(Return to index)

Classics Editor's note: The numbers appearing in parentheses throughout the text refer to the items
in the bibliography at the end of the volume.
 

THE experiments in taste and smell dealt with the following subjects:

A. Taste.
    1. Threshold of presence [1] for sweet, salt, sour, and bitter.
    2. Threshold of recognition [1] for sweet, salt, sour, and bitter.
    3. Discriminative sensibility at T2.
    4. Discriminative sensibility for strong tastes (viz., those of series B of Table VIII).

B. Smell.
    1. Threshold of presence for cloves and violet.
    2. Threshold of recognition for cloves and violet.
    3. Discriminative sensibility at T2.
    4. Discriminative sensibility for strong odors (viz., those of series B of Table IX).

The substances used for the four tastes were saccharin, chemically pure salt, sulphuric acid, and sulphate of quinine. Two series of solutions in distilled water were prepared from each substance. Series A began below the normal threshold of presence and extended above the average threshold of recognition. Series B consisted of solutions which were all strong to the normal taste. The limits of the series, and the gradations necessary in each one were determined experimentally. The bottles containing the solutions [p. 51] were all alike in appearance. The series of solutions, in percentages, are given in Table VIII.

No attempt was made to control the temperature of the solutions any more closely than the temperature of the room.

Since taste is a sense which, like smell, is peculiarly subject to illusions at the threshold, the subject was given two bottles at each test, one of which contained distilled water, and the other a weak solution. When a difference from the distilled water of the cup was reported, the subject was asked in which bottle he noticed it. If it was an illusion, it was quite as likely to be referred to the distilled water bottle as to the solution. Often the subject said that both bottles [p. 53] were different from the distilled water. If they seemed equally different, it was again counted as an illusion. If the bottle containing the solution was reported more different from the standard than the bottle containing distilled water, the discrimination was regarded as genuine, but a threshold was determined only after three correct judgments out of four.

The order of procedure was as follows: The subject was provided with a cup of distilled water and a teaspoon. A jar was placed beside him, and he was told not to swallow the solution unless he wished. Two bottles just alike in appearance were placed before him, and he was directed to taste the distilled water in the cup first, and then taste half a teaspoonful of the liquid each bottle. He was told that he must make the solution touch all parts of the tongue in tasting, since not all parts were equally sensitive. After tasting each solution once, he was required to tell which one, if either, differed from the distilled water of the cup. Both bottles were then removed, and two more given him. [p. 54] The tests always began with the weakest solutions, and worked up to the place where the discrimination from distilled water could be made. This procedure is particularly necessary in taste and smell, because the nerves become fatigued so rapidly that it would be impossible for most subjects to detect the weaker solutions when the stronger ones had just been perceived. After reaching the point at which the subject was sure he could detect something in the solution (T1), the same process was continued until he was able to identify the taste (T2).

The curves for the threshold of presence (Figs. 24-27) show a lower threshold for the women in all four tastes. The difference is most marked in bitter, second in sour, third in salt, and least in sweet.

As regards the threshold of recognition (Figs. 28-31) the women are unquestionably more sensitive to sour and bitter. In salt the women's curve is slightly better. It has more entries in the region of very low thresholds, and no cases which fall beyond the limits of the series. The curve for sweet averages about the same for both sexes. Both the best records and the worst are those of women.[p. 55]

3 and 4. Discriminative sensibility. -- The discriminative sensibility for taste was tested with two standards. The first was the solution marking each subject's threshold of recognition, the second was identical for all subjects, viz., the first solution in series B of Table VIII. The subject sat as for the previous test. Two bottles were set before him, and he was required to judge which of the two solutions was the stronger. The mouth was rinsed with distilled water after each discrimination.

The discriminations in the second series of tests, since an arbitrary standard was adopted, offer material which is comparable. The standard gave a strong taste to all subjects, except those abnormally obtuse. The method of making the discrimination was the same as that described for the previous series.

