Investigating fecundity and egg loss using a non-invasive method during brooding in European lobster (Homarus gammarus)

Abstract

This article examines two important components of measurement of fecundity in the European lobster Homarus gammarus: (i) comparing the traditional gravimetric dry weight fecundity method against two non-invasive depth gauge methods initially developed for Homarus americanus and (ii) utilizing the depth gauge method to determine egg loss during the brooding period and its impacts on effective fecundity estimates. No significant difference was observed between fecundity estimates derived using either the traditional or depth gauge methods. Derived fecundity estimates from the two depth gauge methods differed by −0.31% (±2.7 s.e.) for cylinder and −1.1% (±2.4 s.e.) for ellipsoid fecundity estimates compared with the traditional method. This highlights the utility of the depth gauge method for providing fast, reliable and low-cost estimates without sacrificing lobsters or their egg masses. Egg loss is estimated to be as high as 44% from initial extrusion to hatching. The application of the non-invasive methods for estimating fecundity to other fisheries and stocks is discussed along with the importance of understanding egg loss in this commercially valuable fishery.

egg loss, fecundity, Homarus gammarus, non-invasive

Introduction

European lobster Homarus gammarus provides lucrative and important fisheries throughout its natural range, but there remain significant knowledge gaps on the basic biology of this species. Knowledge of basic life-history parameters is crucial for effective management of sustainable fisheries. This article focuses on methods to determine egg production in H. gammarus, aiming to provide essential information for strategies to conserve spawning potential in exploited stocks.

Fecundity estimates in commercially exploited species have been shown to fluctuate through time, suggesting its potential use as an indicator of fisheries-induced evolution on reproductive parameters (Kuparinen and Merilä, 2007; Lambert, 2013). In addition to this however, the effect of environment on egg production, development, larval survival and future recruitment cannot be ignored (Wright, 2013; Green et al., 2014; Koopman et al., 2015). In clawed lobsters, fecundity is seen to be influenced by sea temperature, with Homarus americanus showing broadly increasing fecundity with northerly latitude along the east coast of North America (Currie and Schneider, 2011), whereas H. gammarus shows an eastwards trend of increasing fecundity in northern European waters (Ellis et al., 2015).

Typically fecundity is assessed through the use of gravimetric methods, entailing the measurements of either the whole or a sample of the ovary and subsequent counting and re-weighing of a subsample of oocytes to determine overall fecundity (Kraus et al., 2002; Miller and Kendall, 2009). This method is both labour-intensive and costly. Methodological advancements however present options for undertaking fecundity estimates that are more cost-effective, more practical and less invasive, even for relatively large sample sizes (Thorsen and Kjesbu, 2001; Klibansky and Juanes, 2008; Alonso-Fernández et al., 2009). Such examples include the development of an automatic diametric method for estimating fecundity in Atlantic Cod Gadus morhua (Thorsen and Kjesbu, 2001). In the case of clawed lobsters, non-invasive techniques for estimation of fecundity have been recently developed for H. americanus by Currie et al. (2010). They used a number of simple volumetric measurements of the brood mass in situ to estimate fecundity. No significant difference was observed between fecundity estimates using traditional gravimetric and this volumetric method. This outcome reveals the opportunity for its trial with other similar decapod species.

