Current research in amphibians: Studies integrating endocrinology, behavior, and neurobiology

Abstract

Amphibian behavioral endocrinology has focused on reproductive social behavior and communication in frogs and newts. Androgens and estrogens are critical for the expression of male and female behavior, respectively, and their effects are relatively clear. Corticosteroids have significant modulatory effects on the behavior of both sexes, as does the peptide neuromodulator arginine vasotocin in males, but their effects and interactions with gonadal steroids are often complex and difficult to understand. Recent work has shown that the gonadal hormones and social behavior are mutually reinforcing: engaging in social interactions increases hormone levels just as increasing hormone levels change behavior. The reciprocal interactions of hormones and behavior, as well as the complex interactions among gonadal steroids, adrenal steroids, and peptide hormones have implications for the maintenance and evolution of natural social behavior, and suggest that a deeper understanding of both endocrine mechanisms and social behavior would arise from field studies or other approaches that combine behavioral endocrinology with behavioral ecology.

Introduction

Current amphibian behavioral neuroendocrinology research combines the approaches originated with Frank Moore’s establishment of the newt Taricha granulosa as an important subject for studying the hormonal basis of courtship behavior (for review, see, Moore et al., this issue) with two other investigative traditions. One is the neuroethological investigation of anuran acoustic communication that grew out of Robert Capranica’s seminal work on the neural processing of advertisement calls (for review, see Capranica, 1976; Wilczynski and Capranica, 1984). The other involves evolutionary animal behavior studies using anurans for investigating basic mechanisms of sexual selection, mate choice, and the evolution of communication systems (e.g., Ryan, 1985). What all these research directions have in common is an interest in reproductive social behavior. That overlap provides the foundation for a multidisciplinary interplay across several dimensions—endocrinology and neurobiology, physiology and behavior, proximate and ultimate mechanisms—that now characterizes amphibian behavioral neuroendocrinology.

This strong focus on reproductive social behavior has resulted in several lacunae in amphibian behavioral endocrinology, despite a well-developed background of basic amphibian endocrinology (reviewed in Herman, 1992; Jorgensen, 1992). No current research targets seasonal homing or migration, but a few studies have investigated the behavioral endocrinology of energetics as related to mate signaling (Marler and Ryan, 1996) and foraging (Carr, 2002; Crespi and Denver, 2004). The role of thyroxin in triggering metamorphosis and the subsequent reconfiguration of the endocrine system is well known (Burggren and Just, 1992), and there is a substantial body of work on the interaction of environmental stress, corticotrophin-releasing hormone, and thyroxin in adaptively regulating metamorphosis (Denver et al., 2002; Boorse and Denver, 2004), but behavioral correlates (other than sexual development) are not precisely understood. The vast majority of current studies that incorporate a strong emphasis on behavior are concerned with social signaling and the response to those signals.

Not surprisingly, given this focus, much of the current amphibian research centers on gonadal steroids (androgens and estrogens) and the peptide neuromodulator arginine vasotocin (AVT; the nonmammalian homolog of vasopressin, AVP). Similar work continues to be done on courtship behavior in two of the three major amphibian groups, the urodeles (salamanders and newts) and the anurans (frogs and toads). The third group, the apodans, is largely uninvestigated outside of neuroanatomical studies (Pinelli et al., 1997; Hilscher-Conklin et al., 1998; Ebersole and Boyd, 2000). The urodeles and anurans share some basic behavioral characteristics, such as males producing advertisement signals and females using them for mate choice, and the presence of amplexus or stereotyped clasping behavior in which males tightly grasp females as females and males release gametes for external fertilization. The two groups differ significantly in the details of their behavior. Anuran advertisement signals are acoustic, while in urodeles, chemical communication via the release of pheromones is critically important. Furthermore, male anurans commonly form lek-like aggregations during the breeding season (reviewed in Wells, 1977) in which males communally display and engage in agonistic vocal and physical interactions to defend call sites from calling intruders or interception of females by silent satellite males. Despite these differences, there are strong parallels in the effects of gonadal steroids and AVT in the two amphibian groups.

Steroid hormones and amphibian behavior

In the classic vertebrate pattern, gonadal steroids are essential for both the organization of body structures distinguishing the sexes and for the later expression of sexually specific behavior, which in both anurans and urodeles has been studied most often in terms of the (male) production of and (female) response to advertisement (or mating) signals. The largest body of work has focused on males, with interest turning to female behavior only recently.

