I Infant Salt Taste: Developmental, Methodological, and Contextual Factors GARY K . BEAUCHAMP BEVERLY J. COWART JULIE A. MENNELLA Monell Chemical Senses Center Philadelphia, Pennsylvania

ROGER R. MARSH Department of Pediatric Otolaryngology Children’s Hospital of Philadelphia Philadelphia, Pennsylvania

Two studies investigated the human infant’s response to salt during development. In the first study, measures of intake and sucking were obtained from two groups of infants, newborns and 4- to 8-month-olds, in response to brief presentations of two concentrations of salt (0.2 or 0.4 M) and water. For several measures of sucking and for intake, there were significant age and concentration effects. Generally, newborn infants tended to reject saline relative to water more than did 4- to 8-month-old infants. This result, consistent with previously published research, suggests a developmental change in salt acceptability and, probably, sensitivity in the human infant. In the second longitudinal study, the response to salted (0. ISM) versus unsalted formulas was evaluated monthly in infants 2 to 7 months of age. Again, a developmental change was observed: Based on some sucking measures, younger infants appeared to be indifferent to the salted formula relative to the unsalted formula whereas older infants tended to reject the salted formula, presumably because either it was less sweet than the unsalted formulas or because it was novel. These data are consistent with the hypothesis developed from animal model studies that during early human postnatal development, transductive elements sensitive to saltiness mature. 0 1994 John Wiley & Sons, Inc.

Sensitivity to and preference for tastants representing different qualities appear at different times during human ontogeny . Sweet taste sensitivity and preference can be demonstrated in both premature (Maone, Mattes, Bernbaum, & Beauchamp, 1990; Tatzer, Schubert, Timischl, & Simbruner, 1985) and newborn (e .g . , Rosenstein & Oster, 1988; Desor, Maller, & Turner, 1973; Steiner, 1977) infants. Newborns also appear to be able to detect (and reject) at least some bitter and sour substances (Desor, Maller,

Reprint requests should be sent to Gary K. Beauchamp, Monell Chemical Senses Center, 3500 Market Street, Philadelphia, PA 19104, U.S.A.

Received for publication 1 January 1994 Revised for publication 1 April 1994 Accepted for publication 11 April 1994

Developmental Psychobiology 27(6):353-365 (1994) 0 1994 by John Wiley & Sons, Inc. CCC 0012-16301941060353- I3

354 BEAUCHAMP ET AL.

& Andrews, 1975; Rosenstein & Oster, 1988; Steiner, 1977), although sensitivity to at least one bitter compound increases during the first few weeks of life (Kajiura, Cowart, & Beauchamp, 1992).

Salt taste perception appears to follow a different developmental course. The new- born infant has been found to be either indifferent to (Desor et al., 1973, 1975; Rosenstein & Oster, 1988) or to reject (Crook, 1978) salt solutions. This inconsistency in the newborn’s response to salt may reflect the use of different measures: Rejection has been demonstrated by a depression in sucking burst length but not by reduction of relative intake to the diluent, which was either water or a weak sucrose solution. Furthermore, even when a high concentration of salt (albeit a very small amount that might be quickly diluted in saliva) was introduced into the oral cavity of newborn infants, no distinctive facial expression was observed (Rosenstein & Oster, 1988).

Although there is no evidence that newborns prefer salt to water as they do sugars, by 4 months of age infants clearly detect 0.1-0.2 M saline solutions and preferentially ingest them relative to water (Beauchamp, Cowart, & Moran, 1986; Vazquez, Pearson, & Beauchamp, 1982). These observations are consistent with a developmental shift in salt preference and sensitivity, as has been reported in two other mammalian species (see Hill & Mistretta, 1990).

Taken together, these observations suggest a developmental shift in responsiveness to NaCl such that the newborn infant is indifferent to or rejects salt and may be relatively insensitive to the saltiness of NaC1. By 4-8 months of age, this response has apparently shifted to one of a preference and, perhaps, increased sensitivity. This hypothesis finds support in animal model studies (see Hill & Mistretta, 1990) that strongly indicate a pre- and postnatal maturation of sensitivity to and preference for NaCl in both sheep and rats. The animal studies suggest that amiloride-sensitive elements of the Na-transduction process mature postnatally in the rat (Hill & Bour, 1985); a similar phenomenon could be occurring in humans as well.

