j. Soc.Cosmet. Chem.,37, 409-428 (November/December 1986) Odorantsrelatedto humanbodyodor E. P. KOSTER,J. s. JELLINEK,N. D. VERHELST, J. MOJET, and M. R. I. LINSCHOTEN, Psychological Laboratory, Ryksuniversiteit, Sorbonnelaan 16, 3594CA Utrecht,Netherlands (E.P.K., N.D.V.,J.M., M.R.I.L.), and Dragoco,D-3450 Holzminden,WestGermany(J.S.J.). Received July 23, 1986. Synopsis Using a T-shirt test designdescribedby M. Schleidtand co-workerswith youngfemale(n = 29) and male (n = 30) respondents,the relationsbetweenconcentration,odor intensity, odor pleasantness, and sex attributionwerestudiedfor femaleandmaleperspiration,androstenol (I, a naturalingredientof perspiration), for two syntheticandrostenol analogues (II and III), and for cyclopentadecanolide, a syntheticmusk. It wasfoundthat the syntheticanalogueIII and, to a lesserextent, II exhibit patternssimilar to the natural stimuli. Cyclopentadecanolide wasperceivedmore asa feminineperfume. INTRODUCTION There is a notion that bodyodormay play a role in humaninteractions,not only in the senseof repulsionby the smellof unwashed,sweatybodiesandclothingbut alsoin the senseof sexualattraction and of individual recognition.This notion is not new (la). However, the evidencesupportingit has been, until recentyears,largely anecdotal. Practicessuch as the use of a perspiration-drenched handkerchiefas an aphrodisiac, reportedfor peasantpopulationsin severalEuropeancountries,have been basedon traditionand belief(lb). During the pastdecade,a numberof studieshavebeenconducted in which human reactionsto human body odorshave been testedunder controlled conditions.Someof thesewere conductedwith bodyodorsin their natural form; othersinvolvedtwo pure substances which havebeenidentifiedin humanperspiration and urine: 5o•-androst-16-ene-3o•-ol and 5o•-androst-16-ene-3-one (2). Well beforetheserecentstudieswere undertaken,Paul Jellinek (3) postulatedthe central importanceto perfumeryof odorantswith odor qualitiesreminiscentof human body odor. He proposedthe followingtheses: 1. The function of perfumesworn by womenis to enhancethe wearers'sexualattraction to the oppositesex. 2. Perfumesperformthis functionby virtue of their contentof "erogenous" odorants. 3. Theseodorants,which are to be found primarily in perfumematerialsof animal origin (musk, civet, ambergris,castoreum)but also in essentialoils and extracts from plantsand amongsyntheticaromaticchemicals,owetheir erogenous effectto the similarity of their odorto someaspectof humanbodyodor. 4. Theseodorantsexert an erogenous effectonly if usedat low levels. 4O9 410 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Although thesetheseshavenot beenuniversallyacceptedamongperfumers,they have had considerable influenceon perfumerypractice.A few studieshavebeenreportedto test their validity. Paukner(4a) foundthat addingincreasingconcentrations of an "erogenous complexof civet, syntheticambergris,and syntheticmusk"to a lavender-type perfumebroughtits positionin a three factorsemanticdifferentialanalysisincreasinglycloseto the position of "the idealerogenous fragrance."Steiner(4b) presented,in very brief flashes(V•70to •/250sec.), ambiguousslidesin which a pictureof a nude womanand a cityscapewere superimposed,to a group of 25 malesaged 20-39 years.In half of the casesthe subjectsviewingthe pictureswerealsoexposed to a perfumewhichthey hadpreviously selectedas being sexuallystimulating. In the presenceof the perfume, significantly more respondents reportedseeingthe womanthan in the absence of perfume. The purposeof the presentstudywasto studythe reactionsof untrainedsubjectsto the odorsof severalaromaticchemicalsrelatedto human body odor, when thesechemicals were presentedunder conditionswhere they might be taken for actual human body odors.Using actualhumanbodyodorsand 5c•-androst-16-ene-3-ol (I) as comparison stimuli, we hopedto establishin what waysthe syntheticchemicalswere similar and how they differedfrom thesenaturalstimuli. We adoptedthe designof M. Schleidt's T-shirt experiment(5) but replacedpart of the shirtsthat had actuallybeenworn by test subjectsby clean shirts that had been treatedwith androstenol(I) or with the aromatic chemicalswe wanted to study, at variousconcentrations.We selectedthe followingchemicals: 2-Methyl-1(2,6-exo-tricyclo [5.2.1.02'6]dec-4(3)-ene-8-yl)-pent1-ene-3-ol(III), syntheticaromaticchemicalsfirst preparedin the laboratoriesof Dragoco, with odor descriptionssimilar to thoseof androstenol(6); and cyclopentadecanolide (IV), a macrocycliclactonefound in ambretteseedoil and closelyrelatedin structure to key ingredientsof natural musk and civet tincture, which has beenusedfor many decadesin fine perfumeryas a musk-typeodorant(Figure 1). We were particularlyinterestedin exploringthe followingquestions: Is it at all possibleto fool respondents into believingthat the odor of chemicaltreatedshirtsis human bodyodor? Schleidt(5) had foundthat sexattributionwaspossibleand that her subjectscould correctlyidentify the sexof the wearerof a given shirt to a statisticallysignificant degree.Is sexattributionpossiblealsofor the aromaticchemicalsstudied?If so, to what sexare they attributed? Schleidtalsofoundthat strongbodyodorswerepredominantlyjudgedasunpleasant and that they were predominantlyattributedto male wearers.Do the samepatterns prevail with the aromaticchemicalsstudied? METHOD Test subjectswere 30 men and 29 women,aged 18 to 30, all studentvolunteersat the