`REFERENCE TO THEIR EFFECT ON BACTERIAL
`VIABILITYl
`
`C.-E. A. WINSLOW AND ELOISE T. HAYWOOD
`
`Department of Public Health, Yale School of Medicine
`
`Received for publication, February 1, 1931
`
`OBJECT OF STUDY
`
`In previous contributions from this laboratory (Winslow and
`Hotchkiss, 1922; Hotchkiss, 1923; Winslow and Falk, 1923a;
`Shaughnessy and Winslow, 1927; Winslow and Dollofi, 1928;
`Fabian and Winslow, 1929) we have brought forward evidence to
`show that cations exert a highly characteristic effect upon bac-
`terial viability.
`The fact that a low concentration of a given salt stimulates
`biological action and a higher concentration inhibits it has been
`shown by numerous observers and in general all the studies have
`indicated much the same relative potencies of the various cations.
`Among the most important work along this line may be mentioned
`that of Lipman (1909) on the effect of NaCl, KC], MgClg and
`03012 upon ammonification by B. subtilis, of Brown and Hitch—
`cock (1917) on nitrification in soils and of Brooks (1919, 1920,
`1921) on carbon dioxide production by B. subtilis. Brown and
`Hitchcock (1917) present excellent curves for the influence of
`N80] N32804, MgSO4, 03003, NaHCOzg, N32003 and 03003
`upon nitrification in soils. Here, however, calcium was least
`potent of the cations studied, in direct contrast with results in
`simpler media.
`-
`Brooks gives excellent curves for NaCl, KCl and CaCh (1919)
`for MgClz (1920) and for La (N03), (1921) all showing stimulation
`of carbon dioxide production by low concentrations and inhibition
`1 Based on a thesis presented by the junior author in partial fulfilment of the
`requirements for the degree of Doctor of Philosophy in Yale University.
`49
`
`Argentum Pharm. LLC V. Alcon Research, Ltd.
`Case IPR2017-01053
`
`ALCON 2122
`
`
`
`50
`
`C.-E. A. WINSLOW AND ELOISE T. HAYWOOD
`
`by higher concentrations. Branham (1929) presents similar data
`for yeast.
`In the first papers of our own series (Winslow and Hotchkiss,
`1922; Hotchkiss, 1923) it was demonstrated that a wide variety of
`cations stimulate bacterial growth in low concentration and
`inhibit it in high concentration. Winslow and Dolloff (1928)
`showed that the efficiency of each cation (both in the stimulating
`and in the toxic range) may be expressed by a characteristic
`constant and that mixtures of the chlorides of sodium, potas-
`siwm and magnesium exhibit exactly the effects which would
`be predicated if their components acted in a purely additive
`Fabian and Winslow (1929) found that the sodium
`fashion.
`ion exerted its characteristic effect in combination with a wide
`variety of anions including the hydroxyl ion, the result being
`determined by the combination of two factors,-concentration
`of sodium and pH.
`It seems reasonable to conclude from these results and those of
`other workers that all cations exert upon bacterial viability a
`certain influence (aside from other more specific influences)
`which is qualitatively the same. The quantitative effect of
`different cations varies very widely but each has a specific
`efficiency, both as regards stimulation and inhibition.
`This
`characteristic, we propose to designate as "specific potency."
`The effect of mixtures of salts appears to be determined (aside
`from differences in pH) largely by the arithmetical sum of their
`specific potencies.
`The present study was designed to test this postulate of specific
`potency by a careful study of salts and salt mixtures involving a
`larger group of cations than those reported upon by Winslow and
`Dolloff (1928).
`
`TECHNIQUE
`The organism used in these studies was the same strain of
`Escherichia coli (communis type) used in all the previous work of
`It was originally isolated from water in 1916
`this laboratory.
`and is unusually well-adapted to such studies because it maintains
`itself in distilled water in almost undiminished numbers for a
`
`
`
`SPECIFIC POTENCY OF CERTAIN CATIONS
`
`51
`
`period of 24 hours or more. The stock culture was maintained on
`nutrient agar with occasional passages through nutrient broth.
`For the actual study of viability we used Dolloff's synthetic
`This medium (Dolloff, 1926) consists of 5 grams of
`medium.
`recrystallized amnmonium tartrate, 5 grams of Pfanstiehl lactose
`and 0.02 gram of dibasic ammonium phosphate in a liter of water.
`Our stock solution was made up two and a half times this strength
`and sterilized at 15 pounds pressure for 20 minutes,-to be later
`added to the salt solutions to produce the standard concentration
`of the final medium.