In an attempt to further understand the course of development of human salt sensitivity and preference, and to resolve some of the apparent inconsistencies evi- denced in the aforementioned studies, a study was designed to measure both intake and various sucking parameters when newborn and 4- to 8-month-old infants are pre- sented with salted water and plain water. The results of this study support the hypothesis that there is a developmental shift in response to salt in water, and provide direct evidence that the inconsistency among studies with newborns is due to methodological differences. A second study was then designed to evaluate the infant’s perception of salt in the context of a familiar food, infant formula. The results of this study add further support to the hypothesis that there is a developmental change in human sensory responsiveness to salt during the first few months of life.

Experiment 1

Methods

Subjects

Two groups of infants were studied. The first (newborns) group consisted of 36 infants (18 males, 18 females), I to 4 days of age (median = 3 days). Eighteen were African American, 17 Caucasian, and 1 Asian. These infants were in the newborn nursery of Pennsylvania Hospital in Philadelphia, and all had been judged healthy by the pediatric staff of that hospital. A second group of 4- to 8-month-old infants

INFANT SALT TASTE 355

(median = 189 days), most of whom had also been born in Pennsylvania Hospital, were tested in the Well Baby Clinic of the same hospital. This group of 36 infants consisted of 15 males and 21 females, all but 3 of whom were African-American. A parent gave informed consent prior to testing and was paid a modest amount for the infant’s participation.

Apparatus The device described by Maone et al. (1990) was used to record pressure during

the infants’ sucking. Latex nipples (Enfamil, single-hole nipple, Evansville, IN) were modified to restrict the quantity of solution that could be acquired during each suck and to measure the sucking pressures produced by the infant. A transducer that responded to pressure changes produced by the infant’s sucking responses was attached to each nipple. As the infant sucked, the pressure changes were recorded on a FM cassette tape deck. The signals from the tape were then digitized and analyzed by specifically developed computer software, and a time-pressure curve describing the sucking re- sponses was generated. A deflection from baseline was considered a suck if it exceeded 25% of the amplitude of the third largest negative-pressure deflection that occurred during the two trials for each subject.

The amount of solutions consumed was determined by weighing the bottle before and after testing. Three measures were recorded for each suck: suck amplitude in mmHg, latency to the onset of a suck in seconds, and latency to the termination of a suck in seconds (see Maone et al., 1990, for more details). From these quantities, three parameters were calculated to reflect the frequency and strength of sucking responses as well as the overall patterning of sucking: total number of sucks, mean area under the sucking curve (representing both the strength and the duration of the suck) and number of sucking bursts (a burst is defined as a series of sucks not interrupted by more than 1 s). Although all measures of sucking may, theoretically, be interrelated, the three chosen for analysis were thought most likely to represent relatively indepen- dent aspects of infant sucking.

Solutions Two saline (0.2 M and 0.4 M) solutions were prepared from deionized, sterile water

and reagent grade NaC1. The 0.2 M concentration was chosen because it has been reported that this concentration suppresses sucking burst length in newborns (Crook, 1977) but not intake in either newborns (Desor et al., 1973) or infants 2-4 months of age (Beauchamp et al., 1986). On the other hand, the same concentration of saline is preferred to water by 4- or 6-month-old infants as measured by relative intake (Beau- champ et al., 1986). The higher concentration was chosen to test whether the absence of preference in newborns could reflect relative insensitivity to NaCl (Le., an elevated threshold for the recognition of saltiness).

Procedures Each group (newborns and 4- to 8-month-olds) was randomly divided into two

subgroups of 18 each. One subgroup was tested with water and 0.2 M NaCl, whereas the other was tested with water and 0.4 M NaC1.