A related, poorly researched topic is that concerning the level and impact of egg loss on overall fecundity. In previous studies of egg loss in H. gammarus (Latrouite et al., 1984; Free, 1994; Lizárraga-Cubedo et al., 2003; Agnalt, 2008) this has usually been measured by comparison of fecundity estimates made between newly extruded and later developed eggs, with egg loss between first extrusion and hatching being estimated at 27% in H. gammarus (Latrouite et al., 1984). Egg loss is estimated to be higher in H. americanus, with egg loss documented as ranging up to 40% (Factor, 1995). Factors affecting final spawning full term fecundity can vary widely however; inter-annual variation in environmental condition can effect egg quality and number (Koopman et al., 2015), while geographic extent of populations (Savoie and Maynard, 1991) and species’ environmental tolerance ranges can influence egg loss (Wickins et al., 1995). Furthermore, sample acquisition techniques can introduce uncertainty into these estimates. Estimates of egg loss in H. americanus reported by Perkins (1971), were based on samples initially collected via otter trawl. Trawling is known to induce egg loss in other crustaceans. High degrees of egg loss in the Norway lobster Nephrops norvegicus are seen to occur in trawl samples, ranging from 11 to 22% (Chapman and Ballantyne, 1980; Briggs et al., 2002). Not accounting for egg loss in estimating fecundity can provide inaccurate estimates of egg production and consequently misleading outcomes of egg per recruit analysis, a technique used in the management of some lobster stocks (Tully, 2004; Miller and Hannah, 2006).

In order to investigate potential egg loss the development of egg staging criteria is required. Previous studies have predominately categorized embryo development by measuring the size of the pigment crescent in the eye [eyespot index (ESI)], criteria originally developed by Perkins (1972) for H. americanus and subsequently applied to H. gammarus (Tully et al., 2001; Lizárraga-Cubedo et al., 2003). Such a technique however relies on the measuring of egg pigments utilizing laboratory equipment and is not suited to field classification. Pre-existing classification of H. gammarus egg development based on visual categorization has been developed defining five development stages (Pandian, 1970); however, these have not been aligned with ESI categories. Assignment of ESI categories and visual staging criteria has been undertaken by Latrouite et al. (1984), assigning embryo development across three stages based on both ESI and visual staging. The assignment of development across just three stages however, leads to the potential obscuring of true egg loss during brooding due to the broadness of the criteria. Embryo development of H. gammarus described by Hepper and Gough (1978) indicates the potential of four stages based on changes in ESI, indicating that the four stage method of (Pandian, 1970) is preferable as a basis for categorizing egg development. By comparison, existing field visual staging guides used for H. americanus employ the use of four development stages, which are further categorized by ESI estimates (University of New Hampshire, 2004).

The overall aim of this study is to develop a non-invasive method of determining fecundity in H. gammarus. Specifically, the objectives are to trial and adapt the method of Currie et al. (2010), validate its application through comparison with fecundity estimates derived through a traditional, invasive technique, and to apply the method to determine egg loss during development. The latter objective also requires the development of criteria for determining development of egg developmental stages. By applying this method in this species for the first time, we will evaluate its potential application in the assessment of fecundity variations over the full distributional range of H. gammarus

Material and methods

Study site

One hundred and sixteen ovigerous females ranging in size from 89- to 151-mm carapace length (CL), were sampled from commercial landings and observer trips during October 2016–August 2017 from the Orkney inshore lobster fishery, Scotland (Figure 1). Females with visible signs of egg loss resulting from capture and transportation were not sampled. Signs of egg loss or brood damage were clearly identifiable from correspondence with one or more of three criteria: (i) presence of loose eggs surrounding the brood mass, (ii) irregular shape of brood mass segments, and (iii) no eggs on first pleopod. Evidence of these criteria is commonly the results of stress induced tail slapping during handling and storage.

Figure 1.

Distribution of reef habitat around the Orkney Archipelago and the extent of the 3 nautical mile inshore limit, which encompasses the major lobster grounds. [Habitat prediction—European Marine Observation Data Network (EMODnet)].

Of the 116 females, 30 individuals ranging from 89-to 151-mm CL classified as having eggs at development stage 1 (Figure 2) were collected from October to December 2017 and were used to assess the application of the depth gauge fecundity method against the traditional invasive method.

Figure 2.

Field and laboratory egg staging guide for H. gammarus [adapted from University of New Hampshire (2004)] stage 1: 1, 1a; stage 2: 2-, 2a; stage 3: 3, 3a; stage 4: 4, 4a.

