Introduction

The Crustacea is a speciose clade of marine invertebrates (Bauer 2004), many have evolved symbiotic relationships [symbiosis herein defined sensu De Bary (1879) as dissimilar organisms living together] with a wide variety of host species, and several of these relationships are parasitic (Baeza 2015; Shields et al. 2015). Our knowledge about parasitic partnerships has increased substantially over the last few decades (Rohde 2005; Baeza 2015; Shields et al. 2015), but most remain poorly characterized. Furthermore, little is known about the reproductive strategies exhibited by parasites. Our aim in this study was to focus on the parasite and improve our understanding of its mating system and reproductive performance.

Monogamy [defined as pairs of conspecifics spending extensive periods of time together sensu Wickler and Seibt (1981, 1983)] is common in marine organisms (Knowlton 1980; Vincent and Sadler 1995). Various hypotheses have been proposed to explain the adaptive value of monogamy [e.g., “territorial cooperation” hypothesis (Wickler and Seibt 1981); “mate guarding” hypothesis (Parker 1974)]. According to Baeza and Thiel (2007), monogamy is adaptive in symbiotic crustaceans inhabiting scarce and morphologically simple hosts that attain a small size (relative to the symbiont) in tropical environments where predation risk away from hosts is high. Under these conditions, movement among hosts is constrained by predation risk and host-monopolization is favored in both sexes. Thus, both sexes optimize their reproductive success by sharing a host for long periods. Under monogamy, little sexual dimorphism in body size or weaponry should exist due to the low intensity of sexual selection (Shuster and Wade 2003; Baeza and Thiel 2007). However, if these species depend strongly upon the host for nourishment, natural selection might favor extreme sexual dimorphism in size. Small size in male parasites is favored in this case, because it diverts host-derived nutrients to female partners for offspring production (i.e., a form of indirect paternal care; Andersson 1994; Shuster and Wade 2003), potentially resulting in ‘giant’ females and ‘dwarf’ males. Various studies have reported monogamous symbiotic crustaceans inhabiting relatively small and scarce hosts in environments in which off-host predation risk is high (e.g., Baeza 1999, 2008; Baeza et al. 2013). On the contrary, other studies have shown a degree of promiscuity in symbiotic crustaceans living in environments that should favor monogamy (Baeza et al. 2016 and references therein). Certainly, more studies on symbiotic crustaceans are needed before the evolutionary drivers behind their mating systems can be generalized.

Parasites are also predicted to exhibit augmented reproductive performance (= increased fecundity and reproductive output) compared to free-living relatives (Jennings and Calow 1975; Poulin 1996; Whittington 1997). Increased fecundity in parasites is thought to be a strategy for increasing transmission from host-to-host or a benefit of the stable, nutrient rich environment provided by the host (Jennings and Calow 1975; Poulin 1996; Whittington 1997). In marine parasites with indirect development, small larval stages often experience high mortality (Morgan 1995), so augmented fecundity can increase the chances of finding a host individual, as previously suggested for terrestrial trematodes with complex life cycles (Poulin 1996; Trouvé and Morand 1998). Furthermore, some marine parasite clades exhibit complex life cycles using intermediate hosts before infesting final hosts in which reproduction takes place (Sherman and Curran 2014). The use of multiple host species further decreases the chances of locating a suitable final host if that host is not abundant. Thus, increased fecundity and reproductive output would seem selectively advantageous if reproductive opportunities (i.e., hosts) are uncommon. There are few empirical studies exploring these hypotheses (Poulin 1995a, b), but in agreement with the theoretical considerations, parasitic female mites of Macrocheles subbadius have been shown to have 2.4 times the lifetime fecundity of females in a closely related free-living species (Luong and Subasinghe 2017). Similarly, endoparasitic flatworms (Platyhelminthes) experience heightened fecundity compared to their free-living relatives (Whittington 1997). In general, we expect parasites to exhibit augmented fecundity and reproductive output compared to free-living relatives.

