INTRODUCTION
Caligid copepods, commonly referred to as ‘sea lice’, are ectoparasitic on fish and, occasionally, on other marine metazoans. They are most commonly found on the skin and gills of fish, and are a source of concern for the fish farming industry because of the negative impact they have in aquaculture. For instance, Costello (Reference Costello2009) estimated the global cost associated with sea lice in salmonid aquaculture to be around 6% of the value of production. Such an impact is significant and justifies the focus of studies on sea lice biology on the few species that infect salmonids, i.e. Lepeophtheirus salmonis (Krøyer, 1837), Caligus rogercresseyi (Boxshall and Bravo, 2000) and Caligus curtus O. F. Müller, 1785 (see Boxaspen, Reference Boxaspen2006).
The study of other species of caligid copepods is, however, equally important, not only for the broad host range they present, but also because of the high infection levels they frequently exhibit. Among such species is Lepeophtheirus pectoralis (Müller O. F., 1776), a parasite commonly found infecting flatfish species, such as Platichthys flesus (Linnaeus, 1758) (European flounder), Pleuronectes platessa Linnaeus, 1758 (plaice) and Limanda limanda (Linnaeus, 1758) (common dab) (see Boxshall, Reference Boxshall1974a ). Lepeophtheirus pectoralis has been reported from most of the Atlantic distributional range of P. flesus, with records from localities as far north as Icelandic waters, from the western part of the Baltic Sea, into the English Channel (see e.g. Boxshall, Reference Boxshall1974a ), and south as far as the Portuguese coast (see e.g. Marques et al. Reference Marques, Santos and Cabral2006a , Reference Marques, Teixeira and Cabral b ; Cavaleiro and Santos, Reference Cavaleiro and Santos2007, Reference Cavaleiro and Santos2009). This species displays a particular spatial distribution pattern on the host, with the gravid females typically found on the posterior surface of the pectoral and pelvic fins, and the males and immature individuals over the rest of the body surface and fins (Boxshall, Reference Boxshall1974a ). The high infection levels reported in the literature (see e.g. Cavaleiro and Santos, Reference Cavaleiro and Santos2007, Reference Cavaleiro and Santos2009) suggest that the infection with L. pectoralis may result in significant economic losses for fish farm owners, as fish are usually stocked at high densities in aquaculture systems. Moreover, infection has been associated with skin laceration, haemorrhaging and fin erosion (see Scott, Reference Scott1901; Boxshall, Reference Boxshall1977). Despite the extensive body of knowledge already available on this parasite, namely on morphology (Boxshall, Reference Boxshall1974a ), life cycle (Boxshall, Reference Boxshall1974b ), ecology (Boxshall, Reference Boxshall1976), population dynamics (Boxshall, Reference Boxshall1974c , Reference Boxshall d ) and associated histopathology (Boxshall, Reference Boxshall1977), the reproductive biology, apart from mating behaviour (Anstensrud, Reference Anstensrud1990a , Reference Anstensrud b ), is still poorly characterized.
It is known that the males establish precopula with postchalimus larvae (Anstensrud, Reference Anstensrud1992), and that copula takes place once the female has moulted and developed a hardened exoskeleton (Anstensrud, Reference Anstensrud1990b ). The male then seals the genital openings of the female (Anstensrud, Reference Anstensrud1990a ). The females, which settle in the pectoral and pelvic fins of the host, shed several egg sacs (up to 5 in the laboratory, Boxshall, personal communication) before dying.
From his sampling of plaice, collected off the coast of Yorkshire, UK, Boxshall (Reference Boxshall1974c ) reported marked seasonal variations in the infection levels of L. pectoralis. Prevalence and intensity followed a very similar pattern during the 2 years of sampling, with the minimum recorded for winter and early spring and the maximum for mid late summer. This, along with differences in the maturation times and longevity between summer and winter females, and the existence of two peaks per year in the percentage of females that carried egg sacs, led him to propose a system of alternation of generations for L. pectoralis. Accordingly, overwintering females had a maximum life-span of 10 months, the majority producing eggs in spring, while females that hatched in May were able to produce eggs by July or August, implying a growth and maturation time for these summer females of 2 or 3 months. Females of both generations would die after producing several egg sacs, possibly of physiological exhaustion.
