Territorial aggression, circulating levels of testosterone, and brain aromatase activity in free-living pied flycatchers

Testosterone (T) is a critical endocrine factor for the activation of many aspects of reproductive behavior in vertebrates. Castration completely eliminates the display of aggressive and sexual behaviors that are restored to intact level by a treatment with exogenous T. There is usually a tight correlation between the temporal changes in plasma T and the frequency of reproductive behaviors during the annual cycle. In contrast, individual levels of behavioral activity are often not related to plasma T concentration at the peak of the reproductive season suggesting that T is available in quantities larger than necessary to activate behavior and that other factors limit the expression of behavior. There is some indication from work in rodents that individual levels of brain aromatase activity (AA) may be a key factor that limits the expression of aggressive behavior, and in agreement with this idea, many studies indicate that estrogens produced in the brain by the aromatization of T may contribute to the activation of reproductive behavior, including aggression. We investigated here in pied flycatcher (Ficedula hypoleuca) the relationships among territorial aggression, plasma T, and brain AA at the peak of the reproductive season. In a first experiment, blood samples were collected from unpaired males holding a primary territory and, 1 or 2 days later, their aggressive behavior was quantified during standardized simulated territorial intrusions. No relationship was found between individual differences in aggressive behavior and plasma T or dihydrotestosterone levels but a significant negative correlation was observed between number of attacks and plasma corticosterone. In a second experiment, aggressive behavior was measured during a simulated territorial intrusion in 22 unpaired males holding primary territories. They were then immediately captured and AA was measured in their anterior and posterior diencephalon and in the entire telencephalon. Five males that had attracted a female (who had started egg-laying) were also studied. The paired males were less aggressive and correlatively had a lower AA in the anterior diencephalon but not in the posterior diencephalon and telencephalon than the 22 birds holding a territory before arrival of a female. In these 22 birds, a significant correlation was observed between number of attacks/min displayed during the simulated territorial intrusion and AA in the anterior diencephalon but no correlation was found between these variables in the two other brain areas. Taken together, these data indicate that the level of aggression displayed by males defending their primary territory may be limited by the activity of the preoptic aromatase, but plasma T is not playing an important role in establishing individual differences in aggression. Alternatively, it is also possible that brain AA is rapidly affected by agonistic interactions and additional work should be carried out to determine whether the correlation observed between brain AA and aggressive behavior is the result of an effect of the enzyme on behavior or vice versa. In any case, the present data show that preoptic AA can change quite rapidly during the reproductive cycle (within a few days after arrival of the female) indicating that this enzymatic activity is able to regulate rapid behavioral transitions during the reproductive cycle in this species.     

Rhythms,rhythmicity and aggression

The relationships between biological rhythms and human aggressive behavior are addressed and discussed in this article: First, circadian rhythms and aggression are considered. Studies of sleep/waking cycle disturbances in aggression are reported. Severe aggression is associated with profound changes in sleep architecture. Causal link is difficult to establish given that sleep disturbance and aggressive behavior could be the symptoms of the same disorder. Specific aggressive behavior developed during sleep is also described. In addition, hormonal circadian rhythm studies are reported. Thus, low cortisol levels, in particular low cortisol variability, are associated with aggressive behavior, suggesting an inhibitory role of cortisol. Testosterone has daily and seasonal fluctuations, but no link with aggression has been established. Neurophysiological underlying mechanisms are discussed in the last part of this article, with a focus on the relationship between brain rhythm and aggression. Increase of slow-wave EEG activities is observed in individuals with aggressive behavior. Epilepsy, as a disease of brain rhythm could be associated with aggressive behavior, in pre, post and inter ictal periodes. Incidence of aggression is not likely more prevalent in epileptic individuals compared to those with other neurological conditions. Ictal changes take the form of profound behavioral changes, including aggressive behavior which has been interpreted as the emergence of “archeical” or innate motor patterns. In this multidisciplinary approach, the main difficulty is the categorization of the differents types of aggression. Finally, taken together, these studies suggest that biological rhythms, especially circadian rhythms, could provide therapeutic benefits to human aggressive behavior. Biological rhythymicity seems to be a necessary permanent training offering interesting perspectives for the adaptation to changes in the field of aggression.     

