ORIGINAL ARTICLE

Mate quality and the temporal dynamics of breedingin a sex-role-reversed pipefish, S. typhle

Sarah P. Flanagan1,2& Gunilla Rosenqvist3,4 & Adam G. Jones1

Received: 14 September 2016 /Revised: 6 December 2016 /Accepted: 8 December 2016# Springer-Verlag Berlin Heidelberg 2016

AbstractThe spatiotemporal dynamics of receptivity and breeding date,coupled with individual-level quality and attractiveness, arecentrally important to mating system dynamics. These topicshave been investigated in some detail in birds, but much lesswork has been devoted to other taxonomic groups, and almostno work has addressed spatiotemporal factors and individualquality in sex-role-reversed taxa. The broad-nosed pipefish,Syngnathus typhle, provides an excellent opportunity to inves-tigate these ideas in a sex-role-reversed fish. Here, we ad-dressed three questions related to mating dynamics inS. typhle: (1) Do higher-quality males arrive earlier on thebreeding grounds and mate first? (2) Are early-breeding malesin better condition than later-breeding males? And (3) do mat-ing events involving higher-quality males produce betterclutches than mating events involving lower-quality males?We collected data from a field study and a laboratory breedingexperiment to address our hypotheses. Our results show thatlarger males mate earlier than smaller males and that pregnantmales have higher measures of condition compared to non-

pregnant males. Moreover, our laboratory results demonstratethat pairings between larger males and preferred femalesyielded more offspring than pairings involving smaller males.In summary, the spatiotemporal dynamics of S. typhle breed-ing patterns, combined with variation in individual quality,play an important role in shaping mating systems and shouldbe incorporated in future analyses of mating behavior andsexual selection in this interesting sex-role-reversed pipefish.

Significance statementThe breeding patterns of a species can fluctuate over time dueto a number of factors, one of which is individual quality.Although the effects of both the timing of reproduction andfemale quality on mating systems have been studied in somespecies, they have been investigated primarily in isolation.Here, we demonstrate that individual quality and the timingof reproduction interact to affect reproductive success in awild population of sex-role-reversed fish.

Keywords Syngnathus typhle . Reproductive success .

Condition . Breeding date . Mating systems

Introduction

Characterizing and explaining variation are central for ourunderstanding of mating systems. Studies of mating systemsare often logistically constrained to study a single populationat a single time point, but natural mating systems are known tobe dynamic and fluctuate over time. Over the course of amating season, a variety of biotic and abiotic factors can causefluctuations in mating behavior. For example, the operationalsex ratio can change over the course of a breeding season,which causes fluctuations in the opportunity for sexual selec-tion and changes in mating patterns (Reichard et al. 2008;

Communicated by J. Lindström

* Sarah P. Flanaganspflanagan.phd@gmail.com

1 Biology Department, Texas A&M University, CollegeStation, TX 77843, USA

2 National Institute for Mathematical and Biological Synthesis,University of Tennessee, Knoxville, TN, USA

3 Centre for Biodiversity Dynamics Department of Biology,Norwegian University of Science and Technology,7491 Trondheim, Norway

4 EBC, Uppsala University, Campus Gotland, SE-62167 Visby, Sweden

Behav Ecol Sociobiol (2017) 71:28 DOI 10.1007/s00265-016-2255-3

Olsson et al. 2011; Cunha et al. 2014). Abiotic factors such astemperature (Silva et al. 2007; Milner et al. 2010), precipita-tion (Robinson et al. 2012), and photoperiod (McAllan andGeiser 2006) can also cause dynamic breeding systems.Additionally, biotic and abiotic factors can interact to affecttraits relevant to breeding systems such as mate preferences ormate quality (Jann et al. 2000; Cratsley and Lewis 2005;McAllan and Geiser 2006; Kasumovic et al. 2008; Sefcet al. 2009; Milner et al. 2010; Cunha et al. 2014).

Mate quality can interact with temporal aspects of breedingsystems in a variety of ways. For instance, in many migratorybird species, quality is related to the arrival dates on the breed-ing grounds (e.g., barn swallows, Møller 1994; Americanredstarts, Smith and Moore 2003) and the nesting date offemales (e.g., common terns, Wendeln 1997; tree swallows,Hasselquist et al. 2001). Individual quality may also impactthe timing of reproduction if quality affects receptivity or ifmate choice is condition dependent (reviewed in Cotton et al.2006). In many species, early-breeding individuals havehigher mating or reproductive success (e.g., birds, Oring andLank 1982; Møller 1994; Aebischer et al. 1996; Mitrus et al.2012; Rockwell et al. 2012; Saino et al. 2012; Laczi et al.2013; fish, Dickerson et al. 2005; Parkos et al. 2011).Although these studies show that the timing of reproductionis closely related to mate quality, most examples come frombirds. To understand the generality of the relationship betweenmate quality and the temporal dynamics of breeding, thesetwo components of mating systems should also be studied inother groups of animals.