We used the isopod Parabopyrella lata (Epicaridea: Bopyridae) and its shrimp host Lysmata boggessi (Caridea, Lysmatidae) to test our hypotheses of monogamy and heightened reproductive performance. The ‘peppermint’ shrimp L. boggessi lives in the sub-tropical eastern Gulf of Mexico and is heavily targeted by the marine aquarium ‘ornamental’ industry (Baeza et al. 2014; Baeza and Behringer 2017). The genus Lysmata is also of significant biological interest because of its unusual sexual system: protandric simultaneous hermaphroditism (Bauer 2000, 2004; Baeza 2013). In this sexual system, juvenile shrimps first develop as functional males and later develop female gonads, in addition to their male gonads (Bauer 2000).

Lysmata boggessi often harbors the parasitic isopod P. lata within its branchial chamber (Boyko 2006). The family Bopyridae, to which P. lata belongs, contains over 500 species, all of which are considered parasites (Markham and Dworschak 2005; Williams and Boyko 2012; Boyko et al. 2013). Due to the small size of the branchial chamber of L. boggessi and the low abundance of hosts in the natural environment (10–80 shrimp km−2, Baeza et al. 2014), we expect P. lata to live in monogamous pairs (see Baeza and Thiel 2007). To test this, we examined the population distribution, male–female association pattern, and parasite–host body size relationship. We also expected P. lata to exhibit high fecundity and reproductive output compared to free-living isopod species.

Bopyrid isopods exhibit a complex life cycle in which calanoid copepods serve as intermediate hosts and decapod crustaceans as the final hosts (Sherman and Curran 2014). This complex life cycle likely further decreases the chances of locating a suitable final host, especially if that host is not abundant. We compared fecundity and reproductive output estimates for P. lata with those from other closely related free-living isopod species to evaluate whether or not a parasitic lifestyle increased reproductive performance. Lastly, nothing is known about the effect of P. lata on L. boggessi reproduction, so we also tested whether or not P. lata was a parasitic castrator (a parasite capable of partially or completely halting host reproduction) as reported previously in other bopyrid isopods.

Materials and methods

Study site, sampling protocol, and shrimp measurements

Lysmata boggessi were collected at night and as by-catch once per month from December 2012 through November 2013 via roller-frame bait-shrimp trawlers in a shallow subtidal region off Homosassa Springs, west coast of Florida. During sampling, the trawler simultaneously deployed two roller-frame trawls (4.27 m height × 0.61 m width) 8–10 times between sunset and 02:30 h. Trawl durations were 30–45 min. A mesh size of 2.54 cm was used near the mouth of the trawl net and throughout the tapered body, with a finer mesh catch bag (1.91 cm) woven into the tailing end. The number of L. boggessi landed during each deployment was counted, and ~ 30 individuals from alternating deployments were haphazardly sub-sampled for further laboratory processing. The sub-sampled shrimps were fixed onboard the fishing vessel in 10% neutral buffered formalin and after 48 h, they were washed with tap water and transferred to 70% ethanol for preservation.

In the laboratory, preserved specimens were measured under a stereomicroscope (Leica MC 170 HD) using a reticule scale (carapace length, CL, precision = 0.1 mm), and sex phase and embryo development were assessed. Sex phase was determined by removing the second pair of pleopods and determining the presence (in males) or absence (in hermaphrodites) of appendices masculinae (Bauer and Holt 1998). Hermaphrodites were considered gravid if found carrying embryos underneath their abdomen. Lastly, the presence or absence of the parasite P. lata was determined by looking at the branchial chamber of each shrimp through the semi-translucent carapace. In total, 1–23 shrimps month−1 were preserved for further study of the host–parasite relationship.

Population distribution of Parabopyrella lata on its host Lysmata boggessi

A sub-sample of 52 infested shrimps was used to describe the use pattern of P. lata on its host, which includes a description of its population distribution, male–female association pattern, and host shrimp size relationships.