The results of the tests on the discrimination of strong tastes (Figs. 32-35) show that the men have a finer discrimination in all tastes but salt, in which the women discriminated somewhat better. The general result agrees very well with that for thresholds.

The tests for smell were analogous to those for taste as to apparatus and method. They were made with two series of solutions, one designed to determine the two thresh-[p. 58] olds of presence and of recognition, and the other to test the fineness of discrimination in the strong odors. Two substances were used: violet water (Roget and Gallet violette de Parme), and oil of cloves. The violet was simply diluted the required amount with distilled water. As a basal mixture for the cloves, an emulsion was made by shaking 1 part of oil of cloves in 99 parts of distilled water. This mixture was then diluted to form the series, being shaken thoroughly at each step. The odors were both so persistent that great care was necessary in preparing the weaker solutions. The utensils which had been used for stronger solutions had to be thoroughly cleansed with alcohol and distilled water before being used to make the weaker ones. The solutions were placed in glass-stoppered bottles all alike, being prevented from touching the necks of the bottles when put in. The distance between the surface of the liquid and the mouth of the bottle was made constant for all the series. The series of solutions are given in Table IX.


In the tests on smell (unlike those on taste and all others in the present set of experiments where judgment between two stimuli formed the modus operandi) the subject was allowed to have as many stimulations as he wished from each of the two bottles given him, in the determination both of the thresholds, and of the discriminative sensibility. The reason for this departure in the case of smell is that it is the only sense in which the contact between external stimulus and nerve-ending is produced so indirectly. The actual stimulation of the nerve-ending depends upon [p. 59]

1 and 2. Thresholds of presence and of recognition. -- The determination of the smell thresholds was made by a method like that used for the taste thresholds, but differing in two respects. The first modification was that common to all the smell tests stated above; the second was that the subject was not provided with a bottle of distilled water which he knew to be such, corresponding to his standard of reference in the experiments on taste. This did not seem to be neces-[p. 60]sary, because distilled water showed no tendency to stimulate the nerves of smell in any definite direction, as it stimulated those of taste.
For determining the threshold of presence two bottles were given to the subject, one of which each time contained distilled water. He then reported whether or not he could distinguish any odor in either bottle. The point at which he could select the right bottle three times out of four was taken as the threshold. To avoid the fatigue effects which are so marked in the sense of smell, the series began with the weakest solutions and advanced to the stronger.

The curves for the threshold of presence (Figs. 36 and 37) show a lower threshold for the women, though the difference is slight. It is indicated chiefly by the greater number of women in the regions of extremely low thresholds for both sexes.
In the tests for determining the threshold of recognition the subjects were not required to name the substance used as stimulus, but simply to name the class to which the odor belonged.[p. 61] "Spicy" was called a recognition for cloves, and "perfume" for violet. The effects of practice would have been a disturbing factor if a more definite recognition had been required, but the general classes of spice and perfume are familiar to all.

The threshold of recognition, like that of presence, is a little better in the women than in the men (Figs. 38 and 39). The women are somewhat more numerous in the region of low thresholds, and the men in the region of high. Again the difference is slight. The objection might be made that the two odors selected, cloves and violet water, are more likely to be familiar to women than to men; but since the recognition required was merely of spice or perfume, it does not seem probable that the greater familiarity of the women with the odors could have been a factor in the result. The subject was told that he need not name the substance, but merely describe it as best he could, or name the class of substances to which it belonged.

3 and 4. Discriminative sensibility. -- Like the corre-[p. 62]sponding series for taste, the first series of tests on discrimination of odors was made at the threshold of recognition; and as in the case of taste, so in that of smell the standards are so scattered that they do not afford material for comparison. In the second series of tests under the present head, as in the second series on taste discrimination, stronger stimuli were used; and as in that case, so in this the standard was arbitrary, viz., the first solution of series B in Table IX. This second series of tests yielded results capable of comparison.