Steroid hormones and courtship behavior in male amphibians: androgens

The influence of androgens on male courtship signaling begins postmetamorphically with their effect on the development of the muscular systems responsible for the behavior, leading to larynges that are larger in males (Sassoon and Kelley, 1986; McClelland and Wilczynski, 1989; Ryan and Drewes, 1990; McClelland et al., 1997) with sexually dimorphic muscle fibers (Sassoon and Kelley, 1986; Sassoon et al., 1987), innervation (Robertson et al., 1994), and androgen binding characteristics (Segil et al., 1987; Boyd et al., 1999). Similar androgen-dependent sex differences are found in the oblique muscles of the body wall (Tiagen et al., 1985; Emerson et al., 1999; Girgenrath and Marsh, 2003), and in the flexor carpi radialis, the forelimb flexor muscle used by males to clasp females while they oviposit (Herrera and Regnier, 1991; Regnier and Herrera, 1993; Dorlöchter et al., 1994; Sidor and Blackburn, 1998).

Androgens, specifically testosterone and dihydrotestosterone, have activating effects on male calling as well, which is the most well-studied behavioral effect of the hormone. Androgens (and testes function in general) fluctuate seasonally in concert with reproductive social behavior (Herman, 1992), following the standard vertebrate model for seasonal breeders. The production of advertisement signals is androgen dependent (Wada et al., 1976; Wada and Gorbman, 1977; Wetzel and Kelley, 1983; Solis and Penna, 1997; Iwata et al., 2000; Burmeister and Wilczynski, 2001) and experimentally eliminated upon castration (Dodd, 1960; Schmidt, 1966; Palka and Gorbman, 1973; Kelley and Pfaff, 1976; Deviche and Moore, 1988; Burmeister and Wilczynski, 2001). Just as androgens stimulate call production in anurans, they stimulate pheromone synthesis in male urodeles (see Kikuyama et al., 2002, for review). The specific contribution of androgens in regulating the expression of advertisement signals in anurans is not always clear, however, as investigations of natural patterns in the field have shown. For instance, studies that correlate calling behavior in male anurans with androgen levels report conflicting results. In some cases, androgens are lower in calling anurans relative to non-callers (Mendonça et al., 1985), whereas in other species, androgens are higher in calling males (Townsend and Moger, 1987; Marler and Ryan, 1996). Evoked vocalization rate and plasma androgens were not correlated in a laboratory study of male Hyla cinerea while spontaneous calling rates in the same animals were (Burmeister and Wilczynski, 2001), but androgen concentration and evoked vocalization rate were correlated in a field population of Batrachyla taeneiata (Solis and Penna, 1997).

Clearly, androgens are at least necessary for the initiation and maintenance of seasonal calling behavior. Nevertheless, laboratory and field studies indicate that other factors contribute to calling behavior and interact with androgens in complex ways. Social control seems capable of entraining calling rates once calling is enabled by adequate androgen levels (Burmeister and Wilczynski, 2001). Interspecific variation in androgen control of calling may be a consequence of life history strategies such as expression of aggressive, territorial, or parental behavior (Townsend and Moger, 1987; Townsend et al., 1991; Emerson et al., 1993; Emerson and Hess, 1996; Leary et al., 2004), flexibility in seasonality (Houck and Woodley, 1995; Harvey et al., 1997), or energy required for the species-specific advertisement call (Emerson and Hess, 1996, 2001; Leary et al., 2004). The contributions of other hormonal systems need also be considered, as do brain peptide neuromodulator systems. One such hormone, highlighted for many years by Frank Moore and colleagues (see Moore et al., this issue), is the adrenal stress steroid corticosterone. A second possible influence, much less understood, is the pituitary peptide gonadotropin itself, which a few intriguing studies suggest may have behavioral effects independent of its regulatory role on gonadal steroid secretion.