356 BEAUCHAMP ET AL.

Table I Statisticul Evalrration of Resrrlts of Two Experiments

Experiment 1 Experiment 2

Age Group (G)“ Concentration (C) G x C Age

Response Measures F P F P F P F P

Volume consumed 5.55 0.02 10.48 0.002 0.04 n.s . 0.64 n.s. Number of sucks 6.87 0.01 8.62 0.005 0.30 n.s. 3.70 0.041 Suck area 2.93 11,s. 0.60 n.s. 0.00 n.s. 0.80 n.s. Number of bursts 4.66 0.04 5.83 0.02 0.00 n.s. 7.55 0.003

* Not all 36 subjects contributed data for all measures. The numbers of subjects contributing data are as follows: Intake: newborn (nb), 0.2 M NaCl (0.2) = 18; nb, 0.4 = 18; older (ol), 0.2 = 18, 01, 0.4 = 18. Number of sucks: nb, 0.2 = 17; nb, 0.4 = 13; 01, 0.2 = 17; 01, 0.4 = 16. Suck area = same as number of sucks. Number of bursts: nb, 0.2 = 15; nb, 0.4 = 7; 01, 0.2 = 16; 01, 0.4 = 14.

Tests consisted of successive presentations of 20 ml of water and 20 ml of one of the two saline concentrations. Each solution was presented for 30 s; the order (diluent first or saline first) was counterbalanced. Each infant received only two taste solutions, one saline and one diluent. The experimenter was blind to the solution being presented.

When testing newborns, the experimenter lifted the infant’s back and inclined it slightly while the baby was in the crib. In tests of 4- to 8-month-old infants, the experi- menter held the infant in her arms in the mother’s presence. In both tests, the experi- menter put the nipple into the infant’s mouth, rubbing it against his or her upper palate for a few seconds to stimulate sucking. This was done only once at the beginning of each trial.

Statistical Analyses Four response measures were evaluated statistically: volume consumed, number

of sucks, suck area and number of sucking bursts. To compare the two age groups, a proportional score for each parameter was calculated by dividing each infant’s responses to the saline solution by the sum of its responses to both the saline and the water diluent. In this way, absolute differences in response levels that may be due to maturation of sucking ability were eliminated. Arcsin transformations of proportional values were performed prior to parametric analysis, as recommended by Winer (1971). Separate two-way analyses of variance (ANOVA) were conducted on each measure [ 2 X 2 (Age x Concentration)]. A significant age effect would reflect a change in the relative response to salt with age, whereas a significant concentration effect would reflect a differential response to the saline concentrations.

In addition to these analyses, t tests were conducted to determine whether the proportional response to saline versus water was significantly different from 0.50 in each group (i.e., whether there was significant rejection or preferential acceptance of saline).

For some of the variables analyzed, not all subjects contributed data because some infants did not suck on all trials (see footnote to Table 1). For all but one analysis, at least 13 of the 18 infants’ data could be used. The exception to this is for the number of bursts where only 7 of the 18 newborn infants tested with 0.4 M NaCl provided analyzable data. This was likely due to the strong inhibition of sucking induced by this concentration of salt.

INFANT SALT TASTE 357

Results

Table 1 presents the results of the statistical analyses comparing the infant groups on the four parameters evaluated. Preliminary analyses indicated no effects of test stimuli order, thus, this was not considered further. Significant age and concentration differences were observed for volume consumed, number of sucks, and number of bursts, but not for suck area. There were no significant interactions.

The data for number of sucks and volume consumed are illustrated in Figure 1. In terms of the relative number of sucks elicited (Figure la), newborn infants appeared to reject 0.2 M saline, t(16) = -2.45, p < .05, whereas older infants tended to prefer it, as we have observed in previous studies (although this trend was not statistically significant, t (16) = 1.73, p = .lo). Both newborns and older infants appeared to reject 0.4 M saline, t(12) = -2.62, p < .05; t ( l 5 ) = - 2 . 6 3 , ~ < 0.05. The patterns of response were similar when volume consumed was considered (Figure lb), although rejection of 0.2 M NaCl by newborns was not significant, t(17) = - 1.53; p > 0.1) (in agreement with previous reports, Beauchamp et al., 1986; Desor et al., 1973). For relative number of bursts (not shown), newborns exhibited relatively fewer bursts in response to saline than did older infants, and there were relatively fewer bursts with 0.4 M compared with 0.2 M saline in both groups.

Discussion The human newborn infant can clearly detect the presence of saline because mea-

sures of sucking are suppressed relative to water as shown here and previously reported by Crook (1978). Why this sucking suppression does not manifest as significant suppres- sion of intake is unclear, but this discrepancy could reflect the lesser sensitivity of intake measures (because they are susceptible to inadvertent loss of solution at the beginning and end of the test as well as loss due to expectoration rather than actual ingestion), and the subtlety of the newborn’s response to this stimulus. It should be noted that, as in the present study, relative intake of 0.2 M NaCl by newborns was somewhat suppressed in the study of Desor et al. (1973), although this was apparently not statistically significant.

