A -. = Issued September 17, 1912. ?= U. S. DEPARTMENT. OF AGRICULTURE, ==- BUREAU OF ANIMAL INDUSTRY. BULLETIN 154. 8 ~ : 3 . A. D. MELVIN, CHIEF OP BUWAU. "> 8 = *- //'. - ".^.^-^^ 3 E E 2 ~ METHODS OF CLASSIFYING THE LACTIC-ACID BACTERIA. BY LORE A. ROGERS, Bacteriologist, Dairy Division , AND BROOKE ]. DAVIS, Assistant, Dairy Division. WASHINGTON: GOVERNMENT PRINTING OFFICE. 1912. Issued September 17, 1912. U. S. DEPARTMENT OF AGRICULTURE, BUREAU OF ANIMAL INDUSTRY. BULLETIN 154 A. D. MELVIN, CHIEF OF BUREAU. METHODS OF CLASSIFYING THE LACTIC-ACID BACTERIA. BY LORE A. ROGERS, Bacteriologist^ Dairy Division^ AND BROOKE J. DAVIS, Assistant, Dairy Division. WASHINGTON: GOVERNMENT PRINTING OFFICE. 1912. Chief: A. D. MELVIN. Assistant Chief: A. M. FARRINGTOX. Chief Clerk: CHARLES C. CARROLL. Animal Husbandry Division: GEORGE M. ROMMEL, chief. Biochemic Division: M. DORSET, chief. Dairy Division: B. H. BAWL, chief. Field Inspection Division: R. A. RAMSAY, chief. Meat Inspection Division: RICE P. STEDDOM, chief. Pathological Division: JOHN R. MOHLER, chief. Quarantine Division: RICHARD W. HICKMAN, chief. Zoological Division: B. H. RANSOM, chief. Experiment Station: E. C. SCHROEDER, superintendent. Editor: JAMES M. PICKENS. DAIRY DIVISION. B. H. RAWL, Chief. HELMER RABILD, in charge of Dairy Farming Investigations. S. C. THOMPSON, in charge of Dairy Manufacturing Investigations. L. A. ROGERS, in charge of Research Laboratories. ERNEST KELLY, in charge of Market Milk Investigations. ROBERT McADAM, in charge of Renovated Butter Inspection. 2 - LETTER OF TRANSMITTAL. U. S. DEPARTMENT OF AGRICULTURE, BUREAU OF ANIMAL INDUSTRY, Washington, D. 0., March 22, 1912. SIR: I have the honor to transmit for publication as a bulletin of this bureau the accompanying manuscript entitled "Methods of Classifying the Lactic-Acid Bacteria," by Messrs. Lore A. Kogers and Brooke J. Davis, of the Dairy Division of this bureau. There has hitherto been felt a need by dairy bacteriologists and others of a classification of the lactic-acid bacteria into naturally related groups by means of characters that can be determined with reasonable accuracy and in a manner ordinarily available. This paper describes the study of about 150 cultures isolated from milk, butter, and cheese, derived from various parts of the country, with the object of laying the basis for a satisfactory classification. Respectfully, A. D. MELVIN, Chief of Bureau. Hon. JAMES WILSON, Secretary of Agriculture. 3 CONTENTS. Page. Introduction 7 The significant characters of the lactic-acid bacteria 10 Morphology 13 Growth on solid media 14 Growth in milk 15 Growth in broth 17 Reduction of nitrates 17 Reduction of neutral red 18 Liquefaction of gelatin 18 Fermentation of carbohydrates 19 Conclusions 28 References to literature ... 29 ILLUSTRATIONS. Page. FIG. 1. Rate of acid formation in milk at 30 C. by cultures freshly isolated from milk 14 2. Rate of acid formation in milk at 30 0. after 26 generations (2 years) on lactose-agar 15 3. Frequency curve for gelatin liquefaction 19 4. Frequency curves for acid formation by the liquefying cultures 25 5. Frequency curves for acid formation by the nonliquefying cultures. . . 26 6. Grouping of cultures and distribution of characters in groups 27 5 METHODS OF CLASSIFYING THE LACTIC- ACID BACTERIA. INTRODUCTION. / The grouping of bacteria according to their action on any one specific substance usually brings together bacteria related in that one characteristic only but entirely unrelated in other respects. It is, however, sometimes convenient from a technical standpoint to group bacteria in this way. The bacteria concerned in the souring of milk have been so grouped for so long that many people have come to consider them as a division by themselves and their relation to other bacteria has been little considered. The bacteria taking part in the souring of milk may be readily divided into four general groups. Group I includes those bacteria which sour milk without peptoniza- tion or gas formation ; they grow poorly on artificial media and fail to liquefy gelatin. Morphologically they show some variation, usually appearing as a coccus or very short bacillus in pairs or in chains of varying lengths. The bacteria of this group are the ones ordinarily designated as the lactic-acid bacteria and have been described under various names. They have a very general distribution and their presence in milk is so constant that they may be considered as normal inhabitants of this medium. Group II includes the bacteria forming an acid curd with evolution of gas. This embraces varieties of BaciUus coli and Bacterium aerogenes or the Bacillus acidi lactici of Heuppe. The members of this group are readily distinguished from those of Group I by their abundant growth on artificial media, the vigorous evolution of gas, and the marked difference in their morphology. An examination of milk usually reveals their presence in small numbers, but their number is increased by the influence of high temperatures or insanitary con- ditions under which the milk has been collected or held. Group III includes those bacteria forming an acid curd which is subsequently partially peptonized. The bacteria of this group have been little studied in their relation to milk. It will be shown that this description applies to varieties only distantly related to our Group I as well as to some closely connected with this type. Group IV includes the high-acid-forming bacteria of which the Bacillus bulgaricus is the type. This organism is distinguishable from 7 8 CLASSIFYING LACTIC-ACID BACTERIA. those of the preceding groups by its slender rodlike form, its charac- teristic colonies on agar, its inability to grow in ordinary artificial media > and its growth in the presence of free acid. Bacillus bulgaricus has been studied in its relation to the fermented milks extensively used in Turkey and the neighboring countries. It has recently been shown *, 2 , 3 , that it is very widely distributed and may be isolated from almost any sample of mixed milk. Its growth at normal temperatures is so slow that it is improbable that it is a factor in the ordinary souring of milk. It is obvious that these groups are connected only by their ability to ferment lactose to acid and the consequent precipitation of th<> casein. Groups I, II, and IV are evidently related to each other in no other way. The identity of the bacteria of Group III in so far as they occur as milk bacteria has not been established. The need of work on methods of classification rather than on descriptions of new varieties or rearrangement of old names and descriptions is ex- emplified by the confused nomenclature of the bacteria included under Group I. We find in the literature such names as Bacterium lactis acidi, Bacillus lactis acidi, B. acidi lactici, B. guntheri, and Streptococcus lacticus, all of which, so far as can be determined by published descriptions, may be included in Group I. These names are based on variations in morphology, differences in growth on artificial media, on the rate of acid formation in milk, and various other characteristics of doubtful significance and uncertain stability. In the light of recent investigations these names and their accompanying descriptions have little more than a historic interest. They represent the attempt to establish types by the study of an isolated individual organism with little regard to its relation to other similar individual organisms. Approached from the standpoint of the dairy bacteriologists, the lactic-acid bacteria have been considered as a sharply defined group peculiar to milk. The students of pathogenic bacteria have boon inclined to look on them as a variety of some of the pus-forming streptococci. The opinion that the term lactic-acid bacteria covers a group of species or varieties and that the names and descriptions already pub- lished do not represent the true grouping is reflected in the frequent attempts to establish means of separating the group into stable species or varieties. McDonnell 4 attempted to do this, basing his descriptions largely on the effect on milk. A somewhat similar course was followed by Weigmann. 