University of Utrecht. For participatingin the test, they receiveda modestmonetary reward (Fl. 35,--), or someacademiccredit. They were not prescreenedin any way. There was only one among them who had trouble smelling many of the stimuli; his ratingswere countedalongwith thoseof the other participants. HUMAN BODY ODOR 411 II 5o•-And rost- 16-e ne-3o•-o I (Androstenol) 2-Methyl-1(2,6-exo-tricyclo [5.2.1.02,•]dec-4(3)-ene-8-yl)pent-l-ene-3-one Oh III 2-Methyl-1 (2,6-exo-tricyclo [5.2.1.02,6]dec-4(3)-ene-8-yl)- • -0"'/ ' 0 IV Cyclopentadecanolide (CPD) pent-1-ene-3-ol Figure 1. Structuresof materialsusedin thesestudies. The subjectswere providedwith yellow cotton T-shirts and were instructedto wear theseon five consecutivenights. During this period they were instructednot to useany deodorantand to washonly with a very faintly perfumedsoap* with which they were provided. They were alsogiven odorlesspolyethylenebagsin which to storethe shirts between and after wearing them. At the end of the 5-day period, the test subjectsparticipatedin two smellingsessions which were conductedon two consecutivedays. In thesesessions,they were presentedwith 30 T-shirts of the kind they had been wearingand were askedto rate thesefor odorintensity(four replications)and for pleasantnessof odor (three replications)and to try and identify the sexof the wearerof each shirt (threereplications).The respondents wererandomlydividedinto two groups.The sequence of the tasksperformedby the two groupsis shownin Table I. The orderof presentationof the shirtswasrandomized.This wasdone(1) within tasks, * Penatensoap,the samebrandaswasusedin Schleidt'sexperiments.The experimentaldesignof this test replicatedin all essentialaspectsthat usedby Schleidt. 412 JOURNAL OF THE SOCIETY OF COSMETICCHEMISTS Table I Sequence of Task Presentation Group I (n = 29) Stimuli: Group II (n = 30) "Body" shirts1-5 (m and f)* "Body"shirts11-15 (m and f) All All "solution" shirts "solution" shirts Day 1 Tasks: Intensity (1 X ) Sex Attribution (3 X ) Intensity (1 X ) Intensity (1 X ) Pleasantness(3 X Intensity (1 X ) Stimuli: "Body"shirts6-10 (m and f) "Body"shirts16-20 (m and f) All All "solution" shirts "solution" shirts Day 2 Tasks: Intensity (1 X ) Sex Attribution (3 X ) Intensity (1 X ) Intensity (1 X ) Pleasantness(3 X) Intensity (1 X ) * This meansshirts1 through5 wornby a manandshirts1 through5 wornby a woman.Similarfor the other groups. by assigningdifferentstartingpointsto differenttestsubjects,and(2) betweentasks, by rearrangingthe shirts.The subjectsmarkedtheir responses on a ratingscalewhich wasan undividedline scalewith two markersplaced7 cm apart.At both ends,the scale extended1 cm beyondthe markers(Figure2). Respondents wereinstructed to indicate intensityand pleasantness by placinga verticalline acrossthe centersegmentof the scale:nearerto the left markerfor a weakor unpleasant odor,nearerto the right marker for a strongor pleasantodor. For extremelyweak or unpleasantodors,they coulduse the line segmentextendingto the left beyondthe left marker;for extremelystrongor pleasantodors,they couldusethe line segmentto the right of the right marker.The markingswere subsequentlyconvertedinto numericalscoresfrom 0 to 8, using the scoringscaleindicatedin Figure 2. The respondents weremade to believethat the shirt they weresmellingwerethe ones that had beenworn by themselves and by their fellowrespondents. In actuality,only ten of the shirtsin eacharrayhadbeenwornby respondents, five by femalesandfiveby males. These had been selected at random from the 29 female and 30 male shirts available,avoiding, however,shirtsthat were judgedby the experimenters to be extremely low in odor. In this paper,theseshirtswill be referredto as "body"shirts. Different"body"shirtswereusedon day 1 andday 2, andby groupI andgroupII, to avoid the effectsof odor lossdue to exposureand of odor contaminationdue to handling. The remaining20 shirts in the array were cleanshirtsthat were impregnated with differentodorantsolutions by placingthreedrops(ca. 0.06 ml) of solutionat each of the underarmareasof the shirts.Thesewill subsequently be referredto as "solution" shirts. The odorants used and the concentrations of the solutions are summarized Response Scale [ ScoringScale [ 0. 1 . 2 . 3 ß4. 5 . 6. 7 . 8 Figure 2. Scaleand scoring in HUMAN BODY ODOR 413 Table II. The solventin all caseswas 96% ethanol. The solutionswere placedon the shirts 30 minutes beforethe beginningof the sniffing sessions. The "solution"shirts usedon day 2 were cleanshirtsfreshlytreated with odorantsolutions,as on day 1. The concentrations of the odorantsolutionshad beenchosenso that they represented a rangeof intensitylevelscomparable to the rangeshownby the shirtsactuallyworn. The appropriateconcentrations were determinedin a pre-testin which 16 male and 16 femalerespondents participated;thesebelongedto the samestudentbodyasthe participantsof the main test, but theyweredifferentindividuals.In the pre-test,the subjects were screenedfor anosmiato the odorantstested. They were provided with yellow cotton T-shirts and askedto wear them on five consecutivenights, using the same precautions as described for the main test. Reference arraysof the four odorantswere prepared.Eacharrayconsistedof eight cottonswabson which threedropsof alcoholic odorantsolutionhad beenplaced, using eight different odorantconcentrations.Each subjectwasinstructedto compare16 shirts(eight