`The Dolloff medium was selected after a preliminary study in
`which it was compared with distilled water, used by Winslow and
`Falk (1923a) and with 1 per cent peptone water (Difco peptone)
`It was expected that the activity of
`used by Hotchkiss (1923).
`such salts as calcium might be very different in the different media
`since even so low a concentration as 0.005 M CaCl2 formed a
`distinct precipitate in the Dolloff medium.
`In distilled water there were on the average 11 million bacteria
`per cubic centimeter alive at the end of 48 hours, in Dolloff
`medium, 99 millions and in peptone-water, 161 millions. The
`quantitative effect of the salts tested (NaCl and CaCl2) was to a
`slight degree affected by the medium, the nutrient materials
`present in the more complex media exerting a protective effect.
`Thus, maximum stimulation with NaCl was apparent at a con-
`centration of 0.05 M in distilled water at 0.08 M in Dolloff medium
`and at 0.1 M in peptone-water. Marked toxicity appeared at
`With
`0.25 M in distilled water, and at 0.5 M in the other media.
`CaCl2, however, the effect was even less, optima so far as dis-
`tilled water and Dolloff medium were concerned, being between
`0.005 and 0.008 M in both cases and marked toxicity appearing at
`In peptone-water, the toxic effect of CaCl2 was markedly
`0.01 M.
`So far as the Dolloff
`lowered, being insignificant even at 0.1 M.
`medium was concerned, it seemed clear that such precipitation
`as occurred did not seriously affect the relative potency of the
`salts and this medium was therefore used in all succeeding work.
`In the inhibitive range many salts are rendered far less active in a
`peptone medium (Winslow and Dolloff, 1928) so that our results
`cannot be directly compared with those of Hotchkiss.
`
`
`
`52
`
`C.-E. A. WINSLOW AND ELOISE T. HAYWOOD
`
`The salts used were all chlorides and were Baker Analyzed
`products. They were made up in convenient concentrations
`with sterile distilled water and stored in glass-stoppered bottles.
`All glassware, except that used for plating, was Pyrex and was
`allowed to stand at least twenty-four hours in cleaning solution,
`rinsed in hot water and in distilled water and sterilized at 180°C.
`for two hours.
`In making our tests, the organism was grown for twenty-four
`hours on nutrient agar, washed from the slant with distilled
`water and then washed three times by centrifugalization. A
`suspension of the organism was then made up to contain approxi-
`Nine cubic
`mately ten million organisms per cubic centimeter.
`centimeters of the test solution, containing a mixture of Dolloff
`medium and salt solution adjusted to give the desired final con-
`centration, were inoculated with 1 cc. of this bacterial suspension,
`so that the initial concentration of organisms at the beginning of
`an experiment was about one million per cubic centimeter.
`The suspensions thus prepared were incubated for 44 to 48
`hours at 37°C., when plates were made in triplicate and colonies
`counted after 48 hours at 37°C.
`The incubation period of 48 hours was selected after consider-
`able preliminary experimentation with eight different salts.
`In
`these early studies the origimal suspension contained 20 to 50
`million bacteria per cubic centimeter.
`In the Dolloff medium
`without added salts, the number rose to over 200 million after
`24 to 48 hours and then fell to some 80 million after 144 hours.
`In favorable salt solutions the numbers rose to perhaps double
`their respective salt-free controls, while in unfavorable solutions
`the numbers fell off rapidly, in some cases reaching sterility after
`Stimulating effects were manifest in about the same
`48 hours.
`degree at all the different time intervals (24, 48, 72, 96 and 144
`hours); but slightly toxic salt concentrations tended to lose their
`inhibitive power after 48 hours, perhaps as a result of adaptation
`of the organisms to their menstruum.
`This phenomenon was
`marked in 0.05 M CaC12, 0.25 MgCl2 0.1 M LiCl and 0.0005 M
`ZnCl2. For this reason, 48 hours was chosen as our standard
`test period since at this time the salt effects were most sharply
`contrasted.
`
`
`
`SPECIFIC POTENCY OF CERTAIN CATIONS
`
`53
`
`Hydrogen ion determinations were made both before and after
`incubation by the electrometric method, using a Leeds and
`Northrup student's potentiometer with quinhydrone electrodes.
`Differences in reaction were not important under the conditions
`of this study. The Dolloff medium without added salts had a
`pH of 5.5 and remained at about that level. With the added
`salts the pH was a little higher, lying in the range 5.5 to 6.3 in
`72 out of 80 experiments at the beginning and varying somewhat
`Variations in bacterial numbers were
`more widely at the end.
`not, however, correlated with differences in hydrogen ion concen-
`tration.
