Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`A STUDY OF THE BACTERICIDAL ACTION OF ULTRA
`VIOLET LIGHT
`
`III. TuE ABSORPTION OF ULTRA VIOLET LIGHT BY BACTERIA
`
`BY FREDERICK L. GATES
`
`(From the Laboratories of The Rockefeller Institute for Medical Research, New Yark,
`and the Department of General Physiology, Harvard University, Cambridge)
`
`(Accepted for publication, June 10, 1930)
`
`In this study of the bactericidal action of ultra violet light the first
`paper (1) described the reaction of an 18 hour culture of Staphylococ(cid:173)
`cus aureus to monochromatic radiations. It was shown that the course
`of the reaction among large numbers of organisms was approximately
`the same at each wave length studied but that widely different incident
`energies were required at different wave lengths to produce these
`similar effects.
`The second paper (2) discussed the limits of the bactericidal zone,
`showed that the reaction had a low temperature coefficient, (approxi(cid:173)
`mately 1.1), gave evidence that within the variations of the methods
`used no significant errors were introduced by differences in the meas(cid:173)
`ured intensity of the source or in the hydrogen ion concentration of
`the medium, and indicated that plane polarization of the incident
`light had no effect upon the reaction.
`The present paper deals with the absorption of ultra violet light
`by intact bacteria. A final paper in this series will discuss structural
`and chemical units of bacterial protoplasm that may prove to be in(cid:173)
`volved in the reaction which results in the organism's death.
`
`Incident Energy Relationships at Various Wave Lengths
`Text-fig. 1, reproduced from the first paper of this series, shows that
`although the course of the bactericidal reaction was approximately
`the same at each wave length studied, these similar curves were found
`at very different incident energy levels at different points in the ultra
`violet 'Spectrum.
`
`31
`
`The Journal of General Physiology
`
`1
`
`EXHIBIT 1014
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`TEXT-FIG. 1. Incident energies required for bactericidal action at various wave lengths in the ultra violet spectrum
`
`to"'
`
`I
`
`IP '
`.J 14 5 6' 6
`
`,!
`
`f'RG.51'-Rmm.1
`
`,,.S6 41
`
`IQ'
`.J
`
`s
`~
`....
`t'"'
`~ H
`
`~
`~
`~ H
`
`V
`IV
`
`V
`J
`V
`
`/
`
`i/1,/
`
`N
`v,)
`
`~
`
`/.,,.
`
`~
`~
`
`I
`/~
`/
`
`~
`
`V
`
`j
`
`I
`I
`y
`
`/
`
`/.I
`/
`
`J
`
`I
`l/
`,
`
`J
`
`I
`
`~ .,... I,' 'ii' V
`
`I
`
`J
`
`/
`
`~
`
`,
`/
`
`:; ~ •
`.
`~ '<) 9Q ~ ~. -.
`
`..., °') "
`
`-
`
`-
`
`-
`
`-
`
`WAYl' L&l'fUTtr-1 NV ll7,I,[
`
`11\a t\, "' &\I
`
`I
`
`V
`I
`
`/
`
`I
`I V
`/, V I
`A
`,;/ I/ V I
`I/
`/; 1/ / I I
`
`'/; VJ
`/2~ ~v I
`~W/
`II
`
`/
`
`I.~
`
`~/
`
`2
`·/
`/
`
`-l.1'6 81
`
`/
`
`~ A V
`
`IP'
`J
`
`,!
`
`I
`V ~ ~ /
`,I
`I
`J v ,
`I) I I
`,I
`
`/
`
`I
`
`Q
`
`IQ
`
`2Q
`
`~
`..,.
`~ ,4(J
`~
`~ SQ
`\j
`~ 617
`
`JQ
`
`7()
`
`/()
`
`.9()
`
`/()0
`
`2
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`FREDERICK L. GATES
`
`33
`
`For example, if the incident energy required to kill half the exposed
`staphylococci be taken as an index, this energy requirement ranged
`from 3,150 ergs per mm. 2 at 302 m.µ, a wave length near the limit of
`bactericidal action, to 88 ergs at 266 m.µ. At other wave lengths,
`either above or below 266 m.µ, more incident energy was required.
