Hormones and Behavior 58 (2010) 800-807 Contents lists available at ScienceDirect Hormones and Behavior j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y h b e h Short-term exposure to a synthetic estrogen disrupts mating dynamics in a pipe sh Charlyn Partridge a, , Anne Boettcher b, Adam G. Jones a a b Department of Biology, Texas A&M Univ., 3258 TAMU, College Station, TX 77843, USA Department of Biology, University of South Alabama, 124 LSCB, Mobile, AL 36618, USA a r t i c l e i n f o a b s t r a c t Sexual selection is responsible for the evolution of some of the most elaborate traits occurring in nature, many of which play a vital role in competition over access to mates and individual reproductive tness. Because expression of these traits is typically regulated by sex-steroids there is a signi cant potential for their expression to be affected by the presence of certain pollutants, such as endocrine disrupting compounds. Endocrine disruptors have been shown to alter primary sexual traits and impact reproduction, but few studies have investigated how these compounds affect secondary sexual trait expression and how that may, in turn, impact mating dynamics. In this study we examine how short-term exposure to a synthetic estrogen impacts secondary sexual trait expression and mating dynamics in the Gulf pipe sh, a species displaying sex-role reversal. Our results show that only 10 days of exposure to 17?-ethinylestradiol results in adult male pipe sh developing female-like secondary sexual traits. While these males are capable of reproduction, females discriminate against exposed males in mate choice trials. In natural populations, this type of discrimination would reduce male mating opportunities, thus potentially reducing their long-term reproductive success. Importantly, the effects of these compounds on mating dynamics and mating opportunity would not be observed using the current standard methods of assessing environmental contamination. However, disrupting these processes could have profound effects on the viability of exposed populations. (C) 2010 Elsevier Inc. All rights reserved. Article history: Received 22 January 2010 Revised 4 August 2010 Accepted 6 August 2010 Available online 11 August 2010 Keywords: Pipe sh 17?-ethinylestradiol EE2 Endocrine disruptors Sexual selection Introduction Sexual selection is a major evolutionary mechanism that often results in the evolution of secondary sex traits, which play a signi cant role in competition for access to mates. Mate choice mechanisms often lead to the development of elaborate ornaments that are used by individuals to assess mate quality, and experimental alterations of these traits can signi cantly impact mating success. While expression of a secondary sexual trait is normally limited to one sex, the genes in uencing their development are usually present in the genomes of both sexes (Andersson, 1994). Sex-speci c expression of these traits is often the result of the relevant genes being under the control of sexsteroid hormones, testosterone and estrogen (Andersson, 1994; Folstad and Karter, 1992; Parker et al., 2002; McGraw, 2006; McGraw et al., 2006). Because slight changes in hormone levels may affect secondary sexual trait expression without severely affecting primary sex traits, secondary sex traits may be more sensitive to small levels of environmental pollution, particularly when the pollutants are endocrine disrupting compounds. Corresponding author. Department of Anatomy and Neurobiology, Washington University in St. Louis, 660 S. Euclid Ave, St. Louis, MO 63110, USA. Fax: +1 314 362 3446. E-mail address: partridge@pcg.wustl.edu (C. Partridge). 0018-506X/$ - see front matter (C) 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.yhbeh.2010.08.002 Endocrine disruptors are compounds that interfere with normal hormone function and have been shown to greatly impact population viability by interfering with reproduction (Kidd et al., 2007). In addition, studies have shown that exposure to environmental pollutants can affect the development of secondary sexual traits, causing them to be suppressed (Arellano-Aguilar and Garcia, 2008) or expressed in the opposite sex (Ueda et al., 2005; Larsen et al., 2008). Most of the research concerning the impact of endocrine disruptors on population health has focused on their effects on primary sexual organs, particularly the gonads (Allen et al., 1999; Hill and Janz, 2003; Weber et al., 2003; Moncaut et al., 2003; Vandenbergh et al., 2003; Palace et al., 2006; Pettersson et al., 2006; Brown et al., 2008), and reproductive ability (Hill and Janz, 2003; Maunder et al., 2007; Peters et al., 2007; Sch?fers et al., 2007; Hashimoto et al., 2009). While the number of studies concerning the effect of these compounds on precopulatory mating behavior has increased (Bell, 2001; Bjerselius et al., 2001; Oshima et al., 2003; Robinson et al., 2003; Crews et al., 2007; Saaristo et al., 2009a,b), only a few studies have assessed how endocrine disruptors impact secondary sexual trait expression (Ueda et al., 2005; Arellano-Aguilar and Garcia, 2008; Larsen et al., 2008) and how that may, in turn, affect mate choice mechanisms (ArellanoAguilar and Garcia, 2008). This situation is particularly surprising since the ability of individuals to maintain secondary sexual traits signi cantly in uences their mating success. One potential reason for this gap in knowledge is that many of the toxicological model species C. Partridge et al. / Hormones and Behavior 58 (2010) 800-807 801 do not maintain obvious secondary sexual traits involved in mate choice. However, mate choice and sexual selection are key components of reproduction and can signi cantly in uence population viability, so there is a critical need to understand the effects of endocrine disruptors on these evolutionary processes. In this study, we examined how short-term exposure to environmentally relevant concentrations of a synthetic estrogen affected multiple aspects of reproductive tness, including mating opportunity, mating success and reproductive success, in a sex-role reversed pipe sh characterized by strong mating preferences and sexual selection. One endocrine disruptor that has received a large amount of attention is 17?-ethinylestradiol (EE2). EE2 is a synthetic estrogen that is one of the major components of oral contraceptives. A high resistance of EE2 to degradation in the human body is one feature that makes it useful in contraceptives. However, this characteristic also allows EE2 to pass into the environment through domestic wastewater (Hill and Janz, 2003; Lintelmann et al., 2003), and high concentrations of EE2 have been found in wastewater ef uent and rivers within the United States and Europe (Kolpin et al., 2002; Ying et al., 2002; Clouzot et al., 2008). Of the vast array of potential endocrine disrupting chemicals, EE2 is one of the most troubling because it binds to estrogen receptors with high af nity and, compared to other estrogenic compounds, is relatively stable in the environment. The amount of EE2 in wastewater treatment plant ef uent is variable depending upon season and effectiveness of water treatment but can range from 0.1 to 10 ng/L (Kolpin et al., 2002; Clouzot et al., 2008; Vajda et al., 2008). Surface waters around treatment plants commonly show EE2 concentrations of 0-5 ng/L (Clouzot et al., 2008; Vajda et al., 2008). However, a