Role of depleted initial energy reserves in early benthic phase mortality of six marine invertebrate species

Abstract Insufficient energy reserves are widely considered to be a primary factor contributing to high rates of early benthic phase mortality among benthic marine invertebrates, but this hypothesis has been based mostly on indirect, observational evidence, and remains largely untested. We therefore examined the role of initial energy reserves in regulating survivorship and growth during the early benthic phase. Recently settled or hatched individuals of six invertebrate species were collected from natural populations, maintained without food, and their survivorship was monitored. Contrary to expectations, starved individuals of all six species had high survivorship through the critical first 10 days of the early benthic phase, with half of the species experiencing <2% mortality, and the remaining three species experiencing only 6%–12% mortality. For five of the six species, 50% mortality was reached only after ≥50 days of starvation. Additionally, no difference in short‐term survivorship was detected among starved individuals of three different size classes (a proxy for energy reserves) of N. ostrina hatchlings. Finally, the effect of different durations of delayed feeding (0–50 days) on recovery (i.e., growth and survivorship) once food was made available revealed that duration of starvation prior to feeding can nevertheless have significant longer‐term impacts on the proportion of individuals that survive or their ability to grow. Together, these findings suggest that depleted energy reserves are not a primary cause of high mortality at the start of the early benthic phase, as had previously been hypothesized. Levels of energy reserves did influence growth, however, suggesting a possible indirect influence on performance by leaving individuals vulnerable for longer periods.

