Hormonal regulation of chemosignals of female mice that elicit ultrasonic vocalizations from males

HORMONES

AND

BEHAVIOR

20, 60-72 (1986)

Hormonal Regulation of Chemosignals of Female Mice That Elicit Ultrasonic Vocalizations from Males SUSAN BYATT

AND JOHN NYBY

Department of Psychology and Center for Health Sciences, Lehigh University, Bethlehem, Pennsylvania 18015 Two experiments examined the properties of vaginal, facial, salivary, and urinary odors from female house mice to elicit ultrasonic vocalizations from male mice. Experiment 1 demonstrated that facial and salivary secretions from hypophysectomized females were significantly less effective in eliciting ultrasonic vocalizations from male mice than were these same secretions from either intact or ovariectomized females. Thus the hormonal control of chemosignals from these two sources paralleled earlier findings of pituitary rather than ovarian regulation of the urinary chemosignal that elicits ultrasounds. In contrast, ovariectomy and hypophysectomy seemed to have similar depressive effects upon the vaginal cue that elicits ultrasounds. Experiment 2 demonstrated that longterm ovariectomy (8 or 9 months) diminished the effectiveness of female saliva, but not urine, to elicit vocalizations. The apparent dissociation of the hormonal regulation of salivary, vaginal, and urinary chemosignals suggests that multiple chemosignals may possess the property of eliciting male vocalizations. 0 1986 Academic

Press, Inc.

Odors collected from a variety of bodily locations of female mice elicit ultrasonic vocalizations from male mice. Facial, vaginal, and urinary odors (Nyby, Wysocki, Whitney, and Dizinno, 1977) have previously been shown to be effective. In addition, the present study demonstrated that female saliva is also an effective stimulus. Before the present study, the hormonal mechanisms underlying the production of chemosignals which elicit vocalizations from males had been examined only for urine (Nyby, Wysocki, Whitney, Dizinno, and Schneider, 1979). In the present study the hormonal underpinnings of the other 3 odor sources are examined. The ultrasound-eliciting property of female mouse urine appeared to depend upon pituitary factors, possibly gonadotropins. The effectiveness of urine from female mice to elicit vocalizations was not significantly diminished following ovariectomy or ovariectomy combined with adrenalectomy (Nyby et al., 1979). Thus ovarian or adrenal hormones were not necessary to maintain production of the urinary chemosignal. Pituitary factors were suspected after finding that 3 weeks of daily injections of 60 0018-506X/86 $1.50 Copyright All rights

0 1986 by Academic Press, Inc. of reproduction in any form reserved.

CHEMOSIGNALS

OF FEMALE

MICE

61

300 pg of testosterone eliminated the ultrasound-eliciting property of female urine (Nyby et al., 1979). Such an injection regimen should suppress gonadotropin release from the pituitary (Campbell and Schwartz, 1977). Hypophysectomy of females confirmed the suspicion concerning pituitary involvement. The property of female urine to elicit vocalizations was greatly reduced after removal of the pituitary (Nyby et al., 1979). A variety of hormonal mechanisms have been described for the regulation of female-produced chemosignals of rodents (see review by Nyby, 1983). For example, paralleling work on the ultrasound-eliciting chemosignal, the effectiveness of a urinary cue of female mice to cause LH surges in male mice (Johnston and Bronson, 1982) also is diminished following hypophysectomy but not ovariectomy. In addition, the female mouse urinary cue that inhibits puberty onset in other females appears to be dependent upon adrenal factors (Drickamer and McIntosh, 1980), and the maternal pheromone in rats is regulated directly by prolactin (Leon and Moltz, 1973). In contrast, many investigators have found that male rodents are more attracted to the volatile odors of intact or estrous females than to those of diestrous or ovariectomized females. Thus the attractive volatile urinary odors of females appear to be ovarian dependent. The different mechanisms of hormonal regulation may reflect the existence of a variety of chemosignals produced by females that serve different biological purposes. The apparent differences in the hormonal control of female rodent chemosignals stimulated the design of the present set of experiments. If the ultrasound-eliciting chemosensory cues from different body sites are different, then perhaps these cues might be regulated by different hormones. The finding of different hormonal mechanisms would provide the best evidence to date that more than one naturally occurring chemosignal is capable of eliciting ultrasonic vocalizations from males. If the hormonal mechanisms appear similar, this finding would be consistent with either of two hypotheses: (1) the same chemosignal is either spread about the body or excreted from multiple sites, or (2) multiple chemosignals are capable of eliciting vocalizations but all are subject to similar hormonal regulation. GENERAL METHOD