`
`TABLE 1
`Effect of various dilutions of NaC7 upon viability of Es. coli in Dolloff medium
`
`NaCi
`BACTERIA IN MILLIONS PER CUBIC CENTIMETER
`MOLAL-
`ITY
`
`SALT-FREE
`. CONTROL
`
`PER CENT
`S~~~~~~~~~~~
`SURVIVAL AS
`COMPARED
`
`p
`
`1.0
`0.5
`0.25
`0.10
`0.08
`0.05
`0.01
`0.005
`0
`
`0
`59
`
`142
`119
`97
`
`0
`26
`
`108
`164
`87
`
`67
`
`87
`
`0
`30
`129
`164
`195
`226
`178
`202
`185
`
`0
`4
`154
`172
`191
`224
`173
`166
`157
`
`0
`0
`0
`0
`0
`0
`42
`75
`28
`28
`65
`3
`79 130 104 161 129
`146
`147 182 147 110 202 153
`182 149 179 133 240 172
`92 191 140
`81 115 151
`74 134
`82 125 142
`231
`79 114 126
`66 132
`49 114 112
`95
`119 136
`
`0
`37
`115
`137
`154
`125
`127
`113
`
`5.1-5.6
`5.4-5.7
`5.5-6.1
`5.5-6.4
`5.5-6.8
`5.5-6.9
`5.5-6.6
`5.6-6.8
`5.5-5.5
`
`EFFECTS OF THE NINE CATIONS STUDIED
`The type of results obtained may be indicated by a single com-
`ple*te protocol presented in table 1.
`It will be noted that the
`number of duplicate determinations made at a particular dilution
`varied from 6 to 9 in this particular case. With many dilutions
`of other salts the number of duplicate determinations ran up to
`10, 11 or 12. As in most bacteriological work, the variation
`between series is considerable but the general uniformity of the
`average results indicates that these random errors were reason-
`ably well eliminated by the number of series averaged.
`
`
`
`54
`
`C.-E. A. WINSLOW AND ELOISE T. HAYWOOD
`
`The average results, expressed for each concentration of each
`salt as a per cent of the number of bacteria present in the salt-free
`control, are presented in table 2 and in figure 1.
`All tests were
`made in Dolloff medium and the counts were made after 44 to 48
`hours at 37°C.
`
`TABLE 2
`Survival of bacteria in 8alt solutions of various 8trengths as compared with salt-free
`control (per cent)
`
`MOLALITY
`
`NaCI
`
`KCI
`
`IACI
`
`BaCh2
`
`MgCb
`
`CaCIl
`
`MnCh
`
`ZnClg
`
`CdCh
`
`0
`37
`115
`137
`154
`125
`
`127
`
`113
`
`0
`91
`102
`115
`
`155
`140
`127
`
`104
`
`.0O
`16
`51
`66
`111
`156
`121
`
`93
`
`40
`54
`
`123
`187
`199
`
`170
`162
`96
`
`56
`61
`87
`89
`119
`149
`140
`128
`121
`
`114
`
`36
`
`62
`
`159
`192
`176
`
`159
`
`142
`
`93
`
`1.0
`0.5
`0.25
`0.1
`0.08
`0.05
`0.025
`0.01
`0.008
`0.005
`0.0025
`0.001
`0.0008
`0.0005
`0.00025
`0.0001
`0.00008
`0.00005
`0.000025
`0.00001
`0.000005
`0.000001
`
`15
`25
`28
`46
`87
`151
`147
`137
`120
`
`0
`
`57
`135
`154
`208
`191
`142
`105
`86
`
`84
`
`87
`
`102
`104
`111
`121
`117
`91
`83
`
`It will be noted from inspection of the table and the curves that
`all the cations studied show the saxme general phenomena. As
`the concentration of salt increases from a minimum, there is first
`an increasing stimulation of development (as measured by the
`count after 48 hours). As the salt is further increased, the num-
`bers fall off again.
`This we have called the zone of decreasing
`stimulation.
`Finally, as the salt content becomes even higher
`we enter a zone of toxicity in which the number of bacteria is
`
`
`
`________._.
`
`SPECIFIC POTENCY OF CERTAIN CATIONS
`
`55
`
`below that of the control,-reaching a condition of sterility with
`the highest salt concentrations.
`This is the zone of toxicity.
`The point between the zone of diminishing stimulation and the
`zone of toxicity where bacterial counts are approximately the
`same as those of the salt-free control, we have called the cross-over
`point.
