#AUTHOR: Stevens, J. C.; Foulke, E.; Patterson, M. Q.

#TITLE: Tactile Acuity, Aging, and Braille Reading in Long-Term Blindness

#CATEGORY: Tactile Perception; Aging; Braille

#PUBLICATION: Journal of Experimental Psychology: Applied. 1996 Vol. 2, No. 2, 91-106

#ABSTRACT: Tactile acuity came under investigation in 69 blind and 69 age-matched, sighted adults. Measures comprised thresholds for discriminating gaps, length (lines), and orientation (along vs. across the finger). Acuity of blind and sighted participants' index fingertip declined as a function of age at the same rate: roughly 1% threshold rise per annum. Slower braille reading paralleled this decline, as assessed with a tactile adaptation of the Tinker-Carver test. From youths to elders, however, blind participants outperformed sighted participants with the fingertip but not the lip, a site of commensurate acuity. Experimental trials of enlarged braille and development of a tactile Snellen-type chart for screening of acuity-related difficulty suggest themselves as promising applications.

Tactile Acuity, Aging, and Braille Reading

in Long-Term Blindness

Joseph C. Stevens

The John B. Pierce Laboratory and Yale University

and

Emerson Foulke, University of Louisville

and

Matthew Q. Patterson

The John B. Pierce Laboratory and Yale University

This study explores tactile acuity in blind adults as a function of age (compared with sighted controls), seeking implications for the aging reader of braille. Acuity means the capacity to discriminate certain simple, basic spatial features of stimulation. Its most illustrious, if somewhat flawed measure is the two-point threshold, but there are other measures, such as thresholds for discriminating differences in length, orientation, and locus of tactile stimuli on the skin. Several versions of these, including an improved measure for two- point threshold, have recently come under study in sighted persons as a function of age (Stevens & Patterson, 1995); the present study extends some of these measurements to blind persons. The three issues below motivated the study. They are:

1. The punctate nature of the braille alphabet (Foulke, 1991). Some seven published studies show that two-point threshold deteriorates with age, eventually to a size larger than the spacing of the punctate elements of the braille letters. To the extent that reading braille depends on or is facilitated by ability to resolve the dots, many individuals, particularly the elderly, should theoretically have difficulty reading braille and also other tactile reading displays like the Optacon (Craig & Sherrick, 1982). (For a review of two-point resolution and aging, see Stevens, 1992, and Stevens & Patterson, 1995.)

A picture may look worse than it really is. Measurement of two-point threshold, venerable though its history, inclines to methodological short-comings. These include (a) frequent absence of forced-choice procedures, thereby giving free rein to individual criteria for deciding what feels like two, (b) propensity of two-point threshold to confuse true spatial discontinuity with other, spurious bases for discriminating two points from one (Johnson & Phillips, 1981), and (c) the tenuous nature of individual thresholds computed from relatively few judgments, thereby yielding an exaggerated impression of individual differences in acuity.

Nevertheless, even when pains are taken to control criterion and to eliminate spurious two- point thresholds (by a procedure described below), it is clear that the average person over age 40 faces difficulty resolving gaps as small as the separation between the elements of braille (2.3 mm) and the Optacon (2 mm). However, two-point thresholds do not provide a full answer, as a brief explanation of the braille alphabet clarifies.

2. The structure of the braille alphabet and the complex nature of tactile acuity. Other kinds of acuity play roles in reading braille besides two- point discrimination. These include length discrimination (e.g., the difference between braille a, a single dot element, and braille c, two dots side by side) and orientation discrimination (e.g., the difference between braille c and braille b, two dots one above the other). To discriminate these letters does not necessarily entail resolving the two elements of the b and c although Millar (1985) reported that blind persons could more easily discriminate their punctate form than their connected form. Length and orientation (transverse versus longitudinal placement on the fingertip) are among the various kinds of acuity related to age (see Stevens & Patterson, 1995). Compared with two-point discrimination (and of spatial gaps in general), orientation and length have been slighted as acuities, but they are hardly without precedent. Length (and area) were first addressed by Vierck and Jones (1969) and Jones and Vierck (1973); orientation was apparently first addressed (for the forearm) by Gould, Vierck, and Luck, 1979; both length and orientation of Optacon displays were addressed by Schneider, Hughes, Epstein, and Bach-y-Rita (1986) with results commensurate with our own.

Although these other kinds of acuity are not reducible to two-point or gap discrimination, they were shown to age at virtually the same rate (approximately 1% per annum); together they suggest possible difficulties with tactile reading beginning in the middle years. Therefore, when it comes to reading via the skin, blind persons may face the tactile analog of the ubiquitous presbyopia of the aging sighted reader, minus the remedial equivalent of eyeglasses.

3. Aging, disease, and acuity. Age-related diseases, most notably diabetes, can devastate tactile sensitivity. Diabetes could have unfortunate consequences for the aging braille or Optacon reader by gravely accelerating the loss of acuity. Thus, Heinrichs and Moorhouse (1969) and Nakada and Dellon (1989) claimed an association between two-point discrimination and ability to read braille by blind persons with diabetes. Although the data are limited (e.g., no objective tests of braille reading were offered), they suggest that a two- point threshold of 5 mm or larger may preclude ability to read braille at even a fair level. Normally, thresholds as large as 5 mm are uncommon, even in advanced age. To help evaluate such findings, we need to know more about acuity and age of blind persons without diabetes. The same applies to other common age-related disorders, notably hand-arm vibration syndrome, an endemic occupational hazard attributable to hand-held vibrating tools, such as jackhammers and chain saws (Pyykkö, 1986).

There are three basic questions of interest. (a) Do reliance on and continual use of the tactile sense by braille-reading blind people influence the course of aging and deterioration of acuity? The present study suggests that the rate of deterioration in the index fingertip is no different for blind and age-matched sighted persons. (b) Does this loss of acuity with aging affect the reading of braille? (A modest effect was demonstrable, but its exact extent is difficult to assess because other factors could also count.) In this regard, reading braille probably differs little from reading the printed page. Intelligence and practice may compensate for weak eyes or aged skin. (c) Do blind people have a superior sense of touch? Although folklore persists in this notion, the scientific evidence has been largely negative (Hollins, 1989). The present study, with an earlier one by Axelrod (1959), offers evidence that blind persons may indeed have superior spatial acuity in the braille- reading fingertip.

