Introduction

The population dynamics of commensal epibionts is a poorly studied topic in ecology despite these organisms being quite common and diverse, which include a variety of taxa such as barnacles on whales or mites on bird feathers. Entocytherids (Ostracoda, Entocytheridae) constitute a family of ostracods that are epi-commensals on other crustaceans. The components of the main entocytherid subfamily, the Entocytherinae, representing 83% of the 220 species of the group, are native to North and Central America where they live associated with crayfishes belonging to the Astacoidea (Hart & Hart, 1974; Mestre et al., 2012). Recently, some entocytherid species of this group have been found in other continents due to the transport of American crayfish beyond their original area for aquaculture purposes (Smith & Kamiya, 2001; Aguilar-Alberola et al., 2012). This is the case of Ankylocythere sinuosa (Rioja, 1942), recently discovered in the Iberian Peninsula (Aguilar-Alberola et al., 2012), which is associated with Procambarus clarkii (Girard, 1852), a very successful invading crayfish (Geiger et al., 2005). Contributions to knowledge on the population ecology of entocytherids are virtually inexistent, except for the works of Young (1971) and Hart et al. (1985). In addition, although the ecological impact of invasive crayfish has been well-recognised and studied (Rodríguez et al., 2003; Gherardi & Acquistapace, 2007; Holdich et al., 2009), the ecology of their ostracod epibionts outside the American continent remains largely unknown (Aguilar-Alberola et al., 2012).

According to current knowledge on the annual dynamics of entocytherid populations, A. sinuosa maintains a high reproductive rate for all seasons (Young, 1971). However, the entocytherid load per host in this species, that is, the number of entocytherids that inhabit on a crayfish individual, varies considerably all year round: it is minimum in winter, increases in spring and early summer, and peaks in late summer. In another entocytherid species, Uncinocythere occidentalis (Kozloff & Whitman, 1954), all seasons are favourable, except summer (Hart et al., 1985). When focusing on the proportion of juvenile and adult entocytherids, Young (1971) noted that crayfishes with large amounts of A. sinuosa had high proportions of immature individuals, mostly belonging to the first juvenile instar (A-7). Conversely, in crayfishes with low entocytherid loads, the proportion of adults was larger. In A. sinuosa, the sex ratio was biased towards females (Aguilar-Alberola et al., 2012). In many ostracods species, the sex ratio generally shows excess females, while the number of females and males is the same in others, and a larger proportion of males is rare (Cohen & Morin, 1990).

In most free-living ostracods, the growth rate and final size of adults are related to temperature, salinity and to other habitat conditions (Cohen & Morin, 1990). Generally, temperature has been considered the main variable to influence ostracod population dynamics in shallow-water habitats (Horne, 1983), as observed for instance in the cytheroidean Cyprideis torosa (Jones, 1850) or the cypridoidean Heterocypris bosniaca Petkowski et al. 2000 (Mezquita et al., 2000; Aguilar-Alberola & Mesquita-Joanes, 2011). Salinity probably affects the spatial distribution of ostracods more intensely than their population dynamics (Horne, 1983; Mesquita-Joanes et al., 2012). Some ostracods inhabiting constant environments do not show seasonality (Horne, 1983; Cohen & Morin, 1990), but most free-living ostracods show it during their life cycles, with reproduction periods restricted mostly to spring, summer or autumn.