5 Muller 8 found a correlation in the solution of red blood corpuscles and the agglutination of immune sera by the streptococci of sour milk and certain pathogenic streptococci, a The reference figures relate to the list of references to literature at. the end of this bulletin. INTRODUCTION. 9 and considered that this indicated a relationship. Lohnis 7 has made a classification of the lactic-acid bacteria based largely on gas forma- tion, coagulation of milk, formation of slime, and liquefaction of gelatin. Conn, Esten, and Stocking 8 have used in their descriptions the action on milk and the usual culture characteristics. Heinemann 9 as a result of his studies on the bacteria of sour milk, excludes the name of Bacillus acidi lactici, and concludes that aside from the part taken by B. aerogenes and possibly by B. coli the spontaneous souring of milk is brought about by the Streptococcus lacticus of Kruse, an organism identical with the common streptococci of sewage and pathological conditions, lie bases this conclusion on the similarity in morphological, cultural, and pathogenic properties. All of this work and much more of a similar nature adds little to the systematic arrangement of the lactic-acid bacteria. It is generally ad- mitted that it is difficult to identify cultures of any but' the best known and most carefully studied bacteria by the published descriptions. Cultures which seem identical when written descriptions are compared are found to be distinct when they are grown side by side under uniform conditions. This error in identification is due partly to the difficulty of conveying the appearance of an object by words, but in a larger degree to the instability and unessential nature of some of the characters employed in separating one bacterial species or variety from another. The inadequacy of the ordinary methods is partic- ularly felt when one attempts a description of the lactic-acid bac- teria. The cells are small and the morphological differences are uncertain and inconstant. The growth on solid media is scanty and devoid of distinguishing characteristics. While many of the phvsio- logical tests which are found of great value in some groups fail when applied to the lactic-acid bacteria, others, notably the acid fermenta- tions of sugars, offer a possible means of differentiation for members of this group. Variations in the ability to ferment sugars have been observed, but the use of these variations in classifying or identifying cultures has been limited for two reasons. There has been a belief that the fermentative power was not constant; that this property could be lost or acquired so readily that it could not be used to differentiate one culture from another with any certainty. The more common objection, however, is that the separation on the basis of sugar fer- mentation divides the lactic-acid bacteria and others possessing the same general characteristics, not into natural groups, but into innumerable varieties. The use of tests of this kind in the usual way by which the fermen- tation of, or failure to ferment, a certain substance sets the culture so reacting apart from all other cultures gives an endless dichotomy 51194 -12 2 10 CLASSIFYING LACTIC-ACID BACTERIA. limited only by the number of test substances. Consequently, the ordinary use of sugars has increased rather than diminished the con- fusion now existing in the classification of the zymogenic bacteria. It is obvious that what is required in systematic bacteriology is not descriptions of new species or a rearrangement of names, but the establishment of means of classification applicable technically and correct biologically. No one basis of classification can be used for all groups of bacteria, but certain fundamental principles should govern any method of arrangement. Two of the most obvious principles on which the selection of characters for classification should be based may be stated as follows: The characters should be constant; they should be so selected that they show real biological relationships. In other words, bacteria should be arranged by means of characters that can be determined with reasonable accuracy and by means ordi- narily available into groups whose members are related naturally rather than by artificial bonds, and these characters should bo s<> constant and so distinctive that identical organisms can always be placed in the same group. This paper records an attempt to determine which of the charac- ters exhibited by the lactic-acid bacteria fulfill these conditions. No attempt has been made to classify or name any members of this group or to fix its place in the general bacteriological system. THE SIGNIFICANT CHARACTERS OF THE LACTIC-ACID BACTERIA. The morphology, staining reactions, cell grouping, cultural charac- ters, and growth in milk were considered, but more attention was given to the fermentation tests. We studied about 150 cultures isolated from milk, butter, and cheese obtained from various parts of the country. This collection included, in addition to the typical milk- curdling, nonliquefying, lactic-acid bacteria, a number of cultures curdling milk with subsequent digestion and which formed on gelatin plates small saucer-shaped liquefactions surrounding a solid colony. We have determined on these cultures the morphology, Gram's stain, cell grouping, in many cases formation of capsule, the nature and amount of growth on lactose-agar slopes and in gelatin stabs, the rate of liquefaction of gelatin, the nature of growth in broth, growth in milk, the reduction of nitrates and of neutral red, and the formation of acid in broth containing various test substances. In these fermentation tests we have used the sugars lactose, dextrose, galactose, saccharose, and raifinose, the alcohols mannite and glyc- erin, and the polysaccharid inulin. The results of these determina- tions are given in Table 1. SIGNIFICANT CHARACTERISTICS. 11 TABLE 1. Significant characteristics of acid-forming bacteria derived from milk, butter, and cheese. Culture. Agar. Gram stain. Cloudiness in broth. Reduction of neutral red. + + + + + 1 Curdling of milk. Reduction of nitrates. 