study by Kolpin et al. (2002), which evaluated EE2 concentrations in 139 U.S. contaminated rivers, found maximum concentrations of EE2 to be 820 ng/L, and concentrations approaching 35 ng/L in Europe have been reported (Pojana et al., 2007). These levels are of signi cant concern considering that exposure to 0.5 ng/L EE2 can induce vitellogenin production in males of some species (Nash et al., 2004), and a whole-lake experiment showed that 5-6 ng/ L EE2 caused a population of fathead minnows to collapse after only two seasons of exposure (Kidd et al., 2007). This sort of population collapse could result from effects of EE2 on gonads or other primary sexual traits (Bell, 2001; Bjerselius et al., 2001; Robinson et al., 2003; Balch et al., 2004; Nash et al., 2004; Maunder et al., 2007) or by disrupting mating patterns so severely that population viability is reduced. To date, only a few studies have evaluated how EE2 impacts mechanisms of sexual selection and mate choice (Arellano-Aguilar and Garcia, 2008; Coe et al., 2008; Saaristo et al., 2009a,b), so our goal was to study EE2 exposure in a system characterized by ritualistic courtship behaviors and strong sexual selection. The Gulf pipe sh, Syngnathus scovelli, serves as an interesting model system for this type of study for a number of reasons. First, this species is sex-role reversed, in that sexual selection acts more strongly on females than on males (Jones and Avise, 1997; Jones et al., 2001). Theoretically, sex-role reversal can evolve due to a number of factors such as higher potential reproductive rates in females (Clutton-Brock and Vincent, 1991; Clutton-Brock and Parker, 1992; Ahnesj? et al., 2001), female-biased operational sex ratios (Emlen and Oring, 1977; Kvarnemo and Ahnesj?, 1996), higher investment in parental care by males relative to females (Trivers, 1972), or costs of breeding (Kokko and Monaghan, 2001; Simmons and Kvarnemo, 2006). Regardless of the reason for this reversal, availability of males limits female reproductive tness and thus leads to an increase in competition between females over access to males. In the Gulf pipe sh, females transfer unfertilized eggs into a specialized brood pouch on the ventral surface of the male during mating. The male fertilizes the eggs by releasing sperm into the pouch, and then carries the developing embryos for approximately two weeks until they are born as independent juveniles. Gulf pipe sh are polyandrous in that males typically receive eggs from one female per "male pregnancy", but females are capable of lling multiple pouches (Jones and Avise, 1997; Jones et al., 2001). Consequently, males are a limiting resource for reproduction, and females compete for access to males, resulting in strong sexual selection on females and weak sexual selection on males (Jones and Avise, 1997; Jones et al., 2001), a reversal of the usual direction of sexual selection seen in most vertebrates, including teleost shes. As a result of this selection on females, secondary sexual traits have evolved in females rather than males. Adult females are larger than males and possess a number of sex-speci c characteristics that develop during maturation, such as a deeply keeled abdomen, intense silvery-blue lateral stripes, a large dorsal n, and striking temporary breeding coloration. Interestingly, while many studies examining hormonal regulation of secondary sex traits in species with typical sex roles have found that testosterone concentrations in uence ornament expression, a study by Ueda et al. (2005) found that male Gulf pipe sh exposed to high concentrations of EE2 (100 ng/L) developed the silvery-blue lateral strips that are normally only present in females. These data suggest that expression of secondary sexual traits in this species may be affected by estrogen exposure, which consequently may impact mating dynamics. Examining how estrogen exposure impacts reproductive success in the Gulf pipe sh is also interesting in light of the fact that pipe sh may not possess the "typical" sex steroid pro les found in many other teleosts. Sex steroid pro les from two Syngnathus species (Syngnathus acus and Syngnathus typhle) have shown that both breeding and brooding males have higher plasma testosterone (T) levels compared to 11-ketotestosterone (KT) (Mayer et al., 1993), which is the opposite of what is found for many other teleosts. In addition, these males also contain measurable levels of 17ss-estradiol (E2) during both the breeding and brooding period (Mayer et al., 1993). Thus, these differences in steroid pro les may result in differences in the effect of endocrine disruptors on pipe sh mating and reproductive success when compared to studies from other teleost species. The questions we aim to address in this study are: (1) Do environmentally relevant concentrations of EE2 produce males that are feminized with respect to secondary sexual traits? And (2) are these effects of suf cient magnitude to disrupt mating patterns and sexual selection? We addressed these questions by examining the effects of EE2 in several different experiments, including (1) a test of the effects of EE2 exposure on male morphology, (2) mate choice trials involving males and females choosing between exposed and unexposed individuals of the opposite sex, and (3) measurement of the effects of EE2 exposure on the ability of males to mate and to carry their pregnancy to term. Our work provides one of the most comprehensive studies of the effects of an endocrine disrupting compound on reproductive behavior and the only investigation of the effects of such contaminants on a sex-role reversed organism. Methods Pipe sh collection and maintenance A hand seine was used to collect pipe sh from submerged vegetation at the northern end of Mobile Bay in Meaher State Park, Baldwin County, Alabama, USA (30?66624, 87?92731). The breeding period of the Gulf pipe sh in Mobile Bay typically lasts from May to December, with a peak between August and October (Bolland and Boettcher, 2005). For this study, pipe sh were collected between June and November in the years 2006 through 2008 and all exposure experiments were performed between the months of July and November, to ensure that individuals were reproductively viable. Males and females were housed separately in 30 L aquaria connected to a recirculating ltration system. Pipe sh were fed a diet of Artemia nauplii twice daily and their diets were supplemented with live copepods biweekly. All pregnant males were allowed to give birth prior to exposure to control for male reproductive state. All 802 C. Partridge et al. / Hormones and Behavior 58 (2010) 800-807 procedures performed for this study were conducted in accordance with both the University of South Alabama and Texas A&M University IACUC regulations. Mate choice of exposed versus non-exposed individuals Exposure Control (0 ng/L EE2), 1 ng/L EE2 and 100 ng/L EE2 exposure tanks were maintained on separate recirculating systems. These three concentrations were chosen to simulate an EE2-free environment (Control), an environment potentially down-stream from a wastewater treatment plant (1 ng/L EE2) and a highly contaminated area (100 ng/L EE2). The control system consisted of fteen 30 L aquaria, while the two treatment systems consisted of nine 30 L aquaria each. All tanks were divided into two 15 L compartments by a perforated barrier. The control system