Unlike free-living organisms, epibionts can be affected by variables directly related to their host traits or state (Threlkeld & Willey, 1993; Jovani & Serrano, 2004). Within this framework, Young (1971) proposed crayfish moult, size and reproductive cycle to be the main variables that modulate the population dynamics of entocytherid ostracods. According to Young (1971), crayfish moult could explain the broad variability of the entocytherid load that he found among crayfishes of the same size. According to this hypothesis, moulting events would trigger loss of entocytherids, and consequently after the moult event, immigration and reproduction processes would be required to recover the crayfish entocytherid load. This expectation for ostracod entocytherids of Young (1971), with epibiont loads being renewed after the host’s moult, has been described in other works studying epibionts on zooplanktonic crustaceans, such as daphnids (Threlkeld & Willey, 1993; Pérez-Martínez & Barea-Arco, 2000). Furthermore, several studies have found a positive relationship between crayfish size and entocytherid load (Young, 1971; Hart et al., 1985; Aguilar-Alberola et al., 2012). In addition, according to Young (1971), A. sinuosa has lower densities during those periods when the number of juvenile crayfishes increases. He attributes this fact to a dilution effect caused by a spread of the entocytherid population among many immature crayfishes. Therefore, entocytherid density would depend on the host’s reproductive cycle. Host gender, however, does not seem to have any effect on entocytherid load (Hart et al., 1985).

In this work, a monthly monitoring programme was carried out with an A. sinuosa population throughout the year in a Spanish wetland invaded by the American red swamp crayfish P. clarkii. With this monitoring plan, the aim was to describe, for the first time, the population dynamics of an entocytherid beyond its original distribution area, and to analyse the variables that could affect the entocytherid population ecology.

Materials and methods

Sample collection was carried out in a 400-m stretch of the Bovar canal (41°36′21.38″N; 0°45′37.66″W), located inside the Pego-Oliva Wetland Natural Park (Uriós et al., 1993). This canal holds a permanent water table and the red swamp crayfish (P. clarkii) can be found in it all year round. The entocytherid ostracod A. sinuosa has been previously found at this site in a study about the distribution of the species in the Eastern Iberian Peninsula (Aguilar-Alberola et al., 2012).

During monitoring, various in situ water physicochemical parameters were measured. A digital meter (Hanna Instruments HI 98129) was employed to determine temperature (°C), conductivity (mS/cm) and pH. A Winkler test was used to measure the concentration of dissolved oxygen (mg/l). In addition, the Spanish Meteorological Agency (AEMET) provided daily precipitation and air temperature data. Crayfishes were collected monthly over one year between March 2011 and February 2012. We used 40 crayfish baited traps (40 cm × 40 cm × 80 cm) with two circular openings (8 cm in diameter) and 2 mm mesh. They were distributed regularly along the canal, separated by a 10-m distance, and were deposited at the bottom of the canal (ca. 1–2 m deep). Chicken liver was used as bait. For each sampling campaign, baited traps remained at the site overnight. If the crayfish catch did not attain a minimum of 15 individuals the next morning, traps were left at the site and were collected the following morning. After collection, the number of crayfishes per trap was counted, and juveniles and adults were differentiated. Those individuals whose total body length exceeded 6 cm were considered adults (Cano & Ocete, 2000). Smaller, juvenile crayfishes were not used for entocytherid extraction because they do not hold entocytherids, or do so but in very small amounts (pers. obs., and Aguilar-Alberola et al., 2012). Entocytherid extraction was done following the method tested by Mestre et al. (2011), as detailed hereafter. For each monthly sampling, between 15 and 20 adult crayfishes were selected randomly. These crayfishes were submerged individually in carbonated water (Fuente Primavera™) for 5 min. Then, carbonated water was filtered with a 63-μm mesh and the filter was submersed in a 50 ml jar with 96% ethanol. This method removes about 78% of entocytherids from each crayfish (Mestre et al., 2011). The selected crayfishes were weighed and sexed. The intermoult stage was also estimated using ordinal notation based on cephalothorax hardness by taking into account that the harder the cephalothorax, the longer the time since the last moulting event (Reynolds, 2002). This exoskeleton hardening stage (EHS) notation was coded as EHS 0 (very soft exoskeleton, including chelipeds), EHS 1 (soft, especially the cephalothorax lateral areas), EHS 2 (hard, but with some elasticity) and EHS 3 (very hard). Finally, crayfishes were individually marked with a small hole in one uropod. They were released back into the water in order to know if ostracods had already been removed from a collected crayfish in future recaptures so as to restore the crayfish population and to avoid disturbance effects on crayfish density if removed.