`The destruction of 50 per cent of the microorganisms is chosen as the
`index because it is in the most accurately determined part of the curves
`where the mortality rate is least affected by variations in individual
`resistance.
`If the incident energies involved in 50 per cent destruction be plotted
`and joined by a continuous line the resulting curve appears as in Text(cid:173)
`fig. 2A. Parallel experiments on an 18 hour culture of Bacillus coli gave
`bactericidal energy curves similar in trend to those for S. aureus, and
`although complete statistics were not obtained on this bacterium the
`middle or exponential portion of the lethal reaction curves was deter(cid:173)
`mined by repeated observations at each wave length studied. Text(cid:173)
`fig. 3A shows the incident energies involved in the destruction of 50
`per cent of the exposed coli organisms.
`Its essential similarity to the
`corresponding aureus curve is apparent. Both curves would probably
`be somewhat modified in detail if more wave lengths were available for
`study.
`These characteristic curves (Text-figs. 2A and 3A) show clearly that
`less incident energy is required between 260 and 270 m.µ than in any
`other region of the bactericidal zone examined, and point toward a
`second minimum below 230 m.µ. The presence of a sharp peak in the
`energy requirement near 240 m.µ appears to be equally significant.
`Due apparently to the use of but a few wave lengths in the bactericidal
`zone, or to failure to measure spectral intensities, or to crude,methods
`of estimating bacterial destruction, the occurrence of a minimum at
`about 266 m.µ, and of the peak in the curve near 240 m.µ has been
`overlooked by most investigators. Usually they have been content
`with the conclusion that the shorter the wave length the more marked
`the bactericidal action (3).
`
`Bang (4) was apparently the first to observe regions of special bactericidal
`"effectiveness" or of corresponding bacterial susceptibility. Using a 30 ampere
`carbon arc through a spectrograph at 20 m.µ intervals, and estimating bactericidal
`efficiency mainly by relative exposure, he found two regions of maximal action, an
`
`3
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`34
`
`ULTRA VIOLET LIGHT.
`
`III
`
`"inner maximum" between 340 and 360 m.µ, and an "outer maximum" between
`240 and 260 m.µ. The lethal exposure varied from 1920 seconds at 330 to 300
`m.µ through 120 seconds at 300 to 280 m.µ to 4 seconds at 280 to 260 m.µ and 2
`seconds at 260 to 240 m.µ. Then longer exposures were required. The zone from
`240 to 220 m.µ needed 20 seconds, that between 220 and 210 m.µ, 30 seconds, and
`that between 210 and 200 m.µ required 120 seconds exposure to kill the organisms.
`
`I
`8
`6
`'o4
`-3
`2
`
`I
`8
`..,.
`6
`::e
`:::EN 4
`0
`ex: -3
`w
`Cl.. 2
`
`I
`8
`6
`- 4
`Q3
`
`2
`
`I
`I
`V
`
`/
`
`,...--.
`[\_
`
`I
`
`-f--.
`
`✓
`
`'
`
`" ' ~
`
`\
`
`\
`\
`
`I
`8
`6
`4"/
`32
`2
`
`I
`8
`6
`4~
`39
`2
`
`-~2 .23 .24 .25 .26 .27 .28 .29 .30 .31 .22 .23 .24 .25 .26 .27 .28 .29 .30 .3 J'
`IN J.L·
`A WAVE LENGTHS
`8
`TEXT-FIG. 2. A. Curve of incident energies involved in the destruction of 50
`per cent of S. aureus.
`B. Curve of the reciprocals of 2A.
`
`These longer exposures at short wave lengths were evidently due to the rapid
`decrease in intensity of the source employed. Mme. Henri (3) thought that both
`of Bang's maxima should be ascribed to variations in intensity of his carbon arc.
`Newcomer (5) exposed B. typhosus in quartz capillary tubes to narrow bands of
`the iron arc spectrum and counted surviving bacteria plated in agar, after exposures
`of 5 or 10 minutes. His figures, like Bang's, indicate a region of maximum effect
`
`4
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`FREDERICK L. GATES
`
`35
`
`between 254 and 268 m.µ, but his iron arc varied so much in spectral intensity
`that he did not find any significance in this peak of effectiveness. He concluded
`If there is
`that "equalintensities produce equal effects in the regions 2100-2800.
