COMPOSITION STUDIES ON TOBACCO.
`XXI*.-The Headspace Vapours of Leaf**
`
`By A. P. SWAIN, R. F. PETERSON, jun. and R. L. STEDMAN
`
`A procedure for the collection of tobacco headspace vapours without the use of steam, heat, reduced
`pressure or elaborate apparatus is described. Typical results are presented showing the presence of at least
`35 components in the concentrated vapours from Turkish tobacco examined by temperature programmed
`gas chromatography. The use of selective removal techniques to classify components by functional group
`reactions is described, and tentative identifications of some of these components by the use of retention
`indices are reported.
`
`Introduction
`Composition studies on tobacco have generally been
`concerned with large classes of substances grouped by
`solubility in various solvents1 or by the possession of a
`common functional group. 2 Some time ago we initiated
`studies on the composition of the mixture of volatile com(cid:173)
`pounds continually released to the air by cured tobacco,
`which accumulate in the headspace above tobacco stored in
`closed containers. Variations in the composition of this
`aroma-bearing mixture may influence the desirability of
`tobacco for manufacturing purposes and the character of the
`tobacco smoke.
`Previous studies of the volatile compounds of tobacco have
`been concerned with the leaf substances obtained by solvent
`extraction3 or steam distillation at ordinary pressure 4• 5 or
`It was desired to avoid any procedure which
`under vacuum. 6
`might alter the natural composition of the headspace vapours
`during collection or which might include or generate additional
`components not normally present in the headspace itself.
`Also, since the moisture content of a tobacco sample has been
`shown to be related to the total amounts of certain volatile
`carbonyl compounds obtainable by steam distillation, 2 it was
`desirable that the procedure adopted would not alter the
`normal moisture content of the sample. Many alternative
`methods of collection were tried and discarded before a
`procedure was adopted.
`The present paper describes details of a method for the
`collection of tobacco headspace vapours and the chemical
`composition of such vapours.
`
`Experimental
`
`Apparatus
`A commercial cylinder of dry nitrogen equipped with a
`pressure-reducing regulator and needle valve was connected
`to the top of the tobacco container, a gas-washing bottle of
`500 ml capacity with a standard-taper ground-joint and
`fittings for springs to hold the two parts together (Fig. I).
`The gas passed downward through the tobacco to the bottom
`of the central tube, protected by a coarse fritted glass filter,
`then through the central tube to the exit at the top of the
`bottle. Entrance and exit tubes were further protected by
`loose plugs of glass wool. From the exit tube the mixture of
`vapours and nitrogen was led through glass tubing and
`
`* Part XX: J. Sci. Fd Agric., 1964, 15, 774
`** Presented in part at the 18th Tobacco Chemists' Research
`Conference, Raleigh, North Carolina, October 1964
`
`Fig. 1. Photograph of apparatus for collection of headspace vapour
`
`adaptor to one of two 22-gauge hypodermic needles which
`extended through a rubber cap into the neck of a small serum
`bottle (nominal capacity 5 ml) arranged so that its lower
`portion could be immersed in a bath of liquid nitrogen
`maintained at constant level in a Dewar flask. During
`collection of the vapours, the serum bottle contained an
`amount of anhydrous sodium sulphate (0 · 1--0 · 3 g) sufficient
`to saturate the water expected to condense and to be added as
`aqueous classification test reagents. The exit needle was
`attached through glass tubing to a soap-film flowmeter. Gas
`chromatographic analysis of the collected headspace vapours
`was accomplished using a flame ionisation instrument (F. &
`M. Scientific Co. Model 1609*) equipped with a linear
`temperature programmer (Model 240) for the column oven.
`
`Sample Collection
`Initially, several collection methods were used similar to
`those employed for other natural products. A few comments
`on these attempts may be of interest.
`Direct sampling of the unconcentrated headspace vapours
`by means of sampling valves or syringes (see references 7- 9)
`was not very successful because of the extremely small
`amounts present in tobacco headspace.