Later studies with flounder (van den Broek, Reference van den Broek1979; Schmidt et al. Reference Schmidt, Zander, Körting and Steinhagen2003a ) also revealed seasonal variations in infection levels, albeit sometimes slightly different to those reported previously for the plaice. For marine flounders from the Portuguese coast (Cavaleiro and Santos, Reference Cavaleiro and Santos2009), the prevalence and abundance of adult parasites were lowest during autumn, with a slight increase during winter and maximum values during spring and summer. It should also be taken into account that factors such as salinity (Möller, Reference Möller1978; Wichowski, Reference Wichowski1990) or the concentration of xenobiotics or mercury in the muscle of the host (Schmidt et al. Reference Schmidt, Zander, Körting and Steinhagen2003b ) have also been shown to have an influence on infection levels.
Gaining an understanding of these seasonal fluctuations, coupled with information on reproductive effort, i.e. fecundity (F), egg size and total reproductive effort (TRE), is crucial for the successful implementation of sustainable aquaculture systems, and will improve our understanding of when epizootic outbreaks are likely under particular environmental conditions. Moreover, if L. pectoralis is to be used successfully as a bioindicator (as proposed by Schmidt et al. Reference Schmidt, Zander, Körting and Steinhagen2003a , Reference Schmidt, Zander, Körting and Steinhagen b , Reference Schmidt, Zander, Körting and Steinhagen c ), its natural dynamics must also be well-understood. Thus, the present study aimed to characterize the reproductive effort of L. pectoralis and to test the hypotheses on the reproductive biology of this parasite in the literature, namely the existence of alternating generations with differences in maturation time (nearly 10 months vs 2–3 months), longevity and breeding season (spring vs mid-summer) (see Boxshall, Reference Boxshall1974c ). In addition, comparisons were made with patterns in reproductive effort previously reported for a copepod parasitic on an invertebrate host, i.e. Octopicola superba Humes, 1957, a parasite of the common octopus, Octopus vulgaris Cuvier, 1797 (Cavaleiro and Santos, Reference Cavaleiro and Santos2014). Other than evaluating whether L. pectoralis is mainly an r- or K-strategist, the existence of a trade-off between egg number and size, and the possible influence of macroenvironment (the environment of the host) and microenvironment (the environment of the parasite, that is, the host) (sensu Rohde, Reference Rohde1984) in egg production (i.e. egg number and size and TRE) were also analysed, so that a comprehensive perspective could be developed.
MATERIALS AND METHODS
The methodology largely followed that described in Cavaleiro and Santos (Reference Cavaleiro and Santos2014), as one of the main aims was to evaluate to what extent the strategy of egg production of L. pectoralis approached that reported for a copepod parasitic on an invertebrate host. However, some adjustments were made and, for this reason, it is briefly described here.
Collection of ovigerous females
Ovigerous females of L. pectoralis (N total = 120; 30 from each season, i.e. summer, autumn, winter and spring) preserved in 70% ethanol, and each retaining at least one intact egg sac, were analysed. Specimens were isolated from marine European flounder caught in previous studies (Cavaleiro and Santos, Reference Cavaleiro and Santos2009). Host fish were obtained from the city harbour from boats using demersal bottom trawling off Matosinhos (41°10′N, 8°42′W), northern Portugal, and host data, including host total length in centimetres (HTL, from the tip of the snout to the posterior extremity of the extended caudal fin), were recorded. The fish were then frozen until they could be examined for copepods under a stereoscopic microscope. Because sampling relied on the availability of flounders at the harbour, sampling dates were limited to 1 day per season: September 2 for summer, November 17 for autumn, March 2 for winter and May 23 for spring. All samples of parasites were obtained in 2006, except that for the autumn season, which was collected in a preliminary study conducted in 2003 by F. I. Cavaleiro and M. J. Santos. This different date for the autumn sample was due to the lack of available ovigerous females in autumn 2006.
Macroenvironmental data
Based on data from an European Environmental Agency (2008), the salinity was considered approximately constant year round, remaining close to 35‰. The sea water temperature and light intensity values had been obtained for Cavaleiro and Santos (Reference Cavaleiro and Santos2009) from the Physical Oceanography Distributed Active Archive Center (PO.DAAC) at the NASA Jet Propulsion Laboratory (2009), and the Portuguese Meteorology Institute website (Instituto de Meterologia – IP Portugal, 2009).