We concluded two experiments involving blind participants, the first in the Perceptual Alternatives Laboratory at the University of Louisville (Kentucky) in February 1993, and the second at the residence of Emerson Foulke in Louisville in June 1994. The main concern of the first was to assess how four measures of acuity (and their average) vary with age. To that end, we examined participants over the adult age span from youth to advanced years. Although the objectives of the two experiments overlapped, in the first we emphasized rate of change in acuity as a function of age. The second, taking advantage of methodological improvements accomplished in the interim, studied two groups of blind participants: one under age 35 and the other over age 60. Emphasis was given here to differences between blind and sighted participants on three measures of acuity. For both experiments, sighted controls come from a study at the John B. Pierce Laboratory in New Haven, Connecticut, and have been reported in detail by Stevens and Patterson (1995).

In both experiments, the main dependent variable was tactile acuity, measured in millimeters. In all of various acuity tasks studied, the mode of stimulus presentation was passive, as opposed to active scanning or tracing by the participant. Active exploration has been shown in sighted observers to favor tactile recognition of braille and visual letters (Heller, 1986; Loomis, 1985; Phillips, Johnson, & Browne, 1983), though the advantage appears to be secondary (see Dynamic Versus Passive Touching in the Discussion section). As mentioned, the kinds of acuity selected for study relate to features of braille, namely gap, orientation, and length discrimination. Of course, there are, other measures of tactile functioning at the disposal of the psychophysicist and applied in multitudinous con texts (for example, see Schiff & Dytell, 1972, who employed a battery of representative ones to study deaf children). Available measures include sensitivity to vibration of various frequencies (Gescheider, Bolanowski, Hall, Hoffinan, & Verillo, 1994), detection of edges (Johnson & Phillips, 1981), and raised dots (LaMotte & Whitehouse, 1986), and discrimination of texture (Heller, 1989; Lederman & Taylor, 1972). At least some of these may play direct or indirect roles in braille reading, though they are more akin to intensity thresholds than to acuity. A possible interesting exception is texture (roughness), a sensory quality that is salient primarily when the skin is moved over the textured surface or vice versa. Although the separation of braille elements is constant within cells, the density of dots varies somewhat from letter to letter, owing to empty cells, and density therefore does offer a potential cue (Lederman, 1982; Millar, 1994). Density could also play a role in perceiving boundaries between cells (6.35-mm separations). Its exact role in braille seems to be difficult to specify, may depend on degree of acquired skill, and may sometimes even be deleterious (Lederman, 1982; Millar, 1977). Moreover, the effect of aging on texture perception is unfortunately presently unknown.

Experiment 1:

Conventional Two-Point Thresholds,

Thresholds of Detection of Gaps in a Disk,

and Thresholds of Orientation in Two Planes

Method

Participants. Thirty-three blind braille readers between ages 18 and 81 took part in this study. Each was paired with each of 33 sighted controls of comparable age from a pool of 133. Matching was close: 22 pairs were the same age; 10 pairs were different by 1 year; 1 pair was different by 2 years. Mean age (blind = 48.6 and sighted = 48.7) and standard deviation (blind = 16.2 and sighted = 16.1) were nearly identical, and Pearson r relating the ages in the two samples was 0.999. In matching, more attention was given to age than gender because the study of sighted persons yielded large age effects but negligible gender effects. Altogether 24 men and 42 women took part, with approximately the same gender ratio for blind and sighted participants. All participants were living at home and able to come to the laboratory for testing. All were paid and gave informed consent.

All of the blind participants were profoundly impaired, 18 of them totally blind, and the other 15 were able to see unpatterned light only and thus were unable to read visually. The median reported use of braille was 39 years and the median age at which braille was first learned was 6. As expected, the older the participant, the longer on average was the period of braille use; even so, the minimum period was 10 years. Although the causes of blindness were diverse (e.g., accident, cataracts, glaucoma, rubella and other infections, premature birth, retinitis pigmentosa, and detached retina), none named diabetes as a cause. Diabetic retinopathy accounts for about 6.7% of causes of blindness in general and is the leading cause of blindness for persons between ages 20 and 70 (see Hollins, 1989). Diabetes can also cause tactile dysfunction owing to associated neuropathy. One participant had diabetes requiring insulin therapy at the time of testing and neuropathy, but (as so commonly in diabetes) in the foot

Apparatus. For two-point and orientation thresholds, the stimulator was a Mitutoyo Model 500-351 precision steel caliper with digital readout, to whose jaws were attached two steel gauge pins, each 0.4 mm in diameter, which served as the contacting surfaces. On each presentation, these points were touched to the skin for about a second, without movement across the skin with a force of approximately 40 to 60 g. Exact force of contact is unimportant for tactile acuity (Johnson & Phillips, 1981; Jones & Vierck, 1973; Loomis, 1985). For thresholds of detecting gaps in a disk, we used a set of 23 aluminum cylinders, each 12.7 mm in diameter, 43 mm in length, and 9.0 g in mass. One end was a plain disk, and the other the same disk with a notch 5 mm long along its radius, and a variable width at its circumference between 0.25 mm and 8.5 mm. In use, the exact placement of the gap end of the disk on the skin was random from trial to trial. On each trial, the participant had to choose which of two presentations (plain end of the cylinder versus notched end) contained a gap; the disks were placed of their own weight on the skin without lateral movement for about a second at a time.Procedure. In all testing of the finger in Experiments 1 and 2, the stimuli were applied to the index fingertip (approximately midway between the extreme tip and the center of the distal phalanx) of the preferred hand; exact locus was varied from trial to trial so as to wash out possible local irregularities in acuity. In the case of the blind participants, preferred meant which hand the participant would elect to use for braille, if only one hand were permitted; for 8, this was the left hand, although 4 claimed to be primarily right handed. All others claimed to be right-handed and preferred to read with the right hand. A few practice trials were always given with feedback until the participant reported feeling comfortable with the task. Thereafter, trials proceeded without feedback.Likewise, all tasks of Experiments 1 and 2 used a version of up-down tracking by forced choice judgments. A rule in common was that a stimulus step upward (toward an easier discrimination) always followed an incorrect choice, whereas a stimulus step downward (toward a harder discrimination) occurred only after two successive correct choices. This procedure tends to drive the track toward an asymptotic 71% correct responding (Wetherill & Levitt, 1965). A track continued until 12 transitions in direction ("turn arounds") had occurred.

Individual plots of these tracks were routinely made as a check against starting bias, which can result from beginning a track at a stimulus level too far away from the participant's threshold; such bias can plague adaptive procedures (Simpson, 1989). One strategy to avert it, as suggested by Wetherill and Levitt (1965), is sometimes useful: The step size was approximately halved after the first five transitions. The intention was by means of a relatively gross initial step to steer the track to the vicinity of the 71% threshold and thereafter to zero in on the threshold in finer steps. Threshold was defined as the average of the last seven transition steps.