Infants 4-8 months of age tended to prefer the 0.2 M saline solution and, on average, their rejection of 0.4 M saline solution is less pronounced than in newborns. One explanation for this change in response to salt solution is that it reflects a maturation of the peripheral apparatus that specifically detects saltiness. If that is the case, then developmental changes in the ability of salt to modify other tastes in foods might be predicted to follow a similar timecourse. Two lines of argument would lead to the prediction that salt added to a familiar food would have little effect on young (2- to 3- month-old) infants’ responses, whereas it would alter the palatability of food for older (6- to 7-month-old) infants.

First, it is known that salt suppresses other taste qualities including sweetness (e.g., Bartoshuk, 1980). The infant’s sensitivity to this suppression might be expected to change with changes in its responsiveness to salt. Based on this reasoning, it would be predicted that adding salt to a sweet food such as formula would render it less palatable to older infants, for whom the saltiness suppresses sweetness, but not to younger infants, who may not clearly perceive saltiness and for whom sweetness is, therefore, not substantially suppressed.

358 BEAUCHAMP ET AL.

0.6

1 I I

0.2 0.4

CB T 0.6

.l . 0.2 0.4

NaCl Concentration [MI

Fig. 1. Relative response (number of sucks; volume consumed) of newborn and 4- to 8-month-old infants to saline (0.2 M, 0.4 M) and water. The number of subjects contributing to each data point is indicated in Table I . Starred points are different from 50% at p < 0.05.

Second, infants can learn the appropriate contexts for tastes very early in life. For example, Beauchamp and Moran (1982, 1985) found that while almost all newborn infants show a strong preference for sucrose at birth, this preference continues to be manifested strongly only in 6-month-old infants who are repeatedly exposed to sweet- ened water during early development. We suggested that for 6-month-old infants who had no experience with sweetened water, this taste was novel and hence was not more palatable than plain water. Based on these data, salt added to formula may render it unpalatable to older infants (who can taste it as salty) because of its novelty. Thus, if there is a change in salt sensitivity during infancy as Experiment 1 suggested, then both of these arguments lead to the prediction that salted formula will become less palatable as the infant ages from 2-3 months to 6-7 months, even as salted water becomes relatively more acceptable. This prediction was tested in Experiment 2.

Experiment 2

In Part A, we verified that adding salt to commercial formula suppressed its sweet- ness and increased its saltiness as judged by adult panelists. In Part B, we conducted a longitudinal study of infants’ sucking and ingestive responses to unsalted and salted formula to test the prediction made above.

INFANT SALT TASTE 359

Part A

Methods

Panelists The panelists were 13 paid volunteers (9 women, 4 men) ranging in age from 21 to

35 years (median age = 23 years). All were employees of Monell who were naive to the substances used and the purpose of the study.

Solutions Three solutions were prepared by adding 0.075, 0.150, and 0.300 M NaCl to infant

formula (Similac, Ross Laboratories; the most commonly used formula by parents in Part B); the base-formula solution, hereafter referred to as the unaltered formula, was approximately 0.01 M Na'. All solutions were made daily and presented at roam temper- ature.

Procedures Panelists rated the saltiness and sweetness (in that order) of samples of the unaltered

formula and the three formula solutions using a 13-point category scale. The top and bottom ends of the scale (Points 0 and 12) were labeled none and exfremely strong, respectively. The labels slight, moderate, and strong were evenly spaced between the two ends (Points 3 ,6 and 9, respectively). Solutions were presented in three randomized blocks of four (one each of four solutions), with 30 s separating each solution and 5 min separating each block. Ten ml of each solution were placed in polystyrene medical cups. Subjects were instructed to sip the entire contents of the cup, spit it out, and rate the intensity.The subjects rinsed their mouths thoroughly with water after each rating.