Measurements and statistical analysis of data
Each individual female was observed under a stereomicroscope coupled to a computer with version 4.6.3 of Axiovision digital image processing software (Carl Zeiss Microimaging Inc., Thornwood, NY, USA). Parasite dimensions (Fig. 1) recorded were: parasite total length (PTL; excluding setae on caudal rami; in millimetres [mm]); and genital complex length (GCL; defined as the distance between the posterior margin of the free thoracic segment and the anterior margin of the abdomen; in mm). Egg production was characterized with respect to egg sac length (mm) (ESL; measured from the point of insertion in the genital complex to the posterior tip; only one sac [randomly chosen] was considered per parasite), egg sac width (mm) (ESW), egg length (mm) (EL; for 10 randomly chosen eggs per sac), F and TRE (a measure of the total resource investment per egg sac). Mean egg length (MEL) was calculated as the average length of the 10 eggs measured per sac; F, the number of eggs per sac, was estimated by dividing ESL by MEL, as in caligid copepods, eggs within sacs are uniseriate. Both eggs and egg sacs were considered approximately cylindrical; accordingly, egg sac volume (ESV) was calculated using the formula for the volume of a cylinder: ESV = π × (ESW/2)2 × ESL. Mean egg volume (MEV) was estimated by dividing ESV by F. Table 1 shows the acronyms used and the variables they stand for.
Fig. 1. Morphometric measurements recorded for the ovigerous females of Lepeophtheirus pectoralis. Abbreviations: EL, egg length; ESL, egg sac length; ESW, egg sac width; F, fecundity; GCL, genital complex length; PTL, parasite total length.
Table 1. Acronyms used for the morphometric characters and measures of reproductive effort of the ovigerous females
TRE is generally calculated as the product of egg volume (EV) and F (see Caley et al. Reference Caley, Schwarzkopf and Shine2001). However, in L. pectoralis, eggs within sacs are uniseriate and, accordingly, TRE is equivalent to ESV. This means that TRE could be directly assessed, by calculating ESV.
A first set of statistical analyses was performed on the total sample of ovigerous females and two sets of data, i.e. morphometric and reproductive effort data, were used. The following morphometric – i.e. PTL and GCL – and reproductive – i.e. F, MEV and TRE – variables were characterized with respect to mean, standard deviation (s.d.), range interval (R.I., minimum–maximum), coefficient of variation (C.V.), 95% confidence interval for population mean (C.I.), and distribution (one sample Kolmogorov–Smirnov's test testing for a normal distribution). Finally, the reproductive strategy followed by L. pectoralis was evaluated after determining the skewness and kurtosis values for the distributions of F and MEL.
A second set of analyses was performed on the total sample of hosts, and the HTL data were characterized with respect to the same statistical parameters, i.e. mean, s.d., R.I., C.V., C.I. and distribution.
A final set of analyses was conducted to evaluate any seasonal variation in morphometrics and reproductive effort. First, the distributions of PTL, GCL, F, MEV and TRE were characterized using box-and-whisker plots and then, body dimensions and reproductive effort over the four seasons were compared. In the box-and-whisker plots, the horizontal limits of the box represent the 25th percentile and the 75th percentile, the line in the centre represents the median, and the whiskers represent 1·5 times the interquartile range. Outliers more than 1·5 times the interquartile range are represented by circles, and asterisks when values are more than 3 times the interquartile range. A one-way analysis of variance test (one-way ANOVA) was used for multiple sample comparisons for those variables that met the assumptions of normality and homoscedasticity (PTL, MEL and MEV), followed by pairwise sample comparisons through Tukey's test when the first test returned significant results. For variables that followed a normal distribution but did not have homogeneous variances (F, ESL, ESW and TRE, based on Levene's test), Welch's robust test was used in multiple sample comparisons, followed by pairwise sample comparisons using Dunnett's T3 test. A multivariate analysis of variance (MANOVA) was used to conduct a discriminant function analysis (DFA; method: independent variables entered together), which permitted the evaluation of differences in both morphometric and reproductive variables between the females in different seasonal samples. Finally, the constraints of egg production were evaluated using a general linear model (GLM) with type III sum of squares (GLM multivariate analysis). Specifications were as follows: F and MEV – considered, simultaneously, as dependent variables; season – considered as fixed factor; and HTL and PTL – considered as covariates.