Under a high magnification stereomicroscope (Leica MZ16), all entocytherid ostracods were identified and counted following the same procedure: after filtering with a 63 μm mesh the carbonated water where crayfishes were submersed (see above), this filter was washed in a Petri dish with ethanol. The dish was carefully inspected under a stereomicroscope to measure and count all ostracods, classified onto different instars according to their size (see Aguilar-Alberola et al., 2012). Late juvenile instars (A-1) and adults were sexed according to the presence or absence of copulatory apparatus. Number of coupled male–female pairs, either mating or males guarding females (Danielopol, 1977), was noted differentiating between adult males coupled to A-1 females (i.e. ‘biunguis females’ according to Hobbs (1971)) and adult males with A-2 females (Aguilar-Alberola et al., 2012). [N.B.: in this paper, the term ‘population’ refers to the group of entocytherids that lives in the Bovar canal, and not merely the number of entocytherids inhabiting on a single crayfish, which is termed entocytherid ‘load’ per host herein. Entocytherid ‘density’ is used to refer to load per crayfish weight.]

All the graphs and statistical analyses were performed with R version 2.15.0 (R Development Core Team, 2012) using different packages. First, the population dynamics of entocytherids was described as follows: a constrained cluster analysis was done to group entocytherid densities [log10 (number of entocytherids + 1)/log10 crayfish weight] (according to Mestre et al., 2011) throughout the year. To that end, the Rioja package (Juggins, 2012) was employed with the Constrained Incremental Sum of Squares (CONISS) method. In addition, linear correlation was calculated between the juvenile entocytherid proportion and entocytherid density per crayfish. A paired t test was carried out to test for deviations from a balanced sex ratio. We calculated the sex ratio as the proportion of males [males/(females + males)] (Wilson & Hardy, 2002). Second, we analysed which environmental and host variables could be related to entocytherid load (number of entocytherids per crayfish) and the proportion of entocytherids belonging to juvenile instar A-7 (number of A-7 per crayfish/total number of entocytherids). We used Generalized Linear Mixed Models (GLMMs) by means of the lme4 package (Bates et al., 2011) following Bolker et al. (2009). Entocytherid load models were constructed with a Poisson distribution and a logarithmic link function. Models for the proportion of instar A-7 individuals were constructed with a binomial distribution and a logit link function (Quinn & Keough, 2002). In the models with proportions, we merged two vectors (the number of A-7 individuals per crayfish, and the number of entocytherids that are not A-7) by means of the R command cbind into a single object corresponding to the response variable. Initially, general models with all the possible combinations of the following variables were obtained: month, crayfish weight and crayfish EHS. Then those models with lower Akaike Information Criterion (AIC) values—i.e. the best fit with the lowest number of variables—were chosen. In order to facilitate the understanding of the models results, the estimates with standard errors (marginalizing the random effect) were plotted together with the original data to see how well the model fits the data. In addition, the aim was to also know which variable could be related to month effects. To go about this, we constructed two other models for the two response variables with the following explanatory variables: water temperature, conductivity, oxygen concentration, number of captured crayfishes per trap per hour, and proportion of young crayfishes. All the models had a random factor corresponding to individual crayfish.

Results

Changes in the habitat followed a clear seasonal pattern (Fig. 1). The highest air temperatures were recorded in summer, reflected in the highest water temperature measured (25.4°C in July) and the lowest in February (water temperature 10.8°C). Precipitations concentrated mainly in autumn (September, October and November). Conductivity varied between 1.4 and 4 mS/cm, with maximum values recorded in dry summer months. The oxygen concentration was generally low and varied between 2.4 mg/l in August and 6.6 mg/l in February, and the pH average was 7.22 ± 0.28 (1 SD). Throughout the year, crayfish catch peaked twice in late summer (August–September) and winter (February) (Fig. 1). The juvenile crayfishes caught were less abundant than adults. Juveniles were caught all year round, except in spring (March, April and June), and were seen to be more abundant between July and November. Only three marked crayfishes were re-captured, and they were discarded for further statistical analysis of entocytherids population. For entocytherid extraction, 233 crayfishes (P. clarkii) were employed. Only three crayfishes had no entocytherids. Among all other individuals, only one entocytherid species was found: A. sinuosa (Rioja, 1942). The entocytherid load distribution was right-skewed, with 50% of crayfishes hosting less than 32 entocytherids and with a maximum value of 1,113 entocytherids. The majority of the counted ostracods were juvenile individuals, with higher proportions accounted for by the earliest instar A-7.