was initially spiked with 510 L of 95% ethanol, the solvent for the EE2 solution, to account for the 450 L of water within the tanks and a 60 L sump for a total of 510 L. Treatment systems were initially spiked with 330 L of either 1 mg/L EE2 (for 1 ng/L system) or 100 mg/L EE2 (for 100 ng/L system) to account for the 270 L of water within the tanks and 60 L sumps for a total of 330 L per treatment system. Prior to exposure, same sex individuals (for both males and females) were size and color matched and then paired. Size and color matching was performed in order to control for other confounding factors that may in uence mate choice. One of the individuals of the pair was assigned to a 15-L compartment in the control system, while the other individual was randomly assigned to a 15-L compartment in either the 1 ng/L or 100 ng/L EE2 system. At any given point during the experiment, tanks within each system contained at most two individuals (i.e., one individual per 15 L compartment). Males and female were housed on separate shelves within a system so that biases in visual cues among individuals would not occur during the exposure time. All individuals were exposed for a total of 10 days. We performed 2% water changes (including appropriate amounts of either 95% ethanol or EE2 in 95% ethanol for control and exposure tanks, respectively) and cleaned the tanks daily. While the actual concentration of EE2 within these systems was not measured, preliminary studies indicated that this regimen of daily water changes maintained relatively constant EE2 concentrations over time. Body morphology Before exposure, we measured the standard length of each male. We also used a Nikon Coolpix 5000 digital camera to photograph each male in lateral view against a size standard. From the photographs, we used the computer program Image J (NIH, Bethesda, MD, www. rsbweb.nih.gov/ij/) to measure maximum body depth. We then calculated the depth:length ratio by dividing the maximum body depth by the standard length. We measured each individual a second time at the conclusion of each exposure period and mate choice trial. To quantify the effects of EE2 on male coloration, the images taken of the lateral view of each male both prior to and after exposure (see above) were analyzed using Image J. The red color was removed from all pictures in order to obtain images that provided the best color differentiation to detect changes in the banding pattern. The maximum and mean color intensity was then obtained for an area approximately 1.3 cm long and 0.02 cm high along the lateral side of the male, where the iridescent stripes develop. The maximum color intensity, de ned as the value of the brightest pixel within that area, was then divided by the mean color intensity of the area. Because the areas within the iridescent bands produce the highest color intensity, we were able to assess whether areas of iridescence developed in exposed males by evaluating the relationship between maximum color intensity and mean body color intensity along the area where these iridescent stripes develop. Mate preference trials All preference trials were performed in the morning 10 minutes after the beginning of the 12 hour daylight cycle. We used either a Cannon Optura 500 MiniDV or a Sony HDV 1080i MiniDV video camera to videotape all mate choice trials for 1 hour. Behaviors were analyzed using JWatcher version 1.0 animal behavior software (University of California Los Angeles, Los Angeles, CA: http://www.jwatcher.ucla.edu/). Male preference The effect of each treatment on male preference was determined by using a mate-choice design in which the focal male had the ability to assess control and exposed females but the females could not interact with each other (Fig. 1). Each mate-choice tank was divided lengthwise by a transparent barrier. One of the resulting long chambers was divided in half by an opaque divider orthogonal to the transparent barrier. Two females were placed in the tank, with an exposed female in one small chamber (25 W x 28 H x 12.5 D cm) and a control female in the other small chamber (25 W x 28 H x 12.5 D cm). The male was placed into the longer compartment (50 W x 28 H x 12.5 D cm), from which he could see into both the smaller compartments. The females, separated by the opaque divider, could not see each other but they could observe the male. The proportions of time the male spent performing the following behaviors were examined: (1) swimming near the divider close to a female (de ned as the male swimming in a zigzag pattern up and down near the divider), and (2) dancing (de ned as an up and down bobbing type movement with the body in a strict vertical position). In addition, the number of twitches (de ned as abrupt shaking of the individual lasting for less than a second) performed by the focal male toward each female was recorded. For a quantitative measure of mate choice, we lumped the male behaviors of swimming near the divider and dancing into a more inclusive category termed "male response time." We also recorded the proportion of time each focal male spent on the side of the tank with each female. Finally, we recorded and analyzed the behavior of each female (see below) to test for effects of exposure to EE2 on female courtship behavior. Female preference The effect of each treatment on female mate choice was determined by evaluating the relative proportion of time females spent actively courting each male and the number of twitches females Fig. 1. Diagram of mate choice tank. For male choice trials the focal male was placed in the front compartment while choice females were each placed within the separate back compartments. The focal male was separated from choice females by a clear perforated plexi-glass divider to allow for both visual and chemical cues to be detected. Choice females were separated by an opaque plexi-glass divider to prevent visual contact between the individuals. For female choice trials the mate choice design was similar except that the focal female was placed within the front compartment while the choice males were each placed in the separate back compartments. C. Partridge et al. / Hormones and Behavior 58 (2010) 800-807 803 performed toward each male. We used the same experimental setup as in the male-choice trials (see above). Thus, each trial involved a focal female assessing two males, one of which had been exposed to EE2 and one of which was a control male. The female was able to visually assess both males and males were able to assess the focal female; however, males were separated from one another by an opaque divider. The proportions of time the female spent performing the following behaviors toward each male were recorded: (1) color display (de ned as the female displaying a darkened, erect dorsal n and high contrast transverse stripes along the body), (2) posing (de ned as the female moving sharply into a vertical position and assuming a distinct posture with the abdomen protruding), and (3) dancing (de ned above). We also recorded the number of twitches (de ned above) performed toward each male. In S. scovelli, the female color display appears to be a static mating behavior, possibly indicating female receptivity, as this behavior also occurs in the absence of males. However, posing and dancing are behaviors that females actively perform toward other individuals. Thus, the proportion of time females