All statistical analyses were conducted in SPSS for Windows software, version 22.0.
RESULTS
Sea water temperature ranged from a maximum in July and August (slightly over 20 °C) to a minimum in February (less than 15 °C). Light intensity was over 300 h in July, decreased to a minimum in November (less than 150 h), increased slightly to around 200 h until February 2006, to decrease again to 150 h in March, and increased rapidly to a maximum of around 350 h in May.
Morphological traits and reproductive indices are summarized in Table 2, together with host data; Fig. 2 depicts the distributions of F and MEV (total sample of females). The skewness values were 0·461 and 0·106, while the kurtosis values were −0·281 and −0·144, respectively. HTL was the only variable that did not follow a normal distribution (P = 0·012 for one sample Kolmogorov–Smirnov's test testing for a normal distribution, see Table 2). With regard to F, the most prolific females carried about seven times more eggs than the least fecund (97·6/12·2 = 7·07), whereas in the case of MEV, there was a greater than 3-fold variation between maximum and minimum values (15621884/4511830 = 3·46).
Fig. 2. Distributions of F and EV for the total sample of ovigerous females of Lepeophtheirus pectoralis. Abbreviations: EV, egg volume (μ m 3); F, fecundity (number of eggs).
Table 2. Body dimensions and measures of reproductive effort for the entire sample (N = 120) of female parasites, and host total length
C.I., confidence interval; C.V., coefficient of variation; ESL, egg sac length; ESW, egg sac width; F, fecundity; GCL, genital complex length; HTL, host total length; MEL, mean egg length; MEV, mean egg volume; PTL, parasite total length; R.I., range interval; TRE, total reproductive effort.
Statistically significant differences for parasite features were observed for all variables except GCL (Table 3; Fig. 3). The results of the pairwise sample comparisons are presented in Table 3. Seasonal samples were most distinct regarding ESL, F and TRE. Summer females were distinct from those of the other seasons concerning most variables, with the exception of ESW and MEV. For most variables, a clear dichotomy was observed between summer and winter females, with summer females exhibiting a smaller size, shorter egg sacs and a smaller number of eggs per sac compared with winter females; autumn and spring females showed intermediate results between the two extreme seasons. While the only variable that indicated significant differences between autumn and spring was F, data from spring females shows a greater dispersion regarding ESL, F and TRE (Fig. 3).
Fig. 3. Boxplots showing the distribution of PTL, GCL, ESL, ESW, MEL, MEV, F and TRE for each seasonal sample of ovigerous females of Lepeophtheirus pectoralis. Abbreviatons: ESL, egg sac length; ESW, egg sac width; F, fecundity; GCL, genital complex length; HTL, host total length; MEL, mean egg length; MEV, mean egg volume; PTL, parasite total length; TRE, total reproductive effort.
Table 3. Multiple sample comparisons (results of one-way analysis of variance or Welch's robust test, see text) for seasonal samples of parasites, and pairwise sample comparisons (results of Tukey's or Dunnet's T3 tests, see text) among seasonal samples
ESL, egg sac length; ESW, egg sac width; F, fecundity; GCL, genital complex length; HTL, host total length; MEL, mean egg length; MEV, mean egg volume; PTL, parasite total length; TRE, total reproductive effort.
a Significant result (P < 0·05).
Variations in MEL were not accompanied by variations in MEV. While significant differences were observed between most seasonal samples regarding MEL (except between autumn and spring), a statistically significant difference in MEV, i.e. a significant decrease, was found only from winter to spring.
Disparities between summer and winter females are also clear in the two-dimensional DFA plot (Fig. 4): functions 1 and 2 were both statistically significant (function 1: Wilks’ Lambda = 0·227, χ 2 = 167·43, d.f. = 24, P < 0·0001; function 2: Wilks’ Lambda = 0·739, χ 2 = 34·146, d.f. = 14, P = 0·002), accounting, respectively, for 87·0 and 10·9% of the variance. Regarding function 1, summer and winter females were clearly the most dissimilar. Summer females accumulated on the positive end of the function, while winter females tended to be found on the negative end. In total 65·8% of the cases were correctly classified, but percentages varied among seasonal samples. Spring was the season with less successfully predicted membership cases: 80·0% of the summer females, 60·0% of the autumn females and 76·7% of the winter females were correctly predicted to belong to their own group, but only 46·7% of the spring females were correctly classified.