It proved useful to place the choice of stimuli under computer control, with trial-by-trial display to guide the experimenter. The experimenter had only to select the appropriate stimulus, apply it as instructed, and signal whether the participant's choice was correct or incorrect. The program stored the entire track and displayed the calculated threshold immediately.

For two-point discrimination, the participant had to decide which of two configurations (order random) felt more like a dual impression: two points at a given separation versus two points at zero separation. The configuration was always presented along a diagonal midway between the transverse and longitudinal axes of the finger. Starting separation between the two points was 2.5 mm, gross step size 0.5 mm, and fine step size 0.25 mm.

By a similar method, two-point thresholds were also obtained along the long axis of the upper lip, below the hairline, and just to the side (corresponding to the preferred braille finger) of the midline. Compared with most other regions of the body (Weinstein, 1968), the lip and fingertip have superior spatial acuity. Stevens and Patterson (1995) showed a small but significant overall superiority of the lip over the fingertip.

For disk-gap discrimination, starting gap size was 3.0 mm, gross step size was 0.5 mm, and fine step size was 0.25 mm.

For two-point orientation discrimination, the stimulus length was 3.8 mm (total length of the two-point configuration including the diameters of the two points, 0.4 mm each), gross step size was 0.5 mm, and fine step size 0.25 mm. It was impossible for the stimulus length to go below a length of 0.8 mm (zero separation between the two points), thereby imposing a "basement" length below which some participants might have been able to discriminate its orientation, especially when presented in the longitudinal axis, where discrimination is somewhat better than in the transverse axis, as is explained later.

By means of a double-track procedure, two thresholds of orientation were codetermined, one for discrimination in the transverse axis and the other for discrimination in the longitudinal axis. On each trial, the two-point stimulus was presented either in the transverse axis or in the longitudinal axis of the finger; the participant had to decide in which plane the stimulus was presented and was either right or wrong. Whether the presentation was transverse or longitudinal was randomized from trial to trial. On a given trial, if a participant said "transverse" to a longitudinal presentation, this was scored as a wrong choice on the longitudinal track; if the participant said "transverse" on a transverse presentation, this was scored as a correct choice on the transverse track, and so on, until both tracks had at least 12 transitions; threshold was reckoned as the average of the last 7 (or more) transitions. As will become apparent, the average threshold from the longitudinal tracks was significantly smaller than that from the transverse tracks.

Each participant yielded five thresholds in a session: two for two-point discrimination (finger and lip), two for orientation (transverse and longitudinal), and one for gap. These were gathered in quasirandom order for each participant. In the same session the participant was also given a questionnaire covering such questions as general health, history and cause of blindness, smoking history, and usage of braille.

Finally, in the same session, the participant was examined by the braille adaptation of the Tinker-Carver Basic Reading Rate Scale (Duckworth & Caton, 1986). This test consists of silent reading of simple sentences and words, arranged so that one word toward the end of an item is at odds with the meaning of the item; in the visual form of the test, the participant signals the occurrence of such a word by drawing a line through it. However, in the present adaptation, the participant read the word aloud to the examiner (Emerson Foulke) who operated a computer program in Basic that counted the number of correct identifications of the target words, the total number of words read, and the reading rate in words per minute. Depending on the reader's skill, the sampling time varied between 8 mm and 38 mm, but was typically about 18 mm. Because error rate (missing or misreading a target word) was low (less than 4%), the score given to all participants was simply words read per minute.

An entire session, including instructions and brief rests, lasted on the order of 2 hr. Because of scheduling and transportation problems with 4 participants, we were able to obtain reading scores on only 29.

Results

Table 1

#TABLE:1

Regression of acuity on age. For each of the four acuity tasks (for fingertip) and their means, Table 1 lists the slope of the regression equation relating threshold to participant's age. An impressive feature is the near constancy of the slope-about 0.01 for all tasks and for blind and sighted controls alike. This finding is consistent with the conclusions reached by Stevens and Patterson (1995) on the basis of a much larger sample of sighted participants. It suggests that blind and sighted persons change with age at the same rate: to a first approximation at 1% increase in threshold per annum. That there is plenty of room for individual departures from this generalization can be seen in the relatively low associated Pearson coefficients of correlation. Nevertheless, these are all positive in sign, and three of them, including the two rs for mean acuity, are individually significant. The similarity of the slopes for the various acuity tasks suggests that these different types of acuity, though not reducible one to the other, may nevertheless share a common underlying process, such as a network of receptors that grows sparser with age.

Figure 1

#FIGURE:1

Figure 1 shows the mean of the four acuities for blind and sighted participants as a function of the average age of the blind and sighted matched pair in Experiment 1. Figure 1 demonstrates forcefully two important features of the outcome: (a) the similarity of the regression slopes for blind and sighted participants but, at the same time, (b) a clear intercept difference between blinded and sighted participants. That is to say, at all ages, the blind group shows better tactile acuity. The advantage appears to be moderate-about 0.25 mm- but consistent across age at about 15% in favor of the blind.

Blind versus sighted groups compared with t tests.

Table 2

#TABLE:2

The superiority of the blind participants is further illustrated in Table 2, which lists the outcomes of paired t tests (df = 32) for each of the acuities and their means. Although the figures are not overwhelming, they are persuasive. Similar tests on the acuity measures in Experiment 2 leave little doubt that the blind participants outscored the sighted ones in tactile acuity of the fingertip.

Specific acuities of the finger. A word is in order about the two-point threshold and why this measure, despite past usefulness, is in its conventional form flawed. According to Johnson and Phillips (1981), a person can often with a little practice discriminate between single point and two points separated by zero distance or by distances much too small to be realistic measures of spatial discontinuity. When this happens, the basis of the discrimination could be a greater overall intensity of the two-point configuration or some other spatial property of the configuration, such as its length. That length may be a basis was suggested by Jones, Vierck, and Graham (1973). Having participants discriminate between two points and a tactile length equal to the length of the two-point configuration eliminates the spuriously small thresholds and seems to furnish a reasonable measure of spatial discontinuity or "gap" (Stevens & Patterson, 1995). This adaptation of the method, instead of the conventional one used in Experiment 1, was used in Experiment 2.

Although the average blind participant scored better than the sighted on all measures of acuity, only transverse orientation was statistically significant on its own (unlike most of the measures examined in Experiment 2). This failure of Experiment 1 may stem from methodological flaws that became apparent to us later and were accordingly corrected in Experiment 2. For example, conventional two-point discrimination occasionally yielded thresholds that were unrealistically small, as explained by Johnson and Phillips (1981); longitudinal (but not transverse) orientation was plagued by the basement effect explained above. With these flaws corrected, both orientation and two-point discrimination did give significant blind-sighted differences on their own in Experiment 2.