Steroid hormones and courtship behavior in male amphibians: corticosterone

Corticosterone levels vary seasonally (Licht et al., 1983; Pancak and Taylor, 1983; Dupont et al., 1979; Jolivet-Jaudet and Ishii, 1985; Zerani and Gobbetti, 1993), are socially modulated (Burmeister and Wilczynski, 2000), and are thought to contribute to the regulation of male reproductive behaviors in both anurans (Marler and Ryan, 1996) and urodeles (reviewed in Moore and Rose, 2002, Moore et al., this issue). For example, in male rough-skinned newts (T. granulosa), corticosterone rapidly inhibits male clasping behavior. Such rapid effects may be mediated by direct action of corticosterone on membrane receptors (Orchinik et al., 1991). Some of corticosterone’s effects have also been attributed to an inhibition of the HPG axis, as high corticosterone is often associated with low androgen levels (Moore and Rose, 2002; Licht et al., 1983; Marler and Ryan, 1996). However, just as for androgens, corticosterone’s reported relationship to courtship, and to androgen levels, in more natural contexts is inconsistent. In crested newts (Triturus carnifex) corticosterone levels are lower and androgen levels higher in inactive males than in actively courting males (Zerani and Gobbetti, 1993). In some anuran species, corticosterone is elevated in calling males compared to non-callers and does not seem to have an effect on androgen levels (Mendonça et al., 1985; Orchinik et al., 1988; Harvey et al., 1997; Burmeister et al., 2001; Leary et al., 2004).

Emerson (2001) proposed the “Energetics-Hormone Vocalization” model to explain the confusing patterns in corticosterone and androgen levels in naturally behaving animals by incorporating energetic considerations (Fig. 1). Calling is very energetically costly, and corticosterone levels are related to energetic stress. Corticosterone levels in breeding male anurans are in fact higher in species with higher call energy and call rate (Emerson and Hess, 2001). Emerson suggests that corticosterone is elevated in order to meet the energetic demands of advertising but at some threshold, which may vary between species, corticosterone interferes with androgen production thereby inhibiting calling. Calling is reinstated when the male’s energy reserve is restored, corticosterone levels are lowered, and androgen is elevated once again. Leary et al. (2004) tested this model and found that although corticosterone was higher in calling males relative to non-calling satellite males in Bufo woodhousii and B. cognatus, androgen concentrations did not differ. Instead, in order to understand the transition between calling and non-calling behavior, the authors proposed a model that includes energy reserves, androgen and corticosterone concentrations, and direct actions of corticosterone on brain AVT neurons (Leary et al., 2004). Future studies need to examine the interaction between these elements across species, and in both anurans and urodeles, in natural populations in order to understand the diversity of reproductive strategies and their neuroendocrine correlates.

Fig. 1.

Simplified representation of Emerson’s “Energetics-Hormone Vocalization” model (Emerson, 2001). Calling causes corticosterone to increase due to energetic stress; engaging in calling interactions causes androgens to increase; corticosterone has a negative effect on androgen production. The schematic graph below shows that the effects of this interaction can cause corticosterone and androgens to differ in their relationship during a breeding season.

Hormones and courtship behavior in male amphibians: gonadotropin

Most work on the behavioral endocrinology of courtship behavior has targeted the steroid endpoint of the HPG axis. Often, this focus is confirmed experimentally. Gonadotropin injections induce clasping in intact males (Kelley and Pfaff, 1976; Wada and Gorbman, 1977; Schmidt, 1966); however, after castration, gonadotropin is no longer effective in inducing clasping but implantation with T or DHT will reinstate it (Kelley and Pfaff, 1976). Other work, however, suggests that gonadotropin itself, via effects on the brain, may contribute to courtship regulation. Early experiments found that androgen replacement following castration did not fully reinstate male calling in some anuran species (Schmidt, 1966; Palka and Gorbman, 1973; Wada and Gorbman, 1977; Wetzel and Kelley, 1983) but that additional treatment with gonadotropins reinstated courtship behavior to the level of intact males (Wetzel and Kelley, 1983). Such results suggest that gonadotropins themselves may substantially contribute to the initiation of male courtship behaviors in amphibians. In fact, in recent studies, in male X. laevis, localization of LH receptor mRNA indicates that LH receptors are distributed in several brain areas thought to be involved in regulating vocalization (Yang and Kelley, 2004).