Fig. 4. Projection of the cases on discriminant functions 1 and 2, obtained by the discriminant function analysis of the four seasonal samples of ovigerous females of Lepeophtheirus pectoralis.
Concerning the results of the GLM, no statistically significant results were obtained for either main effects (i.e. season, PTL and HTL) or interaction terms (season × PTL, PTL × HTL and season × PTL × HTL).
DISCUSSION
The results reinforce conclusions previously drawn from the study of O. superba, a copepod parasitic on an invertebrate host, O. vulgaris (Cavaleiro and Santos, Reference Cavaleiro and Santos2014). Both the specimens used in the present study and those used in the aforementioned study were collected in the same area, the coast off Matosinhos; thus, these results support the existence of a general seasonal reproductive cycle in parasitic copepods, at least in these temperate latitudes. In both L. pectoralis and O. superba, summer females tended to be smaller and to produce a smaller number of eggs, in comparison with winter females.
Boxshall (Reference Boxshall1974c ) established 4·8 mm as the minimum size at which females became ovigerous. In our study, however, the smallest ovigerous female was considerably smaller, 3·91 mm long (average PTL = 4·77 mm). Though definite conclusions cannot be made because the data span from a single year only, it is possible that latitudinal differences in temperature are responsible for this difference. Indeed, an increasing temperature has been shown to correlate with reduced size in most aquatic ectotherms (Atkinson, Reference Atkinson1995; Angilletta et al. Reference Angilletta, Steury and Sears2004) and ectoparasites, being exposed to external environmental conditions, tend to follow the general rule (Poulin, Reference Poulin2007).
As to females’ size and TRE, two extremes are easily identified in the results found: summer females and winter females, the former being typically smaller and less fecund, and the latter being larger and more prolific. For marine flounder of the Portuguese coast, the prevalence and abundance of L. pectoralis adults during the winter is relatively low, but increases rapidly in spring and summer (Cavaleiro and Santos, Reference Cavaleiro and Santos2009). However, despite more being found in more similar infection levels, spring females were distinct from summer females regarding most variables, and although some were similar in size and F to these, others were clearly more similar to winter females.
These results are consistent with Boxshall's (Reference Boxshall1974c ) hypothesis of alternating generations and the seasonal variations in infection levels described for the Portuguese coast in Cavaleiro and Santos (Reference Cavaleiro and Santos2009). The life cycle of L. pectoralis can be hypothesized to be as follows: during winter, infection levels are relatively low, and only the large and more fecund overwintering females are present. Prevalence and abundance then increase in spring, as the first smaller and less fertile summer generation females become mature and join the surviving overwintering females. The coexistence of both generations may account for the large s.d. among spring females, as well as the lower percentage of correctly classified females obtained in the DFA for this season. As the overwintering females die and are replaced by the rapidly-maturing summer females, infection levels in summer remain similar to those of spring, but the average female becomes smaller and less prolific. Infection levels decrease to a minimum during autumn, probably due to the death of the summer generation. Autumn females were similar to those of winter in size, but were intermediate between summer and winter females for most variables (i.e. ESL and ESW, and TRE), which suggests that they may be mostly precocious overwintering generation females.
Although ovigerous females can be found year-round (Boxshall, Reference Boxshall1974c ; Cavaleiro and Santos, Reference Cavaleiro and Santos2009), Boxshall (Reference Boxshall1974c ) pointed spring (April–May) and summer (July and August) as the main reproductive seasons (i.e. preferential egg shedding season) of L. pectoralis, for winter and summer females, respectively. He supported this inference on the variations in the percentage of ovigerous females, even though some females of the overwintering generation initiate egg production during the winter. Spring females showed great variation in size and measures of reproductive effort. If most overwintering females do indeed breed during spring, it is possible that, of the females found during this season, the larger and winter-like are the overwintering females, finally coming into maturity, while the smaller and summer-like possibly matured from the eggs produced during winter.