The mechanisms responsible for the high rates and extensive variability in EBP mortality are not yet fully understood and remain an area of active research.
A factor that has been proposed as a major cause of mortality during early benthic life is depleted energy reserves, when individuals begin the EBP with minimal energy stores (Gosselin & Qian, 1997;Hunt & Scheibling, 1997;Jarrett & Pechenik, 1997). Energy reserves at the start of independent benthic life are known to affect other relevant aspects of EBP performance, including the subsequent size and growth rate of individuals throughout juvenile life (Emlet & Sadro, 2006;Miller & Emlet, 1999;Miller, 1993;Pechenik et al., 1996Pechenik et al., , 2002Thiyagarajan et al., 2005). Previous studies have also revealed a correlation between initial body size and mortality, with smaller individuals experiencing higher rates of mortality than larger EBP individuals (Moran, 1999;Phillips, 2002Phillips, , 2017Thiyagarajan et al., 2002). Those studies have suggested the relationship between body size and mortality may be due to increased vulnerability to predation, competition, or abiotic stressors, but it might also be due to insufficient initial energy reserves directly causing EBP mortality. While these indirect lines of evidence are consistent with the hypothesis that depleted initial energy reserves directly cause EBP mortality, the hypothesis remains untested.
For the depleted initial energy hypothesis to be correct, there must be variation in initial energy reserves, with some individuals having relatively low energy levels. Accordingly, variation in energy reserves at the onset of the EBP among individuals does occur in at least some species, with some individuals beginning the EBP with relatively low energy reserves (Jarrett & Pechenik, 1997;Moran & Emlet, 2001). This variation is largely due to differences in maternal provisioning or in food availability to the larvae, as well as competent larvae of many species transitioning through energetically demanding periods of development prior to the onset of the EBP (Lucas et al., 1979;Thiyagarajan, Harder, Qiu, et al., 2003). The most energetically demanding of these developmental periods is metamorphosis, which involves extensive rearrangement of tissues and loss or replacement of numerous body structures (Bennett & Marshall, 2005;Bryan, 2004;Thiyagarajan et al., 2002;Wendt, 2000). Additionally, development in some species includes a substantial nonfeeding phase immediately before, during, or following the onset of the EBP (Anderson, 1994; Gosselin & Chia, 1994). During these nonfeeding periods, and until metamorphosis is completed and feeding begins, individuals are solely reliant upon their internal energy stores for resting and active metabolism (Gosselin & Chia, 1994;Lucas et al., 1979;Pechenik et al., 1993). Accordingly, it has been suggested that some portion of individuals may begin the EBP with such critically low energy stores that they are near a minimum threshold necessary to survive (Jarrett & Pechenik, 1997); those individuals would either inevitably die of energy depletion very soon after metamorphosis, or need to quickly feed to replenish energy stores to survive. We illustrate this hypothesis in Figure 1, representing two hypothetical frequency distributions of energy reserves among EBP individuals as well as a theoretical minimum threshold of energy reserves represented by the vertical line. Curve A represents a situation where all individuals have sufficient energy to survive, in which case the depletion of energy reserves would not be a direct cause of EBP mortality.
Conversely, curve B represents the current hypothesis, that a proportion of individuals begin the EBP below the minimum energy reserve threshold, and thus inevitably die during or shortly after metamorphosis. In addition, if the hypothesis represented by curve B is correct, then a substantial proportion of EBP individuals would have energy reserves only slightly larger than the minimum threshold; these individuals would need to feed immediately to survive or would otherwise drop below the threshold and die. Thus, a goal of the present study was to test these two hypotheses to establish the role of energy reserves as a direct cause of EBP mortality.
While EBP mortality rates are often very high, they are also highly variable among and within populations (Gosselin & Qian, 1997;Hunt & Scheibling, 1997;Jarrett, 2000;Phillips, 2017). The hypothesis that variation in energy reserves (bell curves in Figure 1) is a cause of the variability in EBP mortality is partly supported by evidence (Jarrett & Pechenik, 1997;Phillips, 2006) of natural variation among individuals F I G U R E 1 Two hypothetical frequency distributions (bell curves) of energy reserves among individuals at the onset of the EBP. The dashed vertical line represents a minimum energy threshold required for EBP survival in the amount of energy reserves present at the onset of the EBP.
Internal energy reserves can vary depending on the source of energy, and in some cases upon the quantity of food available prior to the beginning of the EBP. For species with planktotrophic larvae, EBP energy reserves originate primarily from particulate food consumed during their larval phase. For these species, energy reserves can vary among individuals depending on the quantity and quality of food they encounter (Jarrett, 2003;Moran & Manahan, 2004), whether they transition through a nonfeeding phase, and whether they encounter conditions that delay settlement (Pechenik et al., 1993;Phillips, 2002;Thiyagarajan et al., 2007). For direct developing species, EBP energy reserves originate entirely from maternal provisioning and thus may vary with the condition of the maternal parent (Moran & McAlister, 2009;van der Sman et al., 2009) or with the genetic programming of the population or species (Allen & Marshall, 2013;Rivest, 1983). In some cases, unfertilized nurse eggs are provided to the embryos to consume as they develop; for these species, energy levels can further vary among individuals of the same cohort depending on the quantity of nurse eggs consumed before hatching (Jablonski & Lutz, 1983;Lloyd & Gosselin, 2011;Marko et al., 2014). All these differences, both interspecific and intraspecific, result in individuals beginning the EBP with vastly different amounts of internal energy reserves, and this might play a substantial role in determining which individuals survive through the first few days of independent benthic life. Additionally, any of these factors has the potential to shift the frequency distribution of energy reserves to the left or right, as depicted by the bell curves A and B in Figure 1, or even alter the shape of these curves and thus also affect the proportion of individuals falling below the minimum threshold for survival.
For individuals starting the EBP with only slightly more internal energy than the minimum threshold, rapidly accessing a new food source to replenish their energy reserves could be vital to their survival. A delay in feeding at this point might drop them below the threshold and thus push them beyond their "point of no return" (commonly referred to as PNR in other publications), a pivotal point beyond which animals are unable to recover even if provided with food (Blaxter & Hempel, 1963;Moran & Manahan, 2004). For individuals that do not feed before this point, both their survivorship and ability to grow are likely to be impaired (Chen et al., 2005;Roberts et al., 2001;Takami et al., 2000). The existence and timing of a point of no return has been studied for the larval phase of several species of intertidal invertebrates (Espinoza et al., 2017;Gebauer et al., 2010;Moran & Manahan, 2004;Yan et al., 2009) but has not been well established for EBP individuals of most species.
Determining the starvation point at which an individual's energy reserves are so heavily depleted that they are unable to recover may be useful for understanding how energy reserves impact EBP survivorship and growth.
The main purpose of this study was to determine the role of energy content in regulating survivorship during the critical first days of independent benthic life, thus helping to better understand the mechanisms controlling EBP mortality, and thus recruitment, of benthic invertebrates. The first specific objective of the study was (1) to establish whether energy reserves of EBP individuals could be experimentally controlled by starvation; this was addressed by determining the relationship between total organic content of EBP individuals and the duration of starvation they had experienced. This was followed by two additional specific objectives: (2) to test the hypothesis (Figure 1, curve B) that a large proportion of individuals begin the EBP with critically low energy stores such that they either die or require rapid access to food to survive the first days of independent benthic life, and (3) to examine the effect of delayed feeding on the performance (growth and survival) of EBP invertebrates over a period of several weeks.