`
`• Mention of company or trade names does not imply endorse(cid:173)
`ment by the U.S. Department of Agriculture over others not named
`J. Sci. Fd Agric., 1966, Vol. 17, August
`
`RJRV EX 1019
`
`

`

`350
`
`Swain et al.: Composition Studies on Tobacco. X XI
`
`The methods of Nawar et al.1°, 11 involving mechanical
`circulation of vapours into various collectors presented many
`complications too numerous to detail. Modifications of the
`collecting traps including some similar to those of Farrington
`et al.12 and West et al. 13 were relatively unsuccessful except for
`the use of a U-tube containing tobacco itself as an adsorbent,
`which proved to be an effective trap although the chromato(cid:173)
`graphic results on the trapped volatiles were difficult to
`reproduce. The method of Hornstein & Crowe14 involving
`collection of volatiles directly on a cooled chromatographic
`column and elution by warming the column was an effective
`method of trapping constituents. However, difficulties in
`obtaining leak-free connections prevented precise and consis(cid:173)
`tent measurements of retention indices.
`An adaptation of an apparatus for total collection of gas
`chromatographic eluates (F. & M. Scientific Co., Model
`TCS-3) was investigated. This apparatus permits collection
`in a large (300 ml) evacuated glass bulb of the total volume
`of carrier gas carrying an eluting gas chromatographic peak.
`After collection, a nipple in the bulb (5 ml volume) is cooled
`and the vapours within the bulb are condensed. This
`apparatus was employed to collect headspace vapours of
`tobacco by repeated filling of the bulb and condensation of
`the vapours. Chromatographic results similar to those
`reported above were obtained; however, several shortcomings
`were apparent, including impractically long waiting periods
`required for condensation of the vapours in successive
`samplings and leakage of laboratory air into the evacuated
`system.
`Many other variations in technique were studied before the
`final method was adopted. Most of the variations failed to
`give satisfactory results and, in many cases, no chromato(cid:173)
`graphic peaks were obtained. To reproduce the findings
`reported with the adopted method, strict adherence to the
`details of the method is required.
`
`Final method adopted
`The gas-washing bottle was filled loosely with cured tobacco
`leaves (250 g) ground to pass a 30-mesh screen and the top was
`attached by carefully working the central tube through the
`tobacco to the bottom of the bottle so that a tight fit at the
`ground joint was obtained and secured by attaching springs or
`rubber bands. Anhydrous sodium sulphate (0 · I g) was
`placed in a serum bottle (5 ml) which had first been flushed
`with dry nitrogen. The bottle was then closed with a rubber
`cap and was connected to the collection system by pushing
`the needles through the cap so that their tips just protruded
`into the wide part of the bottle. The entire system was then
`flushed with nitrogen at room temperature until all air had
`been removed. The bottom third of the serum bottle was
`gradually brought to liquid-nitrogen temperature, care being
`taken to maintain a positive flow of nitrogen through it at all
`times, as judged by the soap-film flowmeter. The upper third
`of the bottle was kept at a temperature above freezing by
`directing a stream of air across it in order to prevent plugging
`of the needles. When temperatures had reached equilibrium,
`the nitrogen flow was adjusted to 30-50 ml per minute and
`was continued at this rate during the 4-h collection period.
`At the end of this time the exit needle was withdrawn and the
`entrance needle was fitted with a gas-tight syringe (capacity
`1 ·0 ml), preferably with the gas-flush modification. The
`bottle was then allowed to come to room temperature and
`was finally immersed in a bath at 60° for 3 minutes before
`removal of a 1 ·0-ml aliquot of the vapour for injection.
`
`Gas-liquid chromatography
`Stainless steel, packed, single columns (10 ft x ¼ in.) were
`used, with helium carrier gas flow rate of 50 ml/min. regulated
`by flow controller and needle valve; helium pressure at
`cylinder gauge, 30 psi; injection port temperature, 225 °;
`detector block temperature, 250°; and flamehead temperature,
`325°. With Carbowax 20M or Silicone SE-30 as the station(cid:173)
`ary phase (25% on Anakrom ABS), the oven temperature was
`programmed at 5° per minute from 75° to 220°. With
`tritolyl phosphate-glycerol (MacDonald & Brunetl5), tri(cid:173)
`/3,/3' -oxydipropionitrile as
`tolyl phosphate alone or
`the
`stationary phase, maximum temperatures were lower; iso(cid:173)
`thermal operation was also used with these phases. Since the
`gas chromatograph used is a single-column instrument,
`temperature programmed runs show a rising baseline. With
`the flame ionisation detector, upper temperature limits for all
`phases were somewhat lower than would be the case with less
`sensitive detectors.
`
`Classification tests
`Component peaks were classified by use of the following
`selective removal reagents :16, 11
`(a) Acidic hydroxy/amine reagent: 1 · 14 g of hydroxylamine
`hydrochloride was dissolved in 10 ml of N/1 sodium
`hydroxide solution and diluted to 50 ml with water.