Transverse versus longitudinal acuity. It is of interest that discrimination of orientation was better for longitudinally than for transversely oriented stimuli for both blind, paired t(32) = 1.77, p < .05, one-tailed test, and sighted participants, t(32) = 3.17,p < .002. This difference was predict able from Stevens and Patterson's (1995) comparison of the two planes. At first glance, this outcome may seem to controvert a history of demonstrated better acuity (two point thresholds) in the transverse plane (Boring, 1942; for fingertip data, see Stevens & Patterson, 1995). This transverse advantage manifests itself also in greater estimated subjective distance between two points oriented in the transverse plane (Green, 1982) and also in the longitudinally elongated shapes of fields over which sensory saltation was able to operate (Geldard & Sherrick, 1983).

A likely neural explanation for this asymmetry was Johansson and Vallbo's (1983) finding that a preponderance of likely human receptors for fine acuity have elongated receptive fields, with their major axis in the longitudinal plane. The geometry of this arrangement implies that the average effective distance among receptors is smaller in the transverse than in the longitudinal plane. In other words, the receptor mosaic may be conceptually idealized as comprising a set of rows (transverse) and columns (longitudinal), with rows separated by larger distance than columns. This means that two points of stimulation separated by a given amount are more likely to engage different columns than they are different rows, accounting for better gap discrimination in the transverse plane. Similarly, rotation of the two points (or a line) from a longitudinal alignment is more likely to engage activity in two adjacent columns (because the columns are tightly packed) than the same rotation from a transverse alignment is likely to engage activity in two adjacent rows (because they are loosely packed). Thus, the same elongated receptor fields can easily reconcile what at first glance appears to be a psychophysical paradox.

Tactile acuity of the upper lip. Two-point threshold in the upper lip yielded no significant dependence on age in either blind or sighted participants, no significant difference between sighted and blind groups, and no difference from two-point thresholds in the finger for either sighted or blind groups.

Braille reading scores. Scores on braille reading rate (n = 29) correlated negatively with the four thresholds of spatial acuity and their mean, as shown in Table 3.

Table 3

#TABLE:3

Three of five coefficients listed there were significant by themselves.

Figure 2

#FIGURE:2

Figure 2 plots the reading scores as a function of the mean finger threshold. It is of considerable interest that this correlation (r = .54, p < .002) is only slightly reduced when the modest correlations between participant's age and mean acuity threshold (r = .35) and reading rate (r = .34) are taken into account. The partial correlation equals - .48 (p < .01), indicating a direct relation between reading speed and acuity threshold regardless of the participant's age. This correlation is moderate but noteworthy, considering the many factors relevant to a person's reading skill.

Experiment 2: Discrimination of Two-Point

Gaps. Line Orientation, and Length

Method

Participants. Thirty-six blind braille readers in two age groups took part; 15 younger persons between age 19 and age 34 and 21 older persons between 60 and 82 were paired with 36 sighted participants in two groups, ages 18 to 34 and 60 to 82. The sighted participants came from a pool of 115 tested by Stevens and Patterson (1995). A close match was obtained: 16 pairs were the same age; 13 pairs were different by 1 year; and 7 were different by 2 years. Mean age (blind = 49.0 and sighted = 49.4) and standard deviation (blind = 21.7 and sighted = 21.3) were nearly identical, and the Pearson r relating the two samples was 0.999. As in Experiment 1, more attention was given to age than to gender for the same reasons. Altogether 28 men and 44 women took part, with approximately the same gender ratio for blind and sighted participants. In addition, all participants were paid, gave informed consent, lived at home, and were able to come to the testing site. (For convenience only, 2 blind participants were tested in their homes.) Of the blind participants, 15 had served about 18 months earlier in Experiment 1. Of the sighted participants, only 1 had served earlier in Experiment 1.

As in Experiment 1, all of the blind participants were profoundly impaired: 16 of them totally blind and the other 20 were able to see unpatterned light only. The median reported use of braille was 47 years, and the median age at which braille was first learned was 7.5. Minimum braille use was 9 years. The causes of blindness were diverse, but nobody cited diabetes. However, 4 blind persons reported diabetes at the time of testing, with 3 not requiring insulin therapy, and none involving neuropathies.

Apparatus. All three acuity tasks were performed with small precision-milled stimulators made of 0.5 mm thick aluminum. For two-point gap discrimination, there were 22 such stimulators. One contacting edge consisted of two points, each 0.5 mm square and the gap between them. The total length of this configuration was matched by another contacting edge having the same length but no gap. On each trial, the participant had to decide which of these two edges (random in order of presentation) felt more like two. Gap size varied from 0.25 to 8.0 mm. For line orientation and length discrimination, there was another set of 26 stimulators having contact edges of different lengths between 0.5 and 9.0 mm.

Procedure. Most of the procedural details were the same as in Experiment 1. However, the following were unique to Experiment 2. Six of the 36 blind participants preferred to read with the left finger and were tested accordingly. Of these, 2 claimed to be mainly right handed. One preferred to read with the right hand but was primarily left handed. All others read with the right hand and claimed to be mainly right handed.

For two-point gap discrimination, the starting gap width was 3.0 mm. Gross step size was 0.5 mm, and fine step size was 0.25 mm. All presentations were made along the transverse axis of the fingertip.

For line orientation, on each trial, an edge of given length was touched four times to the finger tip, three times along longitudinal (I) axis, and one time along transverse (t) axis, in the order l-l followed by l-t, or in the opposite order, l-t followed by l-l The participant had to decide which of the two pairs contained a change in orientation. Starting length was 2.0 mm, initial (gross) step size was 0.3 mm, and fine step size was 0.1 mm. Unlike the procedure of Experiment 1, this procedure yielded only one orientation threshold. It was also designed to avoid the potential basement effect that may have contaminated some measurements in Experiment 1.

For length discrimination, the participant had to decide which was longer, a 5.0-mm edge (L) or an edge longer by some amount (delta L). Starting (delta L) was 0.5 mm (i.e., the first choice was between 5 and 5.5 mm), gross step size was 0.3 mm, and fine step size was 0.1 mm. Thus the finest discrimination ever called for was 5.0 versus 5.1 mm. Because length discrimination is sometimes very fine, this arrangement may occasionally have set a basement effect. Length discrimination, although fine, is also quite variable compared with other acuity measures (Stevens & Patterson, 1995).