`a maximum in this region it is at most only slight and would be in the neighborhood
`of 2600."
`When Mashimo (6) varied the exposure of bacteria in a spectrograph from 15
`seconds to 80 minutes he found the first evidence of bactericidal action at 275 m.µ.
`
`'
`
`T
`
`.
`
`I
`I
`
`)
`
`I
`
`/1.---o. I',.
`
`r
`\
`\ I/
`.,
`
`,
`
`\
`
`I\
`
`I
`
`\
`\
`\
`
`I\ \
`
`l
`
`I
`8
`6
`
`I
`8
`6
`4~
`32
`2
`
`...
`
`I
`8
`6
`~"' 4
`~Q3
`a:
`~ 2
`
`V)
`c.::,
`I
`~ 8
`6
`- 4
`~3
`2
`
`I
`I
`.22 .23 .24 .25 .26 .27. 28 .29 .30 .31
`.2223 24 .25 .26 .27 .28 29 .30 .31
`WAVE LENGTHS IN JJ.
`A
`8
`TEXT-FIG. 3. A. Curve of incident energies involved in the destruction of 50
`per cent of B. coli.
`B. Curve of the reciprocals of 3A.
`
`With somewhat longer exposures the zone widened rapidly so that with a 3 minute
`exposure it extended from below 210 to above 280 m.µ. The marked action at
`275 m.µ was evidently due to a relatively high intensity of his source at this wave
`length, as an examination of his published photograph shows.
`Thus variations in spectral intensity, with no adequate methods of measurement
`or control, made it difficult to interpret the maxima found by Bang,Newcomer, and
`Mashimo, and these authors laid no stress upon them.
`
`5
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`36
`
`ULTRA VIOLET LIGHT.
`
`III
`
`In the present study the measurement of the monochromatic radiant
`energy in absolute units focuses attention upon the marked differences
`in the incident energies required at different wave lengths to kill S.
`aureus and B. coli. The shape of the energy curves for 50 per cent
`destruction immediately suggests that it is the specific absorption of the
`ultra violet radiations that gives the curves significance. This relation
`of incident energy to its specific absorption is made the more striking
`by plotting the reciprocals of these energies (Text-figs. 2B and 3B),
`by which graphs not unlike absorption curves, with maxima at 266 and
`beyond 230 m.µ are produced.
`As a first step in the further analysis of the bactericidal reaction it is
`obviously necessary to compare these curves with those for the absorp(cid:173)
`tion of ultra violet energy by the corresponding bacteria.
`
`The Absorption of Ultra Violet Light by Bacteria
`Apparently studies of the absorption spectra of bacteria have been
`confined hitherto to bacteria in suspension in a fluid medium (7,
`8). Suspensions in liquids are unsuitable for such examinations.
`Reflection and refraction from the bacterial bodies, with consequent
`scattering of light so that bacterial suspensions are opaque even in the
`visible region of the spectrum, and-the difficulty of estimating the num(cid:173)
`ber of organisms traversed, frustrate any attempt to obtain results of
`quantitative significance.
`But a loopful of bacteria may be taken en masse from the surface
`of an agar slant and pressed between quartz plates into a layer so thin
`that is is all but colorless in visible light, and so transparent that
`objects may be seen through it clearly and without distortion. The
`bacteria are in optical contact and form a homogeneous medium for
`the transmission of visible or ultra violet light. Such a film is com(cid:173)
`posed almost entirely of bacterial cells and the immediate products of
`their metabolism. Tests show that films of like thickness of the agar
`medium from which the bacteria were removed absorb no significant
`amounts of ultra violet light.
`Such a film of bacteria may be set up in one optical path of a quartz
`photometer, with similar plates of quartz and a drop of glycerol in the
`other path as a control, and the absorption coefficient of the organisms
`readily obtained in the manner commonly employed for chemical solu(cid:173)
`tions (9).