Testosterone, paternal behavior, and aggression in the monogamous California mouse (Peromyscus californicus)

Testosterone (T) mediates a trade-off, or negative correlation, between paternal behavior and aggression in several seasonally breeding avian species. However, the presence or absence of a T-mediated trade-off in mammals has received less attention. We examined the relationship between paternal behavior and territorial aggression in the biparental California mouse, Peromyscus californicus. In contrast to seasonally breeding birds, T maintains paternal behavior in this year-round territorial species. Castration reduced paternal behavior, whereas T replacement maintained high levels of paternal behavior. We hypothesize that T is aromatized in the brain to estradiol, which in turn stimulates paternal behavior. In contrast to paternal behavior, aggressive behavior was not reduced by castration. Interestingly, only sham males showed an increase in aggression across three aggression tests, while no change was detected in castrated or T-replacement males. Overall, trade-offs between aggression and paternal behavior do not appear to occur in this species. Measures of paternal behavior and aggression in a correlational experiment were actually positively correlated. Our data suggest that it may be worth reexamining the role that T plays in regulating mammalian paternal behavior.     

Variation in fecal testosterone levels, inter-male aggression, dominance rank and age during mating and post-mating periods in wild adult male ring-tailed lemurs (Lemur catta)

In primate species exhibiting seasonal reproduction, patterns of testosterone excretion in adult males are variable: in some species, peaks correlate with female receptivity periods and heightened male-male aggression over access to estrous females, in others, neither heightened aggression nor marked elevations in testosterone have been noted. In this study, we examined mean fecal testosterone ( f T) levels and intermale aggression in wild adult male ring-tailed lemurs residing in three groups at Beza Mahafaly Reserve, Madagascar. Results obtained from mating and post-mating season 2003 were compared to test Wingfield et al. [1990. Am Nat 136:829-846] "challenge hypothesis", which predicts a strong positive relationship between male testosterone levels and male-male competition for access to receptive females during breeding season. f T levels and rates of intermale aggression were significantly higher during mating season compared to the post-mating period. Mean f T levels and aggression rates were also higher in the first half of the mating season compared with the second half. Number of males in a group affected rates of intermale agonism, but not mean f T levels. The highest-ranking males in two of the groups exhibited higher mean f T levels than did lower-ranking males, and young males exhibited lower f T levels compared to prime-aged and old males. In the post-mating period, mean male f T levels did not differ between groups, nor were there rank or age effects. Thus, although male testosterone levels rose in relation to mating and heightened male-male aggression, f T levels fell to baseline breeding levels shortly after the early mating period, and to baseline non-breeding levels immediately after mating season had ended, offsetting the high cost of maintaining both high testosterone and high levels of male-male aggression in the early breeding period.     

The Dynamic of Aggression: How Individual and Group Factors Affect the Long‐Term Interspecific Aggression Between Two Sympatric Species of Dolphin

Interspecific aggression, similar to intergroup conspecific aggression, has been observed in a variety of taxa. The dominant group or individual is determined by multiple aggressive events and can be influenced by the size, age, or group size of the participating individuals. Interspecific aggression between Atlantic bottlenose (Tursiops truncatus) and spotted (Stenella frontalis) dolphins, both resident and sympatric to Little Bahama Bank, the Bahamas has been consistently observed for over two decades. However, it is unclear whether one species is more dominant and little is known about the factors that influence the progression of aggression. For this study, underwater video recordings of 32 aggressive encounters composed of 451 aggressive behavioural events were analysed over a 12‐yr period (1993–2004). These were used to describe the interspecific aggression observed and quantify which factors (the species and age class of the participants or the group size and behaviour of spotted dolphin groups) had the strongest impact on the progression and outcome of aggression. Over the long term, interspecific aggression was bidirectional with neither species being more dominant. During a single encounter, spotted dolphin group synchrony had the strongest impact on the dynamic of aggression, specifically impacting which group (1) initiated aggression, (2) the direction of aggression and (3) the occurrence of dynamic shifts or dominance reversals. This is the first study to quantify the dynamic of aggression for this population, to document bidirectional aggression and dynamic shifts during long‐term interspecific aggression in free‐ranging delphinids, and this study quantifies the role of synchrony during interspecific aggression using underwater observations.     