For this purpose, we first removed all of the parasites observed in the gill chamber of each infested shrimp and placed them in a separate petri dish for closer inspection. The gill chamber (left versus right) from which the parasites were extracted was not recorded. Female parasites were identified by their large, asymmetrical bodies that take up most of the branchial chamber of the host. The presence or absence of a male (or multiple males) was then determined for each female parasite by inspecting the female body for attached males. Male parasites were identified by their smaller, symmetrical bodies (Boyko 2006). Lastly, each host shrimp was also inspected for the presence of any additional parasites, including other much smaller recently settled juvenile males. Next, the total length of each parasite (from the tip of the cephalon to the end of the telson) was measured using a stage micrometer to the nearest 0.01 mm. Using the information above, we determined if the frequency distribution of P. lata among the shrimp population differed from a random distribution. We compared the observed distribution with a truncated Poisson random distribution (Elliott 1983).

Several shrimp hosts contained pairs of isopods (see “Results”). To determine whether the sexes were randomly distributed among pairs inhabiting the same host individual, the observed distribution was compared with the Binomial distribution. The expected random frequencies of the different sexes were calculated based on the proportion of males and females recorded in the entire population. A Chi square test of goodness of fit was used to inspect for significant differences between observed and expected (binomial) distributions (Sokal and Rohlf 1981).

Reproductive performance in Parabopyrella lata

A sub-sample of 11 female parasites carrying early-stage eggs (without eyes or obvious development of appendages) was used to describe individual-level reproductive parameters in P. lata: fecundity, egg size, and reproductive output. The presence or absence of eggs in the brooding chamber of each female was noted. If present, the total egg mass was gently removed with fine-tipped forceps from the parasitic brooding chamber and placed in a petri dish with tap water. Next, photographs were taken of a sample of 10 eggs from each brood. A stage micrometer (precision = 0.01 mm) was used to measure the maximum length (long axis length) and width (short axis length) of each egg. Egg volume was estimated with the formula for the volume of an ellipsoid (see Baeza et al. 2014), EV = 1/6(L S2 π), where L = long axis and S = short axis. Next, all eggs were counted and moved to a separate aluminum weighing boat. The total body length of each brooding female parasite (from the longest portion) was measured using a stage micrometer to the nearest 0.01 mm. The female parasite was then moved to a separate weighing boat, after all the eggs were removed from the brooding chamber. The female isopods and their corresponding embryo mass were then dried at 60 °C in a VWR Gravity Convection Oven (Model 414005-112) for a minimum of 72 h and weighed separately using an analytical balance scale (VWR-164AC) to the nearest 0.0001 g to determine reproductive output (see below).

Reproductive output was estimated as the ratio between dry weight of embryos and dry weight of the females carrying the embryos. We tested whether reproductive output increased linearly (i.e., isometric allocation) with female size. The relationship between egg dry mass and female body dry mass was examined using the allometric model y = a * xb (Hartnoll 1982). The slope b of the log–log least-squares linear regression represents the rate of power increase (b > 1) or decrease (b < 1) of the estimate of reproductive allocation with a unit of increase in female dry mass. A t test was used to determine if the estimated slope b deviated from the expected slope of unity. Before conducting the analysis, assumptions of normality and homogeneity of variances were checked and found to be satisfactory (Zar 1999).