`For quantitative comparison of the individual cations we have
`read off from the curves of figure 1 the concentrations correspond-
`
`.E:~'
`
`cw
`
`MOLALITY
`
`TI.E___
`
`zJ
`
`j
`
`N
`
`/ \-
`Sl_ _a E=
`-.,_
`-0
`
`1=
`
`__~~~~~~~~~~~~~~~~~~~1
`
`I-
`I.- 1.\S
`/1
`z
`
`U I.
`
`l.
`
`wg
`
`0.001
`
`0A
`
`0OJ
`
`0.0000
`
`0.00001X,
`
`900
`MOLLITY
`MiOLALITY
`FIG. 1. RELATION BETWEEN SALT CONCENTRATION AND SURVIVAL OF BACTERIA IN
`DOLLOFF SOLUTION AFTER 48 HOURS
`Bacterial counts are expressed in percentages of the number present in a salt-
`free control (abscissa corresponding to 0.001 molal solution indicated by heavy
`line).
`
`ing to counts equal to 125 per cent of the salt free control in the
`zone of increasing stimulation, to the mid-point of the zone of
`increasing stimulation, to the point of maximum stimulation, to
`points corresponding to counts equal to 150 and 125 per cent
`of the control in the zone of decreasing stimulation, to the cross-
`over point and to points corresponding to 75 and 50 per cent
`of the control in the zone of toxicity. The concentrations having
`
`
`
`56
`
`C.-E. A. WINSLOW AND ELOISE T. HAYWOOD
`
`In computing
`these comparable effects are presented in table 3.
`the mid-point of the stimulating zone for NaCl, the abnormally
`low point at 0.05 M has been omitted.
`It will be noted that the nine salts studied fall naturally into
`four groups, which are indicated in the four subdivisions of figure
`1. The scales in these four subdivisions are the same but the
`actual points on the logarithmic abscissa are different, the abscissa
`corresponding to 0.001 M concentration being indicated by a heavy
`It will be noted that sodium and potassium
`line in each instance.
`
`TABLE 3
`Summary of molal concentrations of various salts producing certain effects upon
`bacterial viability
`
`ZONE OF INCREASING
`
`ZONE OF DECREAS-
`MAXIMUMINSTMLIO
`OVECR
`CROSS-
`ING STMULATIONPOINT
`STIMULATIN
`STIMUATION
`
`ZONE OF TOXICITY
`
`Percentage of salt-free control
`
`125
`
`Midpoint
`
`121-208
`
`150
`
`125
`
`100
`
`75
`
`50
`
`0.17
`0.08
`0.28
`0.36
`0.09
`0.01
`Na. 0.009
`K.
`0.01
`0.29
`0.08
`0.05
`0.05
`0.56
`0.009
`0.07
`0.04
`0.06
`0.03
`0.02
`Li.0.01 0.01
`Ba .
`0.04
`0.06
`0.001
`0.002
`0.08
`0.01
`0.05
`0.07
`0.02
`0.01
`Mg.0.002 0.002
`0.008
`0.008
`0.07
`0.04
`Ca. . 0003
`0.03
`0.02
`0.0006
`0.008
`0.01
`0.0003 0.0003 0.0004 0.0006 0.0008
`0.0002
`0.00006 0.00006
`Mn........
`0.00008 0.0001 0.0003 0.0003 0.0004 0.0005
`Zn........ 0.00002 0.00004
`0.000008 0.00002
`0.0001
`Cd........
`
`0.44
`0.68
`0.10
`0.13
`
`are essentially identical in their effect. Barium and lithium are
`effective in lower dilution (both in stimulation and toxicity),
`their curves lying to the left of those for sodium and potassium.
`Calcium and magnesium form a still more potent pair and man-
`ganese, zinc and cadmium are some hundreds of times more
`powerful than sodium and potassium, their curves lying almost
`wholly below the 0.001 M concentration, while the curves for the
`other cations studied lie almost wholly above this point.
`For the most part, the results obtained check fairly well with
`those obtained in earlier studies.
`
`
`
`SPECIFIC POTENCY OF CERTAIN CATIONS
`
`57
`
`The point of maximum stimulation for NaCl was found at 0.08
`M while the stimulating range lay between 0.005 M and 0.25 M.
`These results correspond with those of Winslow and Falk (1923a)
`who found 0.01 NaCl favorable and 0.7 M NaCl toxic and with
`those of Winslow and Dolloff (1928) who found a range of stimula-
`tion between 0.0001 M and 0.3 M with a maximum at 0.05 M.
`Fabian and Winslow (1929) reported stimulation extending up to
`Hotchkiss (1923) found a higher
`0.3 M with a maximum at 0.1 M.