Sibling Age and Intragroup Aggression in Captive Saguinus midas midas

Different species of Callitrichidae show a consistent pattern in intragroup aggression. Aggression is more frequent between same-sex individuals, with males and older siblings being more aggressive. These findings have been associated with reproductive competition and dispersal events. However, the few studies dealing with intragroup aggression have not fully explored the effects of food availability on aggression or the relationship between age and aggression. I observed 3 family groups of Saguinus midas for 50 weeks and recorded all occurrences of aggression. A total of 138 agonistic events occurred at an average of 1.16 aggressive acts/hour per group, and 52.2% of aggressive acts occurred during a food shortage condition, which accounted for about 12% of all meals in the groups. During the food shortage, aggression was significantly higher, which suggests that food competition promoted an increase in aggression levels. Aggression amongst siblings accounted for 65.2% of all aggressive acts. Aggression from brothers to sisters was significantly more frequent than aggression between brothers. Aggression rates were significantly different between age classes. Young siblings were significantly more aggressive toward old ones than the reverse. Age and aggression received followed a linear relationship, while age and aggression initiated tended to follow a curvilinear relationship. Higher rates of aggression in young siblings probably were prompted by puberty and associated hormonal changes.     

Neurosteroids, GABAA receptors, and escalated aggressive behavior

Aggressive behavior can serve important adaptive functions in social species. However, if it exceeds the species-typical pattern, it may become maladaptive. Very high or escalated levels of aggressive behavior can be induced in laboratory rodents by pharmacological (alcohol-heightened aggression), environmental (social instigation), or behavioral (frustration-induced aggression) means. These various forms of escalated aggressive behavior may be useful in further elucidating the neurochemical control over aggression and violence. One neurochemical system most consistently linked with escalated aggression is the GABAergic system, in conjunction with other amines and peptides. Although direct stimulation of GABA receptors generally suppresses aggression, a number of studies have found that positive allosteric modulators of GABAA receptors can cause increases in aggressive behavior. For example, alcohol, benzodiazepines, and many neurosteroids are all positive modulators of the GABAA receptor and all can cause increased levels of aggressive behavior. These effects are dose-dependent and higher doses of these compounds generally shift from heightening aggressive behavior to being sedative and anti-aggressive. In addition, these modulators interact with each other and can have additive effects on the GABAA receptor and on behavior, including aggression. The GABAA receptor is a heteropentameric protein that can be constituted from various subunits. It has been shown that subunit composition can affect sensitivity of the receptor to some modulators and that subunit composition differentially affects the sedative vs anxiolytic actions of benzodiazepines. Initial studies targeting alpha subunits of the GABAA receptor point to their significant role in the aggression-heightening effects of alcohol, benzodiazepines, and neurosteroids.     

Castration does not decrease nonreproductive aggression in yearling male European starlings (Sturnus vulgaris)

In the nonbreeding season, some bird species express high levels of aggression despite basal plasma testosterone (T) concentrations. Consequently, nonreproductive aggression is believed to be independent of plasma T. In the present study, we investigated the effect of castration on nonreproductive aggressive behavior in yearling male European starlings (Sturnus vulgaris). We paired castrated (Cx) with control (C) males in dyadic trials during three test periods (December, January-February, and March-April), and by using an infrared camera, we defined which male was the most aggressive one when both males were competing over access to a nest box to roost in. During each of the three periods, Cx males behaved more aggressively than C males but differences between groups did not reach significance. When analyzing the results of the three periods together, Cx males were found to be significantly more aggressive than C males. Likewise, when considering only the second and third period (when plasma T levels of C males were significantly higher than those of Cx males) the same result was obtained. Furthermore, in the third period, aggression tended to be lower than in the first and second period, although T concentrations of C males were highest in this period. Our data thus clearly show that nonreproductive aggression in yearling male starlings is independent of gonadal sex steroids and suggest it even increases by castration.     

A quantitative study of serum testosterone,sex accessory organ growth,and the development of intermale aggression in the mouse

Radioimmunoassay of serum testosterone (T) was used to characterize circulating T levels in mice from birth to sexual maturity. Until 25 days of age, serum T levels ranged from 1 to 4 ng/ml. A significant increase in T concentrations was observed in 30-day-old males, followed by a secondary rise in serum T between Days 45 and 50 of life. The latter increment was associated with the appearance of extreme individual variation in circulating T levels which was also observed in adult (120 days) males. The most rapid growth of accessory sex organs occurred between 30 and 50 days of age, the period preceding attainment of peak serum-T levels. The first incidence of intermale aggression coincided with a prepubertal rise in circulating T, but adult levels of fighting were present prior to the secondary increase in T observed between 45 and 50 days of age. Although animals involved in a fight did not differ with respect to weight of the accessory sex organs or serum T concentrations, the male that weighed more than his opponent usually won an encounter. Compared to males in encounters in which no fighting occurred, animals that won or lost an aggressive encounter showed significantly greater accessory sex organ development. While circulating T is required for the initiation and maintenance of intermale aggression, it is apparent that additional factors are related to the onset of fighting and the establishment of dominance/ subordinance relationships in mice.     