The study of mating patterns in sex-role-reversed speciesprovides a unique opportunity to test the generality of matingsystem dynamics (Jones et al. 2000a, 2005). Female fecundityis often the limiting factor in reproduction for species withconventional sex roles. Consequently, size and fecundity areoften important indicators of female quality. In sex-role-reversed species, on the other hand, females are the targetsof stronger sexual selection and female fecundity is not thelimiting factor in mate choice (Berglund et al. 1989; Emlenand Wrege 2004; Rosenqvist and Berglund 2011). Rather,male parental care and mate availability are the limiting re-sources in many sex-role-reversed taxa (e.g., phalaropes,Reynolds 1987; pipefish, Berglund et al. 1989; jacanas,Emlen and Wrege 2004). As in species with conventionalsex roles, mating patterns and sexual selection in sex-role-reversed taxa are affected by temporal aspects of breeding,as has been shown, for example, for the rate of extra-pairfertilizations in red phalaropes (Dale et al. 1999) and withregard to fluctuations in the operational sex ratio and female-female competition in Wilson’s phalaropes (Colwell andOring 1988). The broad-nosed pipefish, Syngnathus typhle,has been a model for studying sex-role-reversed mating sys-tems for decades (see BMethods^ and BStudy species^), but thetemporal dynamics of the breeding season and the influence of

individual quality have not been adequately documented.Here, we present three complementary studies conducted tounderstand how the timing of reproduction and individualquality interact to affect reproductive success in S. typhle.

Methods

Study species

The broad-nosed pipefish, S. typhle, is a sex-role-reversed fishin the family Syngnathidae (seahorses, pipefishes, andseadragons) that occurs in marine waters throughout Europe.In the Baltic and North Seas, where the behavior of S. typhlehas been studied in most detail, the breeding season typicallylasts fromMay until August (Vincent et al. 1995). The broad-nosed pipefish has exclusively paternal care in the form ofmale pregnancy, and male S. typhle can brood embryos frommultiple females (Jones et al. 1999). Females also mate mul-tiply and collectively produce eggs faster than males canbrood them (Berglund et al. 1989).Males prefer larger femalesas mates (Berglund 1993, 1994; Berglund and Rosenqvist1993; Berglund and Rosenqvist 2001; Berglund et al. 1997,2005; Mobley et al. 2011; Aronsen et al. 2013; Sundin et al.2013), probably because larger females produce larger eggsand transfer more eggs per copulation (Berglund et al. 1986a,1988; Jones et al. 2000b; Partridge et al. 2009; Mobley et al.2011; Sogabe and Ahnesjö 2011).

In S. typhle, male quality is likely to impact the fitness of hisoffspring, since males contribute energy to embryos (Berglundet al. 1986b) in the form of both amino acids and glucose(Kvarnemo et al. 2011). Males are also able to absorb nutrientsfrom eggs and use them in brood pouch, liver, and muscletissues (Sagebakken et al. 2010). Previous work suggests thatbody length is an indicator of male quality, at least with regardto reproductive success. Large males with larger embryos in-vest more energy per embryo than smaller males (Berglundet al. 1986a). Although previous studies have shown that malelength does not affect very early embryo survivorship(Partridge et al. 2009; Mobley et al. 2011; Sagebakken et al.2011), larger males produce more newborn offspring (Ahnesjö1992a; Sagebakken et al. 2011), and their offspring better sur-vive predation compared to offspring from small males(Ahnesjö 1992b; Sandvik et al. 2000). Intriguingly, anecdotalobservations imply that larger males breed earlier than smallermales (Ahnesjö 1992b), and some evidence suggests that re-productive males have smaller fat stores than non-reproductive males (Svensson 1988). However, Svensson(1988) scored fat content on a qualitative, binary scale (Blessfat^ versus Bmore fat^) based on visual inspection.

These previous findings are sufficient to permit the formu-lation of a hypothesis that in S. typhle, the highest-quality malesbreed first in the mating season and those early matings result

28 Page 2 of 10 Behav Ecol Sociobiol (2017) 71:28

inmore or higher-quality offspring.We tested this hypothesis ina series of three studies with clear predictions: (1) the highest-quality males are the first to mate, (2) early-breeding males arein better condition than males that do not mate early in theseason, and (3) mating events involving higher-quality maleshave better clutches than matings involving low-quality males.In all three studies, it was not possible to record data blindbecause our study involved focal animals in the field.

Timing of reproduction

To evaluate the role of individual quality in the timing ofreproduction in S. typhle, we measured the body length ofmated (pregnant) and unmated (non-pregnant) males and allfemales at the start of the breeding season. We chose to focuson body length as a measure of quality for this component ofthe study, which allowed us to sample the population withreplacement (i.e., the captured fish were returned to thepopulation).

We sampled a population of S. typhle at the beginning ofthe breeding season. Fish were caught in shallow eelgrass(Zostera marina) beds in Kyllaj, Gotland, Sweden (N 57°35.163′, E 18° 44.749′) inMay and June of 2014 using a smallbeam trawl (mesh size 4 mm) pulled by a boat. The trawl waspulled for 2 h or until 100 fish were captured. Captured fishwere kept in buckets with aerated seawater and artificialseagrass during transportation by a car (25 km) to theUppsala University Research Station in Ar, Gotland,Sweden. Each fish was anesthetized by submersion in cloveoil (100 mg/L) and uniquely marked with an elastomer tag(VIFE; Northwest Marine Technology, Inc., Shaw Island,WA, USA) injected beneath the dorsal skin immediately be-hind the operculum (see Rose et al. 2013a, b; Flanagan et al.2014). Fish were photographed and allowed to recover inaerated seawater with artificial seagrass. Once all fish weretagged, they were transported back to Kyllaj and returned tothe eelgrass bed from which they had been captured. Thissampling, tagging, and returning procedure was repeated ona total of 17 days between 17 May 2014 and 8 June 2014.When a tagged fish was recaptured, it was anesthetized andphotographed again but not re-tagged. To establish whetherpregnancy status could be predicted by date of capture or bodylength, a binomial generalized linearmodel was run with preg-nancy status (pregnant or not pregnant) as the response vari-able and date captured and body length as the predictor vari-ables, with the interaction term included in the model.Recaptured individuals were only included in the analysis ontheir first date of capture.