To assess changes in changes in brood size (egg loss) during development all 116 ovigerous females sampled from the Orkney commercial lobster fishery from October 2016 to August 2017 were used. Broods were visually assessed against egg development criteria (Figure 2) to ensure coverage across all four egg development stages. Fecundity at each of the egg stages was then estimated through the use of the non-invasive method.

Lobster egg development stages were characterized using a combination of macroscopic and microscopic criteria. Egg development of H. gammarus was categorized according to four egg development stages (Figure 2, Table 1), adapting the approach applied to H. americanus by University of New Hampshire (2004) with alteration based on the development rates observed in H. gammarus (Pandian, 1970; Hepper and Gough, 1978). Broods were initially visually assessed and assigned to one of four development stages, 15 eggs per brood were also removed and egg size and eyespot criteria determined microscopically in the laboratory and cross referenced with visual staging. Egg size was calculated by averaging the longest and shortest axes of the 15 eggs (Figure 1), measured to the nearest 0.1-mm under ×3 magnification using a calibrated eyepiece graticule on an Olympus CH2 Microscope. ESI was calculated as outlined by Perkins (1971), calculated by measuring the long and short axes of the developing embryos eye pigment crescent.

Table 1.

Field and laboratory egg development staging criteria for H. gammarus.

Egg stage	Egg description
	Field assessment—macroscopic	Laboratory assessment—microscopic
1	Black in colour	Eye index 0–0.39, Diameter 1.9–2 mm, Eye small/only visible under microscope
2	Two toned black/purple, Black yolk ½ of egg	Eye index 0.39–0.7, Diameter 1.9–2 mm
3	Orange/red in colour, Black yolk ½ of egg	Eye Index 0.6–0.8, Diameter 2–2.1 mm
4	Dark red/brown/green in colour, Black yolk 1/8 of egg, Developed larvae present	Eye index >0.8, Egg Diameter >2.2 mm

Fecundity estimates

Traditional invasive method: gravimetric method

Brood masses were removed from individuals, each being secured ventral side up inside a white polythene tray, and eggs were carefully removed from pleopods using curved tweezers. Removed brood masses were stored in sea water from Stromness harbour and processed within 24 h of capture. Broods were carefully rinsed, blotted dry and then oven dried at 100°C for 24 h or until constant weight was achieved. Brood masses were then sieved over a 250 µm mesh to remove any residual connective tissue, and weighed. The average egg dry weight was determined for five counted subsamples of 30 eggs. All weights were determined to the nearest 0.001 mg using an A&D HR-120 balance. Fecundity was then estimated as the total dry weight of the brood mass divided by average egg dry weight.

Non-invasive sampling: depth gauge methods

Immediately after capture, the brood mass of each individual was measured using a non-invasive depth gauge method based on Currie et al. (2010) to estimate fecundity (Figure 3). First, the entire egg mass length was measured using Vernier callipers (to the nearest mm). The height of each brood mass segment was measured using a ruler (∼1-cm wide) inserted between each segment, recording the height measured from ventral surface of the abdomen to the highest point of the brood mass attached to each segment. To estimate egg volume (EV) the same 15 undamaged eggs that were used to categorize egg stage development were also used to estimate EV. Average EV per brood was obtained by calculating the average egg radius across 15 eggs sampled. Following Currie et al. (2010), EV was calculated by using the formula for a spheroid:

$$\mathrm{EV}=\frac{4}{3\left(\pi r^3\right)}.$$

Figure 3.

Measurements taken to estimate fecundity by the non-invasive technique. (a) Ventral view of ovigerous lobster abdomen, showing length measurement of entire egg mass (A1). (b) Ventral side of ovigerous lobster abdomen, showing measurements of each egg segment depth (A2, A3, A4, A5, and A6). (c) Cross section of brood mass, illustrating placement of ruler in between segments of the egg mass from where height measurements are taken. (d) Long and short axis measurements used to calculate EV. Adapted from Currie et al. (2010).

