Marine protected areas rescue a sexually selected trait in European lobster

Abstract Marine protected areas (MPAs) are increasingly implemented worldwide to maintain and restore depleted populations. However, despite our knowledge on the myriad of positive responses to protection, there are few empirical studies on the ability to conserve species’ mating patterns and secondary sexual traits. In male European lobsters (Homarus gammarus), the size of claws relative to body size correlates positively with male mating success and is presumably under sexual selection. At the same time, an intensive trap fishery exerts selection against large claws in males. MPAs could therefore be expected to resolve these conflicting selective pressures and preserve males with large claws. We explored this hypothesis by contrasting claw size of males and females in three pairs of MPAs and nearby fished areas in southern Norway. By finding that male lobsters have up to 8% larger claws inside MPAs compared to similarly sized males in fished areas, our study provides evidence that MPAs rescue a secondary sexual trait. Recovery from harvest selection acting on claws is the most likely explanation; however, the higher abundance of lobster inside MPAs does not rule out a plastic response on claw size due to increased competition. Regardless of the underlying cause, our study demonstrates (a) the value of protected areas as a management tool for mitigating fisheries‐induced evolution and (b) that MPAs help maintaining the scope for sexual selection in populations with vulnerable life histories and complex mating system.

. In large terrestrial animals subjected to trophy hunting or ivory trade, selective harvest of primarily superior and sexually dominant males has been shown to induce artificial selection and hence evolution towards smaller horn size, reduction in male body size or, in the case of elephants, loss of tusks (Chiyo et al., 2015;Coltman et al., 2003;Martin, Festa-Bianchet, Coltman, & Pelletier, 2016;Pigeon, Festa-Bianchet, Coltman, & Pelletier, 2016). Such conspicuous traits are fundamental to the outcome of competitive interactions and are the results of strong sexual selection (Swain, Sinclair, & Mark Hanson, 2007;Wilber, 1989;Woolmer, Woo, & Bayes, 2013).
Well-functioning mating systems are perceived as the foundation for population resilience and growth rate (Allendorf & Hard, 2009). Recent evidences suggest that reducing the opportunity for sexual selection by removing individuals with higher expression of secondary sexual traits can lower population fitness and increase the extinction risk under environmental change (Cally, Stuart-Fox, & Holman, 2019;Knell & Martínez-Ruiz, 2017;Lumley et al., 2015;Plesnar-Bielak, Skrzynecka, Prokop, & Radwan, 2012). This is because secondary sexual traits are likely to be honest signals of "good genes" reflecting the owner's overall genetic match to the environment (Weatherhead & Robertson, 1979). If the environment changes, the best adapted individuals should afford the highest expression of secondary sexual traits and thus gain mating success (Lorch, Proulx, Rowe, & Day, 2003;Siller, 2001;Whitlock & Agrawal, 2009). Consequently, sexual selection can improve population mean fitness and be able to drive adaptation at a much higher rate than natural selection alone (Lorch et al., 2003;Lumley et al., 2015). Nature reserves, or protected areas, should have the potential to restore sources of individuals that are not affected by harvest selection. In trophy-hunted bighorn rams (Ovis canadensis), individuals harvested near protected areas in Canada had larger average horn size compared to rams shot far from protected areas (Pelletier, Festa-bianchet, Jorgenson, Feder, & Hubbs, 2014), and in Zimbabwe, horn size of impala (Aepyceros melampus) decreased with distance from a national park (Crosmary et al., 2013). In oceans and coastal areas worldwide, marine protected areas (MPAs) are increasingly being implemented to restore depleted populations, improve ecosystem health and benefit fisheries through spillover effects (Hastings & Botsford, 2003;Pendleton et al., 2018). Although there is mounting evidence of how number, biomass, size and age of fish and invertebrate species within MPAs are often much greater than in comparable areas open to fishing (Baskett & Barnett, 2015;Gillespie & Vincent, 2019;Halpern, 2003;Lester et al., 2009;Russ, Cheal, & Dolman, 2006), it is rare to assess the potential for MPAs to preserve or restore secondary sexual traits. Thus, this warrants further investigation since secondary selected traits are likely affected by fishing, especially in the light of many recent studies demonstrating harvest selection on behavioural or morphological traits independently of body size (Alós, Palmer, Linde-Medina, & Arlinghaus, 2014;Arlinghaus et al., 2017;Biro & Sampson, 2015). In salmonid fishes, secondary sexual traits (body depth) have also been shown to correlate with increased catchability, which may affect the opportunity and strength of sexual selection (Hamon & Foote, 2005;Kendall & Quinn, 2013).
The secondary sexual traits of many harvested fish species may be cryptic and sometimes poorly described or identified. However, some commercially important crustaceans have dimorphic chelaeor claws in adults with a major molar-toothed (crusher) claw and a minor incisor-toothed (cutter) claw. In most species, males grow larger and heavier claws than females and are considered secondary sexual traits (Hartnoll, 1974;Mariappan, Balasundaram, & Schmitz, 2000;Stein, 1976;Templeman, 1935). The claws are tools used in foraging and in defence against predators, but are also weapons used in male-male conflicts (armaments) and signals indicating fighting ability and attractiveness towards females (ornaments) (Atema, 1986;Elner & Campbell, 1981;Jivoff, 1997;Sneddon et al., 1997). A recent field study on European lobster (Homarus gammarus) found that large claws increase male mating success. Specifically, sexual selection seems to be acting more strongly on relative claw size (with respect to body size) than on absolute claw size or body size (Sørdalen et al., 2018). Furthermore, the strength of sexual selection in males appeared to be higher inside a marine protected area relative to a nearby, heavily fished area (Sørdalen et al., 2018). A telemetry study conducted in the same fished area found that relative claw size was positively correlated with capture probability and hence mortality in the trap fishery (Moland, Carlson, Villegas-Rios, Wiig, & Olsen, 2019). This means that harvest selection against large claws may become effective as soon as lobsters reach the minimum size limit in the fishery. Thus, both sexual selection and harvest selection have been identified to act on the same trait, but in opposite directions. Hence, MPAs should be able to preserve males with large claw phenotypes, assuming any genetic component underlying claw size is not strongly reduced by past fishing. The effect on females is expected to be smaller; harvest selection has not been studied in female lobsters but is presumably weaker than on males because all egg-bearing females are protected, and they have lower catchability than males (Moland, Ulmestrand, Olsen, & Stenseth, 2013). In this study, we address these hypotheses by comparing the relationship between body and claw size of lobsters inside and outside three lobster reserves established in 2006. By confirming our prediction that lobsters, particularly males, have larger claws relative to body sizes inside protected areas, this study documents the usefulness of MPAs to preserve a trait under sexual selection.