All of the experiments follow a similar plan. Methods common to the experiments are presented here. Subjects. Three different categories of mice were used: (1) subjects, (2) social experience animals, and (3) stimulus donors. Subjects were adult hybrid males bred in our laboratory from a cross between females of the C57BL/6J strain and males of the AKR/J strain. Social experience animals were adult hybrid males and females. Stimulus donors were Swiss Webster and CD-l mice purchased, respectively, from Perfection

62

BYAlT

AND NYBY

Breeders and Charles River Laboratories. The Swiss Webster mice arrived at 50 days of age and the CD-l mice at 39 days of age. Subjects were individually housed while the social experience animals and stimulus donors were housed by sex and surgical condition in groups of three or four. Apparatus. All animals were housed in wire-topped, translucent plastic cages (13 x 17 x 28 cm) with wood chip bedding. The subject’s home cage also served as the test chamber. Odors were presented on 6-in. cotton-tipped surgical swabs. Ultrasonic vocalizations were detected with a QMC SlOO bat detector tuned to 70 kHz with the microphone centered 25 cm above the floor of the test chamber. Procedure. Hypophysectomies were performed at Charles River Laboratory before arrival in our laboratory and were verified at the end of the experiments by observing the animals’ body size. The hypophysectomized females were approximately half the size of normal sham-operated females at this time. Bilateral ovariectomies were performed under pentobarbital anesthesia in our laboratory. All animals were maintained on a 12: 12 light: dark cycle with ad libitum food and water. Behavioral testing occurred during the light portion of the cycle. All subjects received 8 consecutive days of social experience beginning 10 days prior to the first test. Social experience consisted of sequential 3-min presentations of one male and one female social experience animal per day with the order of gender presentation alternated daily. For the first several days each social experience male was presented using the dangling method of Scott (1966). Social experience exposures were terminated early if aggression occurred. The social experience regimen is detailed elsewhere (Nyby and Whitney, 1980). On the day following completion of social experience the prospective subjects were screened for ultrasonic vocalizations using a normal adult female as a stimulus. Vocalizations were quantified by dividing the 3min test into 36 five-set intervals. Intervals containing vocalizations were summed yielding possible scores ranging from 0 to 36. Only males having screening scores greater than 11 were used as subjects in subsequent research. Acceptable subjects were later tested for ultrasonic vocalizations to female odors. Each 3-min odor test was scored identically to the screening test. However, in addition, each odor test was preceded by a I-min habituation period. If any vocalizations were detected during the habituation period, 2 min were required to elapse before presentation of the odor. The stimulus odor was placed on a swab during the minute immediately preceding its use. The odorized swab was then placed in a test tube and the part of the swab touched by the experimenter was discarded. The swab was emptied from the test tube into the test chamber to begin a test. Each stimulus donor provided only one stimulus per day.