What can animal research tell us about the link between androgens and social competition in humans?

A contribution to a special issue on Hormones and Human Competition.The relationship between androgenic hormones, like testosterone (T), and aggression is extensively studied in human populations. Yet, while this work has illuminated a variety of principals regarding the behavioral and phenotypic effects of T, it is also hindered by inherent limitations of performing research on people. In these instances, animal research can be used to gain further insight into the complex mechanisms by which T influences aggression. Here, we explore recent studies on T and aggression in numerous vertebrate species, although we focus primarily on males and on a New World rodent called the California mouse (Peromyscus californicus). This species is highly territorial and monogamous, resembling the modern human social disposition. We review (i) how baseline and dynamic T levels predict and/or impact aggressive behavior and disposition; (ii) how factors related to social and physical context influence T and aggression; (iii) the reinforcing or “rewarding” aspects of aggressive behavior; and (iv) the function of T on aggression before and during a combative encounter. Included are areas that may need further research. We argue that animal studies investigating these topics fill in gaps to help paint a more complete picture of how androgenic steroids drive the output of aggressive behavior in all animals, including humans.     

The effects of habitat complexity on aggression and fecundity in zebrafish (<Emphasis Type="Italic">Danio rerio</Emphasis>)

Female zebrafish housed in aquaria with spatial complexity (plastic plants) over a 13–16-week period showed reduced levels of aggressive behavior compared to females in bare tanks. In tanks with plants, there was no relationship between levels of aggression and fecundity but, in bare tanks, females experiencing the highest levels of aggression showed reduced fecundity. Our results suggest that it may be beneficial, when maintaining zebrafish at moderate to high densities or working with especially aggressive strains, to house them in spatially complex conditions.     

The progesterone challenge: steroid hormone changes following a simulated territorial intrusion in female Peromyscus californicus

There is a growing body of evidence that the rapid but transient increase in male androgens, particularly testosterone (T), following a single social encounter such as a territorial intrusion occurs in a wide array of vertebrate taxa. Yet, this phenomenon, often called the Challenge Hypothesis, has rarely been investigated in females. Moreover, when studying male challenge effects, researchers have rarely investigated other hormones that can be important to the expression of aggression, such as progesterone (P4) and estradiol (E2). We conducted 10-min aggression trials using the resident-intruder paradigm in cycling female California mice, Peromyscus californicus, a species in which both sexes show territorial behavior. By comparing the hormone levels of test females to control females, we found a decrease in P(4) and the P4/T ratio, but no change in T, E2, corticosterone, E2/P4, or E2/T. Interestingly, these hormone changes were observed even when the resident was not aggressive toward the intruder, suggesting that the stimulus cueing the hormone changes was the mere presence of the intruder and not the amount of aggression displayed by the resident. Generally, T has a positive relationship with aggression, whereas P4 inhibits male and nonmaternal female aggression. Thus, decreasing the P4/T ratio following an encounter may serve to increase future aggression in females. These results suggest that females may use different hormonal mechanisms than do males to mediate aggression in a challenge situation.     

Factors that Affect Aggression Among the Worker Caste of Reticulitermes spp. Subterranean Termites (Isoptera: Rhinotermitidae)

Aggression was observed among both inter- and intraspecific combinations of four colonies of Reticulitermes flavipes (Kollar) and Reticulitermes hageni Banks in laboratory assays each month for 4 consecutive months. Termites were most frequently aggressive toward colonies of a different species. Number of individuals that displayed aggression decreased over the study period, from April to July. There was a slight trend toward reduced aggressive behavior as termites were maintained in the laboratory for 3 months. Passive and aggressive individuals were identified and reexamined for display of aggressive or passive behavior toward nonnestmates. Eighty-nine percent of previously aggressive termites displayed aggression a second time. Eighty-eight percent of previously passive termites were passive upon reexamination. Differences in head capsule size between passive and aggressive individuals provided no correlation between the presence of aggressiveness and the head capsule size in the worker caste.     