Quality of mated males

Although body length likely reflects some key aspects of qual-ity in S. typhle (Berglund et al. 1986a; Ahnesjö 1992a, b;

Sagebakken et al. 2011), potential condition-dependent traitssuch as fat content and gonad mass are probably also impor-tant (Tobler 2008; McPherson et al. 2011). To supplement theBTiming of reproduction^ component of this study, we mea-sured fat content, gonad mass, and body length in 28 pregnantmales, 27 non-pregnant males, and 29 females, caught on 8June 2014 in shallow eelgrass beds in Kyllaj as describedabove. The fish were transported to the Uppsala UniversityResearch Station in Ar, Gotland, Sweden, where they werehumanely euthanized and measured. A fin clip was taken fromeach fish and stored in ethanol before the remaining tissue wasplaced in 10% formaldehyde. The gastro-intestinal tract andreproductive organs, as well as embryos from the pregnantmales, were removed from the preserved bodies. Gonadsand embryos were dried at 55 °C for 24 h and the bodies weredried for 72 h. After drying, tissues were weighed. Rather thanusing the gonado-somatic index as a measure of gonad sizerelative to body size, we used the residual gonad mass fromthe linear relationship between body mass and gonad mass(Tomkins and Simmons 2002), calculated separately for malesand females, where the body mass did not include the mass ofthe gastro-intestinal tract. The bodies and gastro-intestinaltracts were then placed in petroleum ether and the petroleumether solution was changed every day for 4 days. After 4 days,the bodies were dried for 24 h at 55 °C and weighed again.The fat content index was calculated as (Massbefore −Massafter)/Massbefore (Tobler 2008). A binomial generalizedlinear model, including interaction terms, was used to estab-lish if the binary pregnancy status of a male (pregnant or not)was predicted by fat content index, standard length, and/ortestis residuals.

Male quality and reproductive success

Finally, we wanted to establish whether higher-quality males,as determined by body length, have better broods or whetherpreviously documented relationships betweenmale length andoffspring quality are a female effect. We conducted matingexperiments on fish caught at Kyllaj, as described above, inJune 2014. The fish were transported to the field station,where they were acclimated to a flow-through system withseawater pumped in from the Baltic Sea. Fish were initiallystored in 200 L barrels, were sorted by sex, and were fed threetimes daily with a mix of live Artemia brine shrimp and frozenmysids. After acclimation, the males were measured andsorted into two size classes, small (115–130 mm standardlength) and large (145–160 mm standard length). Thesegroups were determined by splitting the size distribution ofthe population from the tagging data into three non-overlapping groups and taking the top third and the bottomthird as the large and small size classes. We set up 24 male-female pairs (12 pairs with a large male and 12 pairs with asmall male), with females being evenly distributed by size

Behav Ecol Sociobiol (2017) 71:28 Page 3 of 10 28

among the large and small males (i.e., average female size wasthe same for pairs with large males as it was for pairs withsmall males). Each male-female pair was placed into its own200-L barrel and allowed 2 weeks to mate. Each male andeach female were only provided with one mate.

After mating, the males were kept until day 21 of the preg-nancy, at which time they were euthanized. Females werekilled at the same time as their mates. Embryos were removedfrom pregnant males and counted. Each embryo could beclassified as either Bdeveloping^ or Breduced^ by comparingthe general stage of development within the pregnancy. Evenif the pairs had mated multiply, the male’s pouch closes after afew days (Berglund et al. 1988) so all of the embryos were atroughly the same developmental stage. Gonads and gastro-intestinal tracts were dissected from males and females. Thebody, gonads, and ten developing embryos from pregnantmales were placed in a drying oven at 55 °C for 24 h. Driedtissues were weighed and the residuals from a linear regres-sion of gonad mass on bodymass were calculated as describedabove. Three fitness correlates were quantified: clutch size,embryo weight, and embryo survivorship. Clutch size wasthe number of developing embryos plus the number of re-duced embryos or in other words the total number of eggsinitially transferred to the male. Embryo weight was calculat-ed as the average dry weight of one embryo, calculated as themean of the ten embryos dried and weighed. Embryo survi-vorship was calculated as the ratio of surviving embryos to thetotal number of eggs transferred (number of developingembryos/number of eggs transferred). Linear models wereused to determine whether the three fitness components couldbe explained bymale size class, female length, testis residuals,and ovary residuals. Because the study design treated malesize as dichotomous (either large or small), we analyzed thedata with male size as a categorical variable.

Statistics

All statistical analyses were performed in R 2.3.2 (R CoreTeam 2013). The three datasets were analyzed separately.Missing data were excluded, and all statistical tests weretwo-tailed. Correlations were estimated using the R packageHmisc version 3.14-3 (Harrell et al. 2014). The raw data and Rcode used for analysis are archived on Dryad (http://dx.doi.org/10.5061/dryad.5ff20).