| Study species
This study examined the role of initial energy reserves on survivorship in six species of benthic intertidal invertebrates. All animals used in this study were from wild populations and had developed through Stimpson 1871 are small (~2 cm) porcellanid crabs that inhabit cryptic habitats, primarily the undersides of large rocks (Jensen, 1989), in the low to mid intertidal zone; EBP individuals of these two species could not be readily distinguished, so these were grouped together and designated as Petrolisthes spp.
These species have different sources and quantities of energy reserves at the beginning of the EBP. N. ostrina and N. lamellosa are lecithotrophic and thus depend entirely on maternal provisioning for EBP energy reserves. The four other species, however, have planktotrophic larvae, and a large portion of their EBP energy reserves come from the food they consume during the larval stage.
Additionally, B. glandula and C. dalli have a nonfeeding cyprid larval stage prior to metamorphosis, and thus use up some energy stores prior to the EBP, whereas M. trossulus and Petrolisthes spp. are able to feed up to and throughout metamorphosis. By including species with different developmental histories, this study explored how EBP survivorship is impacted by energy reserves across species with different modes of energy acquisition prior to entering the EBP.

| Study site and collection of animals
Research was conducted at the Bamfield Marine Sciences Centre (BMSC) on the west coast of Vancouver Island from May to September of 2019. Animals were collected from five rocky intertidal shore sites located in Barkley Sound ( Figure 2). Sites were selected based on the availability of EBP animals of each species.
Collected animals were brought to and maintained at BMSC, where all experiments were conducted.
Juvenile barnacles (B. glandula and C. dalli) were collected as described by Sandee et al. (2016) and Hamilton & Gosselin (2020). Small rocks (5-10 cm diameter) were gathered from the intertidal zone and brought to the laboratory where all small settlers were dislodged under a dissecting microscope, and a perimeter was marked on the rock around the cleared area with nail polish. Marked rocks were then returned to the field and left for 48-72 hr (4-6 tidal cycles) to allow for new settlement, after which rocks were recovered and examined for newly settled barnacles. At that age, barnacles would still be reliant upon energy reserves obtained during the larval phase, as juvenile barnacles do not begin to feed until 2-5 days postmetamorphosis (Rainbow & Walker, 1977). Ripe egg capsules of both Nucella species (N. ostrina and N. lamellosa) were collected from intertidal substrata using fine-tipped forceps as in Gosselin and Chia (1995).
Egg capsules were returned to the laboratory and held in small cages in aerated seawater for 72 hr, after which all newly hatched juvenile snails were removed and retained for experiments. EBP M. trossulus were extracted from tufts of the filamentous algae Cladophora columbiana. Extracted mussels were measured and only recently settled juveniles measuring 250-600 μm shell length (Phillips, 2017) were retained for experiments. Petrolisthes spp. megalopae were collected from the underside of large rocks in the intertidal zone.
Megalopae were carefully removed with a pair of soft insect forceps and transported to BMSC in small containers. All animals were returned to the laboratory within 3 hr of collection.