Results and Discussion Average saltiness and sweetness intensity ratings were calculated for each panelist

for each of the four formula solutions. As shown in Figure 2, and verified by repeated measures, ANOVA, the sweetness intensity rating decreased, F(3,36) = 15.25, p < 0.001, and the saltiness intensity ratings increased, F(3,36) = 122.74, p < 0.0001, as the salt concentration increased. Because we wished to use a moderate concentration of NaCl in the subsequent infant study, we specifically tested 0.15 M NaCl for its saltiness and sweet-suppressing activity. Panelists rated the 0.15 M NaCl formula solu- tion (the one used in Part B) as tasting more salty, paired t ( 1 2 dfl = 10.69, p < 0.001, and less sweet, paired t(12 df) = 6.01, p < 0.001, than the unaltered formula solution.

Part B

Methods

Subjects Twelve mothers (9 African Americans, 3 Caucasians) who were formula feeding

their infants (7 girls, 5 boys) were recruited from ads in local newspapers and from the Womens, Infants, and Children Center at North, Inc. in Philadelphia. The mothers'

360 BEAUCHAMP E T AL

-

-

- - 3 v, s 2 (D

3 - - (D

51 - 2 E

B -

- 3 r n v) - - 1

saltiness 0 sweetness -

0

10 [ ( 4

8

2

0

Fig. 2. Adult sensory evaluation of unadulterated formula and formula to which three concentrations of salt have been added. Added salt increases saltiness and decreases sweetness judgments.

ages ranged from 14 to 33 years (median = 22.5 years), and the infants’ ages at the beginning of the study ranged from 59 to 109 days (median = 69 days).

Appa ra t u s

The same apparatus was used as described in Experiment 1.

Solutions The brand of ready-to-feed formula used during the testing procedures was the

same one the infant was currently being fed; 8 of the infants were being fed Similac, 2 were fed Isomil, one was fed Prosobee, and one was fed Nursoy. The taste of the formula was modified by adding 0.87 g NaCl to 100 ml of formula, yielding a final NaCl concentration of approximately 0.15 M. The control, unsalted formula solutions used during these trials, was prepared according to manufacturers’ instructions and contained concentrations of Na’ of approximately 0.008-0.010 M. All formulas were prepared daily and presented at room temperature.

Procedures

To familiarize the infants with aspects of the testing procedure, the mothers were sent a packet of nipples (Enfamil single-hole nipples) and bottles (Volufeed disposable nurser, Columbus, OH) to use during the 3 days that preceded the testing day. Phone contact was made with the mother during these 3 days.

Subjects began participating in the study when they were 2 (n = 9) or 3 ( n = 3) months of age. Each infant was brought by its mother to the Monell Center and, following acclimatization to the room and personnel, was given a preliminary test with mint-flavored versus unaltered formula approximately 30 min to 1 hr before its next scheduled feed. This test was conducted as part of another study designed to evaluate the effects of flavoring formula with mint on sucking and intake. The data from these preliminary tests will not be discussed further- here. Approximately 30 min to 1 hr

INFANT SALT TASTE 361

V bursts .sucks

I I I I

2-3 4-5 6-7 Age of Infant (months)

Fig. 3. Relative response (number of sucks; number of bursts) of 12 infants tested at three ages (2-3. 4-5, and 6-7 months) to infant formula with added (0.15 M) salt and unadulterated formula. Starred points are different from 50% at p < 0.05.

before the next scheduled feed, the test of salted versus unsalted formula was conducted. The testing procedure consisted of four 60-s trials in which the salted formula was alternated with the control unsalted formula in an ABBA or BAAB design. The order of presentation was counterbalanced. Sucking responses and the amount of formula solution consumed during each trial were recorded as previously described for Experi- ment 1. This entire procedure was repeated once each month until each infant was 7 months of age.

Statistical Analyses The same four parameters used in Experiment 1 were evaluated: volume consumed,

number of sucks, mean area under the sucking curve, and number of bursts. Based on previous studies suggesting a developmental change in response to salt at approximately 4 months of age, the data from each infant were averaged across 2-month blocks: that is, 2 and 3 months, 4 and 5 months, and 6 and 7 months. (For those infants who were not tested at 2 months, the values for the 2- to 3-month period reflect data collected at 3 months only.) The data were converted to relative scores [salted/(salted + plain)] as described in Experiment 1. One-way repeated measures ANOVAs (with age as the variable) were conducted following arcsin transformations of the data. Students t tests were used to determine which data points were significantly different from chance (50%).