As in Experiment 1, the order of measurement of the different thresholds was varied randomly from one person to another. Participants also performed the Tinker-Carver test of braille reading. Those participants who had previously done so in Experiment 1 were not retested. Scores were available for 33 of the 36 blind participants tested in Experiment 2.

Results

Figure 3

#FIGURE:3

Regression lines were computed relating each participant's mean of three acuity scores (in millimeters) to age, and these are plotted in Figure 3. Regression analysis on a bimodal distribution of ages may be unconventional, but it nonetheless serves to reveal two main similarities between the outcomes of Experiments 1 and 2: (a) Blind and sighted persons appear to age at the same rate (i.e., the slopes of the two lines are nearly identical), and (b) blind persons on the average show superior finger acuity (i.e., the position of the two lines clearly differ). Thus, although blind persons of all ages consistently yielded smaller thresholds than their sighted controls, their size increased with aging by the same rate of approximately 1% per annum as did those of sighted persons studied by Stevens and Patterson (1995) and the subset of these matched to the blind participants of the present study. The difference in mean acuity between blind and sighted was 0.4 mm. This amounts to 29% better average acuity in the fingertip of the blind participants.

In ability to tell young persons from old and sighted persons from blind, two-point gap discrimination proved to be the most effective, length discrimination was least effective, and line orientation was intermediate. This is shown by t tests (see Tables 4 and 5).

Table 4

#TABLE:4

Table 5

#TABLE:5

The ability to differentiate may depend on the variability of thresholds determined by these three measures. Stevens and Patterson (1995) found that variability was least for gap thresholds, intermediate for orientation, and greatest for length.

Tactile acuity of upper lip. In contrast to the difference (mean of 0.4 mm) between the two-point gap thresholds of the blind and sighted participants' fingertip, the same paired test applied to the upper lip yielded a nonsignificant difference of only 0.16 mm, paired t(32) = 1.076, p < .3. In addition, for the sighted participants (as for the larger group of 115 participants from whom the matched sample was drawn), the lip proved to be significantly more acute than the fingertip, t(32) = 2.54,p < .02; for the blind it did not, t(32) = 1.36, p < .20. The likely answer for this difference rests in the superiority that these blind people have in their fingers. The finding is reminiscent of Axelrod's (1959) report of superior two-point thresholds in the right index finger, but not in the contralateral index finger or the ring finger of blind braille-reading children, as compared with sighted controls.

Braille reading and acuity. Although the three measures of acuity adopted for Experiment 2 seem to have been related to at least three critical features of the braille alphabet as explained above, Experiment 2 was somewhat less successful in demonstrating a direct link between acuity and braille reading rate. Possibly this was in part a result of the bimodal distribution of the participants' ages. All acuity thresholds (and their means) were indeed inversely related to braille reading rate, but none of these was individually significant. One measure that was significant, however, was the two-point gap threshold of the 18 older participants for whom reading scores were available (r = - .48, p < .04). This is depicted in Figure 4.

Figure 4

#FIGURE:4

In these older participants, braille reading was also about equally correlated with participant's age: r = - .53, p < .03. Two-point gap threshold was, however, nonsignificantly related to age (r = .34,p < .14). Thus, the association linking reading skill to aging and deterioration of acuity is weaker than that turned up in Experiment 1 but neverthe less supportive.

Discussion of Experiments 1 and 2

Aging and acuity. Although better-than-average acuity in the blind person's fingertip seemed to mark all ages, acuity still declines at approximately the same rate in aging blind and sighted persons. This is the chief finding of the present study. Experiments 1 and 2 also add evidence to Stevens and Patterson's (1995) thesis that various types of acuity deteriorate at approximately the same rate as a function of age. They theorized that although these types are not reducible one to the other, they may nevertheless be subserved by a common receptor network. A thinning of this network is one possible mechanism that would explain the similar loss of all types of acuity. The same rate constant (approximately 1% per annum) appears to describe both sighted and blind persons. The same mechanism, such as a thinning receptor network, is doubtless at work in the aging of blind and sighted persons alike.

The sameness of aging rates in blind and sighted persons and on several acuity tasks, important as it is, does not extend to different parts of the body. Thus, the fingertip declines faster than the forearm and the lip (Stevens & Patterson, 1995). Studies of olfaction, taste, and touch (Stevens, Cruz, Hoffman, & Patterson, 1995; Stevens & Patterson, 1995) demonstrate that exact rate of physiological change with age also differs significantly among individuals for as yet unknown reasons.

Despite real differences in individual rate, it is uncommon for a person to escape loss altogether. The scatter in Figures 1 and 3 might suggest otherwise (i.e., that there is much overlap among the acuity thresholds of younger and older per- sons). However, recent studies (Rabin & Cain, 1986; Stevens & Dadarwala, 1993; Stevens et al., 1995) have demonstrated that brief adaptive procedures for measuring thresholds, like the tracking method used here, give a seriously exaggerated picture of individual differences. When three to six such threshold tests are averaged to provide a more stable index of individual sensitivity, a picture results of greatly reduced individual differences and minimal overlap of youthful with elderly participants. This includes averaged two-point gap thresholds (Stevens et al., 1995). Thus, the coefficients of correlation in Tables 1 and 3 and the scatter of points in Figures 1-4 fail to tell the whole story about individuals.

Tactile superiority of the blind person: Is it real?

The present study reveals an advantage in favor of the blind person's preferred index finger. Interestingly, no such advantage showed itself in acuity of the lip. Hence superior acuity could be limited to the braille-reading finger. That is essentially the conclusion reached by Axelrod (1959) on the basis of traditional (non-forced-choice) two-point thresholds, in a comparison of 82 early-blind youngsters with 82 matched sighted ones between ages 9 and 19 years (M = 13). Tests were performed on the index fingers of both hands and the ring finger of the preferred hand. In brief, the blind youngsters were significantly superior (by about 7%) only on the right index finger. This argues that the putative advantage may be confined to the locus of heavy usage associated with braille reading. Superiority appeared to be specific in another important way: for simple detection thresholds of touch (aroused by calibrated von Frey-like filaments), the blind and sighted youngsters did not differ in any finger. Thus, a blanket claim for a superior sense of touch as compensation for blindness receives little sup port from Axelrod's findings.

The superiority of our blind participants in the braille finger was hardly trifling. In two-point gap discrimination and discrimination of orientation, it amounted to 20% or more (see Tables 2 and 5). Averaged across various types of acuity it amounted to 15% in Experiment 1 and 29% in Experiment 2. The superior performance persists into advanced age (see Figures 1 and 3) and may to a degree (only) compensate for the dulling impact of aging. Moreover, neither experiment gave a significant difference on tests of the upper lip, though admittedly testing at this site was more limited than at the fingertip.