| Procedures for rearing EBP individuals without food
To experimentally control the energy reserves of EBP individuals, most experiments in this study involved rearing EBP individuals without food. In all experiments described below, the following procedures were used to rear EBP individuals of each species while ensuring they did not have access to food. All animals were held in seawater filtered to 1.0 μm to ensure no food particles were available for feeding. M. trossulus, however, can capture particles <1.0 μm (Strohmeier et al., 2012), so filtered water used for this species was also autoclaved and then vacuum-filtered to 0.45 μm. Rocks containing newly-settled barnacles were kept in 1 L containers, each with an air stone for aeration. All other animals were kept in small cages (1.5-600 ml) with mesh-covered cut-out windows to allow for water flow while preventing escapement. Water was maintained at 15-17°C, which approximates average sea surface temperatures during the summer months (Iwabuchi & Gosselin, 2019). Filtered water was replaced every second day, and holding containers and cages were rinsed and scrubbed to prevent the growth of algae.

| Relationship between organic matter content in EBP individuals and duration of starvation
To determine whether starvation could be used as a method for predictably depleting energy reserves in EBP individuals, we tested the hypothesis that energy reserves decrease as a function of the duration of starvation. This experiment was carried out with N. ostrina hatchlings, and total organic matter (OM) was measured by ash-free dry weight (AFDW). AFDW is an effective, albeit simple, indicator of energy content (Moran & McAlister, 2009). It does not, however, reflect the energetic value of any single type of organic matter, such as triacylglycerols (Podolsky, 2002), which are the primary source of metabolizable energy for marine invertebrates (Sewell, 2005;Whitehill & Moran, 2012). Thus, it is assumed that the first OM to be depleted by starved EBP individuals will be these energetic lipids, These were briefly rinsed in distilled water to remove adhering salts and then stored at −80°C. AFDW of each sample was obtained using a procedure adapted from Pechenik et al. (1993). For each sample of 52-60 frozen snails, the sample (all snails together) was placed in a preweighed foil pan and dried at 65°C for 48 hr. After drying, the pan F I G U R E 2 Map of field sites in Barkley Sound near Bamfield, British Columbia: Ross Islets (R), Wizard Islet (W), Prasiola Point (P), Dixon Island (D), and Grappler Inlet (G). Map modified from Gosselin and Chia (1995) was weighed on a Fisher Scientific accuSeries balance with an instrument precision of ±0.01 mg to quantify total dry weight and then placed in a muffle furnace at 500°C for 5 hr. The pan was then placed in a sealed desiccator containing silica gel desiccant for 20 min to cool, and then reweighed to determine ash weight (i.e., the weight of the sample without the OM). AFDW, a measure of OM content of each sample, was calculated by subtracting the ash weight from the initial dry weight. Ash weight consisted almost entirely of the shell, and shell weight was not expected to decline with starvation, so ash weight was used as an indicator of initial snail body mass prior to starvation. The OM estimate (AFDW) was therefore divided by the ash weight to obtain the OM:ash weight ratio; this ratio standardized results for all snails, accounting for any minor differences in initial body size among individuals. This experiment was carried out twice, using two different batches of hatchlings. Batch 1 was collected from Ross Islets on 17 July 2019, and Batch 2 was collected from Prasiola Point on 26 July 2019.

| Survivorship of EBP individuals unable to replenish energy reserves
This experiment determined, for each of the six species, the proportion of individuals that begin the EBP with energy reserves close to the minimum threshold required for survival. EBP individuals were collected on several dates throughout the spring and summer, and each batch of individuals was tested separately as a replicate trial. For each species, 2-8 batches of individuals were tested, each batch consisting of 9-200 individuals (Appendix S1). The number of batches collected for each species and the number of individuals in each batch were dependent on the availability of EBP individuals in the field. All animals were held in 50-mL cages without food, as described above, for up to 100 days starting immediately after the onset of the EBP. Every 10 days, survival was assessed by examining all individuals for movement under a dissecting microscope. If an animal did not show any signs of movement within 2 min, their tissues were gently contacted with a probe and they were observed for an additional 2 min. If no movement was observed at this point, the individual was recorded as dead and removed from the experiment. The proportion of individuals surviving each 10-day period was recorded.