female,eight male)with eachof the four referencearrays,and to indicate,for eachshirt and eachodorant,with which swab it corresponded most closelyin odor intensity. The concentration representedby the swabindicatedwasnotedby the experimenter.For eachodorant,the referenceconcentration ratings were groupedacrossall respondentsand all shirts. The 12.5% highest and the 12.5% lowest concentrationswere eliminated. The remaining concentration rangewasdivided logarithmicallyinto four equalsteps,giving the five concentration levelswhichwerethen usedin the main part of the test. Table Concentration Solution Odorant "Androstenol" (I) III II Cyclopentadecanolide "CPD" (IV) * Starting from a 1.00% solution. II of the Odorant Solutions Concentration No. (g/l) Dilution Factor* 1 10.000 1:1 2 1.429 1:7 3 4 5 0.204 0.029 0.004 1:49 1:343 1:2401 1 5.000 1:2 2 1.250 1:8 1:32 3 0.313 4 0.078 1:128 5 0.020 1:512 1 1.250 1:8 2 3 4 0.313 0.078 0.020 1:32 1:128 1:512 5 0.005 1:2048 1 10.000 2 3 1.429 0.204 4 5 0.029 0.004 1:1 1:7 1:49 1:343 1:2401 4 14 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS RESULTS CONCENTRATION AND PERCEIVED INTENSITY In the courseof the experiment,eachrespondentrated the intensityof each"solution" shirt four timesand of each"body"shirt twice. Sincean analysisof varianceshowedno systematicdifferencebetweentthe four (or two, resp.)ratings,they werecombinedin the calculationsfor Table III. The correlationbetweenconcentrationand intensitywas examinedusingthe non-parametrictest for grouptrend of Jonckheereand Bower(7). This test combines the rank correlation coefficient T between concentration and indi- vidual ratings to an overallstandardnormal deviate. The z-scores(one-tail test) were highly significantfor all test substances, both for femaleand male respondents (Table IV). As Table III indicates,the T-shirts worn by femaleswere, on the average,rated lower for odor intensitythan were thoseworn by males. Table Ilia Intensity Ratingsof "Solution"Shirts IntensityI Women Substance Men (n = 29) W (n = 30) + M (n = 59) Sol. 2 Mean 3 + SD Mean + SD Mean + SD Av. 1 2 3 4 5 1-5 4.00 3.32 2.78 2.21 2.44 2.98 1.57 1.40 1.29 0.82 1.06 .93 3.66 3.37 2.89 2.52 2.71 3.05 1.52 1.59 1.25 0.84 1.02 1.06 3.83 3.35 2.84 2.37 2.58 3.02 1.53 1.48 1.26 .83 1.04 .99 Av. 1 2 3 4 5 1-5 3.27 2.82 2.83 2.41 2.41 2.77 1.44 1.11 1.11 1.05 0.89 .85 3.41 3.00 2.98 2.38 2.71 2.92 1.45 1.16 1.22 0.88 1.20 .98 3.34 2.91 2.91 2.40 2.56 2.84 1.43 1.13 1.16 .96 1.06 .91 1 2 Av. 1.83 1.54 1.60 1.29 1.37 1.30 5.03 4.17 3 4 5 1-5 4.92 4.31 3.70 3.28 2.93 3.85 3.43 3.18 3.18 3.84 1.91 1.58 1.65 1.24 1.18 1.36 4.98 4.24 3.57 3.23 3.06 3.84 1.85 1.62 1.61 1.25 1.26 1.31 Av. 1 2 3 4 5 1-5 3.13 2.66 2.28 2.64 2.28 2.62 1.15 0.93 0.93 0.87 1.06 .76 3.31 3.02 2.64 2.92 2.57 2.92 1.24 1.17 0.98 1.17 1.05 .95 3.22 2.84 2.46 2.79 2.43 2.77 1.18 1.07 1.00 1.02 1.05 .87 Androstenol III II CPD 0 = extremelyweak;8 = extremelystrong. For concentrations,seeTable II. Fourreplicadons fromeachrespondent enterinto the calculation of thesemeans. HUMAN BODY Table ODOR 415 IIIb Intensity Ratings of "Body" Shirts Women Women's shirts Mean 3 Men + SD Mean W+M + SD Mean + SD 1 2.00 .96 2.40 .03 2.21 .90 2 2.79 1.31 3.00 1.31 2.90 1.29 3 4 5 5.57 2.00 3.00 1.45 .68 1.47 5.87 2.47 3.47 1.25 1.36 1.64 5.72 2.24 3.24 1.33 1.09 1.55 6 7 8 9 10 2.93 3.60 4.07 2.87 2.60 1.22 1.45 1.62 1.73 1.68 2.93 2.67 3.33 2.40 2.07 1.39 1.29 1.76 1.84 1.03 2.93 3.13 3.70 3.63 2.33 1.28 1.43 1.71 1.77 1.40 11 12 13 14 15 3.64 3.43 3.36 3.14 3.00 1.22 1.09 1.08 .66 1.24 3.40 3.60 3.20 3.67 3.53 1.59 1.06 1.26 .98 .83 3.52 3.52 3.28 3.41 1.40 1.06 1.16 .87 3.28 1.07 16 17 18 3.07 3.87 3.47 1.67 1.60 1.55 3.33 3.73 3.00 1.68 1.71 1.81 3.20 3.80 3.23 3.17 2.63 1.65 1.63 1.68 1.46 1.61 Day 1 Day 2 hv. 19 3.07 1.44 3.27 1.53 20 2.53 1.36 2.73 1.07 1-20 2.96 2.96 2.96 3 Two replications fromeachrespondent enterinto the calculation of thesemeans. CONCENTRATION AND PLEASANTNESS The meanpleasantness ratingsfor the "solution"shirtsand the "body" shirtsare listed in Table V. ThesevaluesareBasedon threereplicatedpleasantness judgementsfor each shirt by eachrespondent;groupsI (pleasantness ratingson day 1) and II (pleasantness ratingson day 2; compareTable I) were combinedfor the "solution"shirtsonly. The scaleused(seeFigure 2) ran from 0 = extremelyunpleasantto 8 = extremelypleasant. The correlationBetweenpleasantness and concentrationwas calculatedfor eachof the test substances, using the two-tailed non-parametricgroup trend test (7) (Table VI). The most striking finding is the very high negativecorrelationBetweenconcentration and pleasantness for androstenoland alsofor substances II and III. Cyclopentadecanolide showsa totally different pattern in this respect:pleasantness doesnot correlate significantlywith concentration,and in this casethe correlationcoefficientsare positive. The meanpleasantness rating of the T-shirtsworn by womenwasnearthe midpoint of the scale.The women'sshirtswere, on the average,rated somewhatmore pleasantthan were the shitsworn by men. The differencein rating Betweenthe mostpleasantand the leastpleasantshirt wasmuchhigherfor the shirtsworn by men than for the shirtsworn by women(Table XI). 