`
`6
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`FREDERICK L. GATES
`
`37
`
`Two beams of light from the same source, which have passed through the speci(cid:173)
`men under examination or its control in the photometer, are spread by a quartz
`spectrograph into parallel spectra in the plane of a photographic plate. The
`energy that traverses the control path is subject to quantitative variation by
`means of a sector shutter, and with the shutter set at predetermined openings
`a series of photographs is made. Then the point, or points, of equal blackening
`in each pair of spectral photographs indicate the wave lengths at which the test
`specimen and the sector shutter in the control path have reduced the original
`intensity of the light to the same degree.*
`
`In such experiments it is necessary to know accurately the depth of
`the medium traversed in order to calculate the coefficients of absorp(cid:173)
`tion for a layer of unit thickness. The standard or unit of thickness in
`these observations was chosen as 0.8 µ, the average diameter of S.
`aureus (10), so that the coefficients of absorption were obtained for a
`single layer of bacteria. The shape of B. coli and its wide variations
`in size precluded even so crude an estimate of a "single layer," so the
`coefficients for B. coli were :figured arbitrarily for a layer of the same
`thickness (0.8 µ) to permit a comparison with the results for S. aureus.
`The thickness of these films of bacteria between quartz plates was
`found to lie between 5 and 15 µ. Since the method of measurement
`employed (11) is easy, and is accurate (in microns) to the second or
`third decimal place it may be described in brief. The method is
`based on interferometry, namely, the measurement of the interference
`bands formed by the coincident spectra of white light reflected into a
`wavelength spectroscope from the two quartz surfaces enclosing the
`bacterial film.
`
`A point source of white light (carbon arc, concentrated tungsten filament, or
`even a flashlight bulb with the filament vertical), is set up at a measured angle of
`incidence (60°) to the film specimen, so that its light is reflected from the back
`surface of the first quartz plate and from the front surface of the second (the
`surfaces enclosing the film), into a wavelength spectroscope. The reflecting sur(cid:173)
`faces must be chosen near the bacterial film, but not to include it, since adequate
`reflection occurs only when air is the medium between the two plates. With a
`proper set-up, and films of suitable thickness a series of vertical interference bands
`will cut across the spectrum.
`
`* Measurements of absorption by such films made in 1923 with thermocouple
`and galvanometer (Proc. Soc. Exp. Biol. and Med., 1923, 21, 61), were evidently
`not as accurate as those obtained by the present method and have been discarded.
`
`7
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`38
`
`ULTRA VIOLET LIGHT.
`
`III
`
`Then the distance (t) between the two plates (the thickness of the included
`film) is found by noting the wave lengths (>-1 and >-2) between which any convenient
`number (n) of interference bands is counted for substitution in the following
`equation:
`
`n X1 X2
`t= - - - - - -
`2 µ, cos r (X1 - X2)
`
`p, is the refractive index of the medium, and r is the angle of incidence and reflection.
`Since air is the medium p, = 1, and with rat 60° cos r = 0.5 so that the equation
`becomes
`
`For example, with the apparatus set up as described 5 interference bands are
`counted between X1 = 7241 Au. and X2 = 4638 Au. It is only necessary to multi(cid:173)
`ply these waves lengths together, divide by their difference, and multiply the result
`by 5 to determine that the distance between the plates in the area examined is
`6.45 p, or 0.00645 mm.
`
`In these experiments the films of bacteria, pressed out by hand,
`were not strictly plane parallel, but often thicker at one edge than
`at the opposite one, so four readings were taken at 90° intervals around
`the rim of each film and combined for an average thickness. The dif(cid:173)
`ferences in thickness in the small central area exposed in the photom(cid:173)
`eter were not so great as to make it desirable to obtain a mean expo(cid:173)
`nentially according to Lambert's law. For example:
`
`Film 1. B. coli. 90° readings at edge of film: 5.58, 6.17, 5.63, and 5.66 J.£.
`Average 5.76 µ,
`Film 2. S.aureus. 90°readings at edge of film: 9.21, 11.66, 11.24,and 10.53 P,.