| Influence of initial body mass on survivorship of EBP individuals unable to replenish energy reserves
There is a strong correlation between body mass and quantity of organic matter among EBP invertebrates (Moran & Emlet, 2001).
Specifically, larger individuals have larger stores of lipids (Emlet & Sadro, 2006;Phillips, 2002), which are the primary metabolizable energy source for many intertidal species (Sewell, 2005;Thiyagarajan et al., 2002;Whitehill & Moran, 2012). Thus, body mass was used as a proxy for energy reserves to determine their influence on survivorship. To determine how the ability to survive periods of starvation is influenced by body mass at the start of the EBP, the survivorship of three size classes of newly hatched N. ostrina were examined. Most N.
ostrina egg capsules contain 7-20 fertilized embryos that will hatch into juvenile snails, as well as 500-600 unfertilized nurse eggs that the embryos consume as they develop inside the capsule (Lloyd & Gosselin, 2011;Marko et al., 2014). N. ostrina hatchlings consume different numbers of nurse eggs while in the capsule, and therefore, hatchlings range considerably in size upon emergence, with larger hatchlings having consumed more nurse eggs and thus having greater energy reserves than smaller hatchlings. Newly hatched N. ostrina were sorted into three shell length (SL) categories: small (0.81-1.10 mm SL), medium (1.11-1.40 mm SL), and large (1.41-1.80 mm SL). These SL measurements were then converted to body mass (wet weight, WW) using a regression equation for N. ostrina from Hamilton & Gosselin (2020) ( Table 1). Eight batches of hatchlings were collected throughout the spring and summer; each batch was sorted into the three size classes, and each size class of a given batch contained 11-217 hatchlings (Appendix S1). In some batches, certain size classes were not included in this experiment because too few individuals of the size class were available. Hatchlings were held in 50 ml cages without food for up to 70 days, beginning at the time of hatching (i.e., the onset of EBP). Every 10 days, survival was assessed by examining all individuals under a dissecting microscope, as described above, and the number of survivors of each size class was recorded.

| Ability to recover after delayed feeding in EBP invertebrates
The ability of EBP individuals to recover from periods of starvation was examined in three species: B. glandula, C. dalli, and N. ostrina. EBP individuals of each species were reared without food, beginning at the onset of EBP, for various durations (0, 10, 20, 30, 40 or 50 days) before being fed for 30 days. Growth during the 30-day feeding period was determined by measuring the size (rostro-carinal diameter for barnacles, SL for N. ostrina) of each individual at the end of the starvation period (i.e., just before feeding began), and then again at the end of the 30-day feeding period. The number of individuals that died during the 30-day feeding period was also recorded for each group. All individuals of the same species were within a narrow size range at the beginning of the experiment. The experimental design was as follows: 3 species X 6 duration of starvation treatments per species X 10-15 individuals per starvation treatment.
During the feeding period, B. glandula and C. dalli were held in 1-L aerated containers and fed a mixture of diatoms (Chaetoceros muelleri and C. gracilis) and flagellates (Tertaselmis spp.) that had been reared in the laboratory. Algae were provided at a total concentration of 1x10 6 cells/mL, determined by counting algal cells with a hemocytometer. Water and algae were changed every second day.
To feed N. ostrina hatchlings, each individual was placed in a small (1.5 ml) cage with 10 juvenile mussels (M. trossulus), a preferred prey species for EBP N. ostrina (Gosselin & Chia, 1994). N. ostrina feed by drilling a small hole through the shell of their prey with their radula, making it possible to determine whether the hatchling had fed.
To assess the effect of delayed feeding on the performance of EBP invertebrates, individuals were categorized as either recovering (i.e., alive and showing growth) or not recovering (i.e., dead, or alive but with no growth) following the 30-day feeding period. The percent of individuals that were recovering in each starvation treatment was then calculated for each species. Additionally, the growth (i.e., change in size) of each EBP individual during the feeding period was calculated to assess the effects of delayed feeding on the ability to resume growth when food becomes available.