`(b) Basic hydroxy/amine reagent: I· 14 g of hydroxylamine
`hydrochloride was dissolved in 50 ml of N/1 sodium
`hydroxide solution.
`(c) Mercuric chloride reagent: a saturated aqueous solution
`was prepared at room temperature.
`(d) Potassium permanganate reagent: 2 · 5 g of potassium
`permanganate were dissolved in 50 ml of water.
`When a sample was to be used for classification tests, the
`amount of sodium sulphate in the serum bottle at the start of
`collection was increased to O · 3 g so as to saturate the water
`from the aqueous reagent to be added as well as that from the
`tobacco vapours. After the serum bottle had been allowed
`to come to room temperature with the I · 0 ml gas-tight syringe
`in place as described under 'Sample Collection', the syringe
`and needle were withdrawn and O · 5 ml of the classification
`reagent was added through the rubber cap from a liquid
`measuring syringe. The bottle was shaken for several minutes
`and was then set aside at room temperature for 1 h with
`occasional shaking. The gas-tight syringe and needle were
`then re-inserted, the bottle was immersed in the 60° bath for 3
`minutes and a 1 ·0-ml aliquot of the vapour was withdrawn
`and injected into the gas chromatograph. Comparison of the
`resulting chromatogram with either a previously obtained
`chromatogram or a chromatogram on 1 ·0 ml of vapour
`removed before addition of the classification reagent, revealed
`which peaks were eliminated completely or reduced in size.
`A mixture of nitrogen and the vapours of acetone, C3_5
`n-aldehydes, methyl propionate, methanol, n-propanol and
`benzene was prepared in a serum bottle for use in checking
`the action of classification reagents. Both of the hydroxy(cid:173)
`lamine reagents eliminated completely the peaks from alde(cid:173)
`hydes and ketones and diminished those from alcohols.
`Basic hydroxylamine removed methyl propionate as well.
`Potassium permanganate removed all but acetone, methyl
`propionate and benzene.
`
`Identification
`Kovats18 retention indices as modified for use with linear
`J. Sci. Fd Agric., 1966, Vol. 17, August
`
`RJRV EX 1019
`
`

`

`Swain et al.: Composition Studies on Tobacco. XXI
`
`351
`
`temperature programming by Van Den Dool & Kratzl9
`were used to obtain the tentative identifications reported.
`Normal hydrocarbons and ethyl esters of straight-chain fatty
`acids were used as reference compounds. When ethyl esters
`were used the results were calculated to a n-hydrocarbon
`reference system as described by Van Den Dool & Kratz,19
`whose tables of retention indices were confirmed and supple(cid:173)
`mented as necessary by injections of pure compounds. The
`tables of isothermal retention indices of West et aI.20 who
`used tritolyl phosphate and {:J,(:J' -oxydipropionitrile stationary
`phases were also helpful.
`Known compounds were injected as the vapour, using a
`technique similar to that used for the headspace samples.
`Small amounts of the pure liquids were placed in serum
`bottles from which all air had been removed by flushing with
`nitrogen. After the bottle had been closed with a rubber
`serum cap and time allowed for vapour-liquid equilibrium
`to be established, aliquots of the vapour varying in size from
`I µl to several hundred µl were measured at room temperature
`(or at elevated temperatures with less volatile compounds)
`in gas-tight syringes, the amount of each being adjusted until a
`peak similar in size to that of the corresponding unknown
`component was obtained. Retention temperatures were
`determined by the technique ofVan Den Dool &Kratz,19 the
`pyrometer temperature on the chart being marked at 5 °
`intervals and the chart being as a graph. To correct for the
`rising baseline, an arbitrary line was drawn between inflexions
`at the start and end of well-separated peaks, and the point at
`which this line met a line tangent to the rising portion of the
`peak in question was used to locate the retention temperature.
`This is more desirable than use of the apex of the peak, the
`position of which depends to a larger extent on the amount
`injected. Reproducibility was as good as that reported by
`Van Den Dool & Kratz.1 9
`Co-chromatography was accomplished by adding the
`required volume of vapour of the known substance to the
`serum bottle containing the unknown, agitating for a sufficient
`time to produce adequate mixing, and withdrawing an aliquot
`of the mixed vapours for injection into the gas chromato(cid:173)
`graph.