spent posing and dancing were grouped to represent active female courtship behavior. Finally, we recorded the proportion of time focal females spent on the side of the tank with each male. Mating success and reproductive success Mating and reproductive success were evaluated two different ways. For years 2006 and 2007, after each mating trial, control and exposed males were placed individually in 15 L compartments to which a female that had not been involved in a mating trial was added. We checked males each morning to determine whether or not they were pregnant. Once impregnation occurred, we removed the female. Pregnant males were allowed to brood their offspring, and we recorded the amount of time that elapsed from impregnation until birth. In 2008, we exposed ten non-pregnant males each to control water (spiked with the appropriate amount of 95% ethanol), 1 ng/L EE2 water, and 100 ng/L EE2 water. After ten days, males were moved to clean tanks containing females. Given that females can impregnate multiple males (Jones et al., 2001), we housed two females with ve males in each tank. Males were checked twice daily for pregnancy. Pregnant males were removed and maintained individually in 15 L aquarium compartments, where we monitored the progression of their pregnancies. After three days, females from each tank were rotated to ensure that any male's failure to become pregnant was not attributable to the females with which he was housed. Pregnant males nearing parturition, which occurs after 12-15 days of brooding, were moved to individual 2 L birthing chambers. For each male, we counted the number of offspring born as a measure of his reproductive success. Statistics For our statistical analysis, we rst transformed our data by using arcsine (square root) transformations on all proportion data and square-root transformations on the number of twitches females performed to each male, the number of twitches males performed to each female and the maximum/mean color intensity of males. A repeated measure analysis of variance (ANOVA) showed that male depth:length ratios signi cantly differed between pre-and postexposure periods and that a signi cant interaction existed between EE2 concentration and change in maximum/mean color intensity. In order to determine how each concentration affected these two traits, we compared male body depth:length ratio and maximum/mean color intensity between exposed and unexposed males prior to and after exposure by using paired t-tests. For the behavioral analysis, we initially used a multivariate repeated-measures ANOVA to examine whether focal individuals behaved differently to control and exposed individuals and whether these differences were affected by EE2 concentration. To examine which speci c behaviors differed in response to control and treated individuals, t-tests of the mean were used to determine if the proportion of time focal females spent with each male, the relative proportion of time focal females actively courted males, the proportion of time focal males spent with each female and the relative proportion of time focal males responded to a female differed from a random expectation of 0.5. In the male mate choice experiments (two females with one male per trial), we used paired t-tests to compare control and experimental females with respect to: (1) number of twitches performed, (2) extent of color display, and (3) amount of active courtship. In this case, paired t-tests were used since females within a trial were responding to the same focal male and were therefore not independent. A Wilcoxon signed ranks test was used to determine if the difference in number of twitches control and exposed females performed in male choice tests differed from 0. A chi-square analysis of a 3 x 2 contingency table was used to examine whether or not a male's ability to become pregnant was dependent upon exposure. Finally, we used an ANOVA to compare the mean number of offspring produced by males subjected to different treatments. Results With respect to morphology, we found that short-term exposure to EE2 had a signi cant effect on male phenotype. After only 10 days of exposure, males from exposure treatments showed a signi cant change in body depth:length ratios (ANOVA: F1, 35 = 58.57, p b 0.001) with the body depth:length ratio of exposed males becoming signi cantly greater than their control partners (Table 1). This is particularly interesting considering in natural populations reproductively active female pipe sh have a larger body depth (measured at the thickest part of the abdomen) than males. In addition to body shape, exposure to EE2 may also effect male body coloration. While there was no overall effect of exposure on changes in color intensity (ANOVA: F1, 39 = 2.64, p = 0.07), a signi cant interaction between EE2 concentration and changes in color intensity was observed (ANOVA; treatment*color intensity: F2, 39 = 6.68, p = 0.003). Exposure to 100 ng/L EE2 caused males to develop iridescent lateral stripes, which are normally only found in female pipe sh (Fig. 2) and differences in the maximum/mean color intensity along the banding area were observed after exposure (Table 1). Some of the 1 ng/L EE2 exposed males also developed these female-like iridescent stripes, although the expression was not as pronounced as in males exposed to the higher concentration (Fig. 2). In addition, while no signi cant difference in maximum/mean color intensity was observed between control and 1 ng/L EE2 males after exposure, 1 ng/L EE2 exposed males tended to have areas within the banding area that were brighter compared to control males (Table 1). Given that EE2 exposure affects male morphology, the next question was whether or not such changes potentially impact mating dynamics in this species. Our results clearly show that male attractiveness is affected even by low doses of EE2, whereas female mating behavior and attractiveness are largely unaltered. In female choice trials, where focal females were able to choose between control and exposed males, multivariate repeated-measures ANOVA showed that focal female behavior signi cantly differed in response to control and exposed males (F3, 84 = 32.38, p b 0.001) and that the concentration of EE2 to which males were exposed had a signi cant effect on female behavior (F1, 84 = 11.59, p b 0.001). More speci cally, we found that focal females spent more time actively courting (posing and dancing; Table 2) and performed more twitches (Table 2) toward control males compared to males exposed to either 1 ng/L or 100 ng/L of EE2. Females did not differ in the proportion of time they spent in their color display between control and treatment males (Table 2). However, female color display is likely a general indicator of female receptivity, and not necessarily indicative of choice, since females will 804 Table 1 The effects of EE2 exposure on male body morphology. Control male Pre BD:length ratio Post BD:length ratio Pre max/mean CI Post max/mean CI 0.039 ? 0.002 0.038 ? 0.004 2.33 ? 0.14 2.01 ? 0.15 C. Partridge et al. / Hormones and Behavior 58 (2010) 800-807 1 ng/L EE2 male 0.038 ? 0.002 0.041 ? 0.003 2.37 ? 1.20 2.68 ? 0.35 N 9 9 11 11 Paired t-test |t|8 = 0.82 |t|8 = 3.06 |t|10 = 0.13 |t|10 = 2.01 p-Value 0.38 0.01 0.9 0.07 Control male 0.042 ? 0.001 0.042 ? 0.001 2.56 ? 0.21 2.30 ? 0.19 100 ng/L EE2 male 0.042 ? 0.001 0.050 ? 0.001 2.47 ? 