CHEMOSIGNALS

OF FEMALE

MICE

63

Within-phase design. Both experiments consisted of multiple phases. Within each phase each subject was tested in a repeated measures design. The experimenter recording vocalizations was “blind” to subject and stimulus identity. When more than two trials occurred, the order of stimulus presentation was counterbalanced across subjects by randomly selecting without replacement from among the possible sequences of presentation. No sequence was repeated until all possible sequences had been exhausted. When only two stimuli were presented, the order of presentation was randomly determined. These data were analyzed parametrically. Between-phase design. Because of different base rates of vocalizing in the different phases of the experiment, the raw vocalization scores were inappropriate for comparing across different phases of an experiment. In order to make such comparisons possible as well as to obtain an overall graphical description of the experiment, each individual vocalization score to the odor of a surgically operated female or control odor was transformed to a percentage of the mean response to the odor of the sham-operated females in each of the different phases of an experiment. Parametric statistics were performed on the transformed scores for betweenphase analyses. EXPERIMENT

1

In this experiment, the effects of ovariectomy and hypophysectomy upon the production of facial, vaginal, and salivary odors of females to elicit ultrasonic vocalizations from males were examined. The effects of each of the surgeries upon each of the odors were examined in separate phases of the experiment. Within each of these phases, the effects of the surgery were assessed by comparing the effectiveness of the odor from operated females with that of females from the appropriate shamoperated group. Method

Phase 1: The Effect of Ovariectomy on Facial Secretions Animals. Fifteen hybrid male mice between 157 and 200 days of age served as subjects. Social experience animals consisted of six male and six female hybrid mice, 110-153 days of age on the first day of social experience. The stimulus donors were eight adult Swiss Webster females. Procedure. The stimulus donors had been either bilaterally ovariectomized (OVX-1, N = 4) or sham ovariectomized (SHAM-l, N = 4) at approximately 11 months of age. The first vocalization test occurred 4.5 days following ovariectomy. The stimulus swabs were individually prepared by removing the stimulus donor from its home cage and holding it by the nape of the neck. Facial secretions were obtained by rubbing a cotton

64

BYA’I-I-

AND NYBY

swab under the chin and over the face and snout three or four times. Care was taken to prevent the swab from contacting any other body areas and to keep the animal from biting the swab. The subjects were tested as two groups offset by 1 day. Phase 2: The Effect of Hypophysectomy

on Facial Secretions

Animals. Twenty-one hybrid male mice between 125 and 133 days old on the first day of testing served as subjects. Social experience animals consisted of 8 adult male and 8 adult female hybrid mice, 59 days of age on the first day of social experience. The stimulus donors were 14 CD-l females, 58 days old on the day of the first day of testing. Procedure. Stimulus collection was performed as described in Phase 1. The stimulus donors underwent either hypophysectomy (HYPOX-1, N = 7) or sham hypophysectomy (SHAM-l, N = 7) and provided stimuli beginning 19 days following surgery. Clean cotton swabs (CONTROL1) provided the third stimulus examined. With the exception of the stimulus collection procedure, the CONTROL-l swabs were handled in the same manner as the swabs carrying the facial secretions. Phase 3: The Effect of Ovariectomy

on Vaginal Secretions

Animals. Subjects were 18 hybrid male mice between 56 and 99 days of age on the first day of vocalization testing. Social experience animals were six male and six female hybrids, 198-241 days of age on the first day of social experience. Six of the stimulus donors from Phase 1 provided vaginal secretions and were approximately 14 months old at the time of the first experimental trial. Procedure. The two types of vaginal odors were from OVX-3 (N = 3) and SHAM-3 (N = 3) females. To prepare the stimulus, the stimuls animal was removed from its home cage and held by the nape of the neck. Such handling frequently resulted in the animal’s urinating. If the animal urinated, the urine was shaken from its body and the area was wiped with a paper towel. The tip of a cotton swab was inserted into the vagina and the swab rotated l/4 to l/2 turns three or four times. Care was taken to prevent contact of the swab with any other body areas. Phase 4: The Effect of Hypophysectomy

on Vaginal Secretions

Animals. Subjects were twenty-one hybrid male mice that were 139147 days of age on the first day of testing. Social experience animals were 8 male and 8 female hybrid mice, all 59 days of age on the first day of social experience. Stimulus donors were 14 CD-I females who had been used as stimulus donors in Phase 1 and were 73 days old on the first day of testing. Procedure. The stimulus donors had been either hypophysectomized