Fig. 1
figure 1

Variations of the physicochemical parameters and crayfish catch throughout the sampling period, from 13 March 2011 to 15 February 2012, in the Bovar Canal (Pego-Oliva Wetland). Upper graph daily air temperature (minimum and maximum with grey lines), water temperature (black line with dots) and daily precipitation (bars). Middle graph conductivity and oxygen concentration. Bottom graph crayfish catch (crayfishes/trap * hour)

Full size image

The cluster analysis of entocytherid density per month resulted in three groups of samples: G1 (March–June) with intermediate densities; G2 (July–October) with high densities; G3 (November–February) with low densities (Fig. 2). We detected a positive correlation between the proportion of juvenile entocytherid instars and entocytherid density per crayfish (Pearson’s correlation, r = 0.465; P < 0.001). Throughout the year, the months with the highest densities also showed a higher proportion of juvenile individuals (Fig. 2).

Fig. 2
figure 2

At the top, a constrained cluster dendrogram of entocytherid density [log10 (number of entocytherids + 1)/log10 crayfish weight in g] resulting in three groups of months (G1–G3). Below a boxplot of entocytherid density and the proportion of juvenile entocytherids with a line (juvenile entocytherids/total entocytherids) per month

Full size image

Generally, the sex ratio of entocytherids did not significantly differ from 0.5 (t = 0.12; P = 0.90). Males dominated females only in July (t = 2.35; P = 0.03). Male–female coupled pairs were found throughout the year, but the number of pairs varied (Fig. 3). The highest average percentage of coupled pairs per crayfish (38%) was recorded in April, followed by July (17%) and August (18%). Most observed couples were adult males with A-1 females (N = 241), but some pairs composed of adult males with A-2 females were also observed (N = 31), and no male association with adult females was found. The main hatching period took place from March to September when the highest proportions of early juveniles (instar A-7) were recorded (Fig. 4). This hatching period was followed by an increase in the A-6 and A-5 proportions between May and October, an increase in more advanced juvenile instars (A-4 to A-2) between July and November, and it finally resulted in the dominance of A-1 and adults between November and February. Therefore, although all instars were found throughout the year, the main growth period occurred between summer and early autumn.

Fig. 3
figure 3

Bar graph with the number of males associated with A-1 female, A-2 female and alone for different months. Density of A-7 entocytherids is represented by a line [log10 (number of A-7 + 1)/log10 crayfish weight in g]

Full size image
Fig. 4
figure 4

Bubble chart showing monthly variations of the mean proportion of each entocytherid developmental instar (number of individuals of one instar/total entocytherids per crayfish), from A-7 to adult, all year round

Full size image

The GLMMs analysis of the variables affecting population dynamics revealed that the most relevant combinations that explained entocytherid load were sampling month, crayfish weight and EHS (AIC = 1017.1). Month and EHS (AIC = 520.3) were the most important variables to account for the variation in the proportion of A-7 entocytherids (Table 1).

Table 1 Main results of the GLMMs for response variables entocytherid load and proportion of stage A-7 entocytherids, which were built with combinations of explanatory variables crayfish weight, sampling month and EHS
Full size table

In the entocytherid load model (Table 2), the variance corresponding to the individual crayfishes’ random effects was 0.676. Crayfish weight had a positive effect on the number of entocytherids per crayfish (z = 9.16; P < 0.001). The entocytherid load corresponding to each sampling month became significantly different from March, except for those belonging to the G2 clustering group (July–October). The largest difference corresponded to February when the smallest loads were seen (February; z = −10.59; P < 0.001). In addition, significant differences were detected between EHS 0 and other (harder, i.e. a longer time since the last moult) EHSs, which were related to greater entocytherids loads (EHS 1, z = 2.36; P < 0.05; EHS 2, z = 3.43; P < 0.01; EHS 3, z = 3.80; P < 0.001).