| Species and study system
European lobster (hereafter, lobster) are large, long-lived sexually dimorphic crustaceans in temperate waters distributed from the north of Norway to Morocco in North Africa, including the Mediterranean Sea (Triantafyllidis et al., 2005). Males grow faster, mature at smaller size and have relatively larger greater chelae (hereafter, claws) than females (Debuse, Addison, & Reynolds, 2001;Lizárraga-Cubedo, Tuck, Bailey, Pierce, & Kinnear, 2003). The average age of large (150-170 mm carapace length, CL, measured from rear of the eye socket to the rear of the carapace) males and females is estimated to be 31 and 54 years, respectively (Sheehy, Bannister, Wickins, & Shelton, 1999). One of the largest specimens was estimated to be 650 mm (total length, TL, measured from tip of rostrum to mid-tail; Figure 1) based on recovery of a large crusher claw (360-370 mm long) in Skagen, Denmark (Wolff, 1978). The lobster is one of the most valuable and sought-after species in Northern Europe's commercial and recreational fisheries. In Norway, the lobster catch rates have declined by 65% from the 1950s to 2000s and is today at the lowest record in history with no sign of recovery (Pettersen, Moland, Olsen, & Knutsen, 2009). In response, the fishery is now mostly A ban on the harvest of egg-bearing females was implemented in 2008, along with an increase in minimum legal size to 250 mm total length. In 2017, a maximum size limit at 320 mm TL was introduced for lobster caught along the Norwegian Skagerrak coastline (Sørdalen et al., 2018).