Results

Timing of reproduction

We found that pregnant males were larger than non-pregnantmales and that average male size decreased over time, sug-gesting that the largest males were the first to mate and that

smaller, non-pregnant males continued to move into thebreeding area as the season progressed. A total of 809 males(268 pregnant, 541 non-pregnant), 722 females, and 84 juve-niles were captured with a total of 11 recaptures (5 non-pregnant males, 4 pregnant males, 2 females). No pregnantmales were captured on the first day of sampling (Fig. 1a),but females were captured every day of sampling (Fig. 1b). Alinear model showed that pregnancy was predicted by length(β = 0.04696, z = 2.814, p = 0.0049) and the interaction oflength and date (β = 0.005091, z = 2.953, p = 0.00315). Datecaptured on its own was not significant (β = −0.384058,z = −1.681, p = 0.09273; Fig. 1a). Any changes in averagesize over time could be due to growth, but previous studiesand the recaptured individuals suggest that growth over thecourse of 1 month is likely restricted to a change of less than1 cm (Svensson 1988), a difference which would be unlikelyto impact the overall results of this study.

Quality of mated males

The three measures of quality (fat content, body length, andtestis residuals) were not correlated with each other (fat con-tent and body length r = 0.23, p = 0.2231; body length andtestis residuals r = 0.13, p = 0.5049; fat content and testisresiduals r = 0.13, p = 0.5139), so all three were included inthe generalized linear model. Our fat content analysis of field-collected males showed that pregnancy status (pregnant ornot) was significantly predicted by fat content (β = 76.5287,z = 2.145, p = 0.0319; Fig. 2a) but not by body length(β = 0.0322, z = 1.052, p = 0.2927) or testis residuals(β = −1098.5785, z = −0.814, p = 0.4157). Within only thepregnant males, none of our quality metrics significantly pre-dict the number of eggs a male carried (R2

adj = −0.009683,F3,25 = 0.9105, p = 0.45). Fat content did predict the averagedry mass of embryos (β = 0.0452, z = 2.209, p = 0.0366), andlonge r ma l e s t ended to have heav i e r embryos(β = 0.00004581, z = 1.927, p = 0.0654).

We observed a significant positive relationship betweenfemale length and the dry weight of female gonads(r = 0.62, p = 0.0005; Fig. 2d). However, in the females, nosignificant relationships were observed between fat contentand body length (r = −0.16, p = 0.4319) or the ovary residuals(r = −0.02, p = 0.9400).

Male quality and reproductive success

Of the 24 male-female pairs that were established for the mat-ing experiments (12 pairs with small males, 12 pairs with largemales), 22 of the pairs mated. One large and one small pair didnot mate. All of the pairs that mated did so between 15June 2014 and 26 June 2014, which was approximately1 month after the start of the breeding season. Six of the matedmales (three large, three small) died before the 21-day mark so

28 Page 4 of 10 Behav Ecol Sociobiol (2017) 71:28

those pairs were excluded from the analysis. Females rangedin size from 128 to 194 mm (μ = 146.6 mm). The male lengthand body weight were significantly correlated (r = 0.92,n = 16, p < 0.00001; Fig. 3a), and male body mass was sig-nificantly correlated with testis mass (r = 0.55, n = 16,p = 0.0262). However, male size class, male body weight,and residual testis mass were not correlated with clutch size,embryo weight, or embryo survivorship in this experiment(p > 0.05).

Female standard length was significantly correlated withfemale body mass (r = 0.96, n = 16, p < 0.00001; Fig. 3a)and ovary mass (r = 0.68, n = 16, p = 0.0003). Additionally,the residual ovary mass significantly predicted the number ofeggs in the male’s pouch (t = −2.198, p = 0.0483; Fig. 3b),although there was an interaction with male size category

where large males had large clutches even when their mate’sresidual ovary mass was small (t = 2.651, p = 0.0211; Fig. 3b).Female standard length predicted clutch size, regardless ofmale size (t = 2.59, p = 0.0237; Fig. 3c), and embryo sizewas significantly predicted by female body weight(t = 3.197, p = 0.00767; Fig. 3d). Offspring survivorshipwas not related to female size or residual ovary mass(p > 0.05). Hence, female size is associated with the clutchsize and embryo weight, whereas male size is not.

Discussion

We conducted three studies to investigate how individualquality and the timing of reproduction shape the breeding

Fig. 1 The changes in average size over time in samples of pregnant andnon-pregnant S. typhle males and females. Top panel Pregnant maleswere significantly larger than non-pregnant males, and average sizedecreased significantly over time. The largest males were the first tobecome pregnant, and large males consistently gained matings beforesmall males, but the decline in average pregnant male size demonstratedthat additional smaller males became pregnant as the breeding seasonprogressed. The sample sizes for pregnant males are above the pointsand sample sizes for non-pregnant males are below the points. The bars

around each point represent the 95% confidence intervals. Bottom panelThe size range of captured males did not change dramatically over thecourse of the breeding season, although average male size decreased.Females were captured every day the population was sampled andmean size did not change significantly over time. Female sample sizesare in blue below the violin plots for females, andmale sample sizes are inblack above the violin plots for males. Violin plots show the kerneldensity plots with the mean values as the white circles in the center

Behav Ecol Sociobiol (2017) 71:28 Page 5 of 10 28

system in a sex-role-reversed pipefish, S. typhle, and our re-sults show that higher-quality males, as indicated by largersize and greater fat content, mate earlier in the breeding seasonand produce larger clutches. In laboratory breeding trials,clutch size is explained by female size but not male size.Our results suggest that both the timing of reproduction andthe overall quality of mating pairs are factors that must betaken into account to obtain a complete understanding of mat-ing dynamics and sexual selection in S. typhle.