416 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table IIIb (cont'd) IntensityRatingsof "Body"Shirts Wonhen Men's shirts Men W q- M Mean -+ SD Mean + SD Mean -+ SD 1 2 3 4 5 3.29 3.86 4.93 5.07 3.50 t.38 1.29 1.54 1.44 1.70 3.27 3.60 4.87 5.00 4.40 1.39 1.50 1.64 1.46 1.88 3.28 3.72 4.90 5.03 3.97 1.36 1.39 1.57 1.43 1.86 6 7 8 9 10 4.20 4.93 5.20 3.07 3.33 1.86 1.58 1.78 1.75 1.88 3.27 3.53 4.47 3.13 2.93 1.79 1.88 1.96 1.46 1.58 3.73 4.23 4.83 3.10 3.13 1.86 1.85 1.88 1.58 1.72 11 12 13 4.14 3.43 2.86 1.03 t.45 .77 3.80 3.13 3.13 1.47 .92 .92 3.97 3.28 1.27 1.19 3.00 .85 14 15 5.64 3.50 1.74 1.16 5.07 3.80 1.16 1.01 5.34 3.66 1.47 1.08 16 17 18 2.13 4.46 2.53 1.55 1.64 1.41 2.00 3.80 2.47 1.00 1.61 .99 2.07 1.28 4.13 1.63 2.50 1.20 19 2.53 1.30 2.67 1.59 20 3.40 1.76 3.47 1.77 2.60 3.43 1.43 1.74 1-20 3.55 Day 1 I Day 2 II Av. 3.47 3.39 Androstenoland III, in their highestdilutions, obtainedpleasantness ratingscomparable to the shirts actually worn. II, at all concentrationlevelswas rated lower in pleasantness than the othertestsubstances; at the highestconcentration levelsits mean ratingswere nearthe low end of the pleasantness scale.Cyclopentadecanolide, at all levels, obtainedratingsnear the midpoint of the scaleand comparableto the female T-shirts. The differencesbetweenthe pleasantness ratingsassignedby male and femalesubjects were, in general,slight. The malesubjectstendedto give morefavorableratingsto the shirts actuallyworn, both by womenand by men, and somewhatlessunfavorable ratings to the higher concentrations of II. Table IV Significanceof the CorrelationBetweenConcentrationand PerceivedIntensity Women Men W q- M Substance z p z p z p Androstenol III 6.05 2.51 .0001 .006 4.49 3.21 .0001 .0007 7.40 4.05 .0001 .000 l II CPD 7.96 3.39 .0001 .0004 7.26 3.56 .0001 .0002 10.76 4.91 .0001 .0001 HUMAN BODY Table ODOR 417 V Pleasantness Ratings for "Solution"Shirts Women Men (n = 29) Substance Sol. 2 Androstenol Av. W (n = 30) q- M (n = 59) Mean • + SD Mean + SD Mean + SD 1 2 3 4 5 1-5 2.99 3.03 3.63 4.03 3.93 3.53 1.37 1.49 1.17 1.06 1.09 .93 2.98 3.10 3.41 3.41 3.60 3.30 1.27 1.04 1.01 0.92 0.84 .82 2.98 3.07 3.52 3.72 3.77 3.41 1.31 1.27 1.08 1.02 .98 .88 1 2 3 4 2.85 2.93 3.71 3.78 1.39 1.32 1.00 1.14 2.79 2.79 3.60 3.37 1.21 1.08 0.98 1.03 2.82 2.86 3.66 3.58 1.29 1.20 1.00 1.10 III 5 3.72 1.11 3.73 1.18 3.73 1.13 Av. 1-5 3.40 .83 3.25 .82 3.32 .82 Av. 1 2 3 4 5 1-5 1.62 2.05 2.60 3.07 3.22 2.51 1.53 1.41 1.37 1.18 1.06 1.11 2.10 2.30 2.73 3.11 3.01 2.65 1.58 1.49 1.58 1.46 1.00 1.24 1.86 2.18 2.67 3.09 3.11 2.58 1.56 1.44 1.47 1.32 1.02 1.17 1 2 4.01 4.05 1.43 1.20 3.88 3.71 1.34 l. 10 3.94 3.88 1.37 1.15 3 3.87 1.11 3.72 0.93 3.79 1.00 4 5 1-5 3.55 3.87 3.86 0.96 1.07 .95 3.43 3.69 3.68 1.31 1.04 1.50 3.49 3.78 3.77 1.13 1.06 .97 II CPD Av. 0 = extremely unpleasant, 8 = extremely pleasant. For concentrations,seeTable II. PERCEIVED INTENSITY AND PLEASANTNESS RATING The correlationsbetweenperceivedintensityand pleasantness rating were calculatedfor each concentrationlevel of each of the four test substances,and using the two-tailed Pearsoncorrelationtest on the paired mean pleasantness and intensity ratings of each individual respondent.The findings are presentedin Table VII. They show clearly differentpatternsfor the four test substances, and alsofor male and femalerespondents. With substance II, correlations are high and negative,especiallyat high concentration levels, both for male and female respondents.With androstenoland substanceIII, correlationswere also negativeat all concentrations,but significantonly for female respondentsand only at the two highestconcentrationlevels(and, for substanceIII, at the lowest concentrationlevel). With cyclopentadecanolide there seemsto be no covariation betweenperceivedintensityand pleasantness. With the shirtsactuallyworn, the correlation,althoughnegativein most cases,is significantonly in isolatedinstances. SEX ATTRIBUTION The findingsregardingsexattributionare summarizedin TablesVIII and IX. Table 418 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table V (cont'd) Pleasantness Ratings for "Body" Shirts Women Men's shirts 11 12 W d- M 1.77 --- 1.21 3.69 1.03 3.96 1.11 1.69 1.02 4.06 1.15 3.03 .98 2.40 --- 1.03 3.71 1.36 2.10 + 3.70 1.15 1.19 4.31 2.49 4.41 3.41 1.38 1.16 1.49 1.01 4.14 2.10 4.24 3.23 1.24 1.15 1.32 1.00 4.08 2.68 3.84 1.53 1.56 1.62 3.91 3.16 4.05 .76 1.20 .93 3.40 2.92 3.95 1.19 1.39 1.30 9 10 6-10 4.14 3.35 3.49 1.46 1.38 1.20 4.24 3.77 3.82 .97 1.54 .76 4.19 3.56 3.66 1.22 1.45 1.00 11 12 3.71 4.44 1.41 .78 4.09 4.75 1.17 1.67 3.91 4.60 1.28 1.30 13 3.46 1.•13 4.16 1.30 3.82 1.25 Av. 14 15 11-15 4.40 4.36 4.08 1.10 .79 .74 4.03 4.53 4.31 1.00 1.21 1.05 4.21 4.45 4.20 1.04 1.02 .91 Av. 6 7 8 9 10 6-10 3.59 3.21 3.23 3.38 3.90 3.46 1.19 1.91 1.96 1.55 1.58 1.49 3.69 3.53 3.41 3.63 4.22 3.69 1.28 1.27 1.03 .89 1.24 .90 3.64 3.37 3.32 3.51 4.06 3.57 1.22 1.60 1.54 1.25 1.41 1.21 13 14 15 11-15 Av. 6 7 8 •4xV. Women's Men shirts Shirts 1-5 and 16-20, both men'sand women's,had not beenratedfor pleasantness. VIII indicatesthe mean intensity rating for eachcell alongwith its standarddeviation; Table IX presentsthe mean pleasantness ratingswith their standarddeviations.Onesamplechi squaretestsfor eachsubstance,conductedseparatelyfor femaleand male respondents,lead to the following conclusions: Androsreno/is predominantly judgedto be femaleat all concentration levels,both by men and by women. When calculatedfor combinedconcentration levels,the female attribution was highly significant(chi squaretest). Table VI Significance of the CorrelationBetweenConcentration and Pleasantness Women Men (n = 29) Substance Androstenol III II CPD W (n = 30) d- M (n = 59) z p z p z p -4.02 - 2.87 -7.26 1.17 .0001 .004 .0001 .224 - 3.40 - 3.65 -4.93 1.65 .0006 .0003 .0001 .099 - 5.25 -4.62 -8.60 2.00 .0001 .0001 .0001 .046 HUMAN BODY ODOR Table 419 VII CorrelationBetweenPerceivedIntensity and Pleasantness Rating Substance Androstenol Sol. 1 2 3 4 5 1-5 III 1-5 Sex p p f - 0.48 0.008 m tot -0.29 - 0.38 0.13 0.003 f -0.72 0.001 m tot -0.36 -0.54 0.052 0.001 f -0.036 0.85 m -0.25 0.19 tot - 0.09 0.45 f -0.18 0.33 m -0.008 0.96 tot -0.15 0.25 f -0.24 0.20 m -0.05 0.77 tot - 0.17 0.17 f -0.37 0.046 m -0.20 0.29 tot - 0.28 0.029 f -0.43 0.021 m -0.26 0.17 tot - 0.34 0. 