How is one to understand this highly specific superiority? In a more general evaluation of the putative heightened sensitivities of touch and hearing in blind persons, Hollins (1989) drew a helpful distinction between basic sensitivity (including acuity) and sharpened attention to sensory cues useful to the visually deprived person. The literature he reviewed offers scant support of heightened sensitivity of the other senses. However, it does offer examples of impressive feats of auditory perception, such as superlative voice identification and avoidance of obstacles, the likes of which probably led in the first place to the widespread but erroneous presumption of greater sensitivity. Similarly, blind observers outperformed sighted ones at identification of 80 everyday odors in the face of slightly inferior olfactory sensitivity (Murphy & Cain, 1986).

The line between a measured sensory threshold and a skillful use of sensory information may not be totally sharp. It is conceivable that our blind participants were simply more adept than the sighted ones at utilizing sensory information potentially available to all of them. On this view, the blind person is no better equipped on a sensory- neural basis than the sighted person. However, on this view also, the blind person's greater skill would have to be restricted to the finger used to practice it.

This explanation of our results and Axelrod's (1959) is a reasonable possibility but definitely not the only one. Recent studies of somatosensory plasticity in monkeys raise the entirely plausible possibility that extensive year-after-year use of the braille finger may bring about enlarged cortical representation of the fingertip used to read braille. Such alteration did take place in cortical maps (contralateral somatosensory area 3b) made before and after several months of heavy behavioral engagement of the distal phalanx of one or two digits of adult owl monkeys (Jenkins, Merzenich, Ochs, Allard, & Guíc-Robles, 1990). Compared with adjacent, nonstimulated skin, the stimulated skin areas received greatly magnified cortical representation. Moreover, the individual receptive fields within these magnified areas actually shrank in size. After the cessation of the heavy tactile engagement, the cortical maps eventually returned toward normal size. These findings are provocative because it is known that where cortical representation of the skin is great, tactile acuity is fine (Weinstein, 1968).

Behavioral modification is only one technique used to demonstrate plasticity. Others include cortical restructuring following amputation, fusion of digits, and cortical lesions (Jenkins, Merzenich, & Recanzone, 1990). An underlying condition seems to be long-term time-correlated neural activity from receptors in the skin.

It is of interest to estimate how much tactile engagement sufficed to bring about cortical magnification. The monkeys were trained by an operant technique to make contact for 10- to 15-s periods with a revolving sectored disk that produced tactile pulse indentations at the rate of 20 Hz. These contacts were rewarded by a small food pellet. In the course of a day, the average monkey received about 600 such pellets, amounting to roughly 150,000 tactile pulses per day (75,000 per hour of contact). A braille reader could receive commensurate active tactile stimulation in the braille-reading fingertip. Reading averages 85 words per minute in high school students (see Foulke, 1991), braille letters average about 3.3 dots, and English words about 4.3 letters; this amounts to 72,000 dots per hour of reading time. Years of such engagement of a fingertip may by analogy provide a sufficient stimulation for cortical magnification.

The source and nature of the blind participants' superior acuity thus become a matter of theoretical interest. Thorough comparison of blind and sighted readers seems in order, with greater testing of control sites, especially the nonpreferred index fingertip and the second phalanx of the preferred index finger. Plasticity predicts that cortical magnification is quite specific to the region of stimulation. Another telling control would be to compare the fingertips of age-matched readers and nonreaders of braille. Neurological studies may also prove fruitful. Indeed, by means of cortical evoked potentials and transcranial magnetic stimulation of the fingertip of braille readers and controls, Pascual-Leone and Torres (1993) offered promising suggestive evidence for greater somatosensory cortical representation of the braille-reading finger than of the contralateral finger or the fingers of sighted controls.

Implications for braille. Evaluating what role declining acuity with aging has on braille reading is difficult. That the question has practical meaning, though, is testified to by the modest but persuasive correlations that obtained between reading scores of the blind participants and their various thresholds of atuity. Considering all the other determinants of reading skill any significant correlation at all probably points to a problem, but the exact extent of it is difficult to assess. For example, although all of our blind participants were current braille readers, their usage ranged from occasional notes to daily reading of magazines and books. A person with heavy usage and impaired acuity could outscore one with rare usage and unimpaired acuity. Individual differences in cognitive skills and motor dexterity may also obscure the role of acuity. And, finally, the brevity of the acuity tests and the reading test limits their reliability.

Aside from the reading scores, the acuity results speak to the issue of readability of braille in certain implicit ways. The most important is to show that resolution of spatial discontinuity, as represented by the two-point gap threshold in Experiment 2, is marginal. For the average reader over about age 60 (see Table 4), threshold averaged 2.48 mm for the blind participant-better than the 2.81 mm for the average sighted participant, but worse than the distance that separates the dot elements (2.28 mm). Of the 20 blind elderly participants, 13 had two-point gap thresholds larger than 2.28; of the young blind participants, none did. Although these individual thresholds may have limited value the averages illustrate the problem well. In addition, the figures cited above are for the transverse axis; for the longitudinal axis, thresholds run about 13% larger (Stevens & Patterson, 1995). Experiment 1 also had participants over age 60, and 4 of 11 blind and 7 of 11 corresponding sighted participants gave disk gap thresholds larger than 2.28 mm.

The punctate structure of braille seems to be one of its advantageous features, certainly as compared with raised alphabetic characters (Loomis, 1981). Less certainty attaches to the readability of "connected braille," in which the dot elements are connected by raised lines in such a way as to preserve the characteristic shape of a letter and making them more like roman letters. For example, braille b, two longitudinally stacked dots, becomes a short longitudinal line. Millar (1985) found punctate characters to be more discriminable by blind persons than connected characters. How could this advantage persist in the face of growing inability to resolve the dots? Finally, Stevens and Patterson (1995) showed that discrimination of orientation of two points was significantly better than that of lines of the same overall length. Thus there is excellent reason to believe that the age-related breakdown of spatial resolution must hamper braille reading.