Results

The results of the analyses of variance are presented in Table 1 and the data are shown in Figure 3. The response to salted formula relative to unsalted formula changed as the infants aged, as shown by the decrease in the relative number of sucks and increase in the relative number of bursts. Two- to 3-month-old infants tended to respond indifferently to salted versus unsalted formula whereas the oldest infants (6-7 months of age) sucked less, t(l1) = 2.50, p < 0.05, and both 4- to 5-month-olds, t(l1) = 2.19, p < 0.05, and 6- to 7-month-olds, t(l1) = 3.91, p < 0.05, had more interruptions of sucking (hence, more and shorter bursts) in response to the salted formula. Relative intake did not change with age, remaining around .50 throughout this 5-month period.

362 BEAUCHAMP ET AL.

Discussion

The younger infants in Experiment 2 were indifferent or insensitive to the salt that was added to the formula, whereas for the older infants, added salt depressed sucking and disrupted the pattern of sucking (but did not result in an overall decline in intake). From this we infer that the salted formula was less palatable to the older infants. Given the findings with the adult panel (Experiment 2a) indicating that the sweetness of the formula was significantly reduced by adding salt, and extensive data indicating a pro- found preference for sweet in infants, one explanation for diminished palatability of salted formula as infants age is that they become able to perceive the reduction in sweetness produced by the addition of a salty taste.

An alternative explanation for the apparent decrease in the palatability of salted formula for older infants is its novelty. Based on our previous work on sucrose taste perception (Beauchamp & Moran, 1982), we would suggest that the infants are familiar with unsalted formula and, as they begin to be able to perceive the saltiness, they reject the salted formula because it is novel, and novel flavors are often rejected.

General Discussion Developmental changes in response to salt were evident in Experiments 1 and 2.

The changes occurred between early infancy and 6-8 months of age as previously observed by Beauchamp et al. (1986).

It may appear that the results of Experiments I and 2B are contradictory in that the first study indicated that saline was more acceptable relative to water to older infants than to younger ones, whereas in the second study salted formula became less acceptable as the infants aged. It is likely, however, that there is no contradiction and that the difference lies in the medium in which the salt was placed. Water is a relatively unfamiliar and unpreferred stimulus to infants (Desor et al., 1979, perhaps tasting somewhat bitter as a result of the infant’s mouth being adapted to slightly salty saliva. To the newborn infant, added salt appears to make water even more unacceptable. As sensitivity to salt matures, possibly as a consequence of the addition of amiloride- sensitive channels, salt may begin to elicit what adults refer to as “salty” taste, increas- ing the acceptability of saline relative to water such that nearly isotonic saline tends to be preferred to water. In contrast, when the diluent is a highly preferred and familiar formula, added salt (at an age when it tastes salty) makes the formula less acceptable by suppressing the sweetness of that formula, by creating a novel taste complex, or by both mechanisms. Since other tastes (e.g., bitter) also suppress sweetness in adults, but do not apparently exhibit the same degree of postnatal maturation as does NaCl perception, it is predicted that if the second study were repeated with bitter (e.g., urea) added to formula, the bitter formula would be relatively unpalatable at all ages tested.

Animal model studies have also demonstrated a postnatal maturation of salt sensitiv- ity. Numerous studies by Hill, Mistretta, Bradley, and colleagues over the past 15 years (see Hill & Mistretta, 1990 for a review) have demonstrated major changes in gustatory nerve responses to NaCl during early postnatal development. In sheep, there is a developmental increase in the proportion of fibers maximally sensitive to NaCl during prenatal and postnatal development. In rats, increases in neurophysiological responses to salt occur during the preweaning period. It is thought that at least one major compo- nent of this developmental change is mediated by the acquisition of amiloride-sensitive channels (Hill & Bour, 1985). These channels presumably serve as the transductive elements for the highly specific salt fibers found in adult rats (see HilI & Mistretta, 1990).

INFANT SALT TASTE 363

These neurophysiological studies suggest a postnatal maturation of salt sensitivity. Such a change could manifest as a change in preference. Behavioral studies suggest that younger rats prefer much higher concentrations of salt than do older rats (Midkiff & Bernstein, 1983; Moe, 1986). Both Moe and Midkiff and Bernstein suggest that these age-related behavioral changes may be mediated by the physiological changes in salt sensitivity.