Upon evaluation of the Currie et al. (2010) method, two areas of improvement were highlighted: brood mass and egg packing density (PD). Lobster brood mass shape is not typically cylindrical, pleopod length increases from the seminal receptacle along the abdomen and subsequently decrease towards the uropods, with local egg number positively related to pleopod capacity at each segment of the brood mass (Gunning, 2012), resulting in the creation of an overall domed brood mass shape. The original model used a half cylinder volume arrangement (Figure 4a), but our observation of brood masses indicated that a half ellipsoid profile is likely a better representation of the natural shape of a brood mass in H. gammarus (Figure 4b). Underlying differences in overall egg size will fundamentally influence PD of eggs in the brood and consequently fecundity. In the case of H. americanus eggs are generally smaller at extrusion and hatch than in H. gammarus, contributing to observed greater fecundity in the former. In the case of Currie et al. (2010) a PD of 0.535 was used; due to differences in egg size between Homarus spp. the estimation of a H. gammarus-specific PD is required. Differences between the two assumptions about brood volume shape and PD were investigated.

Figure 4.

Proposed longitudinal view packing arrangement of brood mass assumed within non-invasive method; (a) Cylinder packing arrangement used by Currie et al. (2010). (b) Ellipse packing arrangement.

Cylinder fecundity estimate

Total brood volume was estimated according to a half cylinder volume arrangement (Figure 4a). The volume of the brood mass approximated as a half cylinder (CBV) was estimated as:

$$\mathrm{CBV}=0.5\times \pi \times H^2\times L,$$

where H is average pleopod segment depth (average of A2, A3, A4, A5, and A6 in Figure 1) and L is the entire length of the brood (A1 in Figure 3). The factor 0.5 in this equation represents the approximation of a half cylinder, owing to the egg mass lying against the flat surface of the female’s abdomen. Egg PD for the cylinder approximation (PDC) was estimated as the slope of least-squares regression of fecundity estimates derived from the direct (invasive) technique on CBV calculated for the same lobsters, with the regression constrained to a zero intercept. Fecundity (F) was then calculated as:

$$F=\frac{\mathrm{CBV}\times \mathrm{PDC}}{\mathrm{EV}}.$$

Ellipse fecundity estimates

Total brood volume was estimated according to a half ellipsoid volume arrangement (Figure 4b). The volume of the brood mass approximated as a half ellipsoid was estimated as:

$$\mathrm{EBV}=0.5\times \frac{4}{3}\times \pi \times \frac{\mathrm{AW}}{2}\times H\times \frac{L}{2},$$

where AW is abdomen width, measured at the maximum width of the second abdominal segment. Egg PD for the ellipsoid approximation (PDE) was estimated similarly to PDC from regression of fecundity estimates derived from the direct (invasive) technique on EBV calculated for the same lobsters. Fecundity was then estimated as:

$$F=\frac{\mathrm{EBV}\times \mathrm{PDE}}{\mathrm{EV}}.$$

Note that the only meaningful difference between the two brood volume approximations is in the substitution of AW × H for H² in the ellipsoid formula; all other differences are scaling factors, which are adjusted for in the estimation of egg PD.

Fecundity estimates in relation to estimation method and lobster size

Power law models were used to describe the size-fecundity relationship for H. gammarus:

$$F=a{\mathrm{CL}}^b,$$

where F is fecundity, CL is carapace length, and a and b are coefficients to be estimated. This functional form has been favoured in recent studies (Tully et al., 2001; Lizárraga-Cubedo et al., 2003; Agnalt, 2008; Ellis et al., 2015) due to the volumetric nature of brooding capacity of H. gammarus, where brood capacity increases with both CL and abdomen width. Coefficients of the relationship are estimated by linear regression of natural log-transformed variables:

$$\ln F=\ln a+b\ln \mathrm{CL}.$$

Egg loss during development

A generalized linear model (GLM) was used to describe and estimate levels of egg loss during development. Log-transformed fecundity was the dependent variable; explanatory variables were egg development stages, treated as a factor, log-transformed CL and the interaction between these two variables. The CL term accounts for the power law relationship of fecundity with lobster size, as described earlier. A significant interaction term indicates an allometric change in the size-fecundity relationship over the course of egg development, which is equivalent to saying that rates of egg loss differ between lobsters according to size. If the interaction term is not statistically significant and is removed from the model the coefficients for the factor levels are estimates of the instantaneous rates of change (loss) in egg number between successive developmental stages. All analyses were performed using the GLM procedure in R Core Team (2015).