The present study provides the first quantitative assessmentof mating by broad-nosed pipefish as they arrive on the breed-ing grounds, and our results indicate that the timing of repro-duction plays an important role in the mating dynamics of thisspecies. Combined with other work on S. typhle, our resultspaint an exceptionally clear picture of the sources of variationin reproductive success during the breeding season. Broad-nosed pipefish normally move into shallow seagrass beds tobreed in late May, and their mating season lasts until August(Vincent et al. 1994, 1995). By the end of May, the distribu-tion of female sizes in the breeding area has already stabilized(Fig. 1b). Males show a different pattern, however, in whichthe largest of early arriving males are first to mate, confirmingprevious anecdotal observations (Ahnesjö 1992b). As the

breeding season progresses into its second month (June), theaverage size of males on the breeding ground drops, becauseadditional smaller males continue to arrive. Once the largestmales have mated so their pouches are filled, females turntheir attention to the smaller males, so the mean size of preg-nant males also drops as the smaller males fill their pouches.Eventually, all but the tiniest of males become pregnant.Future work could investigate how these dynamics of theearly-breeding season impact mating patterns later in the sea-son (in July and August).

Several important sources of variation in reproductive suc-cess for both sexes may occur during this early part of theS. typhle breeding season. With respect to females, we knowthat larger females are preferred by males (Berglund et al.1986a), that females have a much higher potential reproduc-tive rate compared to males (Berglund et al. 1989; Berglundand Rosenqvist 1993), and that females display above theeelgrass to potential male suitors (Vincent et al. 1994).Despite the fact that both sexes prefer larger mates, field-based parentage data show no evidence of size-assortativemating (Mobley et al. 2014). This pattern may neverthelessbe consistent with mutual mate choice if large females pro-duce sufficient numbers of eggs to fill all males in the

Fig. 2 The impact of quality ofearly-mating male S. typhle onreproduction. a Wild-caughtpregnant males have higher fatcontent indices than non-pregnantmales. The boxplot shows themean, quartiles, and outliers. bAverage embryo weight ispredicted by fat content index inwild-caught pregnant males. cLargewild-caught pregnant maleshad heavier embryos than smallerpregnant males. d A significantrelationship between body lengthand ovary mass was observed inwild-caught females

28 Page 6 of 10 Behav Ecol Sociobiol (2017) 71:28

population, because all males, regardless of their own size,prefer larger females. Regardless, the main sources of fitnessvariation for female broad-nosed pipefish probably stem fromdifferences in mating success (Jones et al. 2000b, 2005). Ourresults also show that larger females have larger ovaries thansmaller females and produce larger eggs that result in largerembryos (Fig. 3). Females may also derive other fitness ad-vantages by mating with the largest males in the early part ofthe breeding season; our results show that longer males andmales with higher fat content carry heavier embryos (Fig. 2).Additionally, large males may be able to compensate for lowerresidual ovary mass in their mates (Fig. 3b), although thispattern is driven by two outliers, and it is possible that thosefemales started with higher residual ovary masses that werediminished by mating.

The sources of variation in fitness for males differ fromthose of females. One very important difference is that mat-ing success makes a smaller contribution to male fitness(Jones et al. 2000b, 2005). Previous work shows that maleS. typhle derive direct benefits from their mating choices byobtaining larger eggs from large females (Mobley et al.2011), by avoiding parasitized females (Rosenqvist andJohansson 1995) and by producing offspring less

susceptible to predation (Sandvik et al. 2000). These directbenefits may represent the primary fitness benefit of matechoice for males, assuming mate choice is beneficial. Ourresults suggest some other sources of variation in fitness formales of this species. Early-breeding males, if they do matewith the most attractive (i.e., largest) females, can obtainclutches of large eggs and larger embryos (Fig. 3). Ourevidence suggests that the largest males are the first tobreed (Fig. 1) and that larger males who successfully matedwithin the first month of the breeding season did haveheavier eggs (Fig. 2), lending indirect support for the hy-pothesis that high quality early-mating males mated withattractive, higher-quality females. The earliest males to matealso had the highest fat content, which conceivably contrib-utes to their ability to provision offspring, although thishypothesis has not been tested. We speculate that early-breeding males may gain an additional reproductive advan-tage if, by breeding early, they increase the probability thatthey will be able to have additional pregnancies before theend of the breeding season. Males that breed early may alsohave more successful offspring, because their offspring willhave more time to grow during the favorable summermonths. These latter two benefits will have to be confirmed

Fig. 3 Male and female effectson clutch and embryocharacteristics in our S. typhlelaboratory experiment. a Largermales and larger females hadhigher body masses. b Residualovary mass significantlypredicted clutch size in thefemale’s mate, and there was aninteracting effect of male sizecategory. c Longer females hadlarger clutches, regardless of malesize class. d Clutches fromheavier females had heavierembryos, regardless of male sizeclass

Behav Ecol Sociobiol (2017) 71:28 Page 7 of 10 28

with additional work, as we know very little about thedynamics of the latter part of the breeding season, thegrowth rate of pipefish during the summer, and patterns ofoverwinter survival.