CHEMOSIGNALS

OF FEMALE

MICE

65

(HYPOX-4, N = 7) or sham hypophysectomized (SHAM-4, N = 7) at 39 days of age with the first day of testing being 34 days following surgery. In addition, clean cotton swabs (CONTROL-4) served as control stimuli. Phase 5: The Effect of Ovariectomy on Salivary Odors Animals. Sixteen hybrid males, between 194 and 199 days of age on the first trial served as subjects. Social experience animals were 8 male and 8 female hybrid mice approximately 180-260 days of age on the first day of social experience. Stimulus donors were 16 hybrid females between 193 and 274 days of age on the first trial. Procedure. The two stimuli tested came from OVX-5 females (N = 8) and SHAM-5 females (N = 8). Thirty-two days had elapsed following their surgeries. To prepare the stimulus, the donor was removed from its home cage and anesthetized with ether. The anesthetized animal’s mouth was opened and a cotton swab was inserted and rotated three or four times to absorb the saliva. Care was taken to prevent the swab from coming into contact with the animal’s face. Phase 6: The Effect of Hypophysectomy on Salivary Secretions Animals. Subjects were the 21 hybrid male mice used in Phase 2 and were 159-167 days old on the first day of testing. The 14 CD-l females who had served as stimulus donors in Phases 2 and 4 again served as stimulus donors. The stimulus donors were now 92 days of age on the first day of testing. In addition, the 7 hybrid males who also had served as social experience animals served as stimulus donors. These males were 104 days old on the first day of testing. Procedure. In addition to the original 8 days of social experience preceding Experiment 2, the subjects were given an additional 3 days of social experience preceding this experiment. The subjects were then given another screening test to determine their subsequent participation in this experiment. The three stimuli tested were saliva from HYPOX-6 females (N = 7), saliva from SHAM-6 females (N = 7), and saliva from normal males (MALE-6, N = 7). Vocalization tests began 53 days following surgery. The stimuli were prepared as described in the previous phase. Results

Within-Phase Analyses Phase 1. Male vocalizations to the facial secretions of OVX-1 females (Mean t SEM = 19.4 4 2.6) and SHAM-l females (Mean ? SEM = 22.9 t 2.7) were not significantly different (t(14) = 1.54, P = n.s.). Phase 2. The overall differences in vocalizations to facial secretions from the HYPOX-2 (Mean + SEM = 16.2 2 2.7), SHAM-2 (Mean 2

66

BYATT

AND NYBY

SEM = 30.3 + 1.5) and to the CONTROL-2 (MEAN t SEM = 9.9 -t- 2.5) were significant (F(2,40) = 39.33, P < .Ol). Orthogonal comparisons indicated that this difference could be accounted for in large part by the greater response to the SHAM-2 facial stimuli versus the lower response to the HYPOX-2 and CONTROL-2 stimuli (F(1, 40) = 71.49, P < .Ol). However, the subjects did emit significantly more vocalizations to the HYPOX-2 stimuli than to the CONTROL-2 stimuli (F(1, 40) = 7.18, P < .05).

An additional nonorthogonal selected contrast indicated that the amount of male vocalization to the SHAM-2 facial secretions were significantly greater than to the HYPOX-2 facial secretions (F(1, 40) = 35.79, P < .Ol). Phase 3. Vaginal secretions from OVX-3 females (Mean + SEM =