`concentration associated with maximum growth (0.25 M) but,
`as pointed out above, the peptone medium and the 3-day incuba-
`tion period which she used require a higher salt concentration to
`yield a given result.
`With KCl, the stimulating range is the same as that for NaCl
`This maximum
`(0.005 to 0.25 M), the maximum falling at 0.05 M.
`point is lower than that recorded by either Winslow and Dolloff
`(0.2 M) or Hotchkiss (0.25 M).
`The stimulating zone for LiCl begins at about the same point
`as that of Na and K (0.006) but is much narrower, having its
`maximum at 0.025 M and ending at about 0.05 M, beyond which
`These values as usual are slightly lower than
`the salt is toxic.
`those of Hotchkiss.
`The stimulating zone for BaC1 begins at a very low concentra-
`tion (0.001 M), has a maximum at 0.01 M and ends about where
`that for Li does, at 0.06 M.
`This corresponds closely with the
`Hotchkiss figures.
`MgCl2 and CaCl2 are closely alike. Both have stimulating zones
`beginning near 0.0001 M with maxima at 0.008 M and extending
`to about 0.02 M, beyond which point both become toxic.
`For
`Mg, these figures correspond with those of Winslow and Dolloff
`who found maximum stimulation at 0.003 M. As usual, Hotch-
`kiss' results in peptone solution show a limitation of salt potency,
`the range 0.0025 to 0.1 M being stimulating in her study.
`For
`calcium, the constants reported by various observers differ more
`Thus, Hotchkiss reported maxi-
`widely than for any other salt.
`mum stimulation at 0.15 M and stimulation up to 0.25 M while our
`It is
`maximum is at 0.008 M and our cross-over point at 0.03 M.
`true that the peptone medium and the long incubation may
`
`
`
`58
`
`C.-E. A. WINSLOW AND ELOISE T. HAYWOOD
`
`account for this but Winslow and Falk (1923a) using water found
`a cross-over point at 0.14 M and Winslow and Dolloff (1928) one
`at 0.1 M.
`MnCl2 is stimulating between perhaps 0.00003 and 0.0004,
`This corresponds with the Hotch-
`with a maximum at 0.0002.
`kiss results as does the curve for ZnCl2, showing stimulation
`between 0.00001 and 0.0003, with a maximum at 0.00008, CdCl2,
`with a stmulating range between 0.00001 M and 0.0001 M, was
`less toxic than reported by Hotchkiss.
`
`TABLE 4
`Specific potency of various catiors
`
`ZONED OF
`INCREASING
`STIMULATION
`
`MX
`
`MAXU-
`
`ZONE OF
`DECREASING
`STIMULATION
`
`CR088-
`OVER
`POINT
`
`ZONE OF
`
`I
`
`AERE
`AVZRAGZ
`
`Percentage of salt-free control
`
`125
`
`Midpoint
`
`150
`
`125
`
`100
`
`75
`
`50
`
`1
`1
`1
`Na..........
`1
`1.6
`1
`K...........
`4.0
`Li..........
`0.9
`1
`8.0
`5.0
`9
`Ba..........
`Mg..........
`4.5
`5.0
`10
`17
`Ca..........
`10
`30
`170
`400
`150
`Mn..........
`1,000
`250
`450
`Zn..........
`Cd.1,200 4,000
`
`1
`1.8
`3.0
`2.3
`11
`9.0
`300
`900
`
`1
`2.1
`4.3
`3.4
`17
`8.5
`600
`600
`
`1
`1.0
`4.7
`4.7
`14
`9.3
`700
`900
`2,800
`
`1
`0.6
`5.1
`4.5
`5.1
`9.0
`600
`900
`
`1
`0.6
`4.4
`3.4
`
`6.3
`600
`900
`
`1
`1.2
`3.4
`5.0
`9.4
`12.0
`400
`700
`3,000
`
`On the whole these results seem reasonably consistent with
`earlier findings and the relative order of the salts from the stand-
`point of both stimulating and inhibiting effects, is the same in all
`studies.
`In order to bring out more clearly the relative effect of the
`individual cations indicated in table 4 we have computed their
`specific potencies for each of the points included in table 3 as
`reciprocals of the ratios of the concentration of a given salt to the
`amount of NaCl necessary to produce the same effect.
`The average specific potencies are, then, as follows, taking the
`potency of Na as 1; K, 1.2; Li, 3.4; Ba, 5.0; Mg, 9.4; Ca, 12; Mn,
`400; Zn, 700; Cd, 3000.