Discrimination of orientation is another type of acuity relevant to braille. Orientation (longitudinal, transverse, and diagonal) is a salient feature of braille letters. From data in Table 5, the average elderly reader should be less handicapped in this kind of discrimination. Only 1 of the 20 elderly participants in Experiment 2, a woman of 82, gave a threshold (3.41 mm) well in excess of the minimal length needed for a braille letter (approximately 2.3 mm). In contrast, 7 of the 20 corresponding thresholds for sighted participants yielded thresholds larger than 2.3. So did 3 of the 13 blind participants over age 60 in Experiment 1, as well as 6 of the 13 corresponding sighted. Although here the blind person's superiority may make a difference, it is clear that in older persons the discrimination is borderline. Just because a person's orientation discrimination exceeds the bare minimum hardly means that this cue is maximally effective in reading braille. After all, threshold is based on forced choice between two immediate alternatives, is probabilistically defined, and gives no indication of the speed with which a person is able to make a discrimination. Therefore, threshold may be a bare minimum for effective discrimination.

The data on length discrimination project less difficulty for the aging reader. The average threshold is clearly finer than necessary to discriminate particular braille letters, such as a single dot (braille a) and two dots (braille b or c). Nor did any of 72 blind and sighted participants of Experiment 2 yield a length threshold as large as 2.28 mm. Of 89 sighted participants tested on this same task by Stevens and Patterson (1995), only 3 exceeded 2.28mm.

One other kind of acuity threshold known to worsen with age, relative point localization, has potential use in braille, but was not measured here. It too is a fine acuity, and in Stevens and Patterson's (1995) study of 115 sighted persons, only 8 (mostly very old participants) yielded thresholds larger than 2.28 mm. Thus, the acuities that seem to matter most to the aging reader are spatial discontinuity and orientation, and those that matter least are localization and length. Although these four are the only ones studied over age, there could be other features of braille, such as dot density (akin to texture) that depend on age.

Dynamic versus static touching. Given that acuity was measured with static impression of the stimuli and that braille reading utilizes active movement of the finger, the above evaluations must be read as suggestive, not the final quantitative word on exactly how acuity interplays with the microstructure of braille. However, it should be noted that active versus static mode is a less potent variable than is sometimes assumed. The matter has been given much attention in sighted (but not blind) subjects (Heller, 1986; Heller, Scrofano, & Nesbitt, 1989; Loomis, 1985; Phillips, Johnson, & Browne, 1983). Phillips et al. (1983) reported that passive recognition of roman letters became equivalent to active recognition when letter height was increased by 25%. Loomis (1985) reported that changing the mode from static to active improved overall recognition of braille and roman characters from 43.3% to 53.2%. It is unclear just how movement operates to favor recognition.

To the degree that texture comes into play, as discussed above, dynamic touch may be crucial, in that texture is salient only when the skin is moved over the textured surface or vice versa (Heller, 1989; Lamb, 1983; Lederman, 1981). It is immaterial whether the movement is induced by the perceiver or imposed; the essential distinction is between dynamic and static mode, not active and passive.

Potential applications. When it comes to learning and reading braille, acuity is only one aspect of a larger picture. Indeed, Millar (1994) has argued that acuity may be one of the less important limitations on learning and using braille. Her studies show that simply enlarging braille does not facilitate the recognition of braille letters and indeed may slow it. What holds for the young person with good acuity does not tell the whole story. Poor acuity, regardless of cause, may degrade braille reading, just as presbyopia degrades visual reading. As we grow older, dulled acuity may become a more and more potent factor. From the point of view of basic tactile acuity braille is small, even for the younger reader.

If and when acuity does become limiting on account of age, disease or injury, or native inferiority, the question arises what can be done about it. Enlarged braille, like large-letter visual text, is an obvious solution to be tried, especially on persons having grossly impaired acuity. Such text, like large-letter roman type for the impaired visual reader, would doubtless take longer to scan than normal braille, but the gain in acuity could well offset the loss of scanning speed. Enlarged braille might also necessitate a period of practice because there is evidence (Millar, 1977) that the recognition of tactile patterns under size transformation is less flexible than that of visual patterns. Nevertheless, the severely impaired reader may deem it worth the trouble. Enlarged braille might also prove useful to persons who experience great difficulty trying to learn braille. Late-blind persons often find it difficult or impossible to learn to read braille efficiently (Foulke, 1982); weakened tactile acuity may help to explain why, and enlarged braille may then help.

Printing braille in enlarged format, though theoretically plausible, could be initially costly to produce. Programmable electronic devices like the Optacon offer greater flexibility in this regard, especially in the experimental stages. In any case, age and spatial acuity ought to be taken into account in the development of alternative tactile prostheses for sensorially disadvantaged persons.

What is perhaps needed first and foremost is a simple diagnostic test of tactile acuity akin to the Snellen chart for visual acuity. Such a test might prove as simple as impressing sets of braille letters in various degrees of enlargement, to permit determination, by active scanning, the minimal size needed for discrimination. A simple, brief, and inexpensive test of this kind could feasibly be administered to large numbers of braille readers of all ages to help determine the extent to which individual difficulties encountered in learning and using braille stem from limitations on basic acuity.

References

Axelrod, S. (1959). Effects of early blindness: Performance of blind and sighted children on tactile and auditory tasks (Research Series No.7). New York: American Foundation for the Blind.

Boring, E. G. (1942). Sensation and perception in the histoiy of experimental psychology. New York: Appleton-Century-Crofts.

Craig, J. C., & Sherrick, C. E. (1982). Dynamic tactile displays. In W. Schiff & E. Foulke (Eds.), Tactual perception: A source book (pp.209-233). New York: Cambridge University Press.

Duckworth, B., & Caton, H. (1986). Basic Reading Rate Scale, braille edition (directions for administering). Louisville, KY: American Printing House for the Blind.

Foulke, E. (1982). Reading braille. In W. Schiff & F. Foulke (Eds.), Tactual perception: A sourcebook (pp. 168-208). New York: Cambridge University Press.

Foulke, E. (1991). Braille. In M. A. Heller & W. Schiff (Eds.), The psychology of touch (pp.291-233). Hillsdale, NJ: Erlbaum.

Geldard, F. A., & Sherrick, C. F. (1983). The cutaneous saltatory area and its presumed neural basis. Perception & Psychophysics, 33, 299-304.

Geseheider, G. A., Bolanowski, S. J., Hall, K. L., Hoffman, K. E., & Verrillo, R. T. (1994). The effects of aging on information processing channels in the sense of touch: I. Absolute sensitivity. Somatosensory & Motor Research, 11, 345-357.

Gould, W. R., Vierk, C. J., Jr., & Luck, M. M. (1979).Cues supporting recognition of the orientation ordirection of movement of tactile stimuli. In D. R.Kenshalo (Ed.), Sensory functions of the skin of humans (pp.63-78). New York: Plenum Press.

Green, B. O. (1982). The perception of distance and location for dual tactile pressures. Perception & Psychophysics, 31, 315-323.