13.2 + 3.1) were not significantly different from those of SHAM-3 females (Mean ? SEM = 18.9 + 3.0) in their elicitation of vocalizations (t(17) = 1.60, P = n.s.). Phase 4. Males emitted significantly different (F(2, 40) = 18.99, P < .Ol) amounts of vocalizations to the vaginal secretions of HYPOX-4 females (Mean & SEM = 16.7 + 3.0), vaginal secretions of SHAM-4 females (Mean & SEM = 24.0 t 2.6), and CONTROL-4 stimuli (Mean + SEM = 9.6 + 2.5). Orthogonal comparisons indicated that significantly more vocalization was emitted in response to the SHAM-4 vaginal secretions than to the other two stimuli (F(1, 40) = 29.44, P < .Ol) and that more vocalizations occurred in response to the HYPOX-4 vaginal secretions than to CONTROL-4 stimuli (F(1, 40) = 8.54, P < .Ol). An additional nonorthogonal selected contrast indicated that the subjects vocalized more in response to the SHAM-4 vaginal secretions than to those of the HYPOX-4 females (F(1, 40) = 10.47, P < .Ol). Phase 5. Saliva from OVX-5 females (Mean ? SEM = 25.2 t 2.3) was not significantly different from the saliva of SHAM-5 females (Mean +- SEM = 23.6 + 2.7) in eliciting ultrasonic vocalizations from males (t(l5) = 1.43, P = n.s.). Phase 6. Overall, salivas from the HYPOX-6 females (Mean f SEM = 9.0 ? 2.1), the SHAM-6 females (Mean +- SEM = 14.6 ? 2.8), and the MALE-6 males (Mean t SEM = 7.0 + 2.2) were significantly different in eliciting vocalizations (F(2,40) = 11.92, P < .Ol). Orthogonal comparisons indicated that the differences among stimuli could be attributed in large part to the greater vocalizations to the SHAM-6 stimuli versus the HYPOX-6 and MALE-6 stimuli (F(1, 40) = 22.17, P < .Ol). The amounts of vocalization to the HYPOX-6 and MALE-6 stimuli were not significantly different (F(1, 40) = 1.67, P = n.s.). These results also confirm other findings (Wysocki et al., unpublished manuscript) that male mice vocalize more to the saliva of female mice than to that of male mice.

CHEMOSIGNALS

Between-Phase

67

OF FEMALE MICE

Analyses

A 2 between (type of surgery: ovx vs hypox) by 3 between (body site of stimulus: facial vs vaginal vs salivary) ANOVA was performed on the relevant transformed vocalization data (see Fig. 1). The effect of type of surgery was significant (F( 1, 106) = 6.38, P < .OS)reflecting the stimuli from the ovariectomized females genetally being better for eliciting vocalizations than the stimuli from the hypophysectomized females. Neither the effects of body site (F(2, 106) = .99, P = n.s.) nor the interaction of body site with type of surgery (F(2, 106) = 1.52, P = n.s.) were significant. Further analysis indicated that stimuli obtained from ovariectomized females were significantly better at eliciting vocalizations than those from hypophysectomized females for the facial stimuli (F(1) 34) = 4.94, P < .05) and for the salivary stimuli (F(1, 37) = 5.97, P < .05) but not for the vaginal simuli (F(1, 37) = 0.01, P = n.s.). EXPERIMENT 2 In Experiment 1, the property of female saliva to elicit vocalizations was not diminished significantly 1 month following ovariectomy. In the present experiment, the ability of female saliva to elicit vocalizations was assessed first at 8 months (Phase 1) and then at 9 months (Phase 2) following ovariectomy. While the female saliva donors were the same in both phases, the male subjects were different, In Phase 3 the effect of long-term ovariectomy (9 months) upon the ability of urine to elicit

120

A

- n.r.

.05

100 80 AMOUNT OF ULTRASOLINO (PERCENT OF 60 s44AIl STlmJLI) 4.

20 0 Clsm-2

Clean-4

clean swabs

CW-I

HYPOX-2 FaCld

CW-3

WfWX-4

Vaginal

WX-S

HYPOX-6 MALE-5 Sallvwy

TYPE W STltlWJS

FIG. 1. Mean ( 1 SEM) amount of ultrasound by males to facial, vaginal, and salivary odors from ovariectomized (OVX) and hypophysectomized (HYPOX) females. Clean swabs (CLEAN) and male salivary odors (MALE) also were examined. The number following each stimulus type designates the phase of the experiment in which the data were collected. Each raw vocalization score was transformed to a percentage of the mean vocalization score to the stimulus of the sham-operated females of the same phase of the experiment.