Egg size during development

Generalized linear modelling was used to explore the relationship of egg size with development stage and lobster size. The dependent variable was average egg diameter with explanatory variables being development stage as a factor and log-transformed CL. First-order interactions between these variables were considered.

Results

Fecundity estimate results

The total number of eggs per lobster determined by the traditional (invasive) method of measuring fecundity varied from ∼8000 from the smallest female of 89-mm CL to ∼34 000 from the largest at 151-mm CL. Using the depth gauge methods, cylinder fecundity estimates varied from ∼7900 at 89-mm Cl to ∼31 000 at 151-mm CL, with average differences between traditional and cylinder depth gauge estimates being −0.31% (±2.7% s.e.). Ellipsoid fecundity estimates varied from ∼78 000 from 89-mm CL to 35 000 at 151-mm CL, with average differences between traditional estimates and ellipsoid depth gauge estimates being −1.1% (±2.4% s.e.).

Estimated PD estimated from the non-invasive method was PDE = 0.504 for the half ellipsoid and PDC = 0.344 for the half cylinder approximation of brood volume (mm³) (Figure 5).

Figure 5.

Packing density estimates for cylinder and ellipsoid brood arrangements for non-invasive fecundity estimates (n = 30).

The relationship between egg number and CL provided evidence of a strong size-fecundity relationship within H. gammarus; power law functions fitted to data from traditional and non-invasive techniques showed little discernible difference between estimates, and at least 88% of the variation (R²) in egg number was accounted for by lobster CL in each case (Figure 6). Use of the ellipsoid approximation for brood size volume in calculating egg-PD provided the closest estimates to those from the traditional method, and this is identified as the preferred method for non-invasive assessment of fecundity.

Figure 6.

Comparison of traditional and depth gauge fecundity estimate models (n = 30).

The exponent of the power law relationship (b) is close in value to 3 for all three methods, demonstrating that fecundity is indeed a volumetric relationship in H. gammarus.

Egg loss during development

Fecundity was significantly different between egg development stages (F_3,113 = 85.3, p < 0.001), with egg number decreasing by 44% from stages 1 to 4 providing evidence of egg loss during incubation (Table 2). However, no significant difference in the slope of the size-fecundity relationship (power law exponent, b) was found between egg stages (F_3,110 = 1.41, p = 0.2), demonstrating that the shape of the size-fecundity relationship remains the same during the course of development, i.e. proportional egg losses are independent of female size.

Table 2.

Summary of GLM estimates investigating the effect of egg development stage on fecundity; log (fecundity) ∼ egg stage + log(CL).

Coefficients	Estimate	s.e.	t-Value	p
Intercept	−5.035	0.67	−7.490	<0.001
Stage 2	−0.15	0.04	−3.408	0.016
Stage 3	−0.52	0.05	−9.219	<0.001
Stage 4	−0.58	0.05	−10.090	<0.001
log(CL)	3.12	0.14	21.842	<0.001

Egg size during development

As an indicator of development, egg size according to egg stage was analysed. Eggs classed as stage 1 had a mean diameter of 1.9 mm, whereas those at stage 4 had a mean diameter of 2.3 mm. Egg size significantly increased with egg development stage (F_{3, 113} = 88.8, p < 0.001) (Table 3) and lobster size (F_{1, 113} = 13.4, p < 0.001). As would be expected, ESI also increased with both egg stage and egg size (Figure 7). Such changes can be attributed to the swelling of the egg coinciding with embryo development (Perkins, 1972; Roberts, 1992)

Figure 7.