1 Mm. gelatin liquefac- ooooooooooooooooooooooooooooooooooo o | t ion 30 days at 20. Per cent lactic acid in broth after 7 days at 30" C. 1 o j { Lactose. o 0.000 .540 .522 .000 .477 .405 .000 .000 .531 .000 .450 .405 .000 .000 .405 .000 .000 .000 .000 .000 .531 .526 .594 .009 .000 .000 .000 .000 .009 .000 .000 .000 .000 .252 .000 .000 .000 .423 .027 a O 0.000 .000 .000 .000 .000 .200 .000 .000 .000 .000 .027 .000 .000 .018 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .283 .018 .000 .000 .000 .000 .000 .009 .027 .234 .072 .000 .000 Mannite. i o Rafflnose. d J_ 0.000 .009 .018 .009 .000 .000 .009 .009 .009 .000 .009 .009 .000 .009 .009 .009 .018 .000 .009 .009 .000 .018 .018 .000 .009 .000 .009 .000 .000 .009 .009 .000 .009 .009 .009 .009 .009 .000 .009 .009 .009 .036 .000 .000 .018 .009 .000 .000 .000 .000 .000 .000 .009 .018 .216 .009 .000 .000 .000 .027 .387 .000 .000 "."666 .009 .000 .000 .01S .000 .018 6es. . t I 44444+ + 0.360 .612 .351 .382 .405 .288 .364 .668 .558 .400 .378 .693 .365 .324 .274 .369 .493 .423 .378 .360 .360 .585 .383 .306 .594 .504 .288 .540 .000 .522 .396 .351 .567 .432 .603 .594 .468 .585 .630 .360 0.319 .360 .450 .423 .378 .315 .316 .315 .378 .432 .414 .378 .414 .414 .387 .396 .333 .387 .378 .360 .369 .405 .324 .324 .378 .369 .171 .306 .162 .351 .360 .369 .351 .378 .360 .270 .306 .495 .369 .281 0.000 .000 .009 .000 .000 .279 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .018 .252 .000 .252 .225 .000 .324 .000 .360 .283 .000 .000 .000 .000 .000 .207 .000 .324 .315 .018 .081 0.198 .234 .216 .324 .378 .081 .288 .270 .297 .342 .283 .306 .315 .342 .054 .342 .405 .423 .397 .351 .189 .3% .242 .324 .369 .315 .108 .225 .378 .216 .306 .099 .252 .000 .414 .153 .297 .234 .234 0.000 .009 .000 .000 .000 .171 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .009 .000 .000 .000 .009 .000 .000 .000 .000 .000 .000 .009 .000 .000 "."666" .009 .000 .000 .000 .000 .000 .027 .003 .030 .000 .000 .000 .009 .000 .018 .018 .000 .018 .018 .018 .000 .000 .000 .009 .000 .000 .000 .009 .000 .000 .000 .000 .045 .000 .009 .000 .000 .000 6ex 6ez 6fa... 4. 6fb + 4 6fl 6fn 4 4 I 1 + 1 + 1 + + + + + + + + + + + 1 1 + 1 + + + + + .... j 6fr 7b 4 7c... + 4 7d 7f 4 + + + 7g... 7?. i i 7k... 4- 7n 4 7o 4- 7p 4- 7q 7s. . 4 7t 4. 7u 4- 7w 4 7x 4. 7y 4. 7z... 4. 7aa + 7ab. . . 1 7ad. 4 4 7ae... 4 7af 4 + J 7ag + 7ah 7ai i 7ai i 7ak 4- + + 7ao 7ap... 4 4 i 7aq... 7ar 4. i 4. .279 .380 .423 .363 .171 .279 .342 .324 .189 .207 .036 .342 .396 .378 .000 .000 .054 .009 .387 .000 .000 .000 .000 .000 .000 .000 .000 .000 .018 .009 .693 .000 .027 .288 .009 .225 .009 .288 .000 .324 .018 .099 .036 .063 .000 .072 .009 .000 .342 .036 .009 .009 .009 .009 .000 .009 .000 .000 .027 .054 .000 .036 .000 .000 .009 .000 .018 .009 .153 .000 .036 .018 .036 .045 .000 .342 .003 .009 .018 .000 .045 .000 .063 .000 .063 "."666" .315 .000 .018 .009 .018 .000 .036 .018 .397 .018 .009 .018 .486 .018 .378 .000 .009 .018 .009 .216 .234 .189 .243 .171 .252 .360 .396 .288 .306 .234 .378 .396 .360 .333 .288 .441 .000 .306 .324 .351 .414 .261 .324 .333 .414 .423 .243 .302 .315 .306 7as. 4. + 4 4 1 :. - "6" .702 .513 .702 .288 .180 .396 .396 .306 .294 .306 .270 :332 7ay... + 4. + 7ba 7be 4. 7bg... 4, 7bi. . . 7bj 7bn 7bp , 7bq 7br... 7bs... 7bt - - t ~ 11 .306 .423 .360 .549 .297 .441 .378 .360 .558 .351 .225 .495 .603 .432 .621 .261 .540 .513 .297 .360 .216 .193 .387 .256 .144 .221 .297 .144 .103 .297 .248 .260 .270 .297 .648 .261 Tbv 7bx 7by... 4. 7bz 4 + 7ca I 7cb 7cc 7cd..., 7ce 7cf 7cg \ 4 4 4- - "6" 7ch... 7ci 7ci 7ck... 7cl 7cm... 4 12 CLASSIFYING LACTIC-ACID BACTERIA. TABLE 1 . Significant characteristic* of acid-forming bacteria derived from milk, butter, and cheese Continued . Agar. 5 Gram stain. Cloudiness in broth. 3 I \ Curdling of milk. Reduction of nitrates. Per cent lactic acid in broth after 7 days at 30* C. Culture. 9 j Mm. geiauu nqi t ion 30 days at Dextrose. , j 3 O 1 1 7cn 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- + 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- H 4- H 6 9 14 378 351 576 576 459 522 .684 .261 .351 .316 .423 .495 .387 .405 .316 .279 .369 .333 .369 .423 .225 .171 .396 .405 .396 .288 .216 .432 .351 .288 .414 .334 .378 .594 .531 .378 .522 .477 .486 "."495 .468 .432 .603 .495 .720 .324 .234 .369 .324 .220 .360 .230 .211 .165 .270 .203 .149 .432 .099 .207 .243 .250 .495 .652 .216 .633 .171 .180 .198 .198 234 206 274 225 265 .243 .297 000 099 027 009 009 000 000 000 000 009 018 027 000 009 243 117 036 009 .531 .549 477 342 441 450 522 000 009 225 000 000 0.000 .018 .000 .000 .009 .036 .009 .018 .270 .000 .000 .018 .306 .000 .000 .252 .216 .288 .279 .288 .261 .000 .306 .000 .297 .342 .108 .630 .306 .270 .306 .000 .000 .000 .027 .000 .000 .000 .008 .009 .000 .000 .000 .000 .oos .008 .008 .ooc .01$ .01$ .OK .OOC .01$ .01$ .03< .01$ .00! .00! .00! .00! .011 .001 .001 .001 .001 .01 .001 7cp 7cr (CS 7ct .000 .009 .009 .162 .198 .225 .045 .117 .099 .045 .198 .252 .234 .261 .162 .234 .198 .234 .000 .027 .315 .225 .198 .180 .216 .252 .180 .189 .000 .000 .000 .000 .000 .000 .018 .018 .000 .000 .018 .009 .000 .000 .048 .207 .216 .027 .162 .072 .054 .081 .180 .086 .207 .315 .234 .054 .027 .045 .036 .045 .009 .045 .000 . 121 .072 .036 .405 .108 .279 .216 .081 .279 .234 .036 .378 .252 .279 .324 .252 .279 .243 .261 .279 .153 .288 .198 .288 .558 .234 .270 .288 .198 .270 .288 .351 .216 .351 .270 .171 .315 .342 .360 .143 .369 .225 .036 .063 .099 .239 .234 .099 .175 .036 .072 .144 .146 .198 .203 .135 .144 .058 .063 .081 .256 .261 .075 .279 .054 .090 .081 .287 .000 .000 .630 .630 .648 .018 .603 .054 .000 .630 .648 .630 .603 .567 .487 .639 .648 .703 .414 .676 .639 .261 .676 .756 .657 .630 .648 .000 .000 .000 .000 .000 .000 .000 .000 .009 .000 .000 .009 .000 .000 .000 .000 .000 .000 .000 .000 .000 .000 .027 .018 .018 .000 .000 .009 .045 .000 .027 .000 .018 .000 .000 .036 .054 .018 .108 .414 .441 .018 .372 .423 .432 .396 ."423" .414 .441 .414 .405 .405 .450 .009 .432 .414 .414 .423 .441 .441 .414 .423 .441 .432 .369 .351 .405 .351 .369 .387 .405 .387 .333 .343 .369 .045 .108 .171 .216 .194 .144 .270 .144 .126 .144 .077 .275 .176 .316 .234 .171 .175 .153 .324 .355 .153 .324 .171 .117 .121 .144 .009 .531 .459 .029 .009 .549 .018 .459 .450 .477 .522 .531 .549 .468 .504 .540 .018 .558 .522 .540 .549 .468 .369 .612 .423 .414 .000 .000 .000 .000 .540 .000 .000 .000 .000 .000 .000 .072 .000 .585 .027 .274 .009 .014 .000 .252 .153 .285 .351 .351 .198 .369 .297 .036 .192 .045 .018 .018 .036 .018 .045 .045 .054 .090 i .504 .045 .531 .018 .009 .567 .558 .558 .558 .549 .558 .549 .595 .054 .531 .549 .567 .558 .540 .549 .594 .540 .558 .000 .000 .054 .018 .000 .009 .000 .000 .000 .000 .000 .000 .000 .369 .072 .191 .406 .004 .243 .004 .000 .297 .135 .162 .230 .216 .297 .018 .045 .036 .018 .009 .009 .018 .000 .027 .018 .027 7da 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- + 4- 4- 4- 4- 7dc 7dd 7df 7de 7dfi 7di 7di 7dk 7dl 7dm 7dn 7do 7dp 7ds 7du 7dv 7dw 7dx 7dv 7dr lOav 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- + 4- 4- + 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- lOaw 4- 4- 4- lOax lObb 4- + 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- lObd lObe lObg . . lObn lObi lObo lObr 4- 4- 4- 4- 4- lObt 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- + 4- 4- 4- 4- 4- 4- lObl 4- 4- 4- 4- 4- 13c 4- 13f 4- 4- 13h 26 2' 21 6 10 18 13 15 25 15 1 i 1 l l i 13 1' 4. 13m... . + . 4- 4- 4- 13o 13r 4- 13t i 13u 4- + 13w 4- 13x 4- 4- 13v 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 4- 13aa 4- 13ab 4- 4- 13ac 13ad 4- 4- 13as; 13ah 13ai + 4- ia 4- 4- 13am 4- 13ap 4- MORPHOLOGY. TABLE 1. Significant characteristics of acid-forming bacteria derived from milk, butter, and cheese Continued . Ag ar. .a I 1 1 1- Per c 'lit lac tic in -i> linbr ath aft er 7 da ysat 3 o e c. e 44 o-~ a S Culture. c "3 a 'I l s s E "3 *S S a a-S | . . p W a 3 3 a '1 t?0 . o 1 | 1 i | I a a % f s B-2 K 3 9, ~ "3 S i O <5 O 6 S S O 9 O a 13ar + _ 0.198 0.027 0.027 0.009 0. 