`Frequent blank runs showed that insignificant amounts of
`impurities were contributed by the nitrogen gas used, by the
`short length of plastic or rubber tubing used in connecting
`together the various parts of the apparatus, and by the
`reagents used for selective removal studies.
`It was found,
`however, that unless laboratory air was rigidly excluded from
`the tobacco samples, the collection apparatus, and the gas(cid:173)
`tight syringes by flushing with nitrogen before each use,
`artifacts could be introduced. Use of the gas-flush modifica(cid:173)
`tion of the syringes was extremely helpful in this respect;
`ordinary gas-tight syringes without this modification tended
`to hold traces from previous injections tenaciously. Effi(cid:173)
`ciency of the trapping system was proved by recovering trace
`amounts of known volatile compounds added to an empty
`gas-washing bottle instead of the tobacco and by the absence
`of detectable amounts of headspace vapour components in a
`second trap immersed in liquid nitrogen in series with the first.
`
`Results and Discussion
`Representative results
`Table I and Fig. 2 show the results of classification tests and
`retention indices, and a typical chromatogram using the
`adopted method on cured Turkish (Smyrna) tobacco leaves.
`The standard retention indices reported in Table I are
`J. Sci. Fd Agric., 1966, Vol. 17, August
`
`referred to the n-paraffinic hydrocarbon system in which the
`hydrocarbons with n carbon atoms are assigned the indices
`100 n and other
`indices are calculated from retention
`temperatures using the following relationship,
`I= lOOi X-Mn + 100n
`Mn+i-Mn
`in which I is the standard retention index of the unknown
`component and X, Mn and Mn+1 are the retention tempera(cid:173)
`tures of the unknown component and two marker hydro(cid:173)
`carbons of n and n + i carbon atoms respectively. Over a
`limited range the index varies linearly with temperature.
`Under our conditions, in which programming was started
`simultaneously with sample injection at 75°, the change in
`index per degree was almost constant over the range 95-200°,
`but below 95° this ratio increased rapidly.
`The chromatogram presented in Fig. 2 was made with a
`maximum sensitivity setting of attenuation 8 in range 1 since
`higher sensitivity gave an excessive noise level.
`Components classified as inert in Table I were not removed
`by any reagent. Aldehydes were completely removed by
`both acidic and basic hydroxylamine and by potassium
`permanganate solution; the ketone was unaffected by perman(cid:173)
`ganate. Evidence of carbonyl groups was confirmed by the
`formation of precipitates when the effluents, collected by
`means of a heated stream-splitting device inserted between the
`column exit and the detector, were reacted with 2,4-dinitro(cid:173)
`phenylhydrazine reagent. The ester was removed by basic
`hydroxylamine but not by acidic hydroxylamine or perman(cid:173)
`ganate. Peaks 15, 16, 17, 20-23 and 25-27 were completely
`removed by permanganate but were not removed or were only
`partially reduced in size by the hydroxylamine reagents; these
`represent easily-oxidised, non-carbonyl compounds, which
`would include alcohols, aromatic compounds with oxidisable
`side-chains, unsaturated aliphatic hydrocarbons, etc. Satu(cid:173)
`rated aqueous mercuric chloride did not remove completely
`any peaks in these tests on Turkish tobacco, although in some
`runs on this and other tobaccos evidence of removal of one or
`more peaks by this reagent was obtained. Although no peaks
`above number 27 were completely removed by any reagent,
`some of these may be unreactive only because of high mole(cid:173)
`cular weight or poor solubility in the aqueous classification
`reagents under the test conditions, and others may be inert
`but reduced in size only because of low volatility after treat(cid:173)
`ment with the reagent.
`Peaks 1 and 2 of Table I appeared in all blank runs. They
`represent, at least partially, artifacts caused by pressure
`changes upon injection or possibly trace impurities. The
`possible presence of highly volatile headspace components in
`this region has not been entirely excluded.
`These results show that the sample-collection method
`described permits the concentration and separation of the
`extremely small amounts of volatile compounds occurring in
`tobacco headspace vapours without elaborate trapping
`techniques under conditions closely similar to those obtaining
`in the natural state of the sample, which is not subjected to
`heat, steam, excess water vapour or the oxygen of the air.