0.24 3.28 ? 0.28 N 10 10 10 10 Paired t-test |t|9 = 0.189 |t|9 = 28.28 |t|9 = 0.25 |t|9 = 2.88 p-Value 0.85 0.0003 0.78 0.02 Body depth (BD):length was calculated by dividing maximum body depth by standard length. Maximum/mean color intensity (CI) is the maximum color intensity divided by the mean color intensity of a 1.3 cm long x 0.2 cm high area along the lateral side of the males where the iridescent stripes develop. Initially, repeated measure ANOVA showed that exposure affected both body depth:length ratios and maximum/mean color intensity (see text). Paired t-test were then used to compare BD:length and maximum/mean CI between control and experimental males before (Pre) and after exposure (Post) for both dosages. All data are reported as mean ? S.E of untransformed data. invoke their color display even in the absence of males. Therefore, based upon these results, EE2 exposure clearly affects male mating opportunity by altering their attractiveness to females, as might be expected from the feminizing effects of this compound. In contrast to our results for males, we found no detectable effects of EE2 exposure on female attractiveness. Male pipe sh did not signi cantly differ in their response between control females and EE2 exposed females from either the 1 ng/L or 100 ng/L EE2 treatments (Table 3). Likewise, female mating behavior did not differ signi cantly between the two treatments (Table 3). In male mate choice tests with non-exposed females, female behavior appears to be a major factor in uencing male response time (unpublished data), and thus the lack of an effect of EE2 on female courtship behavior likely explains why exposure did not affect male mate preference. In our nal experiment, we addressed the question of whether or not exposure to EE2 affects mating ability and male pregnancy. When females were given no choice (see methods), they mated as successfully with males exposed to 1 ng/L EE2 (N = 18; number impregnated = 9) as with control males (N = 33, number impregnated = 19). However, at some point exposure to EE2 limits the ef cacy of mating, as we found that males exposed to 100 ng/L of EE2 were signi cantly less likely to become pregnant (N = 19; number impregnated = 4) than control males (for all three treatments, contingency-?2 = 6.66; p = 0.04). Of the 100 ng/L EE2 males that mated, these mating events did not occur until at least four days after removal from the exposure tanks, suggesting that there may be a lag time after removal from treated water during which males are not able to mate. For those males that became pregnant, exposure to EE2 did not affect the number of offspring in their broods (Control: 13.9? 4.4 [mean ? S.E.], n = 8; 1 ng/L: 16.2 ? 3.8, n = 8; 100 ng/L: 16.3 ? 5.2, n = 4, F2,19 = 0.10, p = 0.9). Development of the offspring also appeared to be unaffected, as males that mated within 10 days after removal from treated water produced successful broods with completely developed juveniles, regardless of treatment. It is Fig. 2. Induction of secondary sexual traits in male pipe sh after exposure to EE2. (A) Control male prior to exposure, (B) Control male after 10 days of exposure to 0 ng/L EE2, (C) Treatment male after 10 days of exposure to 1 ng/L EE2, (D) Treatment male after 10 days of exposure to 100 ng/L EE2, (E) Non-exposed female. The black arrows point to the iridescent stripes, a secondary sex trait normally con ned to females. The green bars at the ends of each photograph indicate a spacing of 1.5 cm. This gure shows that exposure of males to EE2, especially the 100 ng/L concentration, induces the development of female-like secondary sexual traits. Thus, EE2 feminizes the males with respect to external morphology. C. Partridge et al. / Hormones and Behavior 58 (2010) 800-807 Table 2 Effects of EE2 exposure on female mate choice. Control vs 1 ng/L EE2 Male Female Behavior Prop of time with male Rel prop of time in color display Rel prop of time active court male Number of twitches to males Control Male 0.58 ? 0.06 0.57 ? 0.06 0.65 ? 0.06 12.1 ? 3.1 1 ng/L Male 0.42 ? 0.06 0.43 ? 0.06 0.35 ? 0.06 3.7 ? 1.2 N 11 11 11 11 t-Test |t|10 = 1.38 |t|10 = 1.19 |t|10 = 2.49 |t|10 = 3.03 p-Value 0.2 0.26 0.03 0.01 Control vs 100 ng/L EE2 Male Control male 0.63 ? 0.06 0.60 ? 0.08 0.80 ? 0.05 7.4 ? 1.7 100 ng/L male 0.37 ? 0.06 0.40 ? 0.08 0.20 ? 0.05 1.3 ? 0.5 N 12 12 12 12 t-Test |t|11 = 2.02 |t|11 = 1.38 |t|11 = 5.04 |t|11 = 4.46 805 p-Value 0.07 0.2 0.0004 0.001 The experiment consisted of a female choosing between two males, one of which had been exposed to EE2 and one of which had not. Female behaviors included the proportion of time female spent on the same side of the tank as each male, relative proportion of time female was in color display in front of each male, relative proportion of time female actively courted each male and the number of twitches females performed to each male. Multivariate repeated measures ANOVA showed that female behavior differed toward control and exposed males and the concentration to which males were exposed also in uenced female behavior (see text). To evaluate which speci c female behaviors differed, t-tests of the mean were used to determine if the proportion of female behavior toward a speci c male differed from an expected value of 0.05. All data are reported as mean ? S.E. of untransformed data. important to note, however, that during the period of brood development all males were housed in EE2-free water. Discussion Overall, our results lead to two important conclusions regarding the impact of estrogen mimics on secondary sexual trait expression and mating dynamics in the Gulf pipe sh. First, our data suggests that expression of the iridescent bars present in female S. scovelli appear to be affected by estrogenic compounds. Males exposed to EE2 developed iridescent bars and expression appears dose dependent. Second, it suggests that short-term exposure to even low levels of EE2 can signi cantly impact male reproductive tness. Male attractiveness was signi cantly affected by exposure to 1 ng/L EE2, even though these males could still become pregnant and successfully carry broods to term. The results were even more pronounced when males were exposed to higher doses of EE2. For example, males exposed to 100 ng/L of EE2 were not only less likely to attract females than control males, but they also experienced dif culty becoming pregnant. These effects occurred after only 10 days of exposure to EE2, and some of the morphological changes persisted well after the males were removed from the contaminated water (at least 22 days). While many studies have addressed how endocrine disruptors impact reproductive success, very few studies have addressed how these compounds impact secondary sexual trait expression, even though, in many cases, trait expression and reproductive success are correlated. Of the secondary sexual traits that have been studied, androgens (namely testosterone) have been shown to in uence trait expression in both males and females (Rand, 1992; Eens et al., 2000; Parker et al., 2002; McGraw, 2006). However, in the present study, we found that exposure to EE2 signi cantly impacted trait expression in male pipe sh, with exposed males developing secondary sexual traits Table 3 The effects of EE2 exposure on male mate choice and female courtship behavior. Control vs 1 ng/L EE2 Male Male Behavior Prop of time with female Rel prop of time responding to female Number of twitches to females Female behavior Prop of time in color display Prop of time actively courting Number of female twitches Control Female 0.48 ? 