The rejection of NaCl solutions by newborn infants probably represents a response to salt that is not mediated by NaC1-specific fibers in the chorda tympani, but instead reflects a sensitivity to NaCl that is much less specific. Most of the animal-model, neurophysiological data on development of salt taste involve recordings from the chorda tympani nerve, which innervates the anterior portion of the tongue. This is the area of the tongue in which fibers that are highly specific to salt and sensitive to amiloride are found (e.g., Boudreau, Hoang, Oravec, & Do, 1983; Brand, Teeter, & Silver, 1985). The human newborn’s response may not be mediated by these sodium-specific fibers (perhaps they are not present yet) but through less-specific taste fibers of the chorda tympani or the glossopharyngeal nerve or both. In this regard, it is interesting that early postweaning rats do not discriminate between the tastes of NaCl, NH,Cl, and KCI, substances among which intact adult rats can easily discriminate (Formaker & Hill, cited in Hill & Mistretta, 1990). Thus, NaCl may taste different to young rats than it does to adults. Perhaps the same is true for humans, and newborn infants reject NaCl not because it tastes salty as adults perceive it, but because it tastes more like KC1-salty and bitter. The development of a preference relative to water may thus reflect addition of amiloride-sensitive elements.

Alternatively, it is conceivable that the newborn’s apparent rejection of 0.2 and 0.4 M NaCl was the result of trigeminal stimulation. This seems unlikely, at least for the 0.2 M concentration, however, because both neurophysiological studies with rats (Kawamura, Okamoto, & Funakoshi, 1968) and psychophysical studies with adult hu- mans (Green & Gelhard, 1989) have indicated little or no irritation from NaCl at concen- trations below 0.4 M. However, the 0.4 M concentration could stimulate trigeminal fibers and this may at least contribute to the rejection of this concentration by newborn (and older) infants. Because very young rats do not reject 0.4 M NaCl or even higher NaCl concentrations (e.g., Moe, 19861, the parallel in development of responses to NaCl between rat and human is not complete. This could be due to methodological differences in testing and/or species differences in physiology.

It is unclear why salt taste preference and (presumably) sensitivity mature postna- tally while taste responsiveness to sweet carbohydrates is evident in term and even preterm infants. One explanation is that there is a coordination of maturation of the amiloride-sensitive channels in the tongue and kidney (Aperia, Broberger, Thodenius, & Zetterstrom, 1975), and perhaps elsewhere in the body. One could argue that there is no “need” either for very young infants to taste NaCl (and thus be able to regulate intake of sodium) or to have a fully developed kidney system (and thus be able to concentrate urine) because the only ingesta available throughout evolutionary times-human milk-has a low but adequate sodium content. Neither fully operable kidneys nor a fully responsive salt-taste apparatus would thus be required prior to weaning and heightened consumption of sodium or solutes.

There is, additionally, a potential role for salivary sodium in the maturation of salt- taste sensitivity. The saliva of newborn infants is relatively high in sodium (Ben-Aryeh, Lapid, Szargel, Benderly, & Gutman, 1984), and in some way this could depress respon- siveness. However, by a few weeks of age, salivary sodium levels approach those of

364 BEAUCHAMP ET AL.

adults so the timecourse for salivary Na changes does not mirror that for changes in salt-taste responsiveness as documented in the present research.

In summary, the data presented here provide further evidence for a developmental change in salt sensitivity that is generally consistent with animal-model studies. Future work needs to investigate whether particular exposure regimens (e.g., very low- or high- salt diets) during this potentially critical maturation period for salt-sensitivity impact on long-term human salt sensitivity and preferences as originally hypothesized by Dahl (1 972).

Notes We thank the staff of the Women, Infants, and Children Center at North, Inc. for helping to recruit

subjects. Monica Firely, Kirsten Burgess Agee, and Tawana Bannerman provided excellent technical assis- tance. This research was supported in part by NIH DC 00882 and by a gift from the Mead Johnson division of Bristol Myers Squibb.

References Aperia, A, . Broberger, O., Thodenius, K., & Zetterstrom, R. (1975). Development of renal control of salt

and fluid homeostasis during the first year of life. Acta Paediatrica Scandinauicu, 64, 393-398. Bartoshuk, L. M. (1980). Sensory analysis of the taste of NaCI. In M. R. Kare, M. J. Fregly, & R. A.