Comparison of reproductive performance between free-living and parasitic isopods

Lastly, we conducted searches in Google Scholar (https://scholar.google.com/), Aquatic Sciences and Fisheries Abstracts (ASFA, Aquatic Sciences and Fisheries Information System [ASFIS], http://www.fao.org/fishery/asfa/faq/en), and Web of Knowledge (https://webofknowledge.com/) using all of the following term combinations: ‘bopyrid’ or ‘Bopyridae’ or ‘isopod’ or ‘Isopoda’ and ‘fecundity’ or ‘egg production’ or ‘reproduction’ or ‘clutch size’. During our search, we retrieved information only from marine isopods to compare reproductive performance between free-living isopods (regardless of the family to which they belong) and those belonging to the family Bopyridae (see Supplementary Tables 1 and 2). For each manuscript, we recorded the following information: isopod size (minimum, maximum, and average, estimated as total length [mm]), fecundity (minimum, maximum, and average, estimated as number of eggs female−1), egg size (minimum, maximum, and average, estimated as maximum diameter or volume) and reproductive output (RO, minimum, maximum, and average [%]). For symbiotic species, we also recorded the identity of the host species and host size (minimum, maximum, and average, estimated as CL, CW [carapace width], and SL [shield length], in shrimps, brachyuran crabs, and hermit crabs, respectively). Additional details can be found in Supplementary Tables 1 and 2. After completion of the data matrix, we used a general linear model (GLM) to examine the effect of lifestyle (free-living vs symbiotic [parasitic]) on the reproductive performance of isopods using fecundity (no. of eggs per female) as a proxy for reproductive performance. During the GLM analysis, fecundity was the dependent variable, lifestyle was the categorical independent variable, and female size was also included as the covariate in the model to control for female body size. Unfortunately, the lack of a detailed and well-resolved molecular phylogeny of the Isopoda, including the family Bopyridae and allies (see Boyko et al. 2013), precluded us from correcting for phylogenetic effects in our analysis (Pennell and Harmon 2013).

Results

Prevalence of Parabopyrella lata on its host Lysmata boggessi

During the sampling period, 73 of 1477 (4.94%) L. boggessi hosted at least one P. lata within their gill chambers, and when there were P. lata, they were invariably within the same gill chamber. The number of L. boggessi collected was 110–122 month−1. Prevalence varied during the year with no parasites observed from November 2012 to February 2013. Prevalence increased from March to June 2013, reaching its peak in June (19%) and then decreased to < 1% in November 2013 (Fig. 1a).

Fig. 1
figure 1

Parasite prevalence and size-frequency distribution of its host. a Prevalence of isopod parasite Parabopyrella lata in monthly samples of Lysmata boggessi. b Size-frequency distribution of male (N = 654) and hermaphrodite (N = 823) L. boggessi. c Size-frequency distribution of L. boggessi with or without P. lata. Top left lateral view of L. boggessi (Photo credit: Arthur Anker)

Full size image

Males (N = 654) and hermaphrodites (N = 823) of L. boggessi ranged between 3.5–10.1 (X ± SD = 6.62 ± 1.24) mm CL and 4.0–11.3 (8.62 ± 1.12) mm CL, respectively (Fig. 1b). Males accounted for 70 of the 73 infested shrimps and the remaining 3 were hermaphrodites. Only 1 of 3 infested hermaphrodites was ovigerous (CL = 79 mm, fecundity = 611 early-stage eggs). Taking into account the number of males and hermaphrodites in the population, isopods clearly infested males more frequently than expected by chance alone (Chi square test of independence: χ2 = 82.9, df = 1, P < 0.001). No isopods were found in shrimps < 40 mm CL. Prevalences in shrimps with CL 40–100 mm were similar, with slightly greater prevalence among shrimps 60–80 mm CL than in the other size classes. No shrimp > 100 mm CL harbored a parasite (Fig. 1c).

Population distribution of Parabopyrella lata on its host Lysmata boggessi

There were 0–5 isopods host shrimp−1 with a mean of 2 ± 0.62 in our sub-sample of 52 infested shrimps. The population distribution of P. lata was not random (observed versus truncated Poisson distribution: Chi square test of goodness of fit, \(\chi_{ 2}^{ 2}\) = 59.29, P = 0.0001). This was explained by the larger number of hosts harboring two parasites compared to the number expected by chance alone (decomposition of the Chi square test of goodness of fit; \(\chi_{ 1}^{ 2}\) = 56.83, P = 0.0001, sequential Bonferroni α = 0.025) and due to the smaller number of hosts harboring one parasite compared to the number expected by chance alone (\(\chi_{ 1}^{ 2}\) = 11.24, P < 0.0001, sequential Bonferroni α = 0.05) (Fig. 2a).