008 f -0.49 0.007 m -0.29 0.12 tot - 0.39 0.002 f -0.20 0.30 m -0.03 0.88 tot - 0.11 0.40 f -0.15 0.44 m -0.09 0.72 tot - 0.11 0.41 f -0.43 0.02 m -0.015 0.94 tot - 0.17 0.17 f -0.27 0.15 m tot -0.03 - 0.116 0.88 f -O.84 0.001 m -0.57 0.001 tot - 0.68 0.001 f -0.62 0.001 m -0.74 0.001 tot - 0.68 0.001 f -0.66 0.001 m -0.55 0.002 tot - 0.60 0.001 O.379 420 JOURNAL OF THE SOCIETY OF COSMETICCHEMISTS Table VII (cont'd) Substance Sol. 4 5 1-5 1 CPD Sex p p f -O.43 0.018 m tot -0.38 - 0.40 0.036 f -0.26 0.17 m -0.42 0.02 tot - 0.34 0.008 f m tot -0.75 -0.62 - 0.67 0.001 f m tot 2 3 4 5 1-5 Women's shirts 11 12 0.10 15 6 0.001 0.60 0.94 -0.02 0.03 0.78 -0.43 0.02 m -0.26 0.17 tot - 0.35 0.007 f 0.069 0.72 m -0.06 0.75 tot - 0.006 O.95 f -0.46 0.01 m -0.32 tot - 0.37 0.086 0.004 f 0.10 0.60 m -0.08 tot - 0.001 0.67 0.99 f -0.21 0.27 m -0.22 0.23 tot - 0.229 0.08 f -.24 .41 m .45 .10 tot .11 .57 .18 .55 f -- .02 .93 .04 .82 f - .32 .27 m tot -.31 - .32 .26 f .27 .35 m .38 .17 tot .21 .27 tot 14 0.001 f m 13 0.002 f -.33 .09 .25 m .53 .04 * tot .15 .44 f - .49 .06 m -.13 .65 tot -. .13 28 HUMAN BODY ODOR 421 Table VII (cont'd) Substance Sol. 7 8 Sex p f - .26 .35 m - .35 .20 tot -.31 .10 f - .40 .14 .22 .44 m 9 10 Men's shirts 11 12 p tot -. 17 .37 f - .47 .08 m - .01 .98 tot -.29 .12 f m tot - .72 .23 - .43 .003* .41 .02* f -.10 .72 m tot -.13 -.14 .65 .46 f - .08 .78 m tot -.33 -. 18 .23 .34 f m -.38 .03 .17 .93 tot -. .60 f - .44 .1.1 m - .32 .25 tot -.41 .03* f - .34 .23 .18 .52 tot -. 04 .84 f - .42 .12 .06 .84 tot - .23 .22 f m tot -.16 - .05 .58 .87 -. 17 .37 f m -.37 .20 .17 .47 tot -. 17 .37 9 f m -.43 .42 .11 .12 tot -.13 .51 10 f rn tot -.44 - .46 - .46 .10 .08 .01' 13 14 15 m 6 m 7 8 10 Substance III is judgedto be femaleonlyat lowerconcentration levels(3, 4, and5) both by men and by women. Substance II is judgedto bemaleat highconcentration levels(1 and2--but theassign- 422 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table Villa Concentration,PerceivedIntensity, and SexAttribution--"Solution" Shirts Women (n = 29) Men (n = 30) Substance Sol. 2 Mean• _+SD n4 Mean + SD n Mean ___ SD n Mean _+SD Androstenol 1 2 3 4 5 3.20 2.95 2.49 2.10 2.34 1.33 0.89 1.06 0.94 1.01 16 20 19 17 20 4.98 4.14 3.33 2.35 2.69 1.29 1.97 1.55 0.61 1.21 13 9 10 12 9 3.49 3.15 2.83 2.57 2.61 1.57 1.42 1.18 0.85 1.02 19 17 18 18 18 3.96 3.65 2.98 2.44 2.85 1.46 1.80 1.41 0.85 1.05 1 2 3 4 5 2.73 2.52 2.86 2.32 2.36 1.24 1.54 1.04 0.95 0.82 13 12 21 21 21 3.70 3.03 2.75 2.63 2.53 1.48 0.62 1.38 1.32 1.10 16 17 8 8 8 3.04 2.78 2.97 2.25 2.52 1.44 0.99 1.15 0.70 1.23 12 15 22 17 21 3.65 3.22 3.00 2.56 3.14 1.45 1.31 1.49 1.08 1.05 18 1 2 3 4 5 4.39 2.88 2.87 3.10 2.81 1.72 1.45 1.13 1.34 1.42 7 6 11 17 18 5.09 4.69 4.19 3.52 3.14 1.87 1.36 1.67 1.20 1.25 22 23 18 12 11 4.30 3.52 2.99 2.78 3.00 1.76 1.10 1.50 1.06 1.15 11 13 17 19 21 5.46 4.66 4.00 3.86 3.58 1.90 1.74 1.73 1.29 1.21 19 17 1 2 3 4 5 3.06 2.78 2.28 2.58 2.25 1.07 0.90 0.94 0.78 1.13 24 24 24 20 20 3.45 2.10 2.30 2.78 2.73 1.56 0.98 0.99 1.09 0.92 5 5 5 9 9 3.44 2.99 2.57 2.83 2.42 1.09 1.16 0.96 1.21 1.03 23 20 22 19 19 2.89 3.08 2.84 3.07 2.82 1.67 1.26 1.08 1.14 1.07 III II CPD 11 12 12 12 15 8 13 9 13 11 9 7 10 8 11 11 •,2 CompareTableIII. 3 •? = shirts attributed to female wearer. d' = shirts attributed to male wearer. 4 n = numberof judgesin thegroupgivingthesexattributionindicated. ment is significantonly with femalerespondents) and femaleat low concentrations (3, 4, and 5--but this assignment is significantonlywith malerespondents). Cyc/opentadecano/ide is judgedto be femaleat all concentrations, both by men and by women. Shirtswornbywome,are judgedto be female,both by men and by women. Shirtswornbymenarejudgedto be malebothby menand by women(TableVillA) SEX ATTRIBUTION, INTENSITY, AND PLEASANTNESS The perceivedintensityof anygiven shirt is higherand the pleasantness ratingis lower in those cases where the shirt is attributed to a male wearer than in those where it is attributedto a femalewearer.This is true bothforfemaleandfor malerespondents and it holds for the four substances tested at all concentrations. The absolute differences are not great, but with very few exceptionsthey consistentlypoint in the samedirection (Table X). For the T-shirtsactuallyworn, the test designenablesus only to test the correlationbetweensexattributionand intensity.Here the pattern(higherperceived HUMAN BODY ODOR Table 423 VIilb Concentration, PerceivedIntensity,and SexAttribution--"Body" Shirts Women Women's shirts Mean + SD n Mean Men + SD n Mean + SD n Mean + SD n Shirt Men's 1 2.00 1.00 11 2.00 1.00 3 2.33 .65 12 2.67 1.53 3 2 3 4 5 2.60 5.33 2.11 3.00 1.34 2.08 .78 1.63 5 3 9 10 2.89 5.64 1.80 3.00 1.36 1.36 .45 1.16 9 11 5 4 2.88 5.50 1.80 3.82 .64 1.4l .84 1.66 8 8 5 11 3.14 6.29 