153 0.036 13at _1 1 18 162 171 054 054 027 117 13av + _ + _ 11 .171 .145 .045 .045 .000 .063 .036 13aw -f 1 I 7 117 135 090 036 009 13ax -f. , 5 126 181 'ooo 027 000 090 13be , 12 216 135 054 090 000 099 027 13bg , 13 189 136 036 072 009 054 072 13bk 297 207 126 009 009 189 13bl T 234 117 054 036 0^7 108 081 MORPHOLOGY. The cultures examined showed four more or less distinct types. The liquefying group included a number of cultures of micrococci, with frequent grouping in tetrads. This was associated with good growth on agar and certain fermentative reactions which made them easily distinguishable from the liquefying cultures with the morphol- ogy of the typical lactic bacteria. The nonliquefiers, excepting a few cultures of micrococci, showed three variations. Cells may be nearly or quite round. When in this condition they are usually found in chains of four or more cells. Single cells or pairs of cells are almost always oval and sometimes are distinctly of the bacillus type. A slight variation is sometimes found in that the cells are somewhat pointed at one end. All of these types can usually be found in the same culture and not infrequently in the same microscopic field. The question of the classification of the lactic-acid bacteria as cocci or as bacilli has been much debated and is yet in an unsettled condi- tion. The formation of chains, on which much emphasis is placed by some writers, is not pertinent, as the tendency to form chains is as common among the bacilli as among the streptococci. In our present state of knowledge the proper placing of these bacteria is largely a question of opinion or of definition. No satisfactory settle- ment can be reached until we have sufficient knowledge to establish the natural relations of the lactic-acid bacteria with other groups. Even in this case it is probable that close relationship will be traced on the one hand with groups that are distinct cocci and on the other with groups that are unquestionably bacilli. It may be stated in this connection that one of our cultures (7dv), not differing morphologically from other cultures of the typical 14 CLASSIFYING LACTIC-ACID BACTERIA. lactic type, was actively motile. This culture was pronounced by Dr. Heinemann to bo a typical Streptococcus lacticus, except that it did not curdle milk promptly. All of these cultures stain readily and are Gram positive. A cap- sule could usually be demonstrated if the test was repeated under varying conditions, indicating that it is formed only under certain circumstances. The circumstances under which a capsule was found indicated that it was in some way connected with the acidity of the culture. 0/234 FIG. 1. Rate of acid formation In milk at 30* C. by cultures freshly Isolated from milk. GROWTH ON SOLID MEDIA. Little need be said under this he^d, although the growth of the lactic-acid bacteria on agar and gelatin has been'described in great detail. The growth is always scanty, and the variations, which are very slight, are due more to differences in the chemical and physical condition of the medium than to varietal distinction. Some real variation may be observed in the size of mature gelatin colonies, but this is so influenced by the medium and the number of colonies on the plate that it is of little value. Variations of this kind are probably merely expressions of a tolerance or intolerance to certain conditions which could be determined with much greater GROWTH IN MILK. 15 accuracy by, other methods. The reaction of a particular lot of gelatin may produce large colonies of one culture while it limits the growth of another culture to colonies of almost microscopic si/c. GROWTH IN MILK. The time required to curdle milk under definite conditions has been employed almost universally in describing lactic-acid bacteria, although it is generally admitted that this property is variable, especially after the culture has been grown under artificial conditions. This variation is illustrated by figures 1 and 2. .7 .5 .2 ' * 3 4 S FIG. 2. Rate of acid formation in milk at 30 C. after 26 generations (2 years) on actose-agar. Figure 1 shows a variation in the freshly isolated cultures from 7b, which failed to curdle milk in 5 days, to 7k, which curdled milk promptly, forming nearly 1 per cent of acid in 48 hours. Two years' growth under uniform conditions redxiced these differences materially. The weaker cultures changed little or not at all, but the more active ones lost much of their vigor; that is to say, long-continued growth under.uniform conditions tended to reduce these cultures to a common level. It is not easy to restore this lost vigor. Repeated transfers in milk increased the activity of some of the cultures, but failed to bring the more active ones back to the rapid fermentation of the fresh cultures. It is probable that these differences are due to a variation in the vitality rather than to a variation in the particular function 16 CLASSIFYING LACTIC-ACID BACTERIA. of forming acid from sugar. In these studies it was frequently observed that those cultures curdling milk tardily or not at a)l multi- plied slowly and never attained the numbers reached by the cultures curdling milk in a short time. This is illustrated by Table 2, in which is given the acid formation and rate of multiplication of two cultures, one curdling milk in a short time and one failing to curdle milk in '2\ hours. Flasks of milk were inoculated from fresh milk cultures and incubated at 30 C. TABLE 2. Relative rate of multiplication and acid formation in milk. Slow acid former. Kiipidacid former. Hours from Inocula- Bacteria Bacteria tion. Acidity. per cubic Acidity. per cubic centimeter. centimeter. Per cent. Per cent. 0.216 357,000 0.220 350,000 3 .216 586.000 .225 650,000 6 .225 1,690,000 .225 2,900,000 9 .225 16,900,000 .225 46,000,000 12 .230 59,500,000 .261 157,000,000 24 .306 710,000,000 .900 1,830,000,000 It will be observed that in a general way the acidity in each case is proportioned to the number of cells present. This is in accordance with the observation of Rahn, 10 who calculated the amount of acid formed in relation to the number of cells in the culture and found that this ratio was constant, although when only a small number of bacteria were present the amount of acid was so small that it could not be measured by ordinary methods. Schierbeck, 11 who studied this form of variation in the lactic-acid bacteria, found that by replat- ing and making subcultures new cultures could be obtained, some of which followed the active fermentation of the original, while others were slow acid formers. In some of these cultures the rate of acid formation could not be varied by subsequent plating and selection, but cultures in which the ability to ferment lactose rapidly seemed to be fixed could be changed to slow fermenters by growing them in milk containing a small amount of carbolic acid. Buchanan and Truax lf attempted to fix strains of the lactic-acid bacteria by selec- tion and transfer from tubes of lactose broth showing wide differ- ences in acidity. They failed entirely and concluded that "im- pressed variations do not appear to be heritable." It seems propa- ble that they were unable to fix these variations because they were working, not with variations in a function, but with the rate of multiplication, something which may be controlled by an endless chain of circumstances. For the same reasons this characteristic can not be successfully used as a basis for classification. REDUCTION OF NITRATES. 17 GROWTH IN BROTH. The growth in a medium of this nature should be considered as the expression of certain peculiarities not evident on superficial examina- tion. A cloudiness or turbidity is, with the lactic-acid bacteria, an evidence of rapid growth, and it is governed more by external condi- tions than by the nature of the culture. Cultures which remain clear in nonsaccharine broth may be cloudy when certain sugars are added, and others may show cloudiness when dibasic potassium phosphate is added, but not in its absence. Profuse cloudiness is usually fol- lowed by a clearing and the accumulation of a sediment coinciding with the checking of the growth by increasing acidity. Some few cultures, however, never show turbidity, and the broth remains per- fectly clear with no evidences of growth, although a high acidity is produced. This is perhaps due to the tendency to form tangled chains possessing in the aggregate a specific gravity greater than the broth. Muller 13 states that a clouded bouillon is associated with the formation of single cells or pairs, while a clear bouillon is coordinated with the formation of long chains. In his work he found that the cultures that never clouded the bouillon were the "Gait Stamme" found associated with yellow mammitis. When the failure to cloud the broth is fixed and constant it may be of some assistance in clas- sification. REDUCTION OF NITRATES. The reduction of nitrates to nitrites was determined by growth in the following medium for 7 days at 30 C. Peptone gram. . 