Fig. 2
figure 2

Population distribution and host-use pattern of Parabopyrella lata. a Observed and expected (Poisson) parasite frequency distribution on shrimps. b Observed and expected (binomial) distribution of male and female parasites. c Relationship between paired male and female parasite TL (mm). d Relationship between male and female parasite TL and host shrimp CL (mm)

Full size image

A total of 51 male and 52 female isopods were retrieved from the 52 infested shrimps. The isopod sex ratio was not biased toward either sex (considering only sexed isopods: sex ratio = 0.49; Fisher’s exact test, P = 0.539). A total of 6 out of 52 infested shrimps harbored a single, adult female isopod. All singleton females were found brooding embryos (N = 4 and 2 carried early [without ocular development] and late [with well-developed eyes and appendages] embryos, respectively).

A total of 43 (82.7%) of 52 infested shrimps harbored a pair of parasites. Invariably, pairs of parasites comprised one adult male and female and inhabited the same branchial chamber (Fig. 2a). All paired females were brooding embryos (N = 35 and 8 carried early and late stage embryos, respectively). No shrimp host harbored two adult males or two adult females (Fig. 2b). Taking into account the binomial distribution and the relative abundance of males and females in the studied population, the number of shrimp hosts harboring parasite heterosexual pairs expected by chance alone would have been 21. Therefore, paired shrimps were found to be heterosexual more frequently than expected by chance alone (Fig. 2b).

Paired males were 0.85–2.05 mm TL and females 4.68–9.17 mm. There was a positive correlation in TL between males and females of P. lata living in pairs (R2 = 0.54, F1,43 = 50.11, P < 0.0001) (Fig. 2c). Also, there was a significant, positive relationship between male isopod TL and host CL (R2 = 0.36, F1,43 = 24.72, P = 0.0001) and female isopod TL and host CL (R2 = 0.72, F1,49 = 98.85, P < 0.0001) (Fig. 2d).

A total of 3 of 52 infested shrimps harbored more than two isopods. Invariably, these isopods were found in the same branchial chamber. A single shrimp hosted three isopods: one adult female, one adult male, and one juvenile male. A single shrimp hosted four isopods: one adult female, one adult male, and two juvenile males. Lastly, a single shrimp hosted five isopods: one adult female, two adult males, and two juvenile males (Fig. 2a).

Reproductive performance in Parabopyrella lata

Fecundity in P. lata ranged between 2181 eggs female−1 in a specimen with TL = 5.71 mm and 5998 eggs female−1 in a specimen with TL = 6.83 mm with an average of 3660 (± 1146) eggs female−1. There was a significant, positive correlation between female TL and fecundity in P. lata (R2 = 0.62, P = 0.0022) (Fig. 3a). Average egg volume was 0.0311 mm3 (± 0.0048). The relationship between egg volume and female TL was not statistically significant (R2 = 0.071, P = 0.401) (Fig. 3b). Average female mass and egg mass were 2.10 mg (± 0.899) and 1.21 mg (± 0.413), respectively. A significant, positive correlation was found between female and egg mass (dry weight) (after log–log transformation, R2 = 0.42, P = 0.0228) (Fig. 3c). The average reproductive output was 0.61 ± 0.19, 44.4–83.8%. Reproductive output decreased with female mass, i.e., RO exhibited negative allometry, as the slope (b = 0.56, SE b  = 0.14) of the line describing the relationship between these two variables (after log–log transformation) was significantly < 1 (t test: t = 4.2, df = 1.10, P < 0.002; Fig. 3c).