68

BYATT

AND NYBY

vocalizations was assessed to determine whether the property of urine to elicit vocalizations would be similarly affected. Method Animals. Thirty-two adult hybrid mice served as subjects, 16 in each phase. The saliva donors were 24 hybrid females. In Phase 1, the donors were between 273 and 319 days of age on the first trial and 224-245 days had elapsed since ovariectomy. In Phases 2 and 3, the donors were between 309 and 355 days of age on the first trial and 260-281 days had elapsed since surgery. Adult male and female hybrid animals served as social experience animals. The 16 males of Phase 2 plus 2 additional males from Phase 1 served as subjects in Phase 3. The urine donors in Phase 3 were the same animals that had provided saliva in Phases 1 and 2. Procedure. The stimuli in Phases 1 and 2 were saliva from OVX females (N = 16) and SHAM OVX females (N = 8). Stimulus preparation and presentation were identical to that described in Phase 5 of Experiment 1. The two stimuli of Phase 3 were urine from OVX and SHAM OVX females housed in groups of three overnight in metabolic cages (Maryland Plastics, EllO). The urine was collected in a syringe, and during the Imin habituation prior to use, 0.1 ml was placed on a cotton-tipped surgical swab. Presentation of the swab and vocalization quantification were as before. Results Within-Phase Analyses In both Phases 1 and 2 the saliva from OVX females was significantly less effective in eliciting vocalizations than was the saliva from SHAM OVX females: Phase 1 (t(l5) = 3.99, P < .Ol), Phase 2 (t(l5) = 3.80, P < .Ol). The means 2 SEM during Phase 1 for the ovariectomized females were 9.7 +- 2.3 and for the sham females, 18.4 k 2.4. During Phase 2 the means k SEM were 15.7 k 2.4 and 25.3 + 1.5 for the ovariectomized and sham females, respectively. Thus 8 or 9 months following surgery, the saliva of ovariectomized females was not as potent as the saliva from normal females for eliciting vocalizations. Long-term ovariectomy had no significant effect during Phase 3 upon the property of urine to elicit vocalizations from males (t(16) < 1, P = n.s.). Thus the depressive effect of long-term ovariectomy upon the ultrasound-eliciting property of saliva was not seen for urine. Between-Phase Analyses After transforming the vocalization data to a percentage of mean response to the stimuli of sham females, a l-between ANOVA on the transformed

CHEMOSIGNALS

69

OF FEMALE MICE

data indicated that long-term ovariectomy had significantly different effects upon the three stimuli depicted in Fig. 2 (F(2, 47) = 7.17, P < .Ol). Orthogonal comparisons indicated that the urinary stimuli elicited more vocalizations than did the two salivary stimuli (F(I, 47) = 20.5, P < .Ol) while the two salivary stimuli did not differ (F( 1, 47) = .070, P = n.s.). GENERAL DISCUSSION

Previous studies had demonstrated that saliva possesseschemosignalling properties in gerbils (Block, Volpe, and Hayes, 1981), rats (Teicher and Blass, 1977), and hamsters (Gray, Fischer, and Meunier, 1984). The present study extends this general finding to mice by demonstrating that male mice were able to use the saliva of female mice for gender identification and for the elicitation of male courtship vocalizations. Prevous work (Nyby et al., 1979) was consistent with the vocalizationeliciting chemosignal in female mouse urine being regulated directly by pituitary gonadotropins. Hormones of the ovary, on the other hand, were not essential for urinary chemosignal production (Nyby et al., 1979). The first experiment reported here similarly found that hypophysectomy of female mice also had a more disruptive effect than ovariectomy upon the vocalization-eliciting properties of facial and salivary secretions. Thus the vocalization-eliciting properties of urine, facial secretions, and saliva all appear to be more dependent upon pituitary than ovarian hormones. The relative effects of ovariectomy and hypophysectomy upon the .Ol

100

Amount of ultrasound (percent of Shamstimuli)

00

60

40

20

0 oa

-0rlo.

cw Salivary

-9rlo.