`
`
`
`SPECIFIC POTENCY OF CERTAIN CATIONS
`
`59
`
`It seems clear from the table that the concept of a specific
`potency, characteristic of each cation, is a valid one. The explana-
`tion of this phenomenon is still obscure; but it seems highly
`probable that the action of the cations may be explained on
`Bancroft's theory of disinfection (Bancroft and Richter, 1931)
`as related to coagulation of colloids. The effects of cations upon
`the coagulation of such colloids as sulphur and mastic show
`specific potencies for the various cations varying in somewhat
`the same orders of magnitude (Bancroft, 1921) although the
`relative potency of the various cations is widely different from
`Bancroft, however, states that in such
`that observed by us.
`coagulations "the fundamental rule is that the adsorption is
`specific both as regards the adsorbing substance and the ion
`adsorbed."
`In the Winslow and Dolloff study comparison was made not on
`the basis of molality but of ionic activity.
`Similar computations
`were made in the present study, using the tables of Lewis and
`Randall (1923). The differences in specific potency as computed
`on the basis of molality and of ionic activity did not, however,
`differ materially. The figures for calcium and magnesium and
`barium were slightly increased, those for zinc and cadmium
`Since the differences were insignificant and
`slightly decreased.
`the application of the Lewis and Randall constants seems of
`doubtful validity in the relatively complex medium used, we have
`considered comparison on the basis of molality the soundest basis
`available.
`
`EFFECTS OF MIXTURES OF THE CATIONS STUDIED
`As a check on the soundness of the theory of specific potencies
`we planned a second series of experiments in which mixtures of
`various salts were prepared and the effect upon bacterial viability
`determined in order to see whether the actual results would con-
`Five concentrations of
`form to those predicted from the theory.
`each salt were chosen, which lay on the descending side of the
`curves of figure 1 and which would by themselves give counts
`corresponding to 150 and 125 per cent of the salt-free control in
`the zone of diminishing stimulation, to the cross-over point, and
`
`
`
`60
`
`C.-E. A. WINSLOW AND ELOISE T. HAYWOOD
`
`to 75 and 50 per cent of the control in the zone of toxicity.
`For
`NaCl and KCl five mixtures were prepared as follows; one con-
`taining the 150 per cent concentration of NaCl (0.09 M) and the
`50 per cent concentration of KCl (0.68 M); one containing the
`125 per cent concentration of NaCl (0.17 M) and the 75 per cent
`concentration of KCl, (0.56 M); one containing the cross-over
`
`TABLE 5
`Survival of bacteria in salt mixtures expressed as per cent of number in salt-free
`control
`
`Bacterial count
`corresponding NaCI...
`150
`to concentra- Second salt-.. 50
`tion of
`
`125
`75
`
`100
`100
`
`SALTS PRESENT
`
`NaCl, Kl.84 87
`NaCl, Lil.87 91
`NaCl, BaC2.108 97
`NaCl, MgC1 .117 103
`NaCl, CaCl2
`117
`133
`.
`NaCl, MnC2.
`65
`89
`NaCl, ZnCl2
`87
`86
`.
`
`75
`125
`
`109
`83
`98
`111
`108
`92
`94
`
`75
`125
`
`125
`77
`135
`88
`99
`49
`
`50
`10
`
`124
`86
`87
`105
`91
`90
`94
`
`50
`150
`
`98
`68
`118
`96
`84
`39
`
`100
`99
`97
`113
`109
`98
`88
`
`100
`100
`
`105
`67
`155
`89
`101
`37
`
`125
`75
`
`Bacterial count 1
`corresponding L CaCs.150
`to concentram- Second salt
`50
`j
`tion of
`CaCl2, K .l 108
`109
`CaCl2, LiCl. 46
`59
`CaCl2, BaC .167 155
`CaCi2, MgCl .84 97
`CaCi2, MnC12
`108
`103
`CaC12, ZnCl..43
`65
`Figures which deviate by more than
`boldface.
`
`25 per cent from the expected value are
`
`concentration of each salt (Na, 0.28M, KO.29M); onecontainingthe
`75 per cent concentration of NaCl (0.36 M), and the 125 concen-
`tration of KCl (0.08 M) and one contaiping the 50 per cent concen-
`tration of NaCl (0.44M) and the 150 per cent concentration of KCl
`(0.05 M).
`In each case the two solutions were mixed in equal pro-
`portions so that if the theoryof specific potencies held, and no other
`
`
`
`SPECIFIC POTENCY OF CERTAIN CATIONS
`
`61
`
`phenomena intervened, the final counts should in all cases approxi-
`mate that of the salt-free control.