| Sampling design and lobster data
We sampled lobster as part of a standardized capture-mark-recapture sampling programme conducted annually by the Norwegian  were sexed, and total length (TL) was measured to the nearest millimetre as a measure of body size. We measured the width of the major crusher claw as the widest part of the crusher claw (across the "palm" beneath the top of the ridge of dactyl) to obtain a measure of claw size (CW; see Figure 1). Each lobster was individually tagged with externally visible T-bar tags (TBA2, 45 × 2 mm; Hallprint) and released at the sampling site.

| Statistical analyses
The catch-per-unit-effort (CPUE) was plotted for a visual comparison of population density between reserves and fished areas over the last 13 years. To provide a measure of mean density for the years with claw measures, we averaged the CPUE over the last 3 years (2017, 2018 and 2019). To investigate the effects of protection on claws, we first tested whether the probability of missing one or both claws differed between areas (Status), adjusted for body size (TL). Injuries to and loss of appendages are common among decapod crustacean and are exacerbated by fishing methods (Juanes & Smith, 1995). General linear models were fitted separately for males and females with claw loss as the binomial response variable: Clawloss ∼ 0 + 1 Status + 2 TL F I G U R E 5 Marine protected area (MPA) effect on male claw size. The curves indicate the percentage difference in predicted male claw size (claw width, CW, in mm) in MPA relative to fished area in Vestfold County (see Table 2 for model summary Note: Summary of the linear models between claw size (crusher claw width, CW, in mm) and body size (total length, TL, in mm) of male and female lobster. The fished area and the Aust-Agder County are set as reference levels (ref. Table 1).

Significant values are indicated in bold.
Abbreviation: MPA, Marine protected area.

TA B L E 2 Model estimations
Second, we used a linear model to test the prediction that male lobster in MPAs have larger claws than conspecifics in the contrasted fished areas. Again, we fitted the same model to the data on female lobsters. The following a priori-defined general linear model structure was applied: CW (crusher claw width) is the response variable, with the factor County accounting for spatial differences among the three counties (Aust-Agder, Vestfold and Østfold) as levels. As a first step, we censored any lobsters that had very small crusher claws, most likely resulting from a regeneration of a lost claw. Claw loss inhibits growth in crustaceans (Moriyasu, Landsburg, Wade, & Maynard, 1999), and in the American lobster (Homarus americanus), regenerated claws are typically smaller than intact claws, even after multiple moults (Cheng & Chang, 1994). In order to distinguish lobster with regenerated claws from those with naturally small claws, we conducted the following analysis: we assumed the residuals from model 1 to be normally distributed with a mean of zero. We then sequentially removed the individual with the largest negative residual value and refitted the model until the largest negative residual was equal to or smaller than the largest positive residual value. This method of classifying lobster with regenerating claws would ensure that we are conservative in identifying individuals (with regenerating claws) that should be excluded in our final analysis (34 females and 19 males were thus excluded; see Figure S1). We focused our analysis on lobster larger than the minimum size limit (250 mm TL) since harvest selection is assumed to only operate on legal-sized lobster (Fernández-Chacón et al., 2020). Further, the fished areas had a truncated size distribution; of the 306 lobsters above the maximum legal size limit (320 mm TL), only 4.6% were caught in the fished areas. Thus, we restricted our models to compare only the overlapping size range between fished areas and MPAs (TL max males = 362 mm, TL max females = 355 mm; Table 1). The data and predictions from the model also including the large MPA lobsters are shown in Figure S2. We reran the models with the final dataset and focused on the model terms involving Status (MPAs or fished area). If harvest selection acts on relative claw size, a significant interaction effect can be expected due to cumulative selective mortality that should lead to increasing difference with age (which is assumed to be closely correlated with body size).
The interaction was dropped if nonsignificant (p > .05). The three sampling years were pooled as a preliminary model revealed no year effect on claw width (results not shown). All statistical analyses were performed in R 3.5.1 (R Core Team, 2018).