0.12 0.51 ? 0.13 1.3 ? 0.06 Control female 0.22 ? 0.12 0.06 ? 0.09 2.0 ? 1.4 1 ng/L Female 0.52 ? 0.12 0.49 ? 0.13 4.3 ? 1.4 1 ng/L female 0.34 ? 0.12 0.15 ? 0.13 6.2 ? 3.6 N 10 10 7 N 8 8 9 t-test normally restricted to females. While previous studies have shown that some male pipe sh species (S. acus and S. typhle) naturally maintain detectable levels of 17-ss estradiol (E2) during the breeding and brooding period (Mayer et al., 1993), neither males nor females of these species possess the permanent iridescent stripes that are present in S. scovelli females and induced with EE2 exposure in males. In S. scovelli females, the initial expression of these stripes occurs during female maturation, also suggesting that estrogen levels may play a role in their development. Thus, this observation suggests that, if male S. scovelli follow the same sex steroid pro le as other Syngnathus species, a minimum concentration of estrogen may be needed to actually induce expression of these iridescent stripes. However, whether these traits are solely estrogen dependent is unclear since females will continue to express this trait even during the non-breeding cycle and the stripes in males persisted after removal from treated water. Given that exposure to EE2 signi cantly impacted secondary sexual trait expression in males, the next question was whether or not this effect in uenced mate choice mechanisms in this species. Clearly, our data show that exposure to EE2 signi cantly impacted male mating opportunity, since females discriminated against exposed males from both 1 ng/L and 100 ng/L EE2 treatments. So, why does EE2 decrease male attractiveness? One possible hypothesis is that the induction of secondary sexual traits in males may mimic femalespeci c visual cues, causing females to incorrectly identify exposed males as other females. Alternatively, EE2 exposure may cause males to release chemicals that contain components that typically are released only by females. However, female pipe sh do not readily discriminate between males and females based on chemical cues alone (Ratterman et al., 2009), suggesting that our results are more likely related to visual cues. Males exposed to high doses of EE2 could experience high mating failure rates for a number of reasons. First, as Control vs 100 ng/L EE2 Male p-Value 0.8 0.92 0.22 p-Value 0.54 0.27 0.4 Control Female 0.40 ? 0.09 0.49 ? 0.14 6.1 ? 3.2 Control Female 0.25 ? 0.09 0.05 ? .03 6.6 ? 3.7 100 ng/L Female 0.60 ? 0.09 0.51 ? 0.14 1.2 ? 0.7 100 ng/L Female 0.11 ? 0.06 0.01 ? 0.009 1.3 ? 1.2 N 9 8 6 N 9 9 9 t-Test |t|8 = 0.98 |t|7 = 0.03 |t|5 = 1.05 Paired test |t|8 = 1.04 |t|8 = 0.86 |z| = 1.08 p-Value 0.36 0.98 0.33 p-Value 0.33 0.41 0.28 |t|9 = 0.27 |t|9 = 0.11 |t|6 = 1.36 Paired test |t|7 = 0.65 |t|7 = 1.20 |z| = 0.841 These data are from an experiment in which a single focal male could choose between two females, one of which had been exposed to EE2 and one of which had not. We measured both male and female behavior in these experiments. Male behavior included the proportion of time males spent on the same side of the tank as female, the relative proportion of time males spent responding to each female and the number of twitches males preformed to each female. Female courtship behaviors included the proportion of time females were in color display, the proportion of time females actively courted males and the number of twitches performed by females. Transformation of the number of twitches performed by females was unable to satisfy the assumption of normality, so a Wilcoxon signed rank test was performed. All data are reported as mean ? S.E. of untransformed data. 806 C. Partridge et al. / Hormones and Behavior 58 (2010) 800-807 was just described, females may avoid mating attempts with males exposed to 100 ng/L EE2 because of the physical similarities between exposed males and female pipe sh. Second, exposure to high levels of EE2 also appears to affect the male's brood pouch, where eggs are received, fertilized and develop. Boisseau (1967) found that the brood pouch of male seahorses (Hippocampus sp.) is under testicular control and that male castration signi cantly retards pouch development. This effect can be reversed by testosterone injections. The brood pouches of male Gulf pipe sh, which are close relatives of seahorses, are likely under similar hormonal regulation. Males exposed to high concentrations of EE2 had brood pouches that appeared thin and resembled the pouches of non-reproductive males. Since testosterone is known to be involved in maintaining the brood pouch, it is likely that EE2 exposure interferes with testosterone levels, preventing the male brood pouch from remaining in a reproductive state. Finally, plasma vitellogenin levels of male pipe sh exposed to 100 ng/L EE2 have been shown to be 10 times higher than those of eld collected females and exposure is known to affect male gonadal tissue (Ueda et al., 2005), which would also impact the reproductive ability of males exposed to this high concentration. After the male is removed from treated water, it may take a few days to completely eliminate EE2 from the body and allow the brood pouch and testes to become functional again, possibly explaining why males in the high EE2 exposure treatment were unable to mate for four days after removal from the exposure treatment. While our study solely focused on how EE2 exposure prior to pregnancy impacted male reproductive success, in natural populations pregnant males also would be exposed. Thus, future studies should not only examine how exposure impacts adult reproductive tness, but should also examine whether continuous exposure to low doses of estrogenic compounds impacts offspring health and viability. While exposure to EE2 clearly affects mating dynamics in this species, one additional troubling implication of our observations is that the effects of this compound may not be restricted to contaminated sites. Some of the female secondary sexual characteristics that are induced in males upon exposure are maintained for some time after removal from treated water. We found that males continued to express the iridescent bars that are normally found on females for at least 19-22 days after removal from treated water (Fig. 3). After this point males were sacri ced, so the amount of time exposed males continue to express altered traits is currently unknown. Thus, males migrating from contaminated sites to relatively pristine locales have the potential to carry the phenotypic effects of EE2, and presumably the effects on mating dynamics, with them to the new site. While our study shows that exposure to EE2 can impact mating dynamics, the effects of exposure at the population level is an equally important issue. The exact effects are dif cult to predict and would be worthy subjects of future research. On the one hand, modest levels of EE2 contamination might increase the strength of sexual selection on females, because females would be competing for a smaller pool of males that would be perceived as potential mates. This increased competition for mates could reduce the effective population size, resulting in a loss of variation due to genetic drift. On the other hand, EE2 