The method of making the discrimination was the same as that usually employed. Two bottles, one of which was the standard, were given to the subject, and he was asked to decide which was the stronger of the two. The only modification has already been stated, viz., that instead of being allowed but one stimulation from each stimulus, as in all other discrimination tests, he was allowed to go back and forth from one bottle to the other as often as he wished. A period of several minutes was allowed between stimulations for the recovery of the nerve.

The results (Figs. 40 and 41) show a somewhat better discrimination in cloves on the part of the women while in violet the difference is too slight to be of any significance; the curves are almost coincident. The difference is probably partly due to the fact that the [p. 63] solution of cloves was much less intense than the violet. Many of the thresholds of recognition for cloves fell within the higher series (see Fig. 38), while those for violet were all far below the series. The reason for the difference between the two series is that cloves increase faster in intensity of odor with increased strength of solution than violet. A 1 per cent solution of cloves is entirely too strong to serve as the standard for a series,, while a 1 per cent solution of violet is usable. In attempting tone down the cloves to a point where the intensity of [?] after image was not sufficient to interfere seriously with discrimination, the stand was made far less in absolute intensity than that of the violet series. The fact that, using these series, we find the women's discrimination better than the men's in cloves, and about the same in violet, accords with the lower thresholds of the women in both smells. We find, as we should expect, the class having the lower thresholds better in the discrimination of odors of medium intensity, but not in the discrimination of very strong odors.

The results of the tests on taste and smell may be summarized as follows: In taste the women have lower thresholds than the men both for presence and for recognition. The difference between the sexes is most marked in sour and bitter, much less so [p. 64] in salt, and very slight in sweet. The discriminative sensibility for strong tastes is finer in the men in all tastes except salt, in which it is slightly better in the women. The differences between the men and the women in smell are less than those of taste, but are of the same order. The women have slightly lower thresholds in smell, both for presence and for recognition. In discriminative sensibility for strong smells, the women are better in cloves, while there is no difference in violet. The difference may be accounted for by the fact that the violet series as absolutely much stronger than the clove series. If this supposition is correct, the results for smell are in accord with those for taste; the women have lower thresholds, but their discriminative sensibility in the strong series is as coarse or coarser than that of the strong series of the men.

Experiments on the comparative keenness of the sense of taste in men and women have been performed by Bailey and Nichols (6), Bailey (7), Lombroso (51, chap. iii), Roncoroni (72), Ottolenghi (63), Dehn (20), and Di Mattei (21). In no case has the method been so exact as that employed in the tests here reported. Bailey and Nichols prepared series of each of five tastes -- sweet, salt, bitter, sour, and alkaline. Each series varied in intensity from a solution below the threshold to a strong solution. All five series were mixed together and the subject was required to sort them by taste. The weakest solution recognized was taken as the measure of the keenness [p. 65] of taste. To obtain a statement for each sex, the results for each were averaged. Neither the method of making the test nor the treatment of results is above criticism. All the disturbing influences of after images, fatigue, and contrast enter into such a procedure as sorting tastes of varying quality and intensity. Any or all of these factors might conceivably vary with sex. Moreover, an average of results is not a fair expression of the ability of one class. One or two very abnormal individuals might change the average unduly. The limits within which the majority of the normal individuals of a class fall is the measurement required. Dehn experimented on the four accepted tastes. He used a single weak solution of each taste and recorded the right and wrong judgments.