Fig. 3
figure 3

Reproductive parameters of Parabopyrella lata. a Female size (TL, mm) and fecundity (egg number female−1) relationship. b Female size and egg volume relationship. c Female log dry mass (mg) and egg log dry mass relationship (after log-log transformation). Top right, view of female P. lata (scale bar = 2 mm, modified from Boyko 2006)

Full size image

Comparison of reproductive performance between free-living and parasitic isopods

A GLM using fecundity as the dependent variable, lifestyle (free-living vs symbiotic [parasitic]) as the categorical independent variable, and female TL as the covariate detected a significant effect of lifestyle on fecundity after controlling for female TL (F = 23.47, df = 1.21, P < 0.0001). The GLM also detected a significant positive effect of female TL on fecundity (F = 17.17, df = 1.21, P < 0.0001). In this analysis, the interaction term was significant (F = 16.71, df = 1.21, P < 0.0001) (Fig. 4). In turn, while RO was limited to 44–84% (average = 61%) of body weight in P. lata, in those free-living isopods in which this reproductive trait has been reported, RO is limited to about 11–17% of body weight (Supplementary Tables 1 and 2).

Fig. 4
figure 4

Fecundity in free-living and parasitic (Bopyridae) marine isopods. a Least square mean (LSM) fecundity (± SE) of parasitic (N = 13) and free-living (N = 12) isopods. b Female TL (mm) and fecundity relationship in parasitic (N = 13) and free-living (N = 12) isopods. See Supplementary Tables 1 and 2 for details

Full size image

Discussion

We hypothesized that the isopod P. lata was socially monogamous, given the host characteristics, and our results confirm this. Paired isopods shared the same host gill chamber with a frequency greater than expected by chance alone and pairs were invariably heterosexual. These results agree with previous reports for other socially monogamous symbiotic crustaceans (other bopyrid isopods: Roccatagliata and Lovrich 1999; Griffen 2009; Cericola and Williams 2015; Romero-Rodríguez et al. 2016, non-bopyrid isopods: Baeza 1999, 2008; Baeza et al. 2013, among others). Below, we discuss the association between P. lata and L. boggessi, and offer evidence for (and against) the notion of long-lasting monogamy in P. lata.

First, the significant and tight relationship found between isopod TL and host shrimp CL and between the TLs of paired male and female isopods sharing the same host individual argue in favor of long-term male–female pairing in P. lata. The conditions determining size-assortative heterosexual pairing and symbiont-host pairing in symbiotic species are not fully understood (Baeza and Thiel 2007; Baeza 2008). However, growth restrictions imposed by the host on long-term symbionts often explains such host-symbiont and male–female parasite body size relationships (Cash and Bauer 1993; Baeza 1999, 2008, Baeza et al. 2013). For instance, in the symbiotic and monogamous brachyuran crab Pinnixa transversalis and caridean shrimp Pontonia margarita, male body size explains 63.8 and 77.6% of variation in female body size, respectively (Baeza 1999, 2008). The values above are similar to those observed in female but not male P. lata inhabiting L. boggessi. The tight, size-assortative isopod pairing and relationship between host CL and isopod TL in P. lata suggest that the association between shrimp and host individuals and between male and female isopod pairs is temporally stable. If male and female isopods were not sharing the same host individuals over long periods of time, the tight correlations above would be difficult to explain.

Second, female P. lata were found cohabiting with males in the same host individual regardless of the stage of development of the eggs they were brooding, which also argues in favor of long-term male–female pairing. In promiscuous or polygamous species, heterosexual pairing is also observed (e.g., Diesel 1988; Van Der Meeren 1994), but pairing is short-term, males abandon females shortly after mating, and males are often found with females close to molting and ready to spawn (Baeza 2010). By contrast, male P. lata associate with females at all times during their brooding and reproductive cycle, regardless of egg stage. If males were not monogamous and were roaming among host individuals in search of extra-pair copulations, females paired with males should only be brooding late embryos or no embryos. The absence of a relationship between female brood stage and the presence of males further supports the theory that P. lata exhibits a monogamous mating system in which the males remain with their sexual partners.