contamination could decrease the strength of sexual selection, as males made arti cially unattractive by EE2 exposure may be discriminated against by attractive females, forcing them to mate with less attractive females, which would normally be excluded from the mating pool due to strong female competition. This decrease in sexual selection could lead to a reduction in population viability as mating with sub-par females could result in fewer direct or indirect bene ts of mate choice and reduced offspring tness. Overall, these considerations lead to the conclusion that the population-level effects of EE2-induced behavioral changes represent an area of critical need for future research. Our study's call for additional research on population-level effects of anthropogenic contaminants that disrupt sexual selection is bolstered by other recent studies. For example, changes in reproductive hierarchies and sexual selection were observed by Coe et al. (2008) in zebra sh (Danio rerio) after exposure to EE2. The proportion of offspring sired by males that were dominant in the hierarchy prior to exposure was signi cantly suppressed after exposure to EE2, allowing subordinate males to increase their relative reproductive tness. A similar effect was observed in the sand goby (Pomatoschistus minutus), where exposure to EE2 decreased the strength of sexual selection on male size, allowing smaller males to gain mating opportunities (Saaristo et al., 2009a,b). In addition, exposure of pregnant female amarillo sh (Girardinichthys multiradiatus) to methyl parathion, an insecticide, suppressed the expression of secondary sexual traits in male offspring (ArellanoAguilar and Garcia, 2008). Similar to our study, this effect resulted in the exposed male offspring being less preferred by females during choice trials. These results, along with our study suggest that solely examining the effect of these compounds on reproductive output may not be enough to determine how they are affecting populations, especially considering that processes involved in pre-copulatory sexual selection may be more sensitive than overall reproductive ability. In summary, we found that even low levels of exposure to the endocrine disruptor EE2 in Gulf pipe sh are suf cient to disrupt mating dynamics. Males exposed to EE2 become more female-like with respect to external morphology and develop secondary sexual traits that normally appear only in females. In addition, these altered traits may persist for long periods of time post-exposure. Females prefer non-exposed males compared to exposed males, and in some cases exposed males have a reduced capacity to become pregnant. Thus, EE2 exposure decreases male mating opportunities, potentially impacting their reproductive tness. Overall, our study provides evidence for a role of endocrine disruptors in disrupting mating dynamics, a topic that remains largely unstudied. In addition, we suggest that the effects of endocrine disruptors on pre-copulatory mate choice mechanisms could have important impacts on natural populations. Acknowledgments We would like to thank Gil Rosenthal for help with behavioral analysis. We would like to thank Delayne Ferguson, Doug Haywick, Nina Joiner, Emily Boone, Dan Martin, Molly Mintz, and Brenna Elmen for help collecting, maintaining and feeding sh. In addition, we would like to thank Doug Haywick for providing the pipe sh illustrations. We also thank the University of South Alabama, Department of Biology for providing laboratory space for all the behavioral studies. This research was supported by an EPA Star Graduate Fellowship (C.P.) and a NOAANERR Graduate Fellowship (C.P.) and through aid provided by Texas Fig. 3. A male pipe sh with iridescent stripes 19 days after removal from treated water. The arrow indicates the iridescent stripes, and the space between the green bars on either end of the photograph is 1.5 cm. This image shows that the feminization of males by EE2 persists after the males are moved to clean water. These traits persist at least several weeks, as they were still present when the males were sacri ced at the end of the experiment (2-3 weeks post-exposure). They may persist much longer, but this question was beyond the scope of our study. C. Partridge et al. / Hormones and Behavior 58 (2010) 800-807 807 A&M University. This research was approved by the Animal Care and Use Committee at both the University of South Alabama (AUP#: 06035) and Texas A&M University (AUP#: 2004-253). References Ahnesj?, I., Kvarnemo, C., Merilaita, S., 2001. Using potential reproductive rates to predict mating competition among individuals quali ed to mate. Behav. Ecol. 12, 397-401. Allen, Y., Scott, A., Matthiessen, P., Hawthorn, S., Thain, J., Feist, S., 1999. Survey of estrogneic activity in United Kingdom estuarine and coastal waters and its effects on gondadal development of the ounder Platichthyes esus. Environ. Toxicol. Chem. 18, 1793-1800. Andersson, M., 1994. Sexual Selection. Princeton University Press, Princeton. Arellano-Aguilar, O., Garcia, C., 2008. Exposure to pesticides impairs the expression of sh ornaments reducing the availability of attractive males. Proc. R. Soc. Lond. B Biol. 275, 1343-1350. Balch, G., Mackenzie, C., Metcalfe, C., 2004. Alterations to gonadal development and reproductive success in Japanese medaka (Oryzias latipes) exposed to 17 alphaethinylestradiol. Environ. Toxicol. Chem. 23, 782-791. Bell, A., 2001. Effects of an endocrine disrupter on courtship and aggressive behavior of male three-spined stickleback, Gasterosteaus aculeatus. Anim. Behav. 62, 775-780. Bjerselius, R., Lundstedt-Enkel, K., Olsen, H., Mayer, I., Dimberg, K., 2001. Male gold sh reproductive behavior and physiology are severely affected by exogenous exposure to 17?-estradiol. Aquat. Toxicol. 53, 139-152. Boisseau, J., 1967. Hormonal regulation during incubation in a vert?brate male: Research on the reproduction of Hippocame (French) Ph.D. thesis. University of Bordeaux. Bolland, J., Boettcher, A., 2005. Population structure and reproductive characteristics of the Gulf pipe sh, Syngnathus scovelli, in Mobile Bay, Alabama. Estuaries 28, 957-965. Brown, K., Schultz, I., Cloud, J., Nagler, J., 2008. Aneuploid sperm formation in rainbow trout exposed to the environmental estrogen 17 alpha-ethynylestradiol. Proc. Natl Acad. Sci. USA 50, 19786-19791. Clouzot, L., Marrot, B., Doumenq, P., Roche, N., 2008. 17?-Ethinylestradiol: an endocrine disrupter of great concern. Analytical methods and removal processes applied to water puri cation. A review. Environ. Prog. 27, 383-396. Clutton-Brock, T., Parker, G., 1992. Potential reproductive rates and the operation of sexual selection. Q. Rev. Biol. 67, 437-456. Clutton-Brock, T., Vincent, A., 1991. Sexual selection and the potential reproductive rates of males and females. Nature 351, 58-60. Coe, T., Hamilton, P., Hodgson, D., Paull, G., Stevens, J., Sumner, K., Tyler, C., 2008. An environmental estrogen alters reproductive hierarchies, disrupting sexual selection in group spawning sh. Environ. Sci. Technol. 42, 5020-5025. Crews, D., Gore, A., Hsu, T., Dangleben, N., Spinetta, M., Schallert, T., Anway, M., Skinner, M., 2007. Transgenerational epigenetic imprints on mate preference. Proc. Natl Acad. Sci. USA 104, 5942-5946. Eens, M., Van Duyse, E., Berghman, L., Pinxten, R., 2000. Shield characteristics are testosterone-dependent in both male and female moorhens. Horm. Behav. 