| RE SULTS
All MPA populations have responded well to protection with notable increases in mean catch-per-unit-effort (CPUE) for legal-sized lobster. From 2017 to 2019, the same years as claw measurements were taken, and the lobster catches in the MPAs were 5.08 times higher in Aust-Agder, 3.58 times higher in Vestfold and 2.87 times higher in Østfold than in their respective fished areas (Figure 3).
In total, 2,656 lobster were fished in the three counties in the scien- the fished areas (3.35%) (see Figure S1). Lobster identified as having regenerated claws were then excluded from the following analysis.
The size of claws increased more rapidly with increasing body size for males in the MPAs than in the fished area ( Figure 4; Table 2;  Table 2).

| D ISCUSS I ON
Here, we show that MPAs are home to lobster with larger claws

| Claw size in relation to protection and abundance
The larger claws of lobster in the MPA are most likely a reflection of the high and selective fishing mortality outside MPAs CW ∼ 0 + 1 TL + 2 Status + 3 County + 4 TL:Status (Fernández-Chacón et al., 2020;Moland et al., 2019). Larger claws in males are known to indicate social dominance (Skog, 2009a) where even a small difference in claw size dictates victory in contests and a superior status in a hierarchy (Atema & Cobb, 1980;Van Der Meeren & Uksnøy, 2000). Locally dominant males will also successfully attract and mate with females in the wild (Karnofsky & Price, 1989), and it has previously been shown that claw size is a sexually selected trait in the same study populations (Sørdalen et al., 2018). Therefore, aggressive behaviour towards conspecifics and defence of resources, such as baited traps, have been suggested to be underlying mechanisms driving the fishery selection.
Individual behavioural traits (boldness, aggression, activity, sociability) are increasingly being recognized as determinants of catchability in fisheries (Arlinghaus et al., 2017;Diaz Pauli & Sih, 2017). In crayfish (Cherax destructor), bolder individuals are more likely to be attracted to and captured in baited traps because they also grow faster and require more food (Biro & Sampson, 2015). Furthermore, correlations between boldness (i.e. the propensity to take risks) and strength in the expression of secondary sexual traits have been shown in some fish (Fabre, GarcÍa-Galea, & Vinyoles, 2014;Godin & Dugatkin, 1996) and lizards (Putman, Azure, & Swierk, 2019). Thus, sexual selection may favour certain personality traits (e.g. boldness) associated with the achievement of strong expression of secondary sexual traits (Fabre et al., 2014), such that claw size may be correlated with dominant behaviour that increases the catchability. Lastly, passive gears will be selective to some extent and may therefore not representatively sample the populations we are studying, although we regard it as unlikely that capture selection related to morphology affected our results because we analysed a restricted size range ters are abundant (Debuse, Addison, & Reynolds, 2003). In other crustaceans, the expression of claws has been shown to develop plastically in response to variation in diet (prey) and temperature (Baldridge & Smith, 2008;Edgell & Rochette, 2009;Smith, 2004;Smith & Palmer, 1994). The relative influence of plasticity or harvest selection on claw size in lobster is unknown, but density-dependent phenotypic plasticity is most likely pulling in the same direction as harvest-induced selection.
The results showed that relative claw size of females also differed between protected and fished areas, although the effect was much weaker compared to that of males. Moland et al. (2019) did not investigate whether the fishery is selective on female claw size, which would have been helpful in elucidating whether harvest selection is acting on the same traits in males and females, which our results indirectly suggest. Female lobster also use their claws in frequent fights and can be even more aggressive and cause more harm than males, but their claws grow slower, and by the onset of sexual maturation, females trade off enlarged claws with a broader abdomen and egg production (Debuse, Addison, & Reynolds, 1999;Skog, 2009b). Sexual selection is therefore likely to favour male claws as a primary male secondary sexual trait (Atema, 1986;Sørdalen et al., 2018), which could also (at least partly) explain the increased natural mortality rate of males compared to females . Indeed, injury has been found to be a significant predictor of shell diseases that affect males more than females off the coast of Devon, UK (Davies et al., 2014). Yet, in our study, claw loss seems to be affecting females in MPAs more than females in fished areas, whereas there were no differences in males. This suggests that density and crowding effects might act differently on the sexes, particularly bearing in mind that fighting among females can be more intense (Skog, 2009b) and more frequent when shelters are abundant (Debuse et al., 2003). Regardless, male lobsters are more catchable than females  and protection of egg-bearing females ensures that females experience lower fishing mortality rates than males (Jury, Pugh, Henninger, Carloni, & Watson, 2019). Consequently, fisheries selection on females, and female claw size, might also be weaker.