Steroid hormones and courtship behavior in female amphibians: estrogens

Studies that have examined hormones and behavior in female amphibians generally concentrate on hormonal induction of receptivity toward males, and have examined female anurans far more than female urodeles. Sexually receptive behavior in female amphibians includes approaching an advertising male (Schmidt, 1984, 1985; Zerani and Gobbetti, 1993), emitting a vocalization in some species (Tobias et al., 1998), or inhibiting release calls or leg extensions (Diakow and Nemiroff, 1981; Kelley, 1982). Early studies found that female American toads (Bufo americanus) will approach a conspecific mate signal when injected with a variety of peptide or steroid hormones, such as human chorionic gonadotropin (HCG) or prostaglandin (Schmidt, 1984, 1985; Weintraub et al., 1985). In anuran species in which the female is able to vocalize, call production occurs when the female has mature eggs (Roy et al., 1995; Bush, 1997; Tobias et al., 1998), which is the same period when gonadal steroids are elevated (Harvey et al., 1997; Gobbetti and Zerani, 1999). Surprisingly, until recently, there have been few attempts to relate female behavior to estrogens.

Female gonadal steroid levels are seasonally modulated (Licht et al., 1983; Pierantoni et al., 1984; Iela et al., 1986). In species with long breeding seasons, females may cycle through breeding (essentially ovulation) stages more than once, and gonadal steroids fluctuate over these cycles (Iela et al., 1986; Harvey et al., 1997; Medina et al., 2004; Lynch and Wilczynski, 2005). Several recent studies of natural populations have now shown that female anurans also vary different aspects of their mate choice decisions throughout different reproductive stages (Lea et al., 2000; Bosch and Boyero, 2004; Lynch et al., 2005; Table 1). Near the point of ovulation, female túngara frogs exhibit their highest levels of receptivity and also become less choosy, that is, they will respond behaviorally to previously unattractive calls (Lynch et al., 2005). This is the same point at which circulating estrogens (E) and progesterone (P) are highest (Lynch and Wilczynski, 2005), leading to the suggestion that these gonadal steroids modulate female responses to male calls just as male gonadal steroids modulate call production.

Table 1.

The frequency of three mate choice behaviors and the concentration of three gonadal steroids during different reproductive stages in female anurans; arrows indicate whether the frequency of behavior or hormone concentration increased or decreased during the stage

Reproductive stage	Hormones			Mate choice behaviors
	Estrogen	Progesterone	Testosterone	Receptivity	Permissiveness	Discrimination
Unamplexed
Amplexed
Post-mated

Results from Lynch et al. (2005) and Lynch and Wilczynski (2005).

This conjecture awaits firm experimental confirmation, but work to date has shown that, as for testosterone in males, female behavioral endocrine control will be complicated. For example, in X. laevis, receptivity can be induced in ovariectomized females with just E and P administration; however, maximal receptivity (i.e., vocalization produced and leg extensions inhibited) requires an additional HCG injection (Kelley, 1982). In R. pipiens, E and P administration alone did not inhibit release calls, an indicator of receptivity, in ovariectomized females (Diakow, 1978), suggesting that E and P are necessary but not sufficient for evoking maximally receptive behaviors in females.

Steroid hormones and courtship behavior in female amphibians: testosterone and corticosterone

Female endocrinology is more complex than that of males, and consequently its relationship to behavior more difficult to understand. Circulating plasma testosterone levels are as high in reproductive females as in males (although in males dihydrotestosterone is higher) and higher than estrogen levels in many anuran species (d’Istria et al., 1974; Licht et al., 1983; Iela et al., 1986; Itoh and Ishii, 1990; Harvey et al., 1997; Wilczynski et al., 2003; Medina et al., 2004; Lynch and Wilczynski, 2005). The behavioral relevance of this, if any, is unclear. It is possible that in species where females vocalize, testosterone may be involved in regulating this behavior (Emerson and Boyd, 1999). Despite evidence of corticosterone’s importance in male behavior, little definitive evidence of a role in female social behavior has emerged. It is intriguing, however, that in crested newts, corticosterone was the only steroid that Zerani and Gobbetti (1993) found to differ in concentration between receptive and non-receptive females.