`The vapours collected are not forcibly removed under the
`drastic conditions of a vacuum, but at a pressure close to that
`of the atmosphere. The moisture content of the sample is
`not changed appreciably during sampling (less than O • 25 ml
`of water is removed from 250 g of tobacco of normal moisture
`content during the specified collection period). Repro(cid:173)
`ducibility is excellent and efficiency of collection is close to
`100% even for very volatile substances at liquid nitrogen
`
`RJRV EX 1019
`
`

`

`352
`
`Swain et al.: Composition Studies on Tobacco. XXI
`
`TABLE I
`Data on possible identity of vapour components in the headspace over Turkish tobacco separated on Carbowax
`
`Peak
`no.
`
`Chemical
`c:haracteristics
`
`Possible identity Retention index
`of peak
`
`Retention index
`of
`known substance
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`:l5
`26
`27
`28
`29
`30
`31
`32
`33
`34
`35
`36
`37
`
`Inert
`
`Ald;byde
`
`Ket~ne
`Aldehyde
`Ester
`Oxidisable
`
`"
`Inert
`Aldehyde
`Oxidisable
`
`"
`Aldehyde
`Oxidisable
`
`"
`Uncertain
`
`,,
`
`Artifact
`"
`n-Pentane
`?
`?
`n-Hexane
`?
`?
`n-Heptane
`Acetaldehyde
`Propionaldehyde
`Acetone
`n-Butyraldehyde
`Methyl propionate
`Methyl alcohol
`Ethyl alcohol
`?
`Benzene
`n-Valeraldehyde
`?
`?
`?
`Toluene
`n-Caproaldehyde
`?
`p-Xylene
`m-Xylene
`?
`?
`o-Xylene
`?
`?
`?
`?
`?
`?
`?
`
`500
`
`600
`
`707
`743
`805
`823
`873
`883
`912
`923
`948
`958
`1000
`1020
`1032
`1043
`1063
`1099
`1121
`1149
`1169
`1203
`1224
`1240
`1256
`1278
`1299
`1318
`1338
`1356
`1371
`
`500
`
`600
`
`700
`?
`796
`823
`866
`893
`911
`927
`
`953
`983
`
`1058
`1089
`
`1151
`1173
`
`1242
`
`l II 12
`'1f
`
`2
`
`41 '~
`
`4 7
`
`10
`
`w
`tJ) z
`0
`0.
`tJ)
`w
`ll::
`ll:: w
`0
`ll::
`0
`u
`w
`ll::
`
`19
`
`22
`
`26
`
`34
`
`37
`
`14
`
`20
`
`24
`
`_t·
`
`31
`
`32
`
`80
`
`100
`
`120
`TEMPERATURE, 0c
`
`140
`
`160
`
`0
`
`8
`
`12
`TIME, min
`
`16
`
`20
`
`Fig. 2. Gas-liquid chromatogram of headspace vapour from 250 g of Turkish (Smyrna) tobacco on Carbowax 20 M column
`IX=range l, attenuation 8.
`
`J. Sci. Fd Agric., 1966, Vol. 17, August
`
`RJRV EX 1019
`
`

`

`Swain et al.: Composition Studies on Tobacco. X XI
`
`353
`
`temperature. Cooling with Dry Ice was shown to be less
`efficient, giving chromatograms in which the early peaks are
`less prominent.
`
`Tentative identifications
`Table I also lists retention indices for known compounds
`calculated from the retention temperatures obtained with the
`pure substances injected under standard conditions on the
`Carbowax 20M column. Values for some, but not all, of
`these are contained in the tables published by Van Den Dool
`& Kratz. 19 Agreement with their values in most cases is
`good enough to confirm their conclusion that, although the
`retention temperature is not an absolute constant and will
`vary depending upon the heating rate, carrier gas flow-rate
`and other operating parameters, the difference between
`retention temperatures of members of a homologous series is
`relatively constant under a variety of conditions. The prob(cid:173)
`lem of determining with certainty to which homologous series
`within the same functional group class an unknown component
`belongs, still remains. Although we have supplemented the
`list of known compounds reported by Van Den Dool &
`Kratz19 and confirmed the retention indices of many of the
`peaks on the SE-30 column, it is obvious that absolute
`identification cannot be achieved by these techniques alone.