Neurobiological correlates premeditated (learned) aggression: seeking new experimental approaches]

Theoretical possibility of experimental modeling of learned (premediated) aggression developing in human after experience of aggression is considered. The sensory contact technique increases aggressiveness in male mice and allows aggressive type of behavior to be formed as a result of repeated experience of victories in daily agonistic confrontations. Some behavioral domains confirm the development of learned aggression in males similar to those in humans. The features are: repeated experience of aggression reinforced by victories; elements of learned behavior after period of confrontations; intent, measured by increase of the aggressive motivation prior agonistic confrontation; decreased emotionality estimated by parameters of open field behavior. Relevant stimuli provoke demonstration of aggression. This review summarized data on the influence of positive fighting experience in daily intermale confrontations on the behavior, neurochemistry and physiology of aggressive mice (winners). This sort of experience changes many characteristics in individual and social behaviors, these having been estimated in different tests and in varied situations. Some physiological parameters are also changed in the winners. Neurochemical data confirm the activation of brain dopaminergic systems and functional inhibition of serotonergic system in winners under influence of repeated experience of aggression. The expression of the neurochemical and behavioral changes observed in winners has been found dependent on the mouse strain and on the duration of their agonistic confrontations. Similarities in mechanisms of learned aggression in humans and mice are considered.     

Dental anxiety in relation to emotional and behavioral problems in Croatian adolescents

The investigation was performed on 113 adolescents in the age between 15 and 18 years (63 boys, 50 girls). Corah Dental Anxiety Scale (CDAS) and Children's Fear Survey Schedule - Dental Subscale (CFSS-DS) were used for evaluation of dental fear and Child Medical Fear Questionnaire (CMFQ) for evaluation of the fear of medical treatment. Achenbach Youth Self Report questionnaire (YSR) was used for evaluation of emotional and behavioral problems. The tests were filled in by children. The aim of the study was to evaluate the level of dental anxiety in adolescents and to assess a cause--consequence relationship between dental anxiety and emotional and behavioral problems in adolescents. The results of CDAS, CFSS-DS and CMFQ tests showed that dental anxiety scores and the total internalizing problems were higher in girls. Girls displayed more physical problems (p < 0.001) and were more prone to anxiety/depression disorders (p < 0.05). Both boys and girls were more aggressive, more prone to delinquent behaviour and had more externalizing problems in comparison with the average values obtained for the Croatian population. Significant correlation coefficients for boys were calculated for age and anxiety/depression, and delinquent behaviour and aggression (p < 0.05). Significant correlations were observed between physical problems and dental anxiety measured by the CFSS-DS test (p < 0.01), and between physical problems and the total internalizing problems (p < 0.05). In girls, the CMFQ scores showed significant correlations between dental anxiety and physical problems (p < 0.05), and anxiety/ depression (p < 0.01) and the total internalizing and externalizing problems (p < 0.05). Significant correlations were calculated for age and the total internalizing and externalizing problems for boys (p < 0.05). According to the results of both CDAS and CMFQ tests, anxiety in girls showed significant correlations with delinquent behaviour (p < 0.01). CDAS scores for girls showed significant correlations with aggression (p < 0.05) and the total externalizing problems (p < 0.01).     

哺乳期棕色田鼠对配偶的识别记忆

观察了分娩后哺乳雌性棕色田鼠（Microtus mandarinus）与配偶分离7d、14d、21d后对配偶与陌生雄鼠的行为反应。结果发现,在频次和持续时间上,与配偶分离7d、14d、21d的雌鼠对配偶的攻击行为显著低于陌生雄鼠（P〈0．05）;对配偶的嗅闻及友善行为显著高于陌生雄鼠（P〈0．05）。处于哺乳前、中、后期不同阶段的雌鼠对陌生雄鼠攻击水平没有显著性差异（P〉0．05）。这些结果说明,在整个哺乳期,雌鼠能够识别配偶,对配偶的记忆不因分离时间的增加而减弱,而且对陌生鼠的攻击水平不会因哺乳阶段的变化而变化,维持高水平的攻击可能与其较强的母性行为有关。     

Social parameters and urinary testosterone level in male chimpanzees (Pan troglodytes)

Studies investigating relationships between social parameters (such as dominance rank, rates of aggressive and sexual behaviors) and androgen (particularly, testosterone) levels in male primates have yielded inconsistent results. In the present study, we address the relationship between androgens, male dominance rank and rank-associated behaviors in two groups of captive chimpanzees, a species characterized by a pronounced dominance hierarchy between adult males. By combining behavioral observations with urinary testosterone (T) measurements, we found that the differences in T concentrations between males were small and not obviously related to their dominance rank. T levels were not related to the rates of initiated aggression and copulatory behavior, but a significant negative relationship between male T level and the rates of strong aggression received was apparent. Our findings, combined with those of others, suggest that any relationship between dominance rank and T depends upon the extent to which individual rank-associated behaviors (e.g. aggressive/sexual) are themselves related to T.     