Peptide neuromodulators and amphibian behavior

Peptide neuromodulators influence a vast array of social behaviors in vertebrates (Goodson and Bass, 2001), and often interact with steroid hormones to do so. In amphibians, the peptide AVT, which acts both as a CNS neuromodulator and a neurohypophyseal peptide, is by far the most widely studied. Experimental studies of AVT effects are generally done using peripheral administration, with the assumption that the behavioral action is via agonist effects on AVT neuromodulator systems of the brain. This is a reasonable assumption based on intra-ventricular cranial injections showing identical effects at lower doses (Moore and Miller, 1983). Experiments describing AVT immunoreactivity (AVT-ir), messenger RNA, local concentrations of the neuropeptides, and putative receptor localization all demonstrate that AVT is found in regions important for many aspects of social signaling from signal perception to production (reviewed in Emerson and Boyd, 1999; Moore and Rose, 2002) (Fig. 2). Still, AVT does have peripheral effects on vasculature and other smooth muscle, which could potentially affect, for example, laryngeal function, and it is part of the stress response system, facilitating the release of corticosteroids. Some attention will eventually need to be paid to potential peripheral effects of AVT treatment.

Fig. 2.

Schematic diagram depicting the neural areas involved in AVT-mediated social behavior in anurans and urodeles, including representative brain regions and their hormonal targets. Although the sensory modalities mediating social behavior in the two groups are different, the basic framework for how the behaviors are controlled is quite similar. Both groups rely on sensory input to brainstem premotor nuclei (with or without mediating input from forebrain regions) which then modulate specific neuromuscular machinery. AVT-R appears to be restricted to neural loci upstream of motor regions, suggesting that AVT might have more of a role in modulating upstream circuitry rather than influencing motor output. Details for the anuran model are based on Emerson and Boyd (1999). Anatomical information for the urodele model is based on Thompson and Moore (2000), steroid hormone receptor information is based on Davis and Moore (1996), and AVT receptor distribution is based on Boyd and Moore (1991). Abbreviations: AR, androgen receptor; AVT-R, arginine vasotocin receptor; ER, estrogen receptor; n. IX–X, motor nucleus of cranial nerve IX–X; POA, preoptic area; PTN, pretrigeminal nucleus; TS, torus semicircularis.

The earliest studies of AVT and social behavior in amphibians were conducted in the newt T. granulosa, in which AVT facilitates amplexus (see Moore et al., this issue), a result seen in other urodeles (Iwata et al., 2000). Blocking AVT receptors in urodeles has a significant detrimental impact on male behavior (Iwata et al., 2000; Moore and Rose, 2002; Toyoda et al., 2003). This suggests that AVT, while perhaps not sufficient for producing male social behavior, is important in some aspect of its production.

Studies in many species of frogs consistently show that AVT stimulates male courtship calling (Penna et al., 1992; Boyd, 1994a; Marler et al., 1995; Propper and Dixon, 1997; Chu et al., 1998). It also changes features of the call produced. Depending on the species, AVT changes call patterning (Marler et al., 1995; Chu et al., 1998), call duration and pulse number (Klomberg and Marler, 2000; Trainor et al., 2003), or maintenance of multiple call types (Tito et al., 1999); aggressive behavior (call-site acquisition) may also increase (Semsar et al., 1998). Interestingly, in gray treefrogs, call duration and pulse number are only stimulated by AVT if another calling male is nearby, suggesting a very interesting interaction between social stimuli and AVT effects (Trainor et al., 2003). As in urodeles, blocking AVT receptors greatly decreases courtship behavior (calling) in one anuran species (Propper and Dixon, 1997). The common theme among these various species is that AVT appears to enhance motivation to call, whether by increasing call rate or call duration, or the likelihood that a male will call. It is unclear why AVT differentially affects certain parameters of male calls. Perhaps AVT enhances the feature of the call that is most salient in attracting females in that species, a hypothesis put forth by some of these studies (e.g., Marler et al., 1995; Trainor et al., 2003). However, it remains unknown whether females prefer the calls of AVT-injected males. Testing this is critical for determining the functional significance of AVT’s modulation of calling in natural settings.