Plot of ESI relative to egg size and corresponding egg development stage (n = 116).

Table 3.

Summary of estimates from GLM investigating the effect of egg development on egg size: egg size ∼ egg stage + log (CL).

Coefficients	Estimate	s.e.	t-Value	p
Intercept	0.850	0.29	2.87	<0.01
Stage 2	−0.033	0.01	1.66	0.09
Stage 3	0.203	0.025	8.08	<0.001
Stage 4	0.394	0.025	15.52	<0.001
log (CL)	0.231	0.06	3.67	<0.001

Discussion

This study has successfully trialled the use of a non-invasive method originally designed for H. americanus (Currie et al., 2010) and adapted it for H. gammarus. We demonstrate that the depth gauge method provides reliable and accurate fecundity estimates compared with previously applied gravimetric methods (e.g. Ellis et al., 2015). We also demonstrate that the assumption of a half ellipsoid brood volume for H. gammarus is an improvement over the half cylinder assumption applied in the development of the method for H. americanus by Currie et al. (2010). In addition, this study provides the first description of egg development staging criteria specific to H. gammarus, providing the basis for further comparison in egg development patterns across its range.

Geographical variation in fecundity has been examined previously over latitudinal and longitudinal scales for Homarus spp. (Currie and Schneider, 2011; Ellis et al., 2015). In the case of H. gammarus, samples were predominantly from central North East Atlantic populations with only limited sample sizes from the southern and northern portions of its distributional range (Ellis et al., 2015). The methodology outlined here significantly reduces the time and financial resources needed address to questions regarding geographical variation in clawed lobster productivity.

The majority of fecundity data on H. gammarus was collected more than 10 years ago [1970s—Hepper and Gough (1978); 1980s—Latrouite et al. (1984); Bennett and Howard (1987); 1990s—Free (1994); Tully et al. (2001); Lizárraga-Cubedo et al. (2003); Agnalt (2008)], or recently as one-off exercises (Ellis et al., 2015). Over this time scale there is potential for significant undocumented changes in productivity to occur (Koopman et al., 2015) without being accounted for in spawning stock biomass estimates (Briggs et al., 2002). The lack of information on current egg production or temporal fluctuations in egg production and subsequent recruitment could potentially lead to inaccurate perceptions of the state and productivity of stocks (Lambert, 2008). The vulnerability of benthic decapods to climatic alterations has been identified (Hare et al., 2016). For clawed lobster, climatic shifts could cause thermal habitat at adult, larva and egg stages to be significantly compromised at current distributional limits (Quinn, 2017). Evidence of such alteration to H. americanus stocks has already been documented, both in declining egg production in Canadian fisheries (Koopman et al., 2015) and recruitment failure in the Southern New England fishery (Hoenig et al., 2015; Wahle et al., 2015). The likelihood of H. gammarus stocks being impacted by similar changes may also be high due to the species’ lack of endogenous mechanisms to delay hatching in response to changing conditions (Schmalenbach and Franke, 2010). Shifts in population distribution (Engelhard et al., 2014) of other species owing to climatic change have already been documented within the range of H. gammarus. Nicolas et al. (2014) documented declines in North Sea Atlantic cod (G. morhua) recruitment during 1974–2011, attributing such declines to changes in sea surface temperature and corresponding environmental factors. This further highlights the need to understand long-term seasonal fecundity fluctuations and subsequent recruitment success of this commercially valuable decapod. The adoption of the non-invasive depth gauge technique therefore provides the opportunity to monitor potential changes in egg production and recruitment owing to climate change.