Third, P. lata exhibits ‘reverse’ sexual dimorphism in TL where males attain average and maximum TLs considerably (~ 80%) smaller than females. In symbiotic and free-living crustaceans with polygamous mating systems, males often attain a much larger size than females, because size often determines competitive superiority (Shuster and Wade 2003). By contrast, the lower intensity of male competition characteristic of monogamous mating systems does not drive major size differentiation between the sexes (Shuster and Wade 2003; Baeza and Thiel 2007; Baeza 2008). The observation of reverse sexual dimorphism agrees with the notion that P. lata is a monogamous species in which males engage in ‘indirect parental care’, freeing resources for females to increase reproductive output by constraining their own growth (Andersson 1994; Shuster and Wade 2003; Baeza and Thiel 2007).

While most characteristics of the association between P. lata and L. boggessi support monogamy in P. lata, solitary females were also found brooding eggs. This suggests that some males may exhibit a promiscuous behavior, roaming among hosts in search of extra-pair mating opportunities. Presumably, if males were not switching hosts, solitary females would not have been observed brooding eggs, as reported for other monogamous crustaceans (e.g., Baeza et al. 2013). However, male P. lata lack swimming setae or well-developed pereopods for swimming or crawling efficiently between host individuals, in contrast to other promiscuous symbiotic species (see Ocampo et al. 2017). If not moving between hosts, the presence of solitary brooding females could also be explained by small, male P. lata suffering higher sex-specific mortality.

Environmental sex determination or sex change could play an important role in the establishment of monogamous pairs in P. lata, as shown for other bopyrid isopods (Tsai et al.1999; Cericola and Williams 2015). For an example of the former, cryptoniscus larvae colonizing a non-infested L. boggessi host might mature as female, while larvae colonizing a host infested with female bopyridium might develop as male, as suggested for Athelges takanoshimensis (McDermott 1991; Cericola and Williams 2015; see also Reinhard 1956). Alternatively, if P. lata undergoes sex change (e.g., sequential protandric hermaphroditism), the pioneering larva might mature as male but transition to female, if the colonized host has not been previously infested by a conspecific, while a larva infesting a host individual already harboring a female might mature and remain a male. The latter has been suggested for the protandric cymothoid Ichthyoxenus fushanensis (Tsai et al. 1999). It must be noted that some bopyrids appear to exhibit early sex determination (Hiraiwa 1936), and we do not know with certainty the mode of sex determination for P. lata. The sexual system, the possibility of environmental sex determination, and the cues used by P. lata to find host individuals and decide gender expression during and after host colonization remain to be experimentally addressed.

Fecundity, but not egg size, increased with TL in P. lata as reported before in every species of bopyrid isopod in which fecundity and egg size have been studied (see Supplementary Table 1). Bopyrid isopods exhibit moderate to high variability in both fecundity and egg size. The average fecundity we report for P. lata is similar to that reported for other species. The mean reproductive output (RO) of P. lata was also similar, albeit slightly, greater than that reported for the mud-shrimp Orthione griffenis (Smith et al. 2008), the only other bopyrid for which this has been measured. Fecundity and RO in P. lata and other bopyrid isopods is considerably higher than that reported for free-living isopods (Supplementary Table 2 and references therein). These comparisons support the notion that parasites are more fecund than their free-living counterparts. However, we argue in favor of additional studies of fecundity, size, and lifestyle coupled with the development of robust phylogenetic hypotheses in closely related free-living and parasitic isopods to improve our understanding of those intrinsic (e.g., size) and extrinsic (e.g., ecological) conditions favoring and constraining fecundity and egg size in parasitic and other symbiotic marine organisms. Whether augmented reproductive performance, as we observed in P. lata, is a strategy for increasing transmission from host-to-host (Poulin 1996; Trouvé and Morand 1998) or results from the ‘stable, nutrient rich environment’ provided by the host (Jennings and Calow 1975; Poulin 1996; Whittington 1997) remains to be examined.