2.80 2.50 1.86 .95 1.48 1.29 7 7 10 4 16 17 3.00 3.50 13 10 2 5 9 7 6 8 2.00 1.89 1.60 9 4 6 2.30 1.29 1.20 8 10 9 3.50 4.63 2.86 3.60 3.50 2.26 1.77 6 11 9 3.22 2.71 3.13 3.10 2.22 1.30 .95 3.67 3.00 2.11 3.50 4.60 3.33 3.25 3.17 2.12 1.34 18 19 20 1.68 1.65 .52 1.34 1.05 1.22 2.07 2.51 7 5 6 3.00 3.78 4.00 3.50 5.00 1.41 1.30 0 .71 2.83 5 9 2 2 2 3.44 4.00 5.08 5.39 3.25 1.42 1.41 1.62 1.37 1.49 9 5 12 12 12 3.13 3.60 2.00 4.60 4.50 1.36 1.14 0 1.14 1.73 8 5 1 5 4 3.43 3.60 5.07 5.20 4.36 1.51 1.71 1.49 1.62 2.01 7 10 14 10 11 16 1.20 1.10 5 2.60 1.58 10 2.25 1.17 8 1.71 .76 7 17 18 3.40 1.00 1.14 1.00 5 3 5.00 2.92 1.63 1.24 10 12 3.40 2.57 1.67 .98 5 7 4.00 2.38 1.63 1.06 10 8 19 2.25 1.28 8 2.86 1.35 9 2.43 1.62 7 2.88 1.64 8 20 3.00 .02 4 3.55 2.02 11 3.71 1.80 7 3.25 1.83 8 shirts Shirt 1 2 3 4 5 Shirts6 15, both men'sand women's,had not beenratedfor sexattribution. intensitygoeswith male attribution)holdsfor femalebut not significantlyfor male respondents. DISCUSSION CONCENTRATION AND PERCEIVED INTENSITY The decrease of perceivedintensitywith increasing dilution is amazinglyslight. With substance II, whereit is mostpronounced,it amountsto 1.92 scalepointson a 9-point scalefor a dilution factor of 256. With androstenoland cyclopentadecanolide, a dilution factorof 2401 leadsto a decrease in perceivedintensityof only 1.25 and 0.79 scale pointsrespectively (TableIII). In the faceof suchslight responses to majordilution steps,one may be led to wonderwhetherthe respondents really smell somethingor whethermanyof them werejust guessing--a doubtwhich is particularlypertinentin view of publishedreportsaboutthe high incidenceof anosmiato androstenol (8), in view, moreover,of the observation,commonlyknownto perfumers,that many people areveryinsensitive to macrocyclic musks,andin viewof themassive taskdemanded of 424 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table IX Pleasantness Rating and SexAttribution--"Solution" Shirts Women (n = 29) •3 Men (n = 30) C•3 • C• Substance Sol. 2 Mean • mSD n Mean _+SD n Mean mSD n Mean _+SD Androstenol 1 2 3 4 5 3.44 3.42 3.84 3.96 3.93 1.19 1.27 0.84 1.03 1.02 16 20 19 17 20 2.44 2.19 3.23 4.14 3.93 1.42 1.65 1.61 1.13 1.30 13 9 10 12 9 3.05 3.43 3.54 3.37 3.56 1.40 0.62 1.03 0.89 0.94 19 17 18 18 18 2.85 2.67 3.22 3.47 3.67 1,01 1.33 0.98 1.00 0.71 1 2 3.10 3.11 1.48 1.36 13 12 2,65 2.80 1.33 1.31 16 17 2.81 2.84 1.35 0.89 12 15 2.78 2.73 1.15 1.28 3 3.37 0.90 21 3.67 1.30 8 3.52 1.08 22 3.83 0.64 8 4 5 3.87 3.78 1.21 1.14 21 21 3.54 3.58 0.96 1.10 8 8 3.63 3.79 1.20 1.20 17 21 3.03 3.59 0.67 1.19 13 1 2 3 4 2.29 3.22 3.30 3.28 1.56 1,44 1.41 1.28 7 q.41 6 1.74 11 2.17 17 2.78 1.49 1.26 1.18 1.02 22 23 18 12 2.97 2.87 3.20 3.72 1.75 1.43 1.69 1.18 11 13 17 19 1.60 1.86 2.13 2.06 1.26 1.42 1.25 1.33 19 5 3.22 1.13 18 3.21 0.97 11 3.11 1.07 21 2.78 0.82 III II CPD 1 12 12 12 18 15 9 17 13 11 9 1 3.82 1.49 24 4.93 0,49 5 4.10 1.30 23 3.14 1.27 7 2 4.04 1.25 24 4.07 1.12 5 3.70 1.09 20 3.73 1.20 10 3 4 5 3.96 3.58 3.95 1.19 0.95 1.14 24 20 20 3.47 3.48 3.70 0.51 1.03 0.95 5 9 9 3.88 3.53 3.54 0.86 1.34 1.13 22 19 19 3.29 3.27 3.94 1.03 1.31 0.84 11 8 11 •,2 CompareTable III. 3 Attributed to female; wearer;attributed to male wearer. No valueswereobtainedfor the "body"shirtsdueto the designof the experiment(seeTableI). the test subjects:smelling30 T-shirts in onesession,five timesin a row, may well have led to fatigue. However,the highly significantgrouptrend z-scores (TableIV) for all four of the test substances, both amongmale and femalejudges,give us firm assurance that the respondents asa groupwerenotguessingbut reportingrealobservations.* The consideration that the respondents weregivenno cluesthat they weresmellingthe same materialsin differentdilutionsgivesaddedweight to thesehigh z-scores. The methodology of this studydoesnot warrantthe assignment of numericalvaluesto the slopeof the psychophysical function(that is, the slopeof the plot of log I versuslog C) nor to the detectionthreshold.However,it may be sat•lysaidthat the studyindicatesthat for all of the materialsstudied,both the slopeand the detectionthresholdare very low. In the caseof the highestdilutionsof androstenol,II and cyclopentadecanolide, approx.10-9 mole of odorantwaspresentat eacharm pit; for III, the lowest testedlevelcorresponded to about5 x 10-9 mole. Taking into accountthe low volatility of thesesubstances, the concentration in the air must havebeenvery low indeed. We havepreviouslyargued(9) that materialswith a low slopeof the psychophysical * We haveno explanationfor the curiousfinding that the meanintensityrating for the highestdilution of both androstenol and III (solution5) is higherthan for solution4. The differences are non-significant. HUMAN BODY Table ODOR 425 X SexAttribution, Intensity, and Pleasantness x2--One sampletestfor intensityandsexattribution "Solution" shirts Women Men X2 = 12.8 X2 = 12.8 df = 1 df = 1 p < .00l p < .001 "Body" shirts Women Men X2 = 5.0 X2 = 1.8 df = 1 df = 1 p < .05 p < .20 x2--One sampletestfor pleasantness andsexattribution* "Solution" shirts Women Men X2 = 5.0 X2 = 5.0 df = 1 df = 1 p < .05 p < .05 * An analysisfor the "body" shirtsis not possibleheredue to the designof the experiment(seeTable I). function and a low thresholdare particularlyvaluableto the perfumer, becausethey tend to be persistentand diffusive. The low slopeof the psychophysical functionof androstenol,coupledwith its low threshold,may