29, 80-82.MOLESWORTH, E. H. 1910 The cultural characteristics of the microbacillus of acne.
Brit. Med. J., 1, 1227-1229.SABOURAUD, R. 1897 La seborrhee grasse et la pelade. Ann. inst. Pasteur, 11, 134-159.SCHICKTANZ, S. T., STEELE, W. I., AND BLAISDELL, A. C. 1940 Analyses of mixtures of
StrDMERSON, H. J., AND THOMPSON, E. T. 1909 The cultivation and biological charactersof Bacillus acnes. J. Path. Bact., 14, 224-229.
TASMAN, A., AND BRANDWIJK, A. C. 1938 Experiments on the metabolism of the diph-theria bacillus. II. J. Infectious Diseases, 63, 10-20.
TORREY, JOHN C. 1916 Bacteria associated with certain types of abnormal lymph glands.J. Med. Research, 34, 65-80.
UNNA, P. G. 1896 Histopathology of the diseases of the skin. From MOLESWORTH, E. H.,Brit. Med. J., 1, 1227, 1910.^
U. S. Office Qf Civilian Defense. A technical manual on citrated human blood plasma, de-tailing its procurement, processing and use. Washington, D. C.
VAN NIEL, C. B. 1928 The propionic acid bacteria. Thesis. Delft.WEISER, R. S., AND GUNTER, SHIRLEY E. 1942 Unpublished observations.WERKMAN, C. H. 1931 Determination of organic acids. IV. A method for the provisional
identification and quantitative determination of two fatty acids in a mixture. IowaState Coll. J. Sci., 5, 121-125.