Heinrichs, R. W., & Moorhouse, J. A. (1969). Touch-perception thresholds in blind diabetic subjects in relation to the reading of braille type. New England Journal of Medicine, 280, 72-75.

Heller, M. A. (1986). Active and passive tactile braille recognition. Bulletin of the Psychonomic Society 24, 201-202.

Heller, M. A. (1989). Texture perception in sighted and blind observers. Perception & Psychophysics, 45, 49-54.

Heller, M. A., Scrofano, D. K., & Nesbitt, K. D. (1989). Effect of tactual scanning mode on braille and shape recognition. Bulletin of the Psychonomic Society 27, 131-132.

Hollins, M. (1989). Understanding blindness: An integrative approach. Hillsdale, NJ: Erlbaum.

Jenkins, W. M., Merzenich, M. M., Ochs, M. T., Allard, R., & Guic-Robles, E. (1990). Functional reorganization of primary somatosensory cortex in adult owl monkeys after behaviorally controlled tactile stimulation. Journal of Neurophysiology, 63, 82-104.

Jenkins, W. M., Merzenich, M. M., & Recanzone, 0. (1990). Neocortical representational dynamics in adult primates: Implications for neuropsychology. Neuropsychologica, 28, 573-584.

Johansson, R. S., & Vailbo, A. B. (1983). Tactile sensory coding in the glabrous skin of the human hand. Trends in Neuroscience, 6, 27-32.

Johnson, K. O., & Phillips, J. R. (1981). Tactile spatial resolution. I. Two-point discrimination, gap detection, grating resolution, and letter recognition. Journal of Neurophysiology, 46, 1177-1191.

Jones, M. B., & Vierck, C. J., Jr. (1973). Length discrimination on the skin. American Journal of Psychology, 86, 49-6O.

Jones, M. B., Vierck, C. J., Jr., & Graham, R. B. (1973). Line-gap discrimination of the skin. Perceptual & Motor Skills, 36, 563-570.

Lamb, G. D. (1983). Tactile discrimination of textured surfaces: Psychophysical performance measurements in humans. Journal of Physiology (London), 338, 551-565.

LaMotte, R. H., & Whitehouse, J. (1986). Tactile detection of a dot on a smooth surface: Peripheral neural events. Journal ofNeurophysiology, 56, 1109-1128.

Lederman, S. J. (1981). The perception of surface roughness by active and passive touch. Bulletin of the Psychonomic Society, 18, 253-255.

Lederman, S. J. (1982). The perception of texture by touch. In W. Schiff & E. Foulke (Eds.), Tactual perception: A sourcebook (pp. 130-167). New York: Cambridge University Press.

Lederman, S. J., & Taylor, M. M. (1972). Fingertip force, surface geometry, and the perception of roughness by active touch. Perception & Psychophysics, 12401-408.

Loomis, J. M. (1981). On the tangibility of letters and braille. Perception & Psychophysics, 29, 37-46.

Loomis, J. M. (1985). Tactile recognition of raised characters: A parametric study. Bulletin of the Psychonomic Society 23, 18-20.

Millar, S. (1977). Tactual and name matching by blind children. Britsh Journal of Psychology, 68, 377-387.

Millar, S. (1985). The perception of complex patterns by touch. Perception, 34, 293-303.

Millar, S. (1994). Understanding and representing space:Theory and evidence from studies with blind and sighted children. Oxford, England: Clarendon Press.

Murphy, C., & Cain, W. 5. (1986). Odor identification: The blind are better. Physiology & Behavior 37, 177-189.

Nakada, M., & Dellon, A. L. (1989). Relationship between sensibility and ability to read braille in diabetics. Microsurgery 10, 138-141.

Pascual-Leone, A., & Tortes, F. (1993). Plasticity of the sensorimotor cortex representation of the reading finger in braille readers. Brain, 116, 39-52.

Phillips, J. R., Johnson, K. O., & Browne, H. M. (1983). A comparison of visual and two modes of tactual letter resolution. Perception & Psychophysics, 34, 243-249.

Pyykkö, I. (1986). Clinical aspects of the hand-arm vibration syndrome. Scandinavian Journal of Work Environmental Health, 12 439-447.

Rabin, M.D., & Cain, W. 5. (1986). Determinants of measured olfactory sensitivity. Perception & Psychophysics, 39, 281-286.

Schiff, W., & Dytell, R. S. (1972). Deaf and hearing children's performance on a tactual perception battery. Perception and Motor Skills (Monograph Supplement 2-V35) 35, 683-706.

Schneider, S. L., Hughes, B., Epstein, W., & Bach-y-Rita, P. (1986). The detection of length and orientation changes in dynamic vibrotactile patterns. Perception & Psychophysics, 40, 290-300.

Simpson, W. A. (1989). The step method: A new adaptive psychophysical procedure. Perception & Psychophysics, 45, 572-576.

Stevens, J. C. (1992). Aging and spatial acuity of touch. Journal of Gerontology (Psychological Sciences), 47, 35-40.

Stevens, J. C., Cruz, L. A., Hoffman, J. M., & Patterson, M. O. (1995). Taste sensitivity and aging: High incidence of decline revealed by repeated threshold measures. Chemical Senses, 20, 451-459.

Stevens, J. C ., & Dadarwala, A. D. (1993). Variability of olfactory threshold and its role in assessment of aging. Perception & Psychophysics, 54, 296-302.

Stevens, J. C., & Patterson, M. O. (1995). Dimensions of spatial acuity in the touch sense: Changes over the life span. Somatosensory and Motor Research, 12, 29-47.

Vierck, C. J., & Jones, M. B. (1969). Size discrimination on the skin. Science, 163, 488-489.

Weinstein, S. (1968). Intensive and extensive aspects of tactile sensitivity as a function of body part, sex and laterality. In D. R. Kenshalo (Ed.), The skin senses (pp.195-222). Springfield, IL: Charles C. Thomas.

Wetherill, G. B., & Levitt, H. (1965). Sequential estimation of points on a psychometric function. British Journal of Mathematical and Statistical Psychology 18, 1-10.

Received April 27, 1995

Revision received October 13, 1995

Accepted October 19, 1995.

The Authors

Joseph C. Stevens and Matthew O. Patterson, The John B. Pierce Laboratory, New Haven, Connecticut, and Yale University; Emerson Foulke, Department of Psychology, University of Louisville.

This research was supported by National Institutes of Health Grant AG 10295. We thank Julianne M. Hoffman and Kenneth K. Choo for technical assistance.

Correspondence concerning this article should be addressed to Joseph C. Stevens, The John B. Pierce Laboratory, 290 Congress Avenue, New Haven, Connecticut 06519.