Our study also demonstrates that there are multiple mea-sures of quality that influence reproductive success inS. typhle. For instance, fat content was not associated withlength (see BResults^) but was related to pregnancy status inwild-caught males (Fig. 2a). These findings suggest high-condition males are more likely to be in a position to mateearly. One possibility is that males with fewer fat stores are notyet ready to mate, especially since male pipefish provisiontheir young with nutrients during pregnancy (Ripley andForan 2006; Braga Goncalves et al. 2010). A relationshipbetween fat stores and reproduction is not unheard of in fish;salmon juveniles do not mature in the spring if fat contentlevels are too low (Rowe et al. 1991; Thorpe 1994), and fe-male walleyes become reproductively active only if they havesufficient visceral fat (Henderson et al. 1996). Other than fatstores, male reproductive quality may also influence reproduc-tive success, since we found that residual testis mass is posi-tively associated with average embryo weight. Testis masswas significantly and positively correlated with body mass(see BResults^), and those individuals with higher residualtestis mass carried heavier embryos than other males of asimilar size. These observations again suggest that other mea-sures of condition than body length contribute to reproductivesuccess. Since S. typhle males are guaranteed paternity (Joneset al. 1999), sperm competition cannot explain our results. Amore likely explanation is that some males invest more indeveloping their testes than in growing, and those males pro-duce larger embryos (Reznick 1983; Roff 2011). An intrigu-ing alternative hypothesis is that testes could be involved inthe provisioning of offspring observed in S. typhle and otherpipefish species (Ripley and Foran 2006; Kvarnemo et al.2011).

In conclusion, the results from this study show that thelargest males of S. typhle arrive first on the breeding grounds,are the first to breed, and have a reproductive advantage in thewild over smaller males in the first month of the breedingseason by obtaining clutches with heavier embryos. Thesetemporal aspects of the early portion of the breeding seasoncontribute to fitness variation for both sexes of this species.Hence, the interaction between the timing of reproduction andindividual quality is an important factor shaping reproductivesuccess and breeding patterns in this species and virtually anyother species with seasonal breeding.

Acknowledgements Thanks to A. Berglund for contributing to projectdesign and to S. van Leeuwen, C. Seger, and R. Höglund for the assis-tance with fish care and laboratory work. A. Berglund, K. Wedemeyer-Strombel, E. Rose, and A. Anderson and four anonymous reviewerscommented on previous drafts of this manuscript. We thank UppsalaUniversity-Campus Gotland for being able to use Ar research station.

Compliance with ethical standards All applicable international, na-tional, and/or institutional guidelines for the care and use of animals werefollowed. All handling of fish was done under license Dnr 118-2008 fromthe Swedish Board of Agriculture. This work was partly supported by theResearch Council of Norway through its Centres of Excellence fundingscheme, project number 223257. SPF was supported by the NationalScience Foundation Graduate Research Fellowship under Grant No.DGE-1252521, supplemented by a GROW supplement in connectionwith the Norwegian Research Council. This work was conducted in partwhile SPF was a Postdoctoral Fellow at the National Institute forMathematical and Biological Synthesis, an Institute sponsored by theNational Science Foundation through NSF Award #DBI-1300426, withadditional support from the University of Tennessee, Knoxville. Thiswork was funded by grant number DEB-1119261 from the NationalScience Foundation to AGJ and by grant number DEB-1401688 toAGJ and SPF. The authors declare that they have no conflict of interest.

References

Aebischer A, Perrin N, Krieg M, Studer J, Meyer DR (1996) The role ofterritory choice, mate choice and arrival date on breeding success inthe Savi’s warbler Locustella luscinioides. J Avian Biol 27:143–152

Ahnesjö I (1992a) Consequences of male brood care: weight and numberof newborn in a sex-role reversed pipefish. Funct Ecol 6:274–281

Ahnesjö I (1992b) Fewer newborn result in superior juveniles in thepaternally brooding pipefish Syngnathus typhle L. J Fish Biol 41:53–63

Aronsen T, Mobley KB, Berglund A, Sundin J, Billing AM, RosenqvistG (2013) The operational sex ratio and density influence spatialrelationships between breeding pipefish. Behav Ecol 24:888–897

Berglund A (1993) Risky sex: male pipefishes mate at random in thepresence of a predator. Anim Behav 46:169–175

Berglund A (1994) The operational sex-ratio influences choosiness in apipefish. Behav Ecol 5:254–258

Berglund A, Rosenqvist G (1993) Selective males and ardent females inpipefishes. Behav Ecol Sociobiol 32:331–336

Berglund A, Rosenqvist G (2001) Male pipefish prefer ornamented fe-males. Anim Behav 61:345–350

Berglund A, Rosenqvist G, Svensson I (1986a) Mate choice, fecundityand sexual dimorphism in two pipefish species (Syngnathidae).Behav Ecol Sociobiol 19:301–307

Berglund A, Rosenqvist G, Svensson I (1986b) Reversed sex roles andparental energy investment in zygotes of two pipefish(Syngnathidae) species. Mar Ecol-Prog Ser 29:209–215

Berglund A, Rosenqvist G, Svensson I (1988) Multiple matings andpaternal brood care in the pipefish Syngnathus typhle. Oikos 51:184–188

Berglund A, Rosenqvist G, Svensson I (1989) Reproductive success offemales limited by males in two pipefish species. Am Nat 133:506–516

Berglund A, Rosenqvist G, Bernet P (1997) Ornamentation predicts re-productive success in female pipefish. Behav Ecol Sociobiol 40:145–150

Berglund A, Widemo MS, Rosenqvist G (2005) Sex-role reversalrevisited: choosy females and ornamented, competitive males in apipefish. Behav Ecol 16:649–655

Braga Goncalves I, Mobley KB, Ahnesjö I, Sagebakken G, Jones AG,Kvarnemo C (2010) Reproductive compensation in broad-nosedpipefish females. Proc R Soc Lond B 277:1581–1587

Colwell MA, Oring LW (1988) Sex ratios and intrasexual competition formates in a sex-role reversed shorebird, Wilson’s phalarope(Phalaropus tricolor). Behav Ecol Sociobiol 22:165–173