The prevalence and abundance of adult parasites on marine flounder off the Portuguese coast is relatively low during the winter, when temperature and light intensity are lower (Cavaleiro and Santos, Reference Cavaleiro and Santos2009). If the smallest females amongst the spring sample originate from the eggs shed during winter, this may explain why the only statistically significant difference in EV was a decrease from winter to spring: females developing during winter may have experienced less favourable conditions (i.e. lower temperature and light intensity) during their development, leading to a reduced capacity to produce large eggs. Alternatively, overwintering females producing eggs in winter may contribute fewer resources into egg production as a result of the lower temperatures, resulting in progeny with an impaired capacity of producing large eggs. The lower EV observed in spring may also be caused by the survival until spring of some of the overwintering females that shed eggs during the winter, as death is thought to be caused by physiological exhaustion from egg production (Boxshall, Reference Boxshall1974c ).
However, the interpretation that the smaller spring females were the result of winter egg production would imply a similar maturation time to the summer generation females hatched during spring, i.e. 2 months (Boxshall, Reference Boxshall1974c ), although the temperatures and photoperiod they are exposed to during their development are different. In L. salmonis, temperature seems to influence egg string length and F to a small extent (Heuch et al. Reference Heuch, Nordhagen and Schram2000), although females reared at higher temperatures were significantly smaller (Nordhagen et al. Reference Nordhagen, Heuch and Schram2000). However, a slight increase in light intensity was observed between December and February for the Portuguese coast (Cavaleiro and Santos, Reference Cavaleiro and Santos2009). It is possible that this variation, rather than the light intensity levels themselves, has triggered the development of a small summer-like generation.
Other than physical factors, another possible factor regulating the alternation of generations is the ‘crowding effect’. Considering the site specificity of the ovigerous females of L. pectoralis (Scott, Reference Scott1901; Boxshall, Reference Boxshall1974a , Reference Boxshall1976), and the relatively small area available on the pelvic and pectoral fins, competition for space is a clear possibility. Unlike overwintering females, which find few competitors when they settle on their infection sites (i.e. prevalence and abundance are lower during the winter, see Cavaleiro and Santos, Reference Cavaleiro and Santos2009), summer or spring females settle on hosts already occupied by their overwintering counterparts. Thus, their smaller size, lower F and fast growth rate may be a response to competitors, rather than to other factors. Such ‘crowding effect’ has been found in several groups of parasites, such as cestodes (see, for example, Heins et al. Reference Heins, Baker and Martin2002), nematodes (e.g. Szalai and Dick, Reference Szalai and Dick1989) or digeneans (e.g. Jones et al. Reference Jones, Breeze and Kusel1989).
Concerning egg sac length, the seasonal variations are similar to those recorded for L. salmonis (Ritchie et al. Reference Ritchie, Mordue, Pike, Rae, Boxshall and Defaye1993): an increase from autumn onwards (October) until a maximum was reached during winter (February–March), with a subsequent decrease. Changes in egg length, commonly used to estimated egg size in caligids (Ritchie et al. Reference Ritchie, Mordue, Pike, Rae, Boxshall and Defaye1993; Bravo et al. Reference Bravo, Erranz and Lagos2009, Reference Bravo, Pozo, Silva and Abarca2013), may not always be an accurate predictor of egg size. The lack of statistically detectable differences in EV between most seasons implies that decreases in EL were often compensated for by an increase in egg sac width. This may have been caused by the compression caused by the addition of new eggs to the sac to the eggs already in it.
The parasite's overall strategy is difficult to infer from the present data, since there is no data available on the rate of egg sac production and any possible seasonal variations it may have, needed to complement the data on egg number and egg size. Therefore, TRE as presented here is merely an instantaneous estimate, used for comparative purposes. Still, considering the maturation time and the variations of F or TRE, a general picture can be obtained. While they differ in strategy, none of the generations appears to be a pure r- or K-strategist, as summer females mature faster but individually invest less on F, while overwintering females approach K-selected species in having a slow maturation, but invest on larger numbers of offspring, like an r-selected species.