well have biological,evolutionarysignificance.As a metaboliteof a male hormone, as a substancethat occursnaturally in excretionsof man and other mammals,it may well have, or have had, signal functionfor man (1,2). For such substances, it is obviouslyvery usefulthat they be perceptibleat greatdilutionsin the air (i.e., at great distances) without becomingoverpoweringwhensmelledat far higher concentrations (i.e., from nearby). The finding that both male and femalerespondents ratedthe men'sshirts,on average, higher in odor level than the women'sshirts, without knowing what they were smelling, confirmsSchleidt'sfindings(5). CONCENTRATION AND PLEASANTNESS The negativecorrelationbetweenconcentrationand pleasantness for androstenone,II, and III is in line with perfumers'common experienceregarding "animal" odorants (civet, naturalmusk, indol, and others):that they must be usedsparinglyand lead to rejectionif usedat too high levels(comparealsoreferences 2 and 4). The synthetic musk, cyclopentadecanolide, doesnot exhibit this behaviorat all. In a qualitativesenseaswell, cyclopentadecanolide wasfoundto be considerably further removedfrom actualhuman body odor than are the other odorantstested;numerous respondents remarkedthat the shirtswith cyclopentadecanolide smelled"perfumed." An interestingincidentalfindingwasthe wide variabilityin pleasantness ratingsfor the T-shirtsthat hadactuallybeenworn. Both for shirtswornby womenandthoseworn by men, and both with femaleand with male respondents, the differencebetweenthe shirt judgedmostpleasantandthe onejudgedleastpleasant wassignificant(TableXI)--but more highly for men'sthan for women'sshirts. PERCEIVED INTENSITY AND PLEASANTNESS The finding that the correlations betweenperceivedintensityand pleasantness (Table 426 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table XI Comparisonof Mean Pleasantness Rating for Most Pleasantand LeastPleasantT-shirt (Wilcoxon matchedpairssignedrankstest) Most PleasantRating FemaleRespondents Least PleasantRating Mean q-SD 4.44 4.06 .78 1.15 Shirt 13 14 Mean q-SD z p (2-tailed) 3.46 1.69 1.13 1.02 2.48 3.30 .013 .001 I Women's shirts Men's shirts Shirt 12 15 II Women's shirts 10 3.90 1.58 7 3.21 1.91 1.57 .117 Men's shirts 9 4.14 1.46 7 2.68 1.56 2.94 .003 Male Respondents I Women's shirts 12 (.75 1.67 14 4.03 1.00 2.07 .038 Men's shirts 15 4.4l 1.49 11 2.40 1.03 3.35 .001 II Women's shirts Men's shirts 10 9 4.22 4.24 1.24 .97 7 7 3.53 3.16 1.27 1.20 2.51 2.54 .012 .011 VII) are, on the whole, lessconsistentand lower than betweenconcentration and pleasantness(Table VI) may seemsurprisingat first sight. It must be born in mind, however, that Table VII recordsthe relationshipbetweenintensityperceptionand pleasantnessperceptionat a givenconcentration /eve/.'it takes into accountonly the intensity differencesdue to subjectivefactors,not thosedue to the objectivefactor, concentration. This is true also for the "overall" correlations for solutions 1-5. The differentcorrelationpatternsare curious:for substance II one can say, at all concentrationlevelsstudied,that "the more stronglyoneperceivesit, the moreunpleasant one finds it." It is simply an odor that nearly everybodyconsidersinherently unpleasant.With androstenoland substance III, there appearsto be, for many people,a most pleasantlevel that is low but higher than 0. With male respondents,the correla- tionsbetweenintensityand pleasantness are not significantat any concentration level; with female respondents,they are significantand negativeonly at the two highest concentrationlevels. (We have no reasonableexplanationfor the significantnegative correlation,with women,at the lowestconcentration levelof III.) The finding (TableV) that for thesetwo substances the relationshipbetweenconcentration and pleasantness showsa far clearertrend at higher concentrations (solutions1-3) than at lowerconcentrations (solutions3-5) is in line with this interpretation. With cyclopentadecanolide thereis no consistent patternlinking pleasantness ratingsto perceivedintensity at a given concentration(Table VII), nor pleasantness to concentration (Table V). Here, the interpretation"somelike it strong"appearsappropriate.It must be remembered,for this and all other findings of this test, that the odorswere smelledon T-shirts in a contextin which they werejudged"asbodyodors."Different relationshipsbetweenintensityandpleasantness might well haveresultedhad they been smelledon smellingblottersor in perfumebottlesasaromaticchemicals or asperfumes. With the T-shirts actually worn, a pattern is apparentthat is alsopresentwith androstenol and III: the negativecorrelationbetweenperceivedintensityand pleasantness, that is, the feeling that "strongis unpleasant,"is strongerwith femalethan with male respondents.This is particularlypronouncedwith the shirtsworn by maleswhereall shirts give, among female respondents,negative correlations(see Table VII). This HUMAN BODY ODOR 427 finding that womenreject the more intensivebody odors,or body-likeodors,more stronglythan do men confirmsSchleidt'sobservations (5). SEX ATTRIBUTION Perhapsthe most surprisingfinding of the studyis the fact that a significantmajority both of male and of femalerespondents attributed the shirtsthat had beentreatedwith androstenol to female wearers, at all concentration levels tested. This is unexpectedin view of the fact that androstenol and the closelyrelatedandrostenoneoccurin human plasma(10) and