1. Potassium nitrate gram . . 0. 2 Water (distilled) c. c. . 1, 000 Of the 147 cultures tested for the reduction of nitrates, 17 gave a positive reaction. All of these were differentiated from the typical lactic-acid bacteria by one or more coordinated characters. Twelve were of the liquefying tetrad group. Three (6fl, 7aa, and 7ak), were gas formers, and 7ak was in addition a bacillus larger than the lactic type and grew readily on solid media. One (7cc) resembled the lactic type in morphology but the cells were larger, and in addition to liquefying gelatin and giving an abundant growth on agar it made milk slimy. Culture 7dw was a small micrococcus. None of the cultures belonging to Group I gave any evidence of growth in this medium. It is therefore of no value in making sub- divisions of this group, but may be a convenient means for the detec- tion of cultures resembling the type in many features but differing in certain salient points. 18 CLASSIFYING LACTIC-ACID BACTERIA. REDUCTION OF NEUTRAL RED. In making this test the following medium was used: Broth (neutral) c. c. . 1, 000 Dextrose grams. One-half per cent solution Griibler's neutral red . .c. c. 5 10 The neutral broth was made as follows: Beef extract grams. Peptone grams. 4 10 Water c . o.. 1,000 The tubes were examined after incubating at 30 C. for 7 days in an anaerobic jar from which the oxygen was exhausted by absorp- tion with pyrogallic acid. Gordon u considers this test of diagnostic value. It has, however, the disadvantage of not always giving definite results, although with nearly all cultures the reduction was either nil or very evident. Of the 36 cultures reducing neutral red, a large proportion ferment the more resistant test substances, such as sac- charose, glycerin, mannite, and raffinose, and 7, or 19 per cent of the whole, liquefy gelatin. None of those liquefying gelatin ferment raffinose, while of the 29 nonliquefying cultures reducing neutral red 75 per cent ferment raffinose. There is a correlation between the reduction of neutral red, the liquefaction of gelatin, and the fer- mentation of saccharose, glycerin, and mannite in one group and between neutral red, saccharose, glycerin, mannite, raffinose, and inulin in another. These correlations are evident in figure 6. The faculty of reducing neutral red seems to be usually coordinated with other reactions and is therefore of some value in differentiating cultures. LIQUEFACTION OF GELATIN. The value of this test is too generally recognized to need discus- sion. In our work we have used Clark and Gage's method of reducing the rate of liquefaction to mathematical terms, ignoring the appear- ance of the culture. The gelatin tubes were inoculated by spreading a few drops of a fluid culture on the surface of the medium. The line of the surface was marked on narrow strips of paper pasted on opposite sides of the tube, and the cultures were incubated at 18 to 20 C. At the end of 30 days the amount of liquefaction was measured and expressed as millimeters of liquefaction. The results of this test are tabulated in Table 3. TABLE 3. The liquefaction of gelatin. Liquefaction millimeters. . Nuaiber of cultures 108 ItoS 3 6 to 10 10 11 to 15 10 ' 16 to 20 2 21 to 25 3 Over 25 2 Peri3ent of total 78.3, 2 2 7 2 7 2 1 4 2 2 1 4 FERMENTATION OF CARBOHYDRATES. 19 These results are platted in figure 3 to show the frequency <>f occurrence of certain arbitrary types. This curve gives some indication of a division on the basis of gelatin liquefaction into three types, one failing to liquefy, one liquefying 6 to 15 millimeters, and one 20 to 25 millimeters. How- ever, the total number of liquefying cultures was so small that it is safe to make a division into liquefiers and nonliquefiers only, and to depend on other tests for further division of the liquefiers. Even the separation of the liquefiers from the nonliquefiers is not entirely reliable, as it is well known that this character is vari- able and under some conditions may be entirely lost. If physiological tests are of value they should show by correlation or lack of correlation which cultures 70 \ belong properly with the nonliquefiers and which are members of liquefying varieties in which the ability to produce a proteolytic enzym has been lost. FERMENTATION OF CARBOHYDRATES. Mention has already been made of the objections to the use of the fermentation of sugars and similar sub- stances. The question of the constancy of these reac- tions has been the subject of investigation, and while there is some disagreement the opinion of those who have studied the question most carefully seems to be that they are at least as constant as any of the characters ordinarily used in classification. Twort, 15 working with gas- forming cultures, was able to in- duce acid /O /-S 6-/0 fM5 f6-2O 2J-25 m.m, of //yet e faction. tlon fr m sugars FIG. 3.-Frequency curve for gelatin liquefaction. Which the organ- ism originally did not ferment by repeated transfers in a medium in which this sugar was the only carbohydrate furnished. Each transfer was held 14 days to allow the cells to work on the sugar after other sources of food had been exhausted. Kitchie 16 concluded that while cultures of Bacillus coli tested at different times gave constant fermentation reactions, the streptococci were inconstant. Gordon 17 tested the constancy of 11 cultures by passing them through mice. One culture lost ability to reduce neutral red and one gained ability to ferment salicin. All others remained unchanged. In all of this work the fermentative ability was determined by growing the organism in broth, with the addition of the test substance and litmus, and the change of the 20 CLASSIFYING LACTIC-ACID BACTERIA. litmus from blue was taken as a positive reaction. A slight change in the reaction of the medium may change litmus from blue to rod, and this acidity may be formed from some substance other than the sugar. The reduction of the litmus is certainly not an indication of the fermentation of the test substance. The work of MacOonkey 18 indicates that under natural conditions these reactions are constant and of value in differentiation. Among the cultures examined \v<-iv 15 cultures of B. typhosus varying from one freshly isolated to one grown 16 years on artificial media. These gave identical fermenta- tion reactions when tested with various carbohydrates. In another paper the same investigator states that the fermentative reactions of B. coli remained unchanged after a long exposure to un- favorable conditions. He expresses the opinion that one group is not derived from another by the loss of characters. Harding, 19 working with Pseudomonas campestris, an organism pathogenic to certain plants, obtained somewhat similar results. Of the four substances used for fermentation tests this organism attacked only one, but this and all other physiological tests employed were identical for the 44 cultures collected from various