\

hn

-9no.

,

Urinary

Type of Stimulus FIG. 2. Mean (k SEM) amount of ultrasound by males to salivary and urinary odors of females that were ovariectomized for 8 months (Ovx-8 mo.) or 9 months (Ovx-9 mo.). Each raw vocalization score was transformed to a percentage of the mean vocalization score to the stimulus of the sham-operated females of that particular phase of the experiment.

70

BYATT

AND NYBY

ultrasound-eliciting properties of vaginal secretions were more equivocal. While the within-phases analyses seemed to indicate that hypophysectomy had a more significant depressive effect than ovariectomy, the betweenphase analyses indicated that these two surgeries were equivalent (see Fig. 2). In some ways, the results reported here point to conclusions similar to those of Hayashi and Kimura (1974) who examined the effects of both vaginal and urinary odors of female house mice on male mounting. While the effectiveness of vaginal odors for promoting male mounting varied appropriately with the female estrous cycle, the effectiveness of urinary odors was independent of female cyclicity. In other relevant research examining hamsters, Macrides, Singer, Clancy, Goldman, and Agosta (1984) found that ovariectomy and hypophysectomy equally reduced the effects of vaginal discharge on male investigatory and copulatory behavior. While more work examining mouse vaginal secretions is necessary for firm conclusions, it does appear possible that the vaginal vocalization-eliciting odor in mice is regulated directly by the ovary. If true, the hormonal regulation of this odor would be different from the regulation of the other odors of female mice that have been examined. The last experiment appeared to dissociate the hormonal mechanisms underlying salivary and urinary elicitation of vocalizations. While shortterm ovariectomy (1 month) had minimal effects upon either chemosensory cue, long-term ovariectomy (8 or 9 months) significantly suppressedsalivary but not urinary elicitation of vocalizations. The mechanistic explanation of this difference is not clear. One possibility is that perhaps the same chemosignal is present in both body fluids but is present in greater concentrations in urine than saliva. Perhaps following long-term ovariectomy enough of the cue remains in urine but not saliva to elicit normal levels of vocalizations. Very subjectively, urine has appeared to us to be a more reliable stimulus for eliciting vocalizations than saliva in the experiments reported here and elsewhere (Wysocki ef al., unpublished). Another possibility is that perhaps different naturally occurring chemosignals possess the property of eliciting vocalizations and these chemosignals may be regulated somewhat differently. On the other hand, other work (Nyby et al., 1979) found that ovariectomy produced a small deficit in the ultrasound-eliciting property of female urine that reliably emerged only after many trials. Thus the difference between stimuli observed here could have been a chance occurrence. However, other research (Wysocki et al., unpublished) has found other differences between the urinary and salivary elicitation of ultrasonic vocalizations. Previous experience with females of a particular genotype enhanced vocalizations towards saliva of females of only that genotype. Ultrasonic vocalizations to urine, on the other hand, appeared less dependent upon experience with females of a particular genotype. Wysocki et al. (unpublished), speculated that both saliva and urine contain chemical

CHEMOSIGNALS

OF FEMALE MICE

71

cues that can be utilized for gender recognition and individual identity but urinary cues may be biased in favor of gender recognition while salivary cues may be biased in favor of individual recognition. The existence of more than one natural chemosignal that elicits vocalizations would not be surprising. Previous research (Nyby, Whitney, Schmitz, and Dizinno, 1978; Kerchner, Vatza, and Nyby, 1985) demonstrated that novel odors that normally do not elicit vocalizations will do so if reliably associated with female mice. Thus any naturally occurring odor that is specific to females could serve as a stimulus for vocalization elicitation. Multiple chemosignals carrying the same or similar information may exist to provide redundant backup systems. In summary, ovarian hormones appeared less important than pituitary factors in the regulation of facial, salivary, and urinary chemosignals that elicit ultrasonic vocalizations from males. Vaginal odors, however, may be regulated directly by the ovary. Evidence gathered here and elsewhere suggests that the urinary, salivary, and vaginal chemosignals which elicit vocalizations may not be identical. ACKNOWLEDGMENTS This research was supported in part by NSF Grant BNS-811134. Portions of this research partially fulfilled the M.S. requirements of S. B. We thank Charles J. Wysocki for critically reading an earlier version of this manuscript.