It would be very interesting to know if pheromonal signals in urodeles follow a similar pattern. It is already well established that newt courtship behavior, such as clasping, is stimulated by AVT. But are any male signals altered? Would females be more receptive to AVT-treated newts, or is that irrelevant in this amphibian group? What makes these comparative questions interesting is that AVT clearly stimulates male behavior in two amphibian groups with very different signaling strategies. Understanding the parallels may help unravel how AVT actually exerts its effects. The mechanism by which AVT (or, in mammals, AVP) influences social behavior is difficult to grasp. The fact that there are significant species differences in AVT’s effects across vertebrates (Goodson and Bass, 2001) adds to the difficulty. It does not seem as though AVT treatment works indirectly by increasing circulating testosterone; AVT effects are rapid, and in one direct test of this, AVT treatment did not alter plasma T levels (Burmeister et al., 2001). One theory of AVT (or AVP) action is that it increases overall activity or excitability of target cells (Raggenbass, 2001). Effects on sensory processing may be a key part of the process. In newts, AVT increases the activity of medullary neurons in response to a cloacal stimulus that induces clasping (Rose et al., 1995), and auditory midbrain activity in green treefrogs is enhanced with AVT treatment (Penna et al., 1992). AVT increases the appetitive response to both visual and pheromonal cues used in sexual behavior in Taricha, suggesting that the effect of AVT on behavior involves specific sensory modalities rather than a generally increased motivational state (Thompson and Moore, 2000). Furthermore, AVT receptors are located in parts of the brain involved in both signal perception and signal production and hence may modulate both aspects of social behavior (Rose and Moore, 2002).

Virtually all the published work on neuropeptide modulation of amphibian social behavior has been on males. Male-biased sex differences in the neural expression of AVT (Boyd and Moore, 1992; Boyd et al., 1992; Marler et al., 1999) in fact do suggest that AVT has more influence on male than female behavior. AVT can however induce a variety of behaviors in females depending on the hormonal environment. Female release calls are inhibited (Boyd, 1992) and female receptivity increased by AVT treatment (Diakow, 1978). Estradiol treatment combined with AVT injection produces typical female behavior in ovariectomized female Taricha (Moore et al., 1992; Thompson and Moore, 2003). Furthermore, AVT can induce phonotaxis in females of some amphibian species (Schmidt, 1984; Boyd, 1994a). These intriguing findings suggest AVT may increase the expression of reproductive social behaviors in both males and females, but the affected behaviors are completely different. This is consistent with the idea that AVT acts prior to the motor control itself, perhaps through sensitizing sensory systems to social stimuli or gating sensorimotor processing. Studies of peptide modulation of female behavior would be a fruitful avenue of investigation for narrowing in on the still unresolved mechanism by which this peptide influences behavior.

Peptide–steroid interactions

It is well-established that the expression of AVT in the brain is sensitive to steroid hormones. Castration decreases AVT (measured by radioimmunoassay) in the bullfrog amygdala, among other areas (Boyd, 1994b), as well as the level of putative AVT receptors in the newt brain (Boyd and Moore, 1991). AVT and sex steroid receptors are found in many of the same areas controlling courtship signaling (Fig. 2). Behavioral assays have consistently revealed that AVT changes male social signaling in amphibians only when circulating testosterone levels are high. For example, injections of AVT in H. cinerea increase the likelihood of calling only in the presence of testosterone (Penna et al., 1992). AVT also interacts with testosterone to facilitate clasping behavior in male Taricha (Moore and Rose, 2002). Recent preliminary neurophysiological studies of the frog auditory system find the same pattern: AVT treatment lowers midbrain auditory thresholds in testosterone-implanted, but not control, males (Miranda and Wilczynski, 2004). Testosterone combined with AVT also produces male-like behavior in females (Moore et al., 1992; Thompson and Moore, 2003), and this clasping behavior is sensitive to both visual and olfactory cues (Thompson and Moore, 2003). Interestingly, androgenized females injected with AVT do not produce clasping to the degree of males, which suggests an organizational effect of androgen on these sensorimotor pathways.

Social modulation of endocrine state

This review has thus far largely been concerned with the traditional focus in behavioral neuroendocrinology, the ways in which hormonal state regulate behavior. This relationship can be examined in reverse: how does behavior regulate hormonal state? Emerson’s “Energetics-Hormone Vocalization” model is one such example, where calling, through its energetic demands, induces corticosterone changes, which in turn antagonize androgens, leading to a cessation of calling. More directly, though, there is growing evidence from anurans that the simple reception of conspecific calls can lead to elevations of gonadal steroids.