28 Page 8 of 10 Behav Ecol Sociobiol (2017) 71:28

Cotton S, Small J, Pomiankowski A (2006) Sexual selection andcondition-dependent mate preferences. Curr Biol 16:R755–R765

Cratsley CK, Lewis SM (2005) Seasonal variation in mate choice ofPhotinus ignitus fireflies. Ethology 111:89–100

Cunha M, Berglund A, Monteiro NM (2014) The intrinsically dynamicnature of mating patterns and sexual selection. Environ Biol Fish 98:1047–1058

Dale J, Montgomerie R, Michaud D, Boag PT (1999) Frequency andtiming of extrapair fertilisation in the polyandrous red phalarope(Phalaropus fulicarius). Behav Ecol Sociobiol 46:50–56

Dickerson BR, Brinck KW, Willson MF, Bentzen P, Quinn TP (2005)Relative importance of salmon body size and arrival time at breedinggrounds to reproductive success. Ecology 86:347–352

Emlen ST, Wrege PH (2004) Size dimorphism, intrasexual competition,and sexual selection in wattled jacana (Jacana jacana), a sex-role-reversed shorebird in Panama. Auk 121:391–403

Flanagan SP, Johnson JB, Rose E, Jones AG (2014) Sexual selection onfemale ornaments in the sex-role-reversed gulf pipefish (Syngnathusscovelli). J Evol Biol 27:2457–2467

Harrell, FE, Dupont C et al. (2014) Hmisc: Harrell miscellaneous,http://CRAN.R-project.org/package=Hmisc

Hasselquist D, Wasson MF, Winkler DW (2001) Humoral immunocom-petence correlates with date of egg-laying and reflects work load infemale tree swallows. Behav Ecol 12:93–97

Henderson BA, Wong JL, Nepszy SJ (1996) Reproduction of walleye inLake Erie: allocation of energy. Can J Fish Aquat Sci 53:127–133

Jann P, Blanckenhorn WU, Ward PI (2000) Temporal and microspatialvariation in the intensities of natural and sexual selection in theyellow dung fly Scathophaga stercoraria. J Evol Biol 13:927–938

Jones AG, Rosenqvist G, Berglund A, Avise JC (1999) The geneticmating system of a sex-role-reversed pipefish (Syngnathus typhle):a molecular inquiry. Behav Ecol Sociobiol 46:357–365

Jones AG, Rosenqvist G, Berglund A, Arnold SJ, Avise JC (2000a) TheBateman gradient and the cause of sexual selection in a sex-role-reversed pipefish. Proc R Soc Lond B 267:677–680

Jones AG, Rosenqvist G, Berglund A, Avise JC (2000b) Mate qualityinfluences multiple maternity in the sex-role-reversed pipefishSyngnathus typhle. Oikos 90:321–326

Jones AG, Rosenqvist G, Berglund A, Avise JC (2005) The measurementof sexual selection using Bateman’s principles: an experimental testin the sex-role-reversed pipefish Syngnathus typhle. Integr CompBiol 45:874–884

Kasumovic MM, Bruce MJ, Andrade MCB, Herberstein ME (2008)Spatial and temporal demographic variation drives within-seasonfluctuations in sexual selection. Evolution 62:2316–2325

Kvarnemo C, Mobley KB, Partridge C, Jones AG, Ahnesjö I (2011)Evidence of paternal nutrient provisioning to embryos in broad-nosed pipefish Syngnathus typhle. J Fish Biol 78:1725–1737

Laczi M, Hegyi G, Herényi M, Kiss D, Markó G, Nagy G, Rosivall B,Szöllősi E, Török J (2013) Integrated plumage colour variation inrelation to body condition, reproductive investment and laying datein the collared flycatcher. Naturwissenschaften 100:983–991

McAllan BM, Geiser F (2006) Photoperiod and the timing of reproduc-tion in Antechinus flavipes (Dasyuridae: Marsupialia). Mamm Biol71:129–138

McPherson LR, Slotte A, Kvamme C, Meier S, Marshall CT (2011)Inconsistencies in measurement of fish condition: a comparison offour indices of fat reserves for Atlantic herring (Clupea harengus).ICES J Mar Sci 68:52–60

Milner RNC, Detto T, JennionsMD, Backwell PRY (2010) Experimentalevidence for a seasonal shift in the strength of a female matingpreference. Behav Ecol 21:311–316

Mitrus C, Mitrus J, Sikora M (2012) Badge size and arrival time predictmating success of red-breasted flycatcher Ficedula parva males.Zool Sci 29:795–799

Mobley KB, Kvarnemo C, Ahnesjö I, Partridge C, Berglund A, Jones AG(2011) The effect of maternal body size on embryo survivorship inthe broods of pregnant male pipefish. Behav Ecol Sociobiol 65:1169–1177

Mobley KB, Abou Chakra M, Jones AG (2014) No evidence for size-assortative mating in the wild despite mutual mate choice in sex-role-reversed pipefishes. Ecol Evol 4:67–78

Møller AP (1994) Phenotype-dependent arrival time and its conse-quences in a migratory bird. Behav Ecol Sociobiol 35:115–122

OlssonM,Wapstra E, Schwartz T, Madsen T, Ujvari B, Uller T (2011) Inhot pursuit: fluctuating mating system and sexual selection in sandlizards. Evolution 65:574–583

Oring LW, Lank DB (1982) Sexual selection, arrival times, philopatryand site fidelity in the polyandrous spotted sandpiper. Behav EcolSociobiol 10:185–191

Parkos JJ, Wahl DH, Philipp DP (2011) Influence of behavior and matingsuccess on brood-specific contribution to fish recruitment in ponds.Ecol Appl 21:2576–2586