in axillary sweat(11) at far higher concentrations in men than in women and would, therefore,be expectedto be associated with male bodyodor. Actually, the resultsof the behavioralexperimentsthus far reportedin no way conflict with the notion that androstenolis predominantlyperceivedasfemale. In the dentist's chairexperiment(12), onewould expectthe observedpatternof response (selectionby women, avoidanceby men) if the chair had smelledof somefeminineperfume. In the rating of photographs(13), it was the females,not the males, that were judged more attractive and sexyin the presenceof androstenol.Our finding does, however, throw additional doubt (if such were needed) on the claims of mail-order sex aid houses that men's lotions containing androstenolare sexualattractantsto women. Another novel finding was the concentrationdependenceof sexattribution, found both for substanceII and for substanceIII. Although novel, this finding wasnot unexpected in view of the earlierreportedtendencyto associate strongerbodyodorswith men and fainter odors with women. As to the sex attribution of cyclopentadecanolide, we cannotrule out the possibility that the predominantlyfemaleattribution of shirtstreatedwith this substance wasdue to an association with women'sperfumesrather than with femalebodies. The predominantlycorrectsexassignmentof the shirtsworn by womenand by men impressivelyreplicatesthe findings of Schleidteta/. (5), as do the tendenciesto attribute strongerand lesspleasantodorsto men and fainter and more pleasantodorsto women. It is fascinatingto observehow two patterns that are commonplacein the field of generalaesthetics arealsoreflectedin the olfactoryfield: (1) the "beautyandthe beast" pattern, i.e., the notion that (young) women are more estheticallypleasingthan men, (2) the greaterprevalenceamongwomenof the tendencyto judge by estheticcriteria: the differencesin intensityare perceivedequallyby men and women,* but the women associatethem more strongly with "pleasing" or "not pleasing." SUMMARY In the presentstudy, it wasshownthat when the syntheticbodyodoranaloguesII and III are presentedto human subjectsunder conditionswhere they might be taken for * Equalability to perceivedifferences doesnot necessarily meanequalawareness of them in dailylife. The designof the experimentinevitablyinducedhigh odorawareness amongall respondents; thereis considerableevidence that normallyodorawareness is higheramongwomenthanamongmen. 428 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS human body odors, reactionsto thesechemicals'are similar to thoseto natural body odorsin a number of respects: 1. They are predominantlyattributed to the one or the other sex. 2. Perceivedintensityand pleasantness rating are negativelycorrelated. 3. When the materialsare perceivedas relatively strong or as relativelyunpleasant, they tend to be attributed to male wearers. 4. At low concentration, the averagepleasantness rating is comparable to that of natural body odor (this is not true for the lesspleasantsubstanceII). For the syntheticmusk, cyclopentadecanolide (IV), observations 2 and 3 do not hold. Furthermore,thesesyntheticodorantsincluding cyclopentadecanolide were similar to the naturalhuman body odorant,androstenol,in the following respects: 1. Low olfactorythreshold. 2. Low slopeof the psychophysical function. A surprisingand importantfinding wasthe attribution of androstenolto femalesrather than to males,at all concentrationlevels,both by male and femalerespondents. The responses to the shirtsthat had actuallybeenworn by test subjectsconfirmedM. Schleidt'sfindingsin all respects: 1. Correctsexattribution by a significantmajority of the respondents. 2. A significanttendencyto attribute strongersmellingshirtsto male wearers. 3. Strongerrejectionof high intensityodorsby femalerespondents. REFERENCES (la) A. Comfort,Dragoco Report,19(11), 226-233 (1972). A. Comfort in Pheromones, M. C. Birch, Ed. (North Holland Publ., Amsterdam 1974), pp 386-396. (lb) I. Bloch, Odorants Sex/•alis(PanurgePress,New York, 1934). (2) For reviewsof this work, cf. RichardL. Dory, Chemical Senses, 6(4), 351-376 (1981);J. N. Labows, K.J. McGinley, and A.M. Kligman, J. Soc.Cosmet.Chem.,34, 193-202 0uly 1982); D. B. Gower, M. R. Hancock,and L. H. Bannisterin Biochemistry of Tasteand Olfaction, R. H. Caganand M. R. Kate, Eds. (AcademicPress,New York, 1981), pp 7-31. (3) P. Jellinek, Die psychologischen Grundlagen derParfiimerie (Dr. A. Hi•thig Verlag, Heidelberg, 1951). (4a) E. Paukner,J. Soc.Cosmet. Chem.,16, 515 (1965). (4b) W. Steiner,E.-F. Hanisch,and D. Schwarz,Parr. undKosm.,58, 189-196 (1977). (5) M. Schleidt, Ethology Sociobiol.,1, 225-231 (1981); and M. Schleidt, B. Hold, and G. Artill, J. Chem.Ecol., 7, 19-31 (1981). (6) J. S. Jellinek, Dragoco Report,34(2), (1987). (7) A. R. Jonckheereand G. H. Bower,Brit. J. Math. and SocialPsychol.,20(2), 163-186 (1967). (8) J. E. Amoore, P. PelosiandJ. L. Forrester,Chem.Senses and Flavor, 2, 401-425 (1977). (9) J. S. Jellinek, Dragoco Report,26(2), 85-87 (1979). (10) D.C. Bicknell and D. B. Gower,J. SteroidBiochem., 7, 451-455 (1976). (11) S. Bird and D. B. Gower,J. Endocrinol., 85, 8P-9P (1980). (12) M.D. Kirk-Smith, D. A. Booth, in Olfaction and TasteVII, H. v. d. Starre,Ed. (IRL Press,London 1980), pp 397-404. (13) M.D. Kirk-Smith, D. A. Booth, D. Carroll, and P. Davies,Res.Commun. in Psychol., Psychiatry and Behavior,3(4), 379-384 (1978).
© Copyright 2024