parts of the country. In our own work no systematic investigation was undertaken to determine the constancy of the fermentation reactions, but all our observations tend to prove that the- property of forming acid from carbohydrates and similar substances is not easily lost or acquired. One culture showing no evidence of ability to ferment saccharose was carried for 100 generations, or a period of about one year, on a saccharose-agar. At the end of this period the culture still showed no fermentation of saccharose and the lactose fermentation remained unchanged. In no case did any of our cultures show any change in fermentative ability on repeated tests. It not infrequently happens that a culture failing entirely to give an acid reaction on the first test showed an active fermentation when the test was repeated, but this was evi- dently due to a failure of the inoculation rather than to a change in the organism. Many of the cultures grew so poorly on artificial media that they were propagated with difficulty and transfers fre- quently failed to grow. A large proportion of the cultures were subjected to these tests two or three times, some of them at intervals of several months. The second test almost always agreed with the first not only in the presence or absence of fermentation but also in the amount of acid formed. This is illustrated by Table 4, which contains results on the fermentation of lactose. These figures were picked at random FERMENTATION OF CARBOHYDRATES. TABLE 4. Showing constancy of fermentation of lactose. 21 Test. Acidity of broth expressed as per cent of lactic acid. Firet 0.112 .117 0.171 .261 0.216 .171 0.387 .306 0.288 .225 0.000 .144 0.180 .261 0.306 .288 0.153 .135 0.108 .009 0.252 .243 Second With some of the test substances the reaction was always very / / positive; that is, the reaction remained unchanged or a considerable acidity was developed. This was especially noticeable with sac- charose, mannite, and raffinose. With others, particularly with glycerin, the acid was developed slowly and in such small quantities that it was sometimes difficult to determine if there was a real fermen- tation or a slight change in the reaction which was independent of the test substance. In doubtful cases a retest usually gave definite results. It is not to be expected that a character of this kind would be absolutely fixed. Indeed, the fact that one culture attacks a certain sugar while similar cultures do not is evidence that this function is or has been a variable one. There is, however, no evidence to show that the tendency toward variation in fermentation is any greater than in any other character used as a basis of classification. The greater difficulty comes in the interpretation of the results. The objection that the number of varieties obtained is limited only by the number of test substances used is valid only when an absolute separation is made on each individual reaction. The usual botanical scheme of dichotomous separation when applied to the classification of bacteria on the basis of fermentation tests leads only to confusion and the rejection of the system. In the card arranged for the classi- fication of bacteria by a committee of the Society of American Bacteriologists, dextrose, saccharose, lactose, starch, and glycerin are used as test substances, and cultures are separated in the usual way on the fermentation of or failure to ferment any one substance. By the use of these test substances and similar methods we could separate our nonliquefying cultures into five varieties. But if it is proper to separate cultures on the basis of the fermentation of dex- trose, saccharose, or lactose we can use also raffinose and galactose, and if glycerin is allowable mannite can not be excluded, while inulin may be as useful as starch. Adding these test substances and follow- ing the same principles of division, we obtain no less than 14 varieties, and even these are not stable, because the introduction of a new test substance would probably subdivide them still more minutely. Gordon 10 was the first to make an extensive use of the fermentation tests. These tests were also used on an extensive scale by Andrewes and Gordon 20 and by Houston. 21 This work shows the possibilites of arranging a large number of cultures in groups around type sets of reactions. Not all of the cultures in each group agreed perfectly 22 CLASSIFYING LACTIC-ACID BACTERIA. with the type. Some failed in one reaction, while others possessed some character not common to the entire group. MacConkey, 18 using similar methods in a study of gas-forming bacteria from milk, was able to separate 112 cultures into 17 groups, and even these groups were sometimes separated by minor differences only. On the basis of the individual reactions it was possible to separate these cultures into 64 varieties. This work was continued and placed on more scientific footing by Andrewes and Horder, 22 and especially by Winslow and Rogers, 23 , 24 who applied the principles of biometry to the study of bacteria. In this way has been supplied a method of utilizing the physiological tests in such a way that bacteria may be collected in natural groups. In tabulating the characters of a large number of cultures, frequency of occurrence of those with certain common characters indicates the type, while the cultures varying from these types occur in smaller numbers and form the connecting links between the types. This represents the state of affairs in nature, while a description based by the ordinary method on the characters of a single culture may or may not agree with the type. In our fermentation tests we have followed Winslow in determining the acidity rather than the mere fact of fermentation or nonfermen- tation. This is more exact and sometimes gives additional informa- tion of value in separating cultures. The medium was made as follows : Per cent. Beef extract 0. 4 Peptone 1.0 Dibasic potassium phosphate 5 Test substance 2. The use of dibasic potassium phosphate is of advantage in that it serves to neutralize the acid and thus permits a more active growth and higher acid formation. The acid phosphate formed evidently checks the growth when a certain concentration is reached. The neutralization of culture media by this means is discussed by Hender- son and Webster. 25 The cultures were incubated 7 days at 30 C., with the exception of glycerin, which on account of the slow fermentation was held 14 days, and were titrated while cold against twentieth-normal sodium hydrate with phenolphthalein as an indicator. The result of the titration is expressed as per cent of lactic acid. Gas formation was determined by using an inverted inner tube in the dextrose broth. The results of the fermentation tests are given in detail in Table 1 and are recapitulated in Tables 5 and 6. The results shown in the two latter tables are given graphically in figures 4 and 5, in which the frequency of occurrence of cultures forming certain arbitrary amounts of acid is platted. FERMENTATION OF CARBOHYDRATES. TABLE 5. Fermentation of test substances by liquefying cultures. 23 Test substance. Per cent of lactic acid. Below 0.100. ve 0.700. it -1 o~ H ?, % i 1 1 1 % 8 1 V V V V V ? V V 2 2 2 i i jj i 8 % 9 a a i 8 & < d d o o o o d o o d d Dextrose: Number of cultures Per cent of total... Lactose: Number of cultures Per cent of total... Saccharose: Number of cultures Per cent of total . . . Glycerin: N umber of cultures Per cent of total . . . Mannite: Number of cultures Percent of total... Galactose: Number of cultures Per cent of total. . . 