REFERENCES Block, M. L., Volpe, L. C., and Hayes, M. J. (1981). Saliva as a chemical cue in the development of social behavior. Science (Washington, D.C.) 211, 1062-1064. Campbell, C. S., and Schwartz, N. B. (1977). Steroid feedback regulation of luteinizing hormone and follicle stimulating hormone secretion rates in male and female rats. J. Toxicol. Environ. Health 3, 61-95. Drickamer, L. C., and McIntosh, T. K. (1980). Effects of adrenalectomy on the presence of a maturation-delaying pheromone in the urine of female mice. Horm. Behav. 14, 146-152. Gray, B., Fischer, R. B., and Meunier, G. F. (1984). Preferences for salivary odor cues by female hamsters. Horm. Behav. 18, 451-456. Hayashi, S., and Kimura, T. (1974). A sex attractant emitted by female mice. Physiol. Behav. 13, 563. Johnston, R. E., and Bronson, F. (1982). Endocrine control of female mouse odors that elicit luteinizing hormone surges and attraction in males. Biol. Reprod. 27, 1174-l 180. Kerchner, M., Vatza, E., and Nyby, J. (1985). Ultrasonic vocalizations by male housemice to novel odors: The roles of infant and adult experience. J. Camp. Psycho/. 99, 479-490.

Leon, M., and Moltz, H. (1973). Endocrine control of the maternal phenomone in the postpartum female rat. Physiol. Behav. 10, 65-67. Macrides, F., Singer, A. G., Clancy, A. N., Goldman, B. D., and Agosta, W. G. (1984). Male hamster investigatory and copulatory responses to vaginal discharge: Relationship to endocrine status of the female. Physiol. Behav. 33, 633-637. Nyby, J. (1983). Volatile and nonvolatile chemosignals of female rodents: Differences in hormonal regulation. In D. Muller-Schwarze and R. M. Silverstein (Eds.), Chemical Signals in Vertebrates III, pp. 179-194, Plenum, New York/London.

72

BYATT AND NYBY

Nyby, J., Whitney, G., Schmitz, S., and Dizinno, G. (1978). Postpubertal experience establishes signal value of mammalian sex odor. Behav. Viol. 22, 545-552. Nyby, J., & Whitney, G. (1980). Experience affects behavioral responses to sex odors. In D. Muller-Schwarze and R. M. Silverstein (Eds.), Chemical Signals in Vertebrates and Aquatic Invertebrates, pp. 173-192. Plenum, New York/London. Nyby, J., Wysocki, C. J., Whitney, G., and Dizinno, G. (1977). Pheromonal regulation of male mouse courtship. Anim. Behav. 25, 333-341. Nyby, J., Wysocki, C. J., Whitney, G., Dizinno, G., and Schneider, J. (1979). Elicitation of male mouse ultrasonic vocalizations. I. Urinary Cues. J. Comp. Physiol. Psycho/. 93, 957-975.

Scott, J. P. (1966). Agonistic behavior of mice and rats: A review. Amer. Zoo/. 6, 683. Teicher, M. H., and Blass, E. M. (1977). First suckling response of the newborn albino rat: The roles of olfaction and amniotic fluid. Science (Washington, D. C.) 198, 63% 636.

Wysocki, C. J., Nyby, J., Bernhard, R., and Byatt, S. (1985). Salivary cues communicate gender and genotypic identity among mice. Unpublished manuscript.