Dehn, whose test is perhaps most closely comparable with the present one, finds women keener than men in all four tastes. Nichols and Bailey, in their tests on American students, find women keener than men in all tastes except salt, in which men are keener than women. Nichols obtains the same result in his experiments on Indians. Ottolenghi, experimenting with sweet, salt, and bitter, finds women somewhat keener than men, but attributes this fact to the use of tobacco by men and concludes that they are probably naturally keener than women. Lombroso, using three tastes, sweet, salt, and bitter, finds women keener in sweet and salt and less keen in bitter. Di Mattei, experimenting with children between the ages of four and twelve, finds the boys more sensitive than the girls in bitter, equal to them in salt and less sensitive in sweet. Roncoroni finds sensibility to sweet keener in women, but sensibility to bitter and salt keener in [p. 66] men. The general result of all these tests is to show that women have lower thresholds for taste than men. The question remains as to whether or not this statement holds for all tastes. Four of the eight series of tests find an exception in the case of salt, and three in the case of bitter. Since there is no agreement about the exceptions, and the most accurate methods show women to be somewhat keener in all tastes, it seems probable that this is a correct generalization.

In discussions on the keenness of taste, the distinction between the threshold and the discriminative sensibility has not always been observed. It is ordinarily assumed that a low threshold means also a fine discriminative sensibility -- an assumption which has no a priori justification, and which receives no support from the present series of tests. There are no other results on the discriminative sensibility for strong tastes to compare with the present series, but if these results are to be trusted, fine discriminative sensibility for strong tastes is to be correlated with a high threshold, rather than with a low one. When it is argued that women cannot have a finer taste than men, because all the professional wine and tea-tasters are men, this distinction is overlooked. The tasting of wine and tea depends on the ability to discriminate strong tastes. Threshold tests throw no light on this question. The tests here reported show that men have a better discriminative sensibility for strong tastes than women, although their thresholds are higher than those of women.

There are on record eight sets of experiments on smell: those by Bailey and Powell (4), by Bailey and Nichols (5), by Ottolenghi (64), by Lombroso (51,[p. 67] chap. iii), by Toulouse and Vaschide (80), by Garbini (28, 28a), and by Di Mattei (21). Lombroso does not state his method. Bailey and Powell, Bailey and Nichols, Ottolenghi, and Di Mattei used a method analogous to that of Bailey and Nichols in their experiments on taste, viz., sorting bottles. Bailey and his co-workers used five different odors and all the bottles were given to the subject at once. Ottolenghi used but one odor, and gave the bottles in groups, beginning with the weaker ones. This procedure diminished the disturbing factor of fatigue which is so important in smell. Di Mattei experimented on children of from four to twelve years. To the younger children he gave the bottles in two groups, while to the older ones he gave all the bottles at once. Both Ottolenghi and Bailey and his co-workers find that men are keener than women in smell, the latter reporting that men are about twice as keen as women. These results apply only to the threshold of smell. They are flatly contradictory to the outcome of our test, which finds what little difference there is in favor of the women.

The work that is most closely comparable to that of the present series of tests in method, is that of Toulouse and Vaschide. They used a single odor -- camphor -- began with the subliminal solutions, used distilled water as a control, and worked up to the threshold. Their subjects were hospital attendants. The outcome of the test is in accord with ours. They find a keener sense of smell in women than in men. Garbini's results, cited by Toulouse and Vaschide, agree with theirs. Di Mattei used the method of arranging intensities of a single odor with children,[p. 68] and found that girls could detect a fainter odor than boys, and could arrange the series more accurately. Observations of Garbini (28a) confirm this result.

It is difficult to explain the contradiction in these two sets of results. Those experiments from which the factors of fatigue and contrast are excluded, show a keener sense of smell in women. Whether the presence of these factors in the other set of tests is sufficient to explain the difference, it is impossible to say.

With reference to the thresholds for taste there is practical agreement among all observers that women have lower thresholds than men. The only tests made on discriminative sensibility for strong tastes indicate that men are somewhat superior to women, a result which is in accord with their higher threshold. There is a decided contradiction in the results of the experiments on smell. Three of the previous tests had indicated a lower smell threshold for men than for women. The tests performed with the greatest rigor of method, however -- those of Toulouse and Vaschide and those of the present series -- show a lower smell threshold for women. No difference in discriminative sensibility was demonstrated.

[1] The term "threshold of presence" is sometimes represented in this chapter by the symbol T1, and the term "threshold of recognition" by the symbol T2.