37, 126-134. Emlen, S., Oring, L., 1977. Ecology, sexual selection and the evolution of mating systems. Science 197, 215-223. Folstad, I., Karter, A., 1992. Parasites, bright males, and the immunocompetence handicap. Am. Nat. 139, 603-622. Hashimoto, S., Watanabe, E., Ikeda, M., Terao, Y., Str?ssmann, C.A., Inoue, M., Hara, A., 2009. Effects of ethinylestradiol on medaka (Oryzaia latipes) as measured by sperm mobility and fertilization success. Arch. Environ. Contam. Toxicol. 56, 253-259. Hill, R., Janz, D., 2003. Developmental estrogenic exposure in zebra sh (Danio rerio): I. Effects on sex ratio and breeding success. Aquat. Toxicol. 63, 417-429. Jones, A., Avise, J., 1997. Microsatellite analysis of maternity and the mating system in the Gulf pipe sh (Syngnathus scovelli), a species with male pregnancy and sex-role reversal. Mol. Ecol. 6, 203-213. Jones, A., Walker, D., Avise, J., 2001. Genetic evidence for extreme polyandry and extraordinary sex-role reversal in a pipe sh. Proc. R. Soc. Lond. B Biol. 268, 2531-2535. Kidd, K., Blanch eld, P., Mills, K., Palace, V., Evans, R., Lazorchak, J., Flick, R., 2007. Collapse of a sh population after exposure to a synthetic estrogen. Proc. Natl Acad. Sci. USA 104, 8897-8901. Kokko, H., Monaghan, P., 2001. Predicting the direction of sexual selection. Ecol. Lett. 2, 247-255. Kolpin, D., Furlong, E., Meyer, M., Thurman, E., Zaugg, S., Barber, L., Buxton, H., 2002. Pharmaceuticals, hormones and other organic wastewater contaminants in US streams, 1999-2000: a nation reconnaissance. Environ. Sci. Technol. 36, 1202-1211. Kvarnemo, C., Ahnesj?, I., 1996. The dynamics of operational sex ratios and competition for mates. Trends Ecol. Evol. 11, 404-408. Larsen, M., Hansen, K., Henriksen, P., Baatrup, E., 2008. Male zebra sh (Danio rerio) courtship behavior resists the feminizing effects of 17?-ethinyloestradiolmorphological sexual characteristics do not. Aquat. Toxicol. 87, 234-244. Lintelmann, J., Katayama, A., Kurihara, N., Shore, L., Wenzel, A., 2003. Endocrine disruptors in the environment: IUPAC Technical Report. Pure Appl. Chem. 75, 631-681. Maunder, R., Matthiessen, P., Sumpter, J., Pottinger, T., 2007. Impaired reproduction in three-spined sticklebacks exposed to ethinyl estradiol as juveniles. Biol. Reprod. 77, 999-1006. Mayer, I., Rosenqvist, G., Borg, B., Ahnesj?, I., Berglund, A., Schultz, R., 1993. Plasma levels of sex steroids in three species of pipe sh (Syngnathidae). Can. J. Zool. 71, 1903-1907. McGraw, K., 2006. Sex steroid dependence of carotenoid-based coloration in female zebra nches. Physiol. Behav. 88, 374-352. McGraw, K., Correa, S., Adkins-Regan, E., 2006. Testosterone upregulates lipoprotein status to control sexual attractiveness in a colorful songbird. Behav. Ecol. Sociobiol. 60, 117-122. Moncaut, N., Lo Nostro, F., Maggese, M., 2003. Vitellogenin detection in surface mucus of South American cichlid sh Sichlasoma dimerus (Heckel, 1840) induced by estraditol-17?. Effects on liver and gonads. Aquat. Toxicol. 63, 127-137. Nash, J., Kime, D., Van der Ven, L., Wester, P., Brion, F., Maack, G., Stahlschmidt-Allner, P., Tyler, C., 2004. Long-term exposure to environmental concentrations of the pharmaceutical ethynlestradiol causes reproductive failure in sh. Environ. Health Perspect. 112, 1725-1733. Oshima, Y., Kang, I., Kobayahsi, M., Nakayama, K., Imada, N., Honjo, T., 2003. Suppression of sexual behavior in male Japanese medaka (Oryzias latipes). Chemosphere 50, 429-436. Palace, V., Wautier, K., Evans, R., Blanch eld, P., Mills, K., Chalanchuk, S., Godard, D., McMaster, M., Tetreault, G., Peters, L., Vandenbyllaardt, L., Kidd, K., 2006. Biochemical and histopathological effects in pearl dace (Margariscus margarita) chronically exposed to a synthetic estrogen in a whole lake experiment. Environ. Toxicol. Chem. 25, 1114-1125. Parker, T., Knapp, R., Rosen eld, J., 2002. Social mediation of sexually selected ornamentation and steroid hormone levels in male junglefowl. Anim. Behav. 64, 291-298. Peters, R., Courtenay, S., Cagampan, S., Hewitt, M., MacLatchy, D., 2007. Effects on reproductive potential and endocrine status in the mummichog (Fundulus heteroclitus) after exposure to 17?-ethynylestradiol in a short-term reproductive bioassay. Aquat. Toxicol. 85, 154-166. Pettersson, I., Arukwe, A., Lundstedt-Enkel, K., Mortensen, A., Berg, C., 2006. Persistent sex-reversal and oviducal agenesis in adult Xenopus (Silurana) tropicalis frogs following larval exposure to the environmental pollutant ethynylestradiol. Aquat. Toxicol. 79, 356-365. Pojana, G., Gomiero, A., Jonkers, N., Marcomini, A., 2007. Natural and synthetic endocrine disrupting compounds (EDCs) in water, sediment and biota of a coastal lagoon. Environ. Int. 33, 929-936. Rand, M.S., 1992. Hormonal control of polymorphic and sexually dimorphic coloration in the lizard Sceloporus undulates erythrocheilus. Gen. Comp. Endocrinol. 88, 461-468. Ratterman, N., Rosenthal, G., Jones, A., 2009. Sex recognition via chemical cues in the sex-role-reversed Gulf pipe sh (Syngnathus scovelli). Ethology 115, 339-346. Robinson, C., Brown, E., Craft, J., Davies, I., Moffat, C., Pirie, D., Robertson, F., Stagg, R., Struthers, S., 2003. Effects of sewage ef uent and ethynyl oestradiol upon molecular markers of oestrogenic exposure, maturation and reproductive success in the sand goby (Pomatoschistus minutes, Pallas). Aquat. Toxicol. 62, 119-134. Saaristo, M., Craft, J., Lehtonen, K., H., Lindstrom, K., 2009a. Sand goby (Pomatoschistus minutus) males exposed to an endocrine disrupting chemical fail in nest and mate competition. Horm. Behav. 56, 315-321. Saaristo, M., Craft, J., Lehtonen, K., Bjork, H., Lindstrom, K., 2009b. Disruption of sexual selection in sand gobies (Pomatoschistus minutus) by 17?-ethinylestradiol, an endocrine disruptor. Horm. Behav. 55, 530-537. Sch?fers, C., Teigeler, M., Wenzel, A., Maack, G., Fenske, M., Segner, H., 2007. Concentration- and time-dependent effects of synthetic estrogen, 17?-ethinylestradiol, on reproductive capabilities of the zebra sh. Danio rerio. J. Toxicol. Environ. Health A 70, 768-779. Simmons, L., Kvarnemo, C., 2006. Costs of breeding and their effects on the direction of sexual selection. Proc. R. Soc. Biol. Sci. Ser. B 273, 465-470. Trivers, R., 1972. Parental investment and sexual selection. In: Campbell, B. (Ed.), Sexual selection and the decent of man, 1871-1971. Heinimann, London, pp. 136-179. Ueda, N., Partridge, C., Bolland, J., Hemming, J., Sherman, T., Boettcher, A., 2005. Effects of an environmental estrogen on male gulf pipe sh, Syngnathus scovelli (Evermann and Kendall) a male brooding teleost. Bull. Environ. Contam. Toxicol. 74, 1207-1212. Vajda, A., Barber, L., Gray, J., Lopez, E., Woodling, J., Norris, D., 2008. Reproductive disruption in sh downstream from an estrogenic wastewater ef uent. Environ. Sci. Technol. 42, 3407-3414. Vandenbergh, G., Adriaens, D., Verslycke, T., Janssen, C., 2003. Effects of 17?ethinylestradiol on sexual development of the amphibod Hyalella azteca. Ecotoxicol. Environ. Saf. 54, 216-222. Weber, L., Hill, R., Janz, D., 2003. Developmental estrogenic exposure in zebra sh (Danio rerio): II. Histological evaluation of gametogenesis and organ toxicity. Aquat. Toxicol. 63, 431-446. Ying, G., Kookana, R., Ru, Y., 2002. Occurrence and fate of hormone steroids in the environment. Environ. Int. 28, 545-551.