Variation in steroid hormone levels among Caribbean Anolis lizards: Endocrine system convergence?

Variation in aggression among species can be due to a number of proximate and ultimate factors, leading to patterns of divergent and convergent evolution of behavior among even closely related species. Caribbean Anolis lizards are well known for their convergence in microhabitat use and morphology, but they also display marked convergence in social behavior and patterns of aggression. We studied 18 Anolis species across six ecomorphs on four different Caribbean islands to test four main hypotheses. We hypothesized that species differences in aggression would be due to species differences in circulating testosterone (T), a steroid hormone implicated in numerous studies across vertebrate taxa as a primary determinant of social behavior; more aggressive species were expected to have higher baseline concentrations of T and corticosterone. We further hypothesized that low-T species would increase T and corticosterone levels during a social challenge. Within three of the four island assemblages studied we found differences in T levels among species within an island that differ in aggression, but in the opposite pattern than predicted: more aggressive species had lower baseline T than the least aggressive species. The fourth island, Puerto Rico, showed the pattern of baseline T levels among species we predicted. There were no patterns of corticosterone levels among species or ecomorphs. One of the two species tested increased T in response to a social challenge, but neither species elevated corticosterone. Our results suggest that it is possible for similarities in aggression among closely related species to evolve via different proximate mechanisms.     

Effects of a short period of isolation in adulthood on the aggressive behavior of dominant and subordinate male mice

Male ddY mice were used to investigate the effect of a short period of isolation in adulthood on aggressive behavior. The relationship between the dominance status previous to isolation and the effect of isolation was investigated. The mice were kept in isolation for 3 weeks from 9 weeks of age, during which intruder tests were conducted once a week. They the went through an encounter test, in which the mice encountered dominant or subordinate mice in a neutral space. The number of the formerly-dominant isolated mice that attacked the intruder mice decreased at first and then increased. The latency to attack also lengthened at first and then shortened. Seven former-dominants continued to show aggression throughout the intruder The number of the formerly-subordinate isolated mice that attacked the intruder mice increased linearly. But 3 former-subordinates did not show aggression through the entire experiment. After 3 weeks isolation, the number of mice that showed aggression and the amount of aggression did not differ between the former-dominants and subordinatcs. Isolation housing was concluded to differentially affect the dominant and subordinate mice during the 3 weeks of isolation. It was also concluded to differentially affect the mice of absolute dominance and relative dominance differentially. The aggressive behavior of the isolated mice appears to occur independently of site.     

Relationships between hormones and aggressive behavior in green anole lizards: an analysis using structural equation modeling

We investigated the relationship between aggressive behavior and circulating androgens in the context of agonistic social interaction and examined the effect of this interaction on the androgen-aggression relationship in response to a subsequent social challenge in male Anolis carolinensis lizards. Individuals comprising an aggressive encounter group were exposed to an aggressive conspecific male for 10 min per day during a 5-day encounter period, while controls were exposed to a neutral stimulus for the same period. On the sixth day, their responses to an intruder test were observed. At intervals, individuals were sacrificed to monitor plasma androgen levels. Structural equation modeling (SEM) was used to test three a priori interaction models of the relationship between social stimulus, aggressive behavior, and androgen. Model 1 posits that exposure to a social stimulus influences androgen and aggressive behavior independently. In Model 2, a social stimulus triggers aggressive behavior, which in turn increases circulating levels of androgen. In Model 3, exposure to a social stimulus influences circulating androgen levels, which in turn triggers aggressive behavior. During the 5 days of the encounter period, circulating testosterone (T) levels of the aggressive encounter group followed the same pattern as their aggressive behavioral responses, while the control group did not show significant changes in their aggressive behavior or T level. Our SEM results supported Model 2. A means analysis showed that during the intruder test, animals with 5 days of aggressive encounters showed more aggressive responses than did control animals, while their circulating androgen levels did not differ. This further supports Model 2, suggesting that an animal's own aggressive behavior may trigger increases in levels of plasma androgen.