1 3.45 21 63. (j 25 75.8 22 66.7 17 53.1 31 100 4 13.8 12 36.4 1 3.0 2 6.1 1 3.0 4 12.5 17.2 11 33.3 3 9.1 1 3.0 2 6.1 5 15.6 19 41.4 4 12.1 4 12.1 3 9.1 3 9.4 1 3.45 2 6.0 5 15.1 3 &1 2 6.3 3 9.1 1 3.0 1 3.0 3 10.3 1 3.0 2 6.1 1 3.45 1 3.45 1 3.45 29 33 33 33 1 3.0 1 3.1 1 3.0 33 32 Raffinose Number of cultures Per cent of total. . . 31 Inulin: Number of cultures Per cent of total. . . 22 95.6 . 1 4.4 23 TABLE 6. Fermentation of test substances by nonliquefying cultures. Test substance. Per cent of lactic acid. Below 0.100. 9 ? 1 ? 8 t ?' Above 0.700. "3 . m !l EH 116 117 ?, i 1 I 8 ** 8 j J j 2 2 ? ? 2 i i 8 $ i 9 S 8 i X o o o o d o d o o o Dextrose: Number of cultures Per cent of total. . . Lactose: Number of cultures Per cent of total. . . Saccharose: Number of cultures Percent of total... Glycerin: Number of cultures Per cent of total. . . Mannite: N umber of cultures Per cent of total. . . Galactose: Number of cultures Percent of total... Raffinose: Number of cultures Per cent of total.. . Inulin: Number of cultures Percent of total. .. 1 0.9 3 2.6 75 64.1 93 78.2 78 67.2 8 6.8 90 78.9 101 86.3 5 4.3 2 1.7 I 0.9 5 4.2 ' 2 1.7 5 4.3 12 16.8 8 6.8 1 0.9 4 3.4 8 6.8 1 0.8 7 5.9 3 2.6 19 16.1 1 0.9 2 1.7 10 8.6 16 13.7 2 1.7 4 3.4 4 3.5 28 23.7 9 7.7 12 10.3 17 14.5 I 0.8 1 0.8 4 3.5 20 16.9 5 4.3 30 25.9 30 25.6 3 2.6 13 11.2 31 26.5 8 6.8 10 8.6 1 0.8 7 5.9 12 10.3 15 12.8 11 9.5 3 2.6 4.2 1 0.8 1 0.8 2.f 2.e 1 0.8 117 119 3 2.6 16 13. 6 11 9.3 1 0.9 1 0.9 1 0.8 1 0.9 11 9.5 1 0.8 11 9.5 1 0.8 1 0.9 116 118 12 10.5 7 6.1 1 0.9 114 117 We have reckoned any acidity of below 0.1 per cent as no fermentation, although this may be an arbitrary distinction. In the work by Winslow previously cited the frequency curves usually showed three modes. In his work, however, a larger num- ber of cultures selected from various sources was used, while ours 24 CLASSIFYING LACTIC-ACID BACTERIA. came from milk only. Our curves usually show only two modes, one at the point of no fermentation and one at a point of acidity varying with the different test substances. There are, however, certain differences in the curves of the liquefiers and the nonliquefiers. The dextrose curve for the liquefiers shows that a large proportion form 0.2 to 0.25 per cent of acid, with a smaller number at 0.35 to Jisy- ~^ST- ~^5T- ~^T- ~ ./oa ,,so too -zso -jwe *> *<> -45 3we> -sso FIG. 4. Frequency curves for acid formation by the liquefying cultures. 0.4, while with the nonliquefiers there is a high point at 0.35 to 0.4 per cent and possibly a second mode at 0.5 to 0.6. With the liquefiers the other test substances show two modes only. Raffinose was not fermented at all, and only one culture formed acid from inulin. The number of liquefying cultures was too small to make many deductions therefrom, but it is easy to separate these cultures into FERMENTATION OF CARBOHYDRATES. 25 two distinct groups. One of these forms a small amount (0.1 to 0.3 per cent) of acid from dextrose, and is a micrococcus usually appear- ing in tetrads. The effect on milk is weak, and the curdling, which is slow, is probably due more to the action of a rennet than to the production of acid. This group, as represented by these cultures, is undoubtedly heterogeneous, and by the application of these methods 80 FIG. 5. Frequency curves for acid formation by the nonliquefylng cultures. to a larger number of cultures would be split up in distinct sub- groups. The second group of liquefiers is interesting in that it evidently is a variation from the typical nonliquefying lactic-acid organism. In its morphology it is identical with the ordinary type, but differs from it not only in the liquefaction of gelatin, but also in usually fermenting glycerin. Its action on milk is characteristic. 26 CLASSIFYING LACTIC-ACID BACTERIA. The milk is curdled promptly with a firm acid curd; digestion begins at once and almost always causes a separation of curd from the whey down the side of the tube. In the nonliquefying group there are with all the test substances only two distinct modes in the curves with the possible exception of mannite. In this case there are three modes, one below 0.1 per cent* one between 0.1 and 0.35, and another between 0.45 and 0.5. Reference to Table 1 shows that of the 8 cultures belonging to the group falling between 0.1 and 0.35 per cent, 2 were gas formers and 1 made milk slimy, while the other 5 apparently did not differ from those , forming a high acidity or failing entirely to ferment this substance. In arranging these reactions on the basis of their correlations, one of the formulas for the expression of correlation, as, for instance, that of Yule, may be used, but for this particular work the determination of the coefficient of correlation is not necessary. Even a casual ex- amination of Table 1 shows that the nonliquefiers may be separated into two groups, ;n one of which the fermentation is usually limited to dextrose, lactose, and galactose, with an occasional culture fer- menting saccharose or mannite. A second group may be formed of cultures in which the fermentative ability is distinctly higher. These groups are illustrated by figure 6, in which each culture is placed in one of four groups and arranged on the positive or negative side of a dividing line, as the case may be, in each of the salient characters. The division of the nonliquefiers is not an arbitrary one, as the distinction between the two groups is marked. Not only is there a general lack of fermentative ability in Group A, but there is no cor- relation in the few cases of fermentation of the more difficultly fer- mentable substances. The fermentation of saccharose is no indica- tion that the culture will ferment mannite. On the other hand, in Group B more of the test substances are fermented and there is a high correlation between certain activities. The fermentation of raffinose is usually correlated with the fermentation of saccharose, mannite, and glycerin, and to a lesser degree with inulin. The high fermentation is also correlated with the reduction of neutral red. In making up these groups it was found that 6 cultures (6fl, 7aa, 7ak, 7cq, 7dw, and 13u) did not belong in this collection, and they were not included in the table. It will also be noticed that 7ci, 7cg, and 7dm are in a transition stage between Groups A and B, either through the loss of properties formerly possessed or the acquisition of new ones. In morphology and general culture characters all of the numbers of this group agree with the typical lactic culture. It should be stated that we obtained all of the cultures of Group B from one locality, Albert Lea, Minn., although not from one sample. However, in the course of another investigation a large number of cultures which could be properly placed in this group have been isolated from Washington milk, indicating that it is widely distrib- GROUPING OF CULTURES. 27 tfl I * I FT> -s- - J ^ ^HBB ' 1 ^ ^ ! 1 I* i ~ | S 3 a + <& J *o ^ i i I GELATIN | 1 LACTOSE \ I ^ I 1 1 as K &. ^ ^? 1 \_ 1 I ^ +P^J V)rv U2-J l_l . 