`Nevertheless, the data in Table I indicate the probable
`occurrence of methanol, ethanol and higher alcohols;
`several aliphatic hydrocarbons, including n-pentane, n-hexane
`and possibly n-dodecane; simple aromatic hydrocarbons
`(benzene, toluene and xylenes); and a number of aliphatic
`aldehydes, ketones and esters. Co-chromatography with
`methanol, ethanol, propionaldehyde, methyl propionate,
`acetone and benzene tended to confirm these identifications,
`since the suspected peaks were enhanced without peak
`broadening. This of course does not exclude the possibility
`that the unknown is another compound of the same functional
`group class with closely similar retention. The presence of
`both methanol and ethanol was confirmed using the glycerol(cid:173)
`tritolyl phosphate stationary phase
`recommended by
`MacDonald & Brunet15 for better separations of low-boiling
`polar compounds; the methanol peak was much the smaller
`of the two. Polar compounds were not completely separated
`by the silicone SE-30 stationary phase, but produced broad,
`tailed peaks due to polar compounds which obscured the
`sharper peaks given by non-polar substances. This made
`interpretation of the results of functional group tests difficult.
`Acetone and a number of the other compounds listed were also
`detected in tobacco headspace vapours using the tables of
`retention times published by West et al. 20 for use with tri(cid:173)
`tolyl phosphate and {J,{J' -oxydipropionitrile stationary phases.
`
`The presence of the above hydrocarbons, n-caproaldehyde
`and methyl propionate in tobacco leaf has not been reported
`previously, although Onishi & Nagasawa3 isolated an unidenti(cid:173)
`fied Co-aldehyde from leaf.
`
`Quantitative considerations
`No attempt to determine absolute or relative amounts of the
`components of tobacco headspace vapours was made, but
`peaks of a size range similar to that observed for the main
`components in the headspace chromatograms were obtained
`when injections of O · 1-2 · 0 µg of benzene as vapour were
`made. Other tobaccos showed minor qualitative differences
`in headspace vapour chromatograms, but each exhibited a
`characteristic quantitative pattern.
`
`Eastern Utilization Research & Development Division,
`Agricultural Research Service,
`U.S. Dept. of Agriculture,
`600 East Mermaid Lane, Philadelphia,
`Pennsylvania 19118, U.S.A.
`
`Received 9th Au11ust, 1965
`
`References
`1. Stedman, R. L., Swain, A. P., & Rusaniwskyj, W., Tob. Sci.,
`1962, 6, 1
`2. Weybrew, J. A., & Stephens, R. L., Tob. Sci., 1962, 6, 53
`3. Onishi, I., & Nagasawa, M., Bull. agric. chem. Soc. Japan, 1957,
`21, 38
`4. Onishi, I., & Nagasawa, M., Bull. agric. chem. Soc. Japan, 1957,
`21, 95
`5. Schmeltz, I., Stedman, R. L., & Miller, R. L., J, Ass. off. agric.
`Chem., 1963, 46, 779
`6. Jones, L. A., & Weybrew, J. A., Tob. Sci., 1962, 6, 194
`7. Rhoades, J. W., Fd Res., 1958, 23, 254
`8. Mackay, D. A. M., Lang, D. A., & Berdick, M., Analyt. Chem.,
`1961, 33, 1369
`9. Buttery, R. G., & Teranishi, R.,J. agric. FdChem., 1963, 11,504
`10. Nawar, W.W., & Fagerson, I. S., Analyt. Chem., 1960, 32, 1534
`11. Nawar, W.W., Sawyer, F. M., Beltran, E.G., & Fagerson, I. S.,
`Analyt. Chem., 1960, 32, 1534
`12. Farrington, P. S., Pecsok, R. L., Meeker, R. L., & Olson, T. J.,
`Analyt. Chem,, 1959, 31, 1512
`13. West, P. W., Sen, B., & Gibson, N. A., Ana/yt. Chem., 1958,
`30, 1390
`14. Hornstein, I., & Crowe, P. F., Analyt. Chem., 1962, 34, 1354
`15. MacDonald, R., & Brunet, P. E., J. Chromat,, 1963, 12, 266
`16. Bassette, R., Ozeris, S., & Whitnah, C.H., Analyt. Chem., 1962,
`34, 1540
`17. Hoff, J.E., & Feit, E. D., Analyt, Chem., 1963, 35, 1298
`18. Kovats, E., Helv. chim. Acta, 1958, 41, 1915
`19. Van Den Dool, H., & Kratz, P., J. Chromat., 1963, 11, 463
`20. West, P. W., Sen, B., Sant, B. R., Mallik, K. L., & Sen Gupta,
`J. G., J. Chromat., 1961, 6, 220
`
`J. Sci. Fd Agric., 1966, Vol. 17, August
`
`RJRV EX 1019
`
`