Table 2 Results of the selected GLMM to account for crayfish entocytherid load depending on crayfish weight, sampling month and EHS
Full size table

In the GLMMs model constructed for the proportion of A-7 juveniles (Table 3), the variance corresponding to the individual crayfishes’ random effects was 0.246. Monthly significant differences were detected only in November (z = −1.92; P < 0.05) and February (z = −3.09; P < 0.01), as compared to other months when A-7 proportions where higher. The EHS effect was similar to the previous model, with a lower proportion of A-7 juveniles in the younger moult stage (EHS 0), if compared to others (EHS 1, z = 2.34; P < 0.05; EHS 2, z = 3.12; P < 0.01; EHS 3, z = 3.24; P < 0.01; Table 3). However, crayfish weight was not selected for the best model in this analysis. For the two previous models, the predicted values are presented together with the original data (excluding the random effect) in Fig. 5. The observed and predicted values showed a close pattern of variation.

Table 3 Results of selected GLMM to account for the proportion of A-7 entocytherid juveniles depending on sampling month and EHS
Full size table
Fig. 5
figure 5

Monthly variation of GLMMs estimated values together with the observed data marginalizing the random effect. Top GLMM results to account for crayfish entocytherid load. Bottom GLMM results to estimate the proportion of A-7 individuals. Monthly means predicted are represented by a solid line and the standard deviations by long dashed lines. In addition, the monthly mean for the observed data (dots) with the error bars (standard deviation) is also shown

Full size image

Finally, the GLMMs analyses allowed us to detect which environmental variables could determine the variation between months on entocytherid load and A-7 proportions (Table 4). Two variables had significant effects on entocytherid load: water temperature, with a positive effect (z = 6.93; P < 0.001), and crayfish catch (z = −2.95; P < 0.01), with a remarkably strong negative effect (coeff = −13.99). Instead the variables for the A-7 proportions with significant effects were water temperature, with a positive effect (z = 3.86; P < 0.001), and conductivity, which had a negative effect (z = −3.42; P < 0.001), whereas crayfish catch had no significant effect. In both models, dissolved oxygen and the proportion of juvenile crayfishes appeared to have no significant effects on entocytherid population abundance and A-7 proportions.

Table 4 Main results of the GLMMs built to test the effects of environmental variables with monthly variation (water temperature, conductivity, dissolved oxygen, crayfish catch and proportion of juvenile crayfishes collected) on entocytherid load and proportion of A-7 juvenile entocytherids
Full size table

Discussion

Throughout this 1-year study carried out in the Pego-Oliva Wetland, only one entocytherid species was found, A. sinuosa, which was associated with the red swamp crayfish P. clarkii, and with a prevalence approaching 100% for adult crayfishes. These results are in agreement with the findings of Aguilar-Alberola et al. (2012) for this site and for other places in the Iberian Peninsula. The situation differs from that of North America, its original area, where one crayfish may have several entocytherids species (Hart & Hart, 1974). The poor entocytherid diversity found herein may be due to the fact that crayfishes only transported to Europe a small representation of the entocytherids inhabiting them in America. This pattern has also been found in other invasive species (Torchin et al., 2003). The high prevalence observed in this particular entocytherid species may have proven essential to explain its success in colonising exotic areas (Aguilar-Alberola et al., 2012).