Partridge C, Ahnesjö I, KvarnemoC,Mobley KB, Berglund A, Jones AG(2009) The effect of perceived female parasite load on post-copulatory male choice in a sex-role-reversed pipefish. Behav EcolSociobiol 63:345–354

R Core Team (2013) R: a language and environment for statistical com-puting. The R Foundation for Statistical Computing, Viennahttp://www.R-project.org/

Reichard M, Smith C, Bryja J (2008) Seasonal change in the opportunityfor sexual selection. Mol Ecol 17:642–651

Reynolds JD (1987)Mating system and nesting biology of the red-neckedphalarope Phalaropus lobatus: what constrains polyandry? Ibis 129:225–242

Reznick D (1983) The structure of guppy life histories: the tradeoff be-tween growth and reproduction. Ecology 64:862–873

Ripley JL, Foran CM (2006) Differential parental nutrient allocation intwo congeneric pipefish species (Syngnathidae: Syngnathus spp.). JExp Biol 209:1112–1121

Robinson MR, van Doorn GS, Gustafsson L, Qvarnström A (2012)Environment-dependent selection on mate choice in a natural pop-ulation of birds. Ecol Lett 15:611–618

Rockwell SM, Bocetti CI, Marra PP (2012) Carry-over effects of winterclimate on spring arrival date and reproductive success in an endan-gered migratory bird, Kirtland’s warbler (Setophaga kirtlandii). Auk129:744–752

Roff DA (2011) An allocation model of growth and reproduction in fish.Can J Fish Aquat Sci 40:1395–1404

Rose E, Paczolt KA, Jones AG (2013a) The contributions of prematingand postmating selection episodes to total selection in sex-role-reversed gulf pipefish. Am Nat 182:410–420

Rose E, Paczolt KA, Jones AG (2013b) The effects of synthetic estrogenexposure on premating and postmating episodes of selection in sex-role-reversed gulf pipefish. Evol Appl 6:1160–1170

Rosenqvist G, Berglund A (2011) Sexual signals and mating patterns inSyngnathidae. J Fish Biol 78:1647–1661

Rosenqvist G, Johansson K (1995)Male avoidance of parasitized femalesexplained by direct benefits in a pipefish. Anim Behav 49:1039–1045

Rowe DK, Thorpe JE, Shanks AM (1991) Role of fat stores in the mat-uration ofmale Atlantic salmon (Salmo salar) parr. Can J Fish AquatSci 48:405–413

Sagebakken G, Ahnesjö I, Mobley KB, Braga Goncalves I, Kvarnemo C(2010) Brooding fathers, not siblings, take up nutrients from embry-os. Proc R Soc Lond B 277:971–977

Sagebakken G, Ahnesjö I, Braga Goncalves I, Kvarnemo C (2011)Multiply mated males show higher embryo survival in a paternallycaring fish. Behav Ecol 22:625–629

Saino N, Romano M, Ambrosini R, Rubolini D, Boncoraglio G, CaprioliM, Romano A (2012) Longevity and lifetime reproductive success

Behav Ecol Sociobiol (2017) 71:28 Page 9 of 10 28

of barn swallow offspring are predicted by their hatching date andphenotypic quality. J Anim Ecol 81:1004–1012

Sandvik M, Rosenqvist G, Berglund A (2000) Male and female matechoice affects offspring quality in a sex-role-reversed pipefish.Proc R Soc Lond B 267:2151–2155

Sefc KM, Hermann CM, Koblmüller S (2009) Mating system variabilityin a mouthbrooding cichlid fish from a tropical lake. Mol Ecol 18:3508–3517

Silva K, Vieira MN, Almada VC, Monteiro NM (2007) The effect oftemperature on mate preferences and female–female interactions inSyngnathus abaster. Anim Behav 74:1525–1533

Smith RJ,Moore FR (2003) Arrival fat and reproductive performance in along-distance passerine migrant. Oecologia 134:325–331

Sogabe A, Ahnesjö I (2011) The ovarian structure and mode of eggproduction in two polygamous pipefishes: a link to mating pattern.J Fish Biol 78:1833–1846

Sundin J, Sagebakken G, Kvarnemo C (2013) Female mate choice is notaffected by mate condition in a fish with male care. Acta Ethol 16:189–194

Svensson I (1988) Reproductive costs in two sex-role reversed pipefishspecies (Syngnathidae). J Anim Ecol 32:929–942

Thorpe JE (1994) Reproductive strategies in Atlantic salmon, Salmosalar L. Aquac Res 25:77–87

Tobler M (2008) Divergence in trophic ecology characterizes coloniza-tion of extreme habitats. Biol J Linn Soc 95:517–528

Tomkins JL, Simmons LW (2002) Measuring relative investment: a casestudy of testes investment in species with alternative male reproduc-tive tactics. Anim Behav 63:1009–1016

Vincent A, Ahnesjö I, Berglund A (1994) Operational sex ratios andbehavioral sex differences in a pipefish population. Behav EcolSociobiol 34:435–442

Vincent ACJ, Berglund A, Ahnesjö I (1995) Reproductive ecology of fivepipefish species in one eelgrass meadow. Environ Biol Fish 44:347–361

Wendeln H (1997) Body mass of female common terns (Sterna hirundo)during courtship: relationships tomale quality, eggmass, diet, layingdate and age. Colon. Waterbird 20:235–243

28 Page 10 of 10 Behav Ecol Sociobiol (2017) 71:28