28 CLASSIFYING LACTIC-ACID BACTERIA. uted, although it does not occur in so large numbers as the Group A type. While the number of cultures included in Group C is too small to permit many positive deductions, it is evident that it rep- resents a type quite distinct from A and B, from which it is differ- entiated not only by the liquefaction of gelatin but also by the cor- related functions of the fermentation of mannite and glycerin and the failure to ferment raffmose and inulin. These 9 cultures include 3 differing somewhat from the others morphologically, .and it is probable that a larger collection would allow a deeper and more posi- tive separation. Group D is made up of cultures which, while they are of common occurrence in milk, have such a low fermentative ability that they probably take little part in the normal souring of milk. CONCLUSIONS. The stability of the fermentation tests is made evident not only by the constancy of the reactions on repeated tests, but also by the marked correlation between different fermentative activities and between the fermentations and other characters. The usefulness of these tests is only apparent when by means of biometrical methods the correlations are established and the cultures are arranged in groups possessing certain characters in common, but in which minor variations from the type are not excluded. The test substances used can not be determined arbitrarily. It is probable that it will be desirable to vary the test substances used with different groups of bacteria. We have found raffinose and glycerin and the gelatin test especially valuable, while saccharose, which has long been used for differential tests, has much less value. All of the groups have many cultures fermenting this sugar, and there is little correla- tion with other reactions. While the determination of the fermen- tation of raffinose or glycerin gives one a good idea of the group in which the culture should be placed, the knowledge that a culture fer- ments or fails to ferment saccharose is of little assistance. It should be remembered that these cultures were all selected on the basis of the possession of a single positive character, the fermen- tation of lactose. If the collection had been made on a broader basis, it is highly probable that the cultures would have formed other groups around types distinct from those we have found but related to them by certain common characters and by transition forms. The results recorded in this paper are too meager to warrant any attempt at filing names or establishing the place of the lactic-acid bacteria in the bacteriological system, but we believe that this work indicates that future efforts in the direction of systematic bacteriology should be toward the determination of those characters that are sig- nificant and enduring rather than in fruitless controversy over the priority or stability of some name based on descriptions so undeter- minative that they convey no meaning. REFERENCES TO LITERATURE. 1. HASTINGS, E. G. A preliminary note on a group of lactic acid bacteria not pre- viously described in America. Science, vol. 28, no. 723, p. 656. New York, Nov. 6, 1908. 2. HASTINGS, E. G., and HAMMER, B. W. The occurrence and distribution of a lactic acid organism resembling the Bacillus bulgaricus of yogurt. Wisconsin Agricultural Experiment Station, Research Bulletin 6, pp. 197-206. Madison, ; June, 1909. 3. HEINEMANN, P. G., and HEFFERAN, MARY. A study of Bacillus bulgaricus. Journal of Infectious Diseases, vol. 6, no. 3, pp. 304-318. Chicago, June 12, 1909. 4. MCDONNELL, MILTON EARLE. Uber Milchsaure-Bakterien. Inaugural Disserta- tion, Kiel, 1899. 5. WEIGMANN, H. Versuch einer Einteilung der Milchsaurebakterien der Molkereige- werbes. Centralblatt fur Bakteriologie, Parasitenkunde und Infektionskrank- heiten, Abteilung 2, vol. 5, no. 24, pp. 825-831; no. 25, pp. 859-870. Jena, Dec. 5, 15, 1899. 6. MILLER, PAUL TH. Uber die Streptokokken der Milch. Archiv fur Hygiene, vol. 56, no. 1/2, pp. 90-107. Munich, 1906. 7. LOHNIS, F. Versuch einer Gruppierung der Milchsaurebakterien. Centralblatt fiir Bakteriologie, Parasitenkunde und Infektionskrankheiten, Abteilung 2, vol. 18, no. 4/6, pp. 97-149. Jena, Mar. 14, 1907. 8. CONN, H. W., ESTEN, W. M., and STOCKING, W. A. A classification of dairy bacteria. Connecticut (Storrb) Agricultural Experiment Station, Eighteenth Annual Report (1906), pp. 91-203. Middletown, 1907. 9. HEINEMANN, PAUL GUSTAV. The kinds of bacteria concerned in the souring of milk. Inaugural Dissertation, University of Chicago. Chicago, 1907. 10. RAHN, OTTO. The fermenting capacity of the average individual cell (Bacterium lactis atidi). Science, vol. 33, no. 849, pp. 539-540. New York, Apr. 7, 1911. 11. SCHIERBECK, N. P. Ueber die Variabilitat der Milchsaurebakterien mit Bezug auf die Canmgsfahigkeit. Archiv fur Hygiene, vol. 38, no. 3, pp. 294-315. Munich, 1900. 12. BUCHANAN, ROBERT EARLE, and TRUAX, ROY. Noninheritance of impressed variations in Streptococcus lacticus. Journal of Infectious Diseases, vol. 7, no. 5, pp. 680-697. Chicago, Oct. 25, 1910. 13. MULLER, LEO. Vergleichende Untersuchungeii liber Milchsaurebakterien (des Typus Guntheri) verschiedener Herkunft, nebst Beitrag zur Frage der Stellung dieser Organismen zu den typischen Streptokokken. Centralblatt fur Bakteri- ologie, Parasitenkunde und Infektibuskrankheiten, Abteilung 2, vol. 17, no. 14/16, pp. 468-479. Jena, Dec. 7, 1906. 14 GORDON, M. H. Report of some characters by which various streptococci and staphylococci may be differentiated and identified. Great Britain. Local Government Board, Thirty-third Annual Report (1903-04), supplement con- taining Report of the Medical Officer (1903-04), pp. 388-430. London, 1905. 15. TWORT, F. W. The fermentation of glucosides by bacteria of the typhoid-coli group and the acquisition of new fermenting powers by Bacillus dysenteric and other micro-organisms. Preliminary communication. Proceedings of the Royal Society of London, series B, vol. 79, no. B532, pp. 329-336. London, June 5, 1907. 30 CLASSIFYING LACTIC-ACID BACTERIA. 16. RITCHIE, JOHN. Notes on experiments as to the constancy of the carboyhdrate reactions of the streptococci. Lancet, vol. 175, no. 4432, pp. 374-376. London, Aug. 8, 1908. 17. GORDON, M. H. A ready method of differentiating streptococci, and some resulta already obtained by it* application. Lancet, vol. 169, no. 4289, pp. 1400-1403. London, Nov. 11, 1905. 18. MACCONKEY, ALFRED. A contribution to the bacteriology of milk. Journal of Hygiene, vol. 6, no. 3, pp. 385-407. Cambridge, July, 1906. 19. HARDING, H. A. The constancy of certain physiological characters in the classifi- cation of bacteria. New York Agricultural Experiment Station, Technical Bulletin 13. Geneva, June, 1910. 20. ANDREWES, F. W., and GORDON, M. H. Report on the biological characters of the staphylococci pathogenic for man. Great Britain, Local Government Board, Thirty-fifth Annual Report (1905-06), supplement containing the report of the Medical Officer (1905-06), pp. 543-560. London, 1907. 21. HOUSTON, A. C. Report on the bacteriological examination of the normal stools of healthy persons. Great Britain, Local Government Board, Thirty-third Annual Report (1903-04), supplement containing Report of the Medical Officer (1903-04), pp. 472-527. London, 1905. 2'2. ANDREWES, F. W., and HORDER, T. J. A study of the streptococci pathogenic for man. Lancet, vol. 171, no. 4333, pp. 708-713; no. 4334, pp. 775-782; no. 4335, pp. 852-855. London, Sept. 15, 22, 29, 1906. 23. WINSLOW, C. E. A. , and ROGERS, ANNE F. A revision of the Coccacese. Science > vol. 21, no. 539, pp. 669-672. New York, Apr. 28, 1905. 24. WINSLOW, CHARLES EDWARD AMORY, and WINSLOW, ANNE ROGERS. The syste- matic relationships of the Coccaceae. New York, 1908. 25. HENDERSON, LAWRENCE J., and WEBSTER, H. B. The preservation of neutrality in culture media with the aid of phosphates. Journal of Medical Research, vol. 16, no. 1, pp. 1-5. Boston, Mar., 1907. ADDITIONAL COPIES of this publication -i~\- may be procured from the SUPERINTEND- ENT OF DOCUMENTS, Government Printing Office, Washington, D. C., at 5 cento per copy University of California SOUTHERN REGIONAL "BRARY FACILITY 405 Hllgard Avenue, Lot Angeles, CA 90024-1388 Return this material to the library from which It was borrowed. NOV 1 4 T99f>