`According to the theory a half
`and half mixture of two equipotent salts should produce the same
`effect as either undiluted salt alone and a similar mixture of two
`salts having equal but opposite effects should have the same effect
`as a solution of either salt of a strength half way between the two
`extremes.
`Mixtures were made in this way of NaCl with each of
`Y OF
`SALT FREE
`
`CAM
`
`NACA
`
`NA LIG
`
`140
`
`Shi120
`
`CAJMCOC:Oo
`
`.
`
`2~~~~~~0
`
`AT:ISX
`I1ST
`S0
`12S
`7S
`100
`2ND SALT. S0
`75
`100
`1SO
`12S
`FIG. 2. SUJRVIVAL OF BACTERIA IN VARIOUS SALT MIXTURES EXPRESSED AS PER
`CENT OF NUJMBER IN SALT-FREE CONTROL
`In each experiment a solution of sodium or calciulm chloride was mixed with a
`solution of the chloride of some other cation. The upper line of figures at the
`bottom of the chart represents the relative count which would be obtained from
`the original solution of the first salt (sodium or calcium chloride); the lower figure
`represents the count which would be obtained in presence of the original concen-
`tration of the second salt used.
`
`the other salts and of CaCl2with each of the other salts (except
`CdCl2) and 5 to 7 duplicate determinations were made for each
`mixture. The average results are summarized in table 5 and in
`figure 2.
`Before discussing the results presented in table 5 it should first
`be made clear just what effect would be expected from the single
`salts used in these mixtures if acting alone. The "Bacterial
`counts corresponding to concentrations of salt" which head table
`
`
`
`62
`
`C.-E. A. WINSLOW AND ELOISE T. HAYWOOD
`
`5 refer to the counts obtained from the solutions before mixture
`but mixture of course diluted each one-half. When we mixed
`cross-over concentrations of NaCl and CaCl2, for example, each
`of the solutions by itself would have produced a count equal to
`the control. The actual solutions mixed were 0.28 M NaCl and
`In mixing, however, each of these solutions was
`0.02 M CaCl2.
`diluted one-half. The actual amount of NaCl used (0.14 M)
`would by itself have produced a count of 130 per cent of the con-
`trol and the actual amount of CaCl2 (0.01 M) a count 135 per cent
`of the control. The individual diluted salt solutions reviewed in
`table 5 would in all but six instances by themselves have produced
`stimulation.
`Thus, a mixture of two stimulating salts in these
`experiments produces no stimulation.
`This is the sort of phenom-
`enon often described as antagonism, but it is clearly only an
`additive effect of two different cations, each exerting its inde-
`pendent specific potency effect.
`According to our theory, on the other hand, the effect of a salt
`mixture should be predicted by multiplying the concentration of
`each cation by its specific potency and then adding the two results.
`Thus, in the mixtures of NaCl and CaCl2 for example the total
`specific potencies in terms of Na would be as follows; correspond-
`ing to the five columns in table 5.
`Most stimulating.Na, most inhibitive Ca:
`0.045mNa+ 12 X 0.035mCa - 0.46mNa
`Stimulating Na, inhibitive Ca:
`0.085mNa+ 12 X 0.02m -0.32 mNa
`Neutral Na, neutral Ca:
`0.14mNa+ 12 X 0.015mCa - 0.32mNa
`Inhibitive Na, stimulating Ca:
`0.18mNa+ 12 X 0.01 Ca = 0.30mNa
`Most inhibitive Na, most stimulating Ca:
`0.22 mNa+ 12 X 0.005Ca = 0.28MNa
`Thus, the combined effect of the two salts in the mixture sbould
`be equal to a NaCl of about 0.3 M strength which would corre-
`spond on the Na graph of figure 1 to a count of about 90 per cent of
`the salt-free control.
`Similarly, we may compute the same salt combinations in
`terms of Ca by dividing each actual concentration of Na by 12.
`
`
`
`SPECIFIC POTENCY OF CERTAIN CATIONS
`
`63
`
`This gives for the five points total potencies corresponding to
`0.023,0.025, 0.027, 0.027, and 0.039 M CaCl2 respectively an average
`of 0.028 Ca, a concentration of the calcium curve in figure 1 corre-
`sponding to a count just equal to the salt-free control.
`Such combined potencies have been computed for all of the
`sixty-odd salt mixtures used and most of them give results lying
`close to the cross-over concentrations of Na and Ca.
`Returning
`to table 5 we note that the results for mixtures of NaCl with other
`salts are about what would be expected from the specific potency
`There are only two figures in the upper half of the table
`theory.
`which fall above 125 or under 75 per cent.