Underlying differences in fecundity between the two Homarus species are highlighted through the estimated egg PD within the brood mass, with H. gammarus being 0.345 (this study) and H. americanus 0.535 (Currie et al., 2010). This can be attributed to the size of eggs at extrusion, with H. americanus being 40% smaller than H. gammarus, which would contribute to a higher PD (Ouellet and Plante, 2004) and increased overall fecundity. An area of research not addressed within non-invasive fecundity estimates was the impact of changing egg shape and resulting impact this has on packaging arrangement and subsequent fecundity estimates between egg development stages. In the estimates provided, overall egg shape is assumed to remain similar throughout development, in line with existing methods used to estimate egg shape and volume as an ellipsoid (Sibert et al., 2004; Rosa et al., 2005). During larval development however, alteration of egg shape results in eggs moving from a typical spheroidal to oblong shape (Herrick, 1911; Helluy and Beltz, 1991). This change in egg shape will impact the egg packing arrangement, resulting in an overestimation of egg loss if not accounted for.

A number of previous studies have investigated the degree of egg loss. Agnalt (2008) measured fecundity of H. gammarus comparing freshly extruded eggs and those close to hatching in Kvitsøy. Norway. No significant differences were observed between samples across a 7-month period, indicating that no egg loss occurred during incubation. Similar relationships were reported for Scottish populations (Lizárraga-Cubedo et al., 2003). A re-analysis of these Scottish data however, highlights a minimum decrease of ∼19% fecundity when original ESI measurements are assigned to comparable egg development stages used in this study (Coleman, Unpublished). Our study identified egg loss to be higher still, at 40% to stage 3 and 44% by stage 4, which may be consistent with the substantial increase in EV from extrusion to point of hatching. These estimates appear high, but they are in line with levels of egg loss observed in H. americanus [36% Perkins (1971) and 47–51% Tang et al.,(2017)], whilst direct larvae counts from hatching broods of H. gammarus from Helgoland, Germany (Mehrtens, 2008) are comparable to the fecundity relationship at stage 4 for the Orkney stock. Such high egg loss has been previously considered an overestimation due to the fragility of egg stalks at later stages of development, or potential influences of sample acquisition technique (Agnalt, 2008). To control for this in our study, samples with any sign of egg loss at capture or loss suffered during fecundity measurements were excluded.

Underlying reasons for egg loss are still unknown. Possible influences include variability in optimum brooding temperature (Ellis et al., 2015), unfertilized ova (M. Coleman, pers. obs.), clutch failure (Tang et al., 2017), egg parasites (Campbell and Brattey, 1986; Simpson et al., 2017) and fungal infection (Alderman and Polglase, 1986). Although we do not fully understand the exact reasons for egg loss, having a robust method to measure the degree of egg loss, is essential for developing criteria for sustainable fishery management (Briggs et al., 2002).

With the exception of the Norway lobster N. norvegicus, management frameworks for clawed lobster fisheries across Europe typically operate under “open access,” with no total allowable catch limits. In the absence of explicitly defined harvest control rules, and associated needs to define and update target and limit biological reference points, there has been little incentive for fishery managers to commission and maintain collection of data on important biological parameters such as fecundity. Baseline research on fecundity and other life-history traits, and sources of their variability, is fundamental to the provision of effective fishery management advice (e.g. Lambert, 2008; Stange, 2016). With current trends towards regionalization of fishery management in Europe (van Hoof, 2015) and market-driven adoption of ecolabels awarded to demonstrably sustainable fisheries (Gutierrez et al., 2016) there is an increasing need for regionally appropriate and up-to-date information on life-history parameters underlying stock productivity. The application of the improved method for assessing fecundity in H. gammarus is relevant to the recent enforcement of a national berried ban within England (The Lobster and Crawfish (Prohibition of Fishing and Landing) (Amendment) (England) Order, 2017). The ultimate aim of this measure is to improve egg production and increase stock spawning biomass. The non-invasive method provides the most useful tool in assessing its impact of such measures without undermining the fundamentals of the conservation measure. Egg per recruit analyses are often used to derive proxy measures of spawning capacity in exploited crustacean stocks (e.g. Mesquita et al., 2016; ICES, 2017). Although these analyses are independent of absolute scaling factors for fecundity, they are nevertheless sensitive to the shape of the size-fecundity relationship, which in the case of a power law model is expressed through the value of the exponent b (see Table 2). Our finding that there is potential for significant egg loss during development in H. gammarus highlights the need to account for egg development stage in deriving fecundity relationships. Some previous studies have used eggs at more than one stage of development in estimating the size-fecundity relationship (Lizárraga-Cubedo et al., 2003; Agnalt, 2008) introducing the possibility of biased estimates being taken forward into modelling of spawning potential.