Reproductive output and female body mass allometry scaled negatively in P. lata, suggesting female isopods did not proportionally increase allocation to their brood mass with increase in body mass. The relationship between reproductive output and size is poorly explored in isopods, but in other crustaceans (e.g., brachyuran crabs, caridean shrimps), brood mass is isometric or nearly isometric to female body mass. The negative  scaling of reproductive output with size in P. lata could be explained by limitations on space available for yolk accumulation in the body cavity of large female isopods, as suggested for some brachyuran crabs (Hines 1992). The importance of food availability and quality for brood production is also well-known in crustaceans (Calado 2008). Food limitation can constrain the relationship with increasing female size (see Bolaños et al. 2012). If nutrition derived from the host or the ability of a P. lata female to acquire food from its host diminishes with increased parasite size, small but not large females might have access to enough resources to produce and accumulate yolk entirely filling their body cavities. Unfortunately, little is known about how bopyrid isopods derive nourishment from their hosts (Walker 1977; Astete-Espinoza and Caceres 2000; Neves et al. 2004), so future studies are needed to understand if nutrition becomes limiting for large females.

Parabopyrella lata was found infesting males with greater frequency than expected by chance and the few infested hermaphrodites did not brood embryos (with one exception). Furthermore, a tight relationship between host and isopod size was also observed. This suggests that P. lata preferentially infests males and is a special type of parasite: a parasitic castrator that halts transition from the male to hermaphroditic phase in L. boggessi. Several life history traits common to other macro-parasitic castrators were also observed in P. lata, supporting the contention that P. lata is a parasitic castrator. Our results show that P. lata establish long-term associations with their host individuals, as occurs in many other parasitic castrators (Kuris 1974). Castration of the host shrimp was ‘intensity-independent’, meaning that a single female isopod (and its diminutive male) completely hampered female reproduction of its shrimp host, as occurs in many other parasitic castrators (Kuris 1974; Lafferty and Kuris 2009; Williams and Boyko 2012). P. lata also attains a large size relative to the host, grows concomitantly with the host, and allocates a large percentage of its energy (61%) to reproduction, as noted for many other macro-parasitic castrators (Kuris and Lafferty 2000).

Bopyrids also negatively impact other host life history traits, including reproductive performance (Reinhard 1956; Oliveira and Masunari 1998). They either increase or decrease growth rate and condition of host individuals (Van Wyk 1982; Somers and Kirkwood 1991; Astete-Espinoza and Caceres 2000; Lester 2005; Hall et al. 2007), alter morphology (i.e., carapace shape) and pigmentation (Williams and Brown 1972; Roccatagliata and Jordá 2002), impact gonad development, molting frequency, and various secondary sexual characters (Beck 1979, 1980; Wenner and Windsor 1979; Courtney 1991; McDermott 1991; Raibaut and Trilles 1993; Oliveira and Masunari 1998; Jarrin and Shanks 2008; Smith et al. 2008). Interestingly, suppression of ovarian development by bopyrid castrators is not permanent (Roccatagliata and Lovrich 1999). For example, after experimental extraction of the bopyrid parasite Pseudione tuberculata, regeneration and maturation of the ovary was observed in the king crab Paralomis granulosa (Roccatagliata and Lovrich 1999). However, the mechanism by which bopyrid isopods affect host traits is far from understood.

Conclusion

Our results support the hypotheses that many symbiotic crustaceans inhabiting relatively small and scarce host species are monogamous (Baeza and Thiel 2007) and that parasitic species have higher fecundity than their free-living close relatives (Hines 1982; Jay 1989; Trouvé and Morand 1998; Salas-Moya et al. 2014). We have also shown that P. lata is a parasitic castrator of L. boggessi. Our findings highlight the importance of integrating studies of host and symbiont life history to fully understand how their interdependence drives their evolution.