`These exceptions,
`which are boldfaced in the table, are a high value for one mixture
`of NaCl and CaCl2 and a low value for one mixture of NaCl and
`MnCl2.
`The mixtures of CaCl2 with KCI, MgCl2 and MnCl1 also run
`close to expectation. The mixtures of CaCl2 with LiCl and with
`ZnCl2 on the other hand show very low values throughout and the
`mixtures of CaCl2 with BaCl2 show very high values throughout.
`The latter may perhaps have been influenced by precipitation but
`the low values with Li and Zn are puzzling and suggest some
`action differing from the usual specific potency effect. We have
`at present no explanation to offer for this phenomenon.
`
`SPECIFIC POTENCY AND SALT ANTAGONISM
`The peculiar effect of mixtures of CaCl2 with LiCl, BaCl2 and
`ZnCl2 furnishes a salutary warning against any generalization
`which tends to over-simplify the phenomena of salt action. Yet
`the general validity of the specific potency principle seems
`established and it is tempting to speculate as to the extent to
`which this principle may explain the effect described as "salt
`antagonism."
`The concept of salt antagonism implies a specific neutralization
`It seems beyond question to
`by one salt of the effect of another.
`occur when the cells and tissues of animals, such as starfish eggs
`and mammalian muscle tissup, are exposed to salt mixtures.
`In interpreting the effect of salt mixtures upon bacteria the
`phenomena of specific potency must however be kept in mind,
`
`
`
`64
`
`C.-E. A. WINSLOW AND ELOISE T. HAYWOOD
`
`and we must also be quite clear as to the difference between
`mixing two salt solutions of known strength (which involves
`dilution of each) and the addition of solid salt to a solution of
`another salt (which does not involve dilution).
`It is obvious that
`if we take two different concentrations of the same salt and mix
`them the effect will be the same as that of an intermediate con-
`centration. What that result will be, however, will depend on the
`particular part of the potency curve at which the concentrations
`used may lie. Thus from figure 1, it appears that if we mix two
`concentrations of NaCl, both of less than 0.05 M strength, we
`shall get a stimulating effect intermediate between that of the two
`concentrations used (since both lie in the zone of increasing stimu-
`lation).
`If, however, we mix a concentration lying in the zone of
`increasing stimulation (say 0.01 M) with a concentration in the
`zone of decreasing stimulation (say 0.1 M) we shall obtain a
`greater stimulation than that given by either primary concentra-
`tion alone since the mixture will correspond to the point of maxi-
`mum stimulation.
`If we mix a concentration lying in the zone of
`diminishing stimulation (say 0.1 M) with a concentration in the
`zone of toxicity (say 0.44 m) we shall obtain a neutralization of
`effects. The last two are just the sort of phenomenon often
`described as antagonism when two salts are used. Yet with one
`salt alone it is clearly not antagonism but addition which is
`taking place. The results reported in preceding pages like those
`of Winslow and Dolloff (1928) show that when different salts are
`used the phenomena often follow the same law and, when they do
`so, the assumption of antagonism is superfluous.
`In other studies of so-called antagonism, instead of mixing two
`salt solutions (and thus diluting each) a second salt is added to a
`solution of the first salt, keeping the concentration of the first
`Here the problem is simpler but the result will still
`unchanged.
`be largely determined by the part of the specific potency curve in
`which the addition takes place.
`Thus, if we start with a salt concentration lying in the zone of
`increasing stimulation the addition, of a small amount of another
`cation will push the total cation concentration up to the point of
`maximnum stimulation. A larger addition will carry the total
`
`
`
`SPECIFIC POTENCY OF CERTAIN CATIONS
`
`65
`
`concentration over to the zone of decreasing stimulation or the
`zone of toxicity.
`If we start with a salt concentration giving
`maximum stimulation the addition of any other cation will carry
`the total concentration into the zone of decreasing stimulation or
`toxicity; so that starting at this point any salt will appear antago-
`nistic to any other salt, even if the concentration of the second
`salt added were itself stimulating in effect.
`The only thing that cannot occur according to the principle of
`specific potency is the neutralization of toxic effect by the actual
`addition of any cation to a solution already toxic (without dilu-
`tion).
`This is the critical test, since according to the theory of
`antagonism each cation exerts its characteristic effect uninflu-
`enced by the other while according to the uncomplicated effects
`of specific potency, the addition of any amount of a second cation
`should increase the toxic effect of the first.
`One of the clearest cases of such true antagonism was presented
`by Winslow and Falk (1918) with regard to Na and Ca.
`Solu-
`tions of 0.6 M NaCl and 0.1 M CaCl2 were highly toxic but a solution
`containing both thes