Previous studies have classified H. gammarus egg development stages using either physiological criteria (Pandian, 1970) or changes in eye pigment development as a proxy of overall larvae development (Hepper and Gough, 1978; Latrouite et al., 1984; Charmantier and Mounet-Guillaume, 1992; Lizárraga-Cubedo et al., 2003). Pandian (1970) classified egg development across five stages in H. gammarus using mean egg size, coloration and yolk content. These stages directly compare with those developed in this study across similar stages, with the difference being the amalgamation of Pandian’s stages 4 and 5 into a single stage 4 to incorporate the presence of near hatching larvae into one stage. Studies tracking egg development using an ESI identified full development and hatching to be occurring at a mean ESI of 650 µm (Charmantier and Mounet-Guillaume, 1992), 600–620 µm (Richards and Wickins, 1979) and 634 µm (Hepper and Gough, 1978) in H. gammarus. These are all substantially lower than ESI observed in this study, with mean ESI of broods containing fully developed embryos being 800 µm. The observed larger ESI in this study may be due partly to our assignment of eggs to more narrowly defined development stages with corresponding ESI ranges; historical ESI measurements correspond to stage 3 eggs under the criteria of this study. Such differences in ESI at hatch can also be attributed in part to these studies characterizing development using just ESI. The use just one metric could increases the possibility for error, whilst also making direct comparison difficult with other studies. Differences of localized environmental conditions between studies can influence egg development in H. gammarus, with even small variations of ±0.5°C in annual average sea temperature influencing development time (Branford, 1978). This is the first study to characterize ESI at the later stage of development of H. gammarus at this latitude (∼59°N), highlighting the potential of geographic adaption to localized conditions resulting in the development of larger larva and larger ESI. Latitudinal differences in larval size have not been explored, with similar localized adaption behind observed geographical differences in H. gammarus fecundity having only recently been examined (Ellis et al., 2015). Differences in larval size depending on latitude gradient however have been documented within other marine species, with flatfish populations exhibiting larger larvae at higher latitude compared with those at lower ones, with drivers related to variability in food regimens (Gibson et al., 2014). Differences in observed ESI therefore present a further area of research into understanding the recruitment dynamics of the H. gammarus, with the use of standardized development criteria aiding in the comparison across its range.

The successful adaptation of the non-invasive technique for estimating fecundity to H. gammarus and the identified differences in egg packing arrangements highlights that this technique could also be applicable to other suitably shaped “cylindrical” decapods (e.g. Nephropidae, Palinuridae). Each species would require calibration, but once completed this would provide the capability to derive and update reliable estimates and address key knowledge gaps within commercial fisheries for these species.

Conclusion

We demonstrate the reliable application of a non-invasive technique to estimate fecundity in European lobster, with improvements to the original model. Based on estimates derived from the technique applied to broods at different stages of development, we conclude that it is important to consider egg loss during incubation. The method presents a cost-effective method for generating time-series of fecundity data, essential for understanding long-term patterns of recruitment in commercially important crustacean stocks. Wider application of this method to other decapods, has the potential to be a powerful tool for research underpinning fisheries management and stock conservation.

Acknowledgements

This study would not have been possible without the generous participation of Orkney fishermen and Orkney Fishermen’s Society. Collaborative funding to Orkney Sustainable Fisheries Ltd was provided from the Orkney Shellfish project and its funders (The Crown Estate, WWF UK, M&S, and Orkney Islands Council).