No evidence of a trade-off between egg number and egg-size was found. The same result was obtained for Lernanthropus cynoscicola Timi and Etchegoin, 1996, also a fish parasite (Timi et al. Reference Timi, Lanfranchi and Poulin2005). In both cases, females seem to invest in egg number without compromising size, with larger females producing more eggs that are not significantly smaller. This contrasts with the results obtained with O. superba, which infects the sedentary invertebrate O. vulgaris (Cavaleiro and Santos, Reference Cavaleiro and Santos2014). Sedentary, usually invertebrate, hosts apparently correlate with large eggs, which could explain the existence of a trade-off in O. superba, while fish hosts tend to select for smaller, more numerous eggs in parasitic copepods (Poulin, Reference Poulin1995). However, both P. flesus (see Skerritt, Reference Skerritt2010) and Cynoscion striatus (Cuvier, 1829) (see Lopez Cazorla, Reference Lopez Cazorla2000), the host of L. cynoscicola (Timi et al. Reference Timi, Lanfranchi and Poulin2005), are, to some extent, migratory, entering estuaries or lower salinity areas. The difficulty associated with finding a motile host that may also be less available to parasites during certain periods of time (due to migration to conditions more hostile to the parasite, e.g. lower salinity), may explain such exceptions to the general rule. In this light, it may be hypothesized that selection may have led, not only to a larger clutch size, in order to allow parasites to compensate for the mortality associated with finding a host, but also to relatively large eggs, with larger energy storages to allow the infective stages to withstand periods of potential host shortage. The migratory behaviour of flounder might also help explain why Lepeophtheirus europaensis Zeddam, Berrebi, Renaud, Raibaut and Gabrion, 1988 from flounder produce larger clutches than those from brill, Scophthalmus rhombus L. 1758, which does not enter freshwater (De Meeûs et al. Reference De Meeûs, Raibaut, Renaud, Boxshall and Defaye1993).
Regarding the constraints of egg number and egg size, the GLM failed to indicate any significant effects. However, this may be explained by the coexistence of both generations in spring, since females of both generations, with their inherent reproductive strategies, could be hypothetically found in the same host during this season. A distinct possibility is that macro- and micro-environmental constraints are stronger for the number of clutches (i.e. the number of eggs sacs) produced during the parasite's lifetime than for egg number or size within the clutch. In fact, clutch size (i.e. egg number) has been shown not to be correlated with the frequency of clutch production (Godfray, Reference Godfray1987).
In a study of the seasonality of ectoparasites in P. flesus, Cavaleiro and Santos (Reference Cavaleiro and Santos2009) concluded L. pectoralis had a greatest epizootic potential during summer, due to higher pre-adult and adult levels of infection. Considering the alternation of generations model, these are associated with the emergence of the more numerous, but less individually fecund, summer generation. Prophylactic measures aimed at preventing epizootic outbreaks in flounder culturing systems may, in light of the present results, benefit from being carried out during winter or spring, when the parasite population is dependent on the larger egg production of a smaller number of copepods, thus preventing the development of the summer generation.
In conclusion, this study indicates that L. pectoralis females show seasonal patterns in their reproductive output. This suggests a reproductive pattern of alternating generations that differ mostly in longevity, size and fecundity. While more studies are required regarding the mechanisms that control the determination of the generation, this understanding opens new opportunities for designing epizootic control procedures in flounder marine aquaculture systems, namely intervening when the copepod population is smaller and dependent on fewer, more prolific females.
ACKNOWLEDGEMENTS
The authors would like to thank Professor Geoff Boxshall for his critical comments on a previous version of the manuscript.
FINANCIAL SUPPORT
This work was partially funded by CIIMAR – Interdisciplinary Centre of Marine and Environmental Research, under the Blue Young Talent internship programme, through a grant to D. G. Frade; partially supported by the Strategic Funding UID/Multi/04423/2013 through national funds provided by FCT – Foundation for Science and Technology and European Regional Development Fund (ERDF), in the framework of the programme PT2020., IDASSMYX Project (FCOMP-01-0124-FEDER-020726/FCT- PTDC/MAR/116838/2010, Portugal) to F. Cavaleiro and M. J. Santos; Project AQUAIMPROV (reference NORTE-07-0124-FEDER-000038), co-financed by the North Portugal Regional Operational Programme (ON.2 – O Novo Norte), under the National Strategic Reference Framework (NSRF), through the European Regional Development Fund (ERDF) to F. Cavaleiro and M. J. Santos.