Brzoska and Obert (1984) first reported evidence for such a phenomenon, showing that male Rana esculenta exposed to conspecific calls maintained mature testes longer than did males exposed to control stimuli or no sound. The neural pathways underlying this effect were later described (reviewed in Wilczynski and Chu, 2001), and more controlled experimental conditions and direct measures of androgens confirmed that indeed male anurans (H. cinerea) exposed to mate choruses for several days show elevated androgens compared to control males (Burmeister and Wilczynski, 2000; Chu and Wilczynski, 2001; O’Bryant et al., 2004) (Fig. 3). The effects are even seen at the level of the brain GnRH cells; immunocytochemistry reveals more GnRH-positive cells in males exposed to conspecific calls, presumably due to increased GnRH production correlated with its increased release and subsequent higher plasma androgens (Burmeister and Wilczynski, 2005). Similar effects are indicated in females. Reproductive condition is altered in midwife toads, Alytes muletensis, after hearing calls (Lea et al., 2001). Recently, Lynch and Wilczynski (2004) reported that call reception also increases gonadal steroids in female túngara frogs (Physalaemus pustulosus). Social elevation of gonadal steroid levels may be important as gonadal steroids decline over a breeding season (Jorgensen, 1992; O’Bryant et al., 2004). Hearing calls may combat the decline and sustain breeding condition in both males and females (Fig. 4).

Fig. 3.

Increase in testosterone and androgens in male H. cinerea after 10 days of nightly exposure to a recording of a conspecific chorus, random tones, or no sound. Data are from Burmeister and Wilczynski (2000).

Fig. 4.

Diagram showing hypothesized function of social signaling in maintaining reproductive state. Gonadal hormones and reproductive capability decline seasonally, eventually falling below some threshold for reproduction to occur. Receiving social signals increases gonadal steroid levels, which would keep reproductive state from falling below threshold until later in the season.

Whether peptide systems such as AVT are modified by social interactions remains an open question. Behavioral status is certainly reflected in brain AVT systems. In cricket frogs (Acris crepitans), calling males have smaller and fewer AVT cells in the nucleus accumbens area than non-calling satellites, presumably reflecting increased release in the callers and sequestering in the satellites when calling is socially inhibited (Marler et al., 1999). In the newt Taricha, AVT content in many brain regions is higher in “sexually-responsive” compared to “sexually-unresponsive” males (Zoeller and Moore, 1988). It is not known, however, whether any of these differences are caused by social cues or are reflections of other mechanisms driving behavioral differences. Controlled experiments that address this would be important regardless of whether they do or do not find that social interactions can change brain peptide systems, as such studies are important for answering whether these neuromodulatory systems are modulated by social cues along with gonadal steroid state and behavior in a coordinated fashion.

The reciprocal, reinforcing influences of hormones and calling are interesting from a mechanistic standpoint, but are perhaps more so from the perspective of sexual selection and the evolution of anuran social communication. Most anurans engage in communal displays (leks) in which males produce calls for long periods. Most interpretations of the function of these advertisement signals consider the behaviors they trigger, and ideas about their evolution relate to competition for attracting mates. If, however, in both sexes, calls also elevate gonadal steroids, hormones that are important for regulating reproductive state including gametogenesis, there is a hidden function of social signals and a competition occurring at the endocrinological level in parallel with the behavioral competition more easily observed. Is this function of communication signals a universal consequence of social signaling, or is it a specialization related to lekking behavior or to long-term advertisement displays? Employing the natural variation among amphibians would provide fascinating insights into this problem. For example, do calls stimulate gonadal steroids in “scramble breeders”, species in which the breeding season is exceptionally short and males clasp any remotely appropriate target they encounter (Wells, 1977)? Do the pheromonal and tactile exchanges typifying urodele social behavior have the same capacity to elicit hormone increases? These are interesting questions from both mechanistic and fundamental evolutionary perspectives.

Behavioral endocrinology and behavioral ecology

The interactions of hormones and behavior suggest that extending mechanistic behavioral and neural endocrinology studies to field investigations or studies that take natural behavior patterns into account could engender a deeper understanding of both endocrine mechanisms and animal social behavior. Understanding the interaction of corticosterone and gonadal steroids, and both hormones with peptide hormones, could similarly benefit from an ecologically based approach. Much work in amphibian behavioral endocrinology has historically had an ethological orientation, having been motivated by interests in animal communication and reproductive social behavior, so such a synthesis is a natural progression for this field. Moreover, the understanding of basic patterns in anurans and urodeles is now at the point where a new and interesting direction could be taken: using the natural variation found in amphibian species and their social and communication systems. Such studies would be interesting as tests of basic behavioral–endocrine relationships, and of the mechanisms that underlie them, and enrich the understanding of social evolution by providing an understanding of the physiological changes that support the diversification of animal social behavior.