`Average 10.66 µ,
`
`Given the average thickness of each film, the number of layers (n)
`of bacteria it contained was then available for the determination of the
`coefficients of absorption for a layer 0.8 µ thick. For this purpose the
`familiar equation of Lambert's law was used, to which Wood (11)
`says that no exception has ever been found that was not attributable to
`experimental error. Thus if a is the fraction of the original intensity
`(Io) transmitted by each unit layer, the intensity (/1) observed after
`passage through n layers is given by the equation:
`
`8
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`FREDERICK L. GATES
`
`39
`
`The relation of Ii to I. is chosen at appropriate intervals on the
`sector shutter of the quartz photometer, as described above, and this
`equation solved for a gives the coefficient of transmission at the
`wave length where the two spectra match. Then 1-a is the corres(cid:173)
`ponding coefficient of absorption for a single layer of bacteria 0.8 µ,
`thick.
`The mean absorption curve from five series of determinations with
`S. aureus is given in Text-fig. 4, and from five series with B. coli in
`Text-fig. 5. The curves are characteristic, and similar to those found
`for various biological tissues and fluids containing proteins or protein
`derivatives (12, 13, 14) on which their main features evidently depend.
`Differences in the curves for the two organisms are apparent, but in
`view of the experimental errors involved in such determinations they
`probably should not be stressed.
`The general similarity of these absorption curves to the reciprocals
`of the curves for bactericidal incident energy is obvious. All four
`curves rise rapidly from low levels beyond 300 m.µ to a maximum
`between 260 and 270 m.µ, then drop to a minimum near 240 m.µ and
`rise again toward a limit beyond the range of experimental observation.
`One is tempted to correct the incident energies involved in the bacteri(cid:173)
`cidal reaction by these absorption coefficients for the entire organisms
`in order to obtain an approximation of the total energies absorbed.
`Yet the two sets of curves show important points of difference also,
`especially in the location of the dip near 240 m.µ, and a closer consider(cid:173)
`ation of their relationship indicates that such a quantitative cor(cid:173)
`rection would be futile. The sum of the absorption coefficients of all
`the chemical entities in the bacterial cell cannot be expected exactly to
`correct the wide differences found in the bactericidal incident energies
`at different wave lengths unless every chemical group is involved
`in the bactericidal reaction in exact proportion to its contribution to
`total absorption. Such a hypothesis is hardly tenable. It seems
`more probable that it is the effect of ultra violet energy on a single
`vital and sensitive structural or chemical unit that results in subse(cid:173)
`quent failure in cell multiplication. Rahn (15) has recently figured
`from the curves of abiotic reactions among large numbers of single cell
`organisms tliat death of the cell probably involves only a single chemi(cid:173)
`cal entity. As Coblentz and Fulton (16) have suggested, it is to be
`
`9
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`40
`
`ULTRA VIOLET LIGHT.
`
`III
`
`presumed that only a small fraction of the total absorbed energy first
`affects such an essential structure, and so leads to the death of the cell.
`Neither the reciprocals of the energy curves, which are undoubtedly
`modified by the absorption of light by elements not involved in the
`
`\
`
`\
`' I
`' I
`
`' I .
`\
`\
`\.
`"V
`
`,, ............... I'...
`
`"'~
`\
`
`9 3
`
`36
`
`33
`
`30
`
`27
`._;
`z
`!j24
`er:
`w2
`l
`Q.
`
`~18
`I-
`~ 15
`0
`(/')
`co 12
`<C
`
`9
`
`6
`
`3
`
`\
`\
`\
`
`'['-.....
`
`0 .22 .23
`
`.24
`
`.29
`.28
`.2 7
`.26
`.25
`WAVE LENGTHS IN )l..
`
`.30
`
`.31
`
`.32
`
`TEXT-FIG. 4. The coefficients of absorption of ultra violet light by a layer of
`S. aureus 0.8 µ, thick.
`
`bactericidal reaction, nor the curves for total absorption give an ac(cid:173)
`curate picture of the absorption curve of this vital substance, whatever
`it may be. Yet their similarities and their differences are alike useful
`in the further search for such an essential and sensitive element in the
`cell's structure and economy.
`
`10
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`FREDERICK L. GATES
`
`41
`
`Thus it may be predicted that there are a number of chemical entities
`or aggregates in the living cell which have rather similar coefficients of
`ultra violet absorption, and that the sum of these similar absorption
`curves largely determines the shape of the curve for the entire cell.