Each crayfish entocytherid subpopulation differs in size and structure (Young, 1971; Hart et al., 1985; Aguilar-Alberola et al., 2012; this study). Across the observed variability, the positive correlation found between entocytherid density and the proportion of youngest juvenile stages was also observed by Young (1971) and Aguilar-Alberola et al. (2012). Furthermore, it might be an effect of the interaction among the reproductive, dispersal and mortality processes in which crayfish moult may be involved. Most probably, the highest entocytherid loads originate from recent reproductive successes that do not result in a proportional immediate increase in adult abundance within a crayfish. There may well be other factors, chiefly differential dispersal or mortality of crayfish juvenile stages, which could regulate entocytherid densities to such an extent that the crayfish moult can act as a generator of this type of processes producing entocytherid losses. Throughout the year, entocytherid density is at its minimum in winter, increases in spring, and peaks in summer and early autumn. As previously mentioned, the juvenile entocytherid proportion per crayfish is higher in those months with higher entocytherids densities, which correspond to favourable reproduction periods with high temperatures. This variation between months matches Young’s (1971) findings in the original distribution area. He suggested that entocytherid reproduction is high and mortality is low in spring and summer. Consequently, even though the reproduction events that generate higher entocytherid loads and juvenile instar proportions dominate during the favourable reproductive periods, the processes that result in entocytherid losses, viz. emigration, mortality and crayfish moult, prevail during unfavourable periods (winter). Nonetheless, this pattern does not seem to be universal among other entocytherids. For example, the Uncinocythere occidentalis population decreases in summer (Hart et al., 1985). This difference could be due to preference for low temperatures and lower heat tolerance in this species associated with the signal crayfish Pacifastacus leniusculus (Dana, 1852). This crayfish species prefers temperatures that are lower than 25°C (Bondar et al., 2005) and its associated entocytherids must acquire a similar preference for low temperatures in order to survive and reproduce.

In general, living as a commensal or parasite is considered as a way of life that may offer two advantages: first, environmental variation may diminish; second, there is always some food available (Poulin, 2007). This may be mostly true for endoparasites. However, commensal ostracods live as external epibionts on crayfish and, therefore, they might suffer equally the effects of water temperature or chemistry changes. Here, reduced predation pressure might be an important advantage in their way of life and not diminished environmental fluctuations. In our study, reproduction in the studied exotic entocytherid occurs all year round, as in its original distribution area (Young, 1971). However, within this uniformity, it seems that certain months are better for reproduction and development. In the Pego-Oliva wetland, the favourable reproduction period for A. sinuosa appears to span from April to August, when the proportion of juvenile instars increases. During these months, the largest numbers of coupled male–female pairs are followed almost immediately or with a 1 month lag by peaks in density of A-7 juveniles. In general term, entocytherid A-1 females form (pre-)mating pairs with adult males, and pairs of males with adult females are rare (Hobbs, 1971). Pairs of males with A-2 females are also infrequent, but have been previously found in the Pego-Oliva wetland (Aguilar-Alberola et al., 2012). According to our results, pairs with A-2 females are numerous when there are more coupled males. Therefore, it is feasible that, under favourable reproduction conditions (adequate temperature, etc.), males might be less selective and not only form complexes with A-1, but also with younger females (A-2). Further research is required to know whether males really copulate with A-1 or A-2 females, or if they merely guard them to warrant subsequent copulation when females reach adulthood, as suggested by Danielopol (1977). The adult sex ratio remains at around 50% of males for most of the year, except in July when there are more males than females. These results contrast with those reported in the work by Aguilar-Alberola et al. (2012), in which females are generally more abundant than males.