Bernard (Eds.), Biological und behavioral aspects ofsalt intake (pp. 83-98). New York: Academic Press. Beauchamp, G. K., & Moran, M. (1982). Dietary experience and sweet taste preference in human infants.

Appetite: Journal of Intuke Research, 3, 139-1 52. Beauchamp, G. K., & Moran, M. (1985). Acceptance of sweet and salty tastes in 2-year-old children. Appetite,

Beauchamp, G. K., Cowart, B. J. , & Moran, M. (1986). Developmental changes in salt acceptability in

Ben-Aryeh. H., Lapid, S. , Szargel, R., Benderly, A., &GutmanD. (1984). Compositionof whole unstimulated

Boudreau, J . C., Hoang, N . K., Oravec, J., & Do, L. T. (1983). Rat neurophysiological taste responses to

Brand, J. G. , Teeter, J . H., & Silver, W. L. (1985). Inhibition by amiloride of chorda tympani responses

Crook, C. K. (1977). Taste perception in the newborn infant. Infant Behavior and Development, I , 52-69. Dahl, L. K. (1972). Salt and hypertension. Americnn Journal of Clinical Nutrition, 25, 231-244. Desor, J . A., Maller. O., & Turner, R. E. (1973). Taste in acceptance of sugars by human infants. Journal

qf Comparative and Physiological Psychology, 84, 496-501. Desor, J . A., Maller, O., & Andrews, K. (1975). Ingestive responses of human newborns to salty, sour, and

bitter stimuli. Journal of Compuratiue and Physiological Psychology, 89, 966-970. Green, B. G., & Gelhard, B. (1989). Salt as an oral irritant. Chemicul Senses, 14, 173-178. Hill, D. L., & Bour, T. C. (1985). Addition of functional amiloride-sensitive components to the receptor

membrane: A possible mechanism for altered taste responses during development. Deuelopmentul Bruin RiJseirrch, 20. 2310-2313.

Hill, D. L. , & Mistretta, C. M. (1990). Developmental neurobiology of salt taste sensation. Trends in Netrrosciences 13, 188- 195.

Kajiura, H., Cowart, B. L., & Beauchamp, G. K. (1992). Early developmental change in bitter taste responses in human infants. Deuelopmental Psychobiology, 25, 375-386.

Kawamura, Y.. Okamoto, J., & Funakoshi, M. (1968). A role of oral afferents in aversion to taste solutions. Physiulogy & Behuvior, 3, 537-542.

Maone, T. R., Mattes, R. D., Bernbaum, J. C., & Beauchamp, G. K. (1990). A new method for delivering a taste without fluids to preterm and term infants. Developmental Psychobiology, 23, 179-191.

Midkiff, E. E., & Bernstein, I. L . (1983). The influence of age and experience on salt preference of the rat. Deuelopmentd Psychobiology, 16, 385-394.

5 , 291-305.

human infants. Developmental Psychobiology, 19, 17-25.

saliva of human infants. Archiues of Oral Biology, 29(5), 357-362.

salt solutions. Chemicul Senses, 8, 13 1-150.

evoked by monovalent salts. Brain Research, 334, 207-214.

INFANT SALT TASTE 365

Moe, K. E. (1986). The ontogeny of salt preference in rats. Developmental Psychohiology, 19, 185-196. Rosenstein, D., & Oster, H. (1988). Differential facial responses to four basic tastes in newborns. Child

Development, 59, 1555-1568. Steiner, J . E. (1977). Facial expressions of the neonate infant indicating the hedonics of food-related chemical

stimuli. In J. M. Weiffenbach (Ed.), Taste and deuelopment: The genesis of sweet preference (pp. 173-188). Washington, DC: U.S. Government Printing Office.

Tatzer, E. , Schubert, M. T., Timischl, W . , & Simbruner, G. (1985). Discrimination of taste and preference for sweet in premature babies. Early Human Development, 12, 23-30.

Vazquez, M., Pearson, P. B . , & Beauchamp, G. K. (1982). Flavor preferences in malnourished Mexican infants. Physiology & Behauior, 28, 513-519.

Winer, B . J. (1971). Sratistical principals in experimental design. New York: McGraw-Hill.