`
`~
`\
`\
`\
`\
`
`39
`
`36
`
`33
`
`30
`
`~27
`Lu u
`o:::24
`Lu a..
`z~I
`0
`i= I 8
`a..
`0:::
`~15
`CJ
`<C
`
`12
`
`9
`
`6
`
`'v' ----:--....
`" '\
`\
`
`~---
`
`\
`\
`
`\
`\
`
`'
`
`" ' - 'o
`
`3
`0 .22 .23 24 25
`.29
`.26 27 28
`WAVE LENGTHS IN )1
`
`.30
`
`31 32
`
`TEXT-FIG. 5. The coefficients of absorption of ultra violet light by a layer of
`B. coli 0.8 µ thick.
`
`Among these substances will be found the one essential element first
`affected by ultra violet light in the bactericidal reaction, and its absorp(cid:173)
`tion curve will be similar to, though not identical with, the reciprocal
`of the lethal energy curve. Finally it seems improbable that this
`sensitive substance is uniformly distributed throughout the cell's pro-
`
`11
`
`

`

`Downloaded from http://rupress.org/jgp/article-pdf/14/1/31/1234125/31.pdf by Reprints Desk on 02 February 2022
`
`42
`
`ULTRA VIOLET LIGHT.
`
`III
`
`toplasm. An examination of the evidence for its concentration in the
`cell nucleus, and the further search for evidence of its chemical charac(cid:173)
`ter are reserved for the final paper of this series.
`
`SUMMARY
`The simple conclusion of former investigators that the shorter
`the wave length of ultra violet light the greater the bactericidal action
`is in error. A study with measured monochromatic energy reveals
`a characteristic curve of bactericidal effectiveness with a striking
`maximum between 260 and 270 m.µ. The reciprocal of this abiotic
`energy curve suggests its close relation to specific light absorption by
`some single essential substance in the cell.
`Methods are described for determining the absorption curve, or
`absorption coefficients, of intact bacteria. These curves for S.
`aureus and B. coli have important points of similarity and of differ(cid:173)
`ence with the reciprocals of the curves of bactericidal incident energy,
`and point the way in a further search for the specific substance, or
`substances, involved in the lethal reaction.
`
`REFERENCES
`1. Gates, FrederickL., J our. Gen. Physiol., 1929, 13,231.
`2. Gates, Frederick L., J our. Gen. Physiol., 1929, 13, 249.
`3. Henri, Mme. V., Compt. Rend. Soc. Biol., 1912, 73, 321.
`4. Bang, S., Mitt. aus Finsen's Med. Lichtinstitut., 1905, 9,164.
`5. Newcomer,H.S.,Jour.Exp.Med., 1917,26,841.
`6. Mashimo, T., Mem. Coll. Sci., Kyoto Imp. University, 1919, 4, 1.
`7. Bayne-Jones, S., and von derLingen, J. S., Bull. Johns Hopkins Hosp., 1923,
`34, 11.
`8. Browning, C.H., and Russ, S., Proc. Roy. Soc. Lond., Series B., 1917, 90, 33.
`9. Judd-Lewis, S., Jour. Chem. Soc., 1919, 115,312; Jour. Roy. Soc. Arts., 1921,
`69, 806.
`10. Hiss, P.H., and Zinsser, H., A Textbook of Bacteriology, D. Appleton and Co.,
`New York, 1928, 318.
`11. Wood, R. W., Physical Optics, Macmillan and Co., New York, 1921, 157.
`12. Henri, V., Etudes de Photochemie, Gauthier-Villars et Cie., Paris, 1919.
`13. Saidman, J., Compt. Rend. A cad. Sci., 1924, 179, 1448.
`14. Judd-Lewis, S., Proc. Roy. Soc. Lond., Series B., 1916, 89,327; 1920, 93, 178.
`15. Rahn, O.,Jour. Gen. Physiol., 1929, 13, 179; 1930, 13,395.
`16. Coblentz, W.W., and Fulton, H. R., Sci. Papers Bureau of Standards, 1924,
`No. 495, 19.
`
`12
`
`