| The potential of MPAs for preserving secondary sexually selected traits
The removal of dominant males is likely to disrupt the hierarchical order and subsequently the mating pattern in clawed lobster.
Claw size and body size are strong predictors of male mating success, yet these traits have shown to have little influence on male success in fished areas (Sørdalen et al., 2018). This is likely be- Additionally, spillover by males with attractive phenotypes to harvested areas (where such males are depleted) can strengthen sexual selection through dispersal and gene flow if they are able to reproduce before they are harvested (Baskett & Barnett, 2015;Pelletier et al., 2014). The capacity of MPAs to buffer fisheries-induced evolution will depend on the amount of interchange between protected and nonprotected areas, which so far has proven difficult to demonstrate (for a review, see Lorenzo, Claudet, & Guidetti, 2016).
One study found a tendency for larger female lobsters to spill-into the protected areas of this study system (Thorbjørnsen et al., 2018).
Since the MPAs house higher quality males, that is males of larger size and with larger claws, such patterns could be due to mate attraction. It would also be necessary to disentangle the harvest selection and the potential density effect on morphology, using more and perhaps larger MPAs that are less likely to be impacted by fishing. Studies on trait heritability should give us helpful insights into the underlying mechanisms governing claw traits and how genetics versus plasticity can shape claws under high population density. From an evolutionary perspective, high plasticity could be beneficial to slow down genotypic change.
The implications of the results from this study are twofold.
First, our study reveals how fisheries-induced selection against a male sexual character can drive population changes in such traits.
Second, it shows that marine protected areas can rescue secondary sexual traits in harvested populations. The prerequisite is that fishing has not effectively eroded the mechanisms driving phenotypic variation in claw size (i.e., genetic diversity). The discrepancy in trait expression between protected and fished areas also serves as a strong warning signal about unintended consequences of selective fishing. MPAs with animals not selected by fishing will exchange individuals and genotypes with surrounding areas and can therefore be effective in curbing undesired phenotypic selection from harvesting (Baskett & Barnett, 2015;Baskett et al., 2005;Dunlop, Baskett, Heino, & Dieckmann, 2009). Fisheries managers have largely focused on abundance, size/age and composition of target species within protected areas, yet monitoring of changes in sexually selected traits can perhaps be an equally good or additional measure of population status. When the phenotypic variance in sexually selected traits increases after harvesting ceases, as we show in this study, it is reasonable to assume that genetic diversity is also maintained (Carr & Reed, 1993;Quinn, Wing, & Botsford, 1993), because it often plays a key role in stabilizing social systems and maintaining sexual selection. Harvesting refuges like marine protected areas, if well designed and managed, should therefore relax or even reverse the effects of harvest selection or curb fisheries-induced selection with evolutionary consequences (Rowe & Hutchings, 2003;Tenhumberg, Tyre, Pople, & Possingham, 2004).

ACK N OWLED G EM ENTS
We are grateful to the three anonymous referees whose constructive and insightful comments much improved the quality of the manuscript. We would also like to thank the lobster crew at Flødevigen Institute of Marine Research who helped us obtaining claw measurements. This study was supported by the University of Agder through funding from the Norwegian Ministry of Education and through Centre for Coastal Research (CCR).

CO N FLI C T O F I NTE R E S T
None declared.

E TH I C A L A PPROVA L
The capture-release and tagging of lobster were carried out under