According to our GLMMs, sampling period (month), crayfish weight and EHS (related to the time since the last moult event) all seem to be important explanatory variables that affect crayfish entocytherid load and population structure. The months with lesser entocytherid abundance and lower A-7 proportion correspond to the colder winter period. Young (1971) attributed the winter-related decrease in entocytherid load to a dilution effect of the entocytherid population because their spread for the colonisation of a new cohort of juvenile crayfishes appeared during this period. Consequently according to Young (1971), the host reproductive cycle may well prove more important than the physicochemical external parameters. However, the present study reveals that the external environmental conditions that do not directly relate to the host may also play an important role in the entocytherid life cycle by producing a direct effect on the reproductive rates. In our models, water temperature positively affects entocytherid load and the A-7 proportion. In most ostracods, temperature is regarded the main variable that affects growth rate and final size (Cohen & Morin, 1990). Yet although the negative effect of crayfish catch on entocytherid load found herein is in agreement with the ‘dilution effect’ idea proposed by Young (1971), the number of crayfish captures might be entangled with temperature. In addition, crayfish catch as an estimator of crayfish density can be modulated by crayfish activity to produce spurious results. In any case, further experiments are required to corroborate that the crayfish reproductive cycle actually affects entocytherid population variability. Conductivity negatively relates to the A-7 proportion; therefore, high conductivity probably disrupts entocytherid reproduction. High salinity can adversely influence A. sinuosa population maintenance and, although adults are tolerant to salinities above those measured in the field (pers. obs.), this condition can negatively affect hatching.

The positive effect of crayfish weight on their ostracod epibiont load has been previously reported in other studies on entocytherids (Young, 1971; Hart et al., 1985; Aguilar-Alberola et al., 2012). The reasons for this relationship can be that larger crayfishes offer more food, larger reproductive surfaces and longer intermoult times (Young, 1971) than smaller ones. However, this relation does not hold in some studies about epibionts of zooplanktonic crustaceans. Some authors have found positive relations (Holland & Hergenrader, 1981; Mohlenberg & Kaas, 1990; Threlkeld & Willey, 1993), while others have not (Kankaala & Eloranta, 1987; Weissman et al., 1993; Pérez-Martínez & Barea-Arco, 2000). Nevertheless, some zooplanktonic epibionts prefer species with larger bodies, such as daphnids (Barea-Arco et al., 2001; Regali-Seleghim & Godinho, 2004; Zalocar et al., 2011).

Our results of the impact of crayfish moult on entocytherid load reduction confirm the assumptions of Young (1971). Crayfish at a more advanced intermoult stage, for which more time has elapsed since the last moult, tend to have more entocytherids. The moult process, which is more frequent in younger crayfish, causes loss of entocytherids; consequently, the number of entocytherids increases as the intermoult time accumulates until the next entocytherid loss event after moulting. In the field, four recent crayfish exuviae were collected, three of which contained living entocytherids (pers. obs.), thus supporting the aforementioned explanation. In addition, the GLMM model for the proportion of A-7 entocytherid stages indicates that recently moulted crayfish (EHS 0) have a lower A-7 proportion than those at older intermoult stages. Entocytherids loss is, therefore, more intense in the early juvenile instars of the ostracod than more advanced stages, including adults, probably because the latter can better avoid remaining on older exuviae by moving to new crayfish exoskeletons (Young, 1971). This could explain the lower A-7 proportion noted in recently moulted crayfishes. Yet there is another possible hypothesis to explain this phenomenon; i.e. that the crayfish moult produces entocytherid loss which similarly affects all the entocytherid developmental stages, but the following recolonisation of ‘empty’ crayfish is produced mainly by older instars. In any case, crayfish moult affects the total loss of eggs attached on its exoskeleton (Young, 1971). Therefore, some time is needed for eggs to be replaced with new ones, which will hatch later in a new cohort of A-7 ostracods.

The Entocytheridae represent an interesting model taxon to study the ecology and behaviour of epibionts, although very few studies are currently available. Presently, there are many questions on their ecology that remain unanswered. For instance, host crayfish moult implies loss of ostracods, but we do not yet know whether this loss is partial or complete, or which ostracod instars are more affected. We do not know many details about their dispersal and colonisation mechanisms, or whether or not exotic species could colonise native European crayfish populations, such as Austropotamobius pallipes (Lereboullet, 1858), and implications of such a transfer on the conservation of the native species remain a mystery. Our results help provide an understanding of the ecology of epi-commensal organisms, such as entocytherids, and the relationships they establish with their hosts, particularly within the invasion ecology framework, as they enable us to understand how combinations of associated alien species respond to recently colonised environments.