| Data analysis
In all experiments, survivorship data followed a binomial distribution.
All survivorship data were analyzed using generalized linear models (GLM) with a binomial link function to determine the starvation LD 50 for each species, and the % survivorship after 5-day starvation and the starvation LD 20 for each size class of N. ostrina. The fit of each GLM model was tested by comparing the residual deviance to the residual degrees of freedom, which indicated goodness of fit for each model (i.e., residual deviance / residual df < 1). One-way ANOVA and the Tukey HSD multiple comparisons test were used to compare the % survivorship after 5-day starvation and the starvation LD 20 among the three size classes of N. ostrina hatchlings. The Shapiro-Wilk normality test and the Bartlett test of homogeneity of variances were used to assess the assumptions of ANOVA. All statistical analyses were conducted using R statistical software (version 3.5.0) (R Core Team, 2018).

| Relationship between organic matter content in EBP individuals and duration of starvation
The relationship between OM:ash weight ratio and duration of starvation starting at the onset of the early benthic phase (EBP; The duration of starvation that is lethal to 50% of individuals (starvation LD 50 ), as determined from the above GLM analyses, varied substantially among species (Table 2). Starvation LD 50 ranged TA B L E 1 Shell lengths (SL) and corresponding body masses (wet weight, WW) for N. ostrina hatchlings

| Ability to recover after delayed feeding for EBP invertebrates
The duration of starvation experienced before being offered food had a significant or nearly significant effect on the percent of individuals able to recover (i.e., survive and grow during the 30 days feeding period) for C. dalli (Linear regression: F = 17.56, n = 6, R 2 = 0.81, p =.014) and N. ostrina (F = 7.08, n = 6, R 2 = 0.64, p =.056) ( Figure 7). The duration of starvation resulting in only 50% of individuals being able to recover (i.e., the recovery LD 50 ) was >20 days in both species, but was somewhat longer in C. dalli than in N. ostrina (C. dalli: 28.6 ± 4.2 days (SE); N. ostrina: 20.9 ± 6.6 days).
In B. glandula, the ability of individuals to recover was not significantly affected by the duration of starvation they experience Individuals starved for longer periods of time grew at a substantially slower rate than those starved for a shorter period once particulate food was introduced.

| Relationship between organic matter content in EBP individuals and duration of starvation
Energy content in early benthic phase (EBP) N. ostrina declined rapidly, and in a predictable way, as a function of duration of starvation. The OM:ash weight ratio, a proportional estimate of energy content (relative to initial body size), significantly decreased as the period of starvation increased. The depletion of energy content was

TA B L E 3
Results of the GLM analysis of percent survivorship as a function of duration of starvation for each size class from each batch of N. ostrina hatchlings particularly rapid in the first weeks of starvation; the OM:ash weight ratio dropped by half after only 10 days of starvation, and decreased by an additional 19% by day 20. In both batches of hatchlings, the OM:ash weight ratio levelled off by day 40 at ~0.14, in other words when the mass of remaining organic matter in the body was 14% of the mass of the inorganic material, the latter mostly consisting of the shell. This continuous reduction in OM content is consistent with hatchlings progressively using up their available energy reserves when not able to feed, and confirms the assumption that AFDW is an effective measure of energetic lipids. The minimum OM:ash weight ratio of 0.14 is likely determined by the amount of organic matter that is difficult to use as an energy source by the hatchlings. Almost all organic compounds that can readily be metabolized to release energy, mainly energetic lipids such as triacylglycerols Lee et al., 2006;Sewell, 2005;Thiyagarajan et al., 2002;Whitehill & Moran, 2012), would thus be utilized within 40-50 days of the onset of the EBP if starved, leaving compounds such as structural lipids and proteins as the bulk of the remaining organic matter. Since energy content predictably decreases as a function of the duration of starvation, this finding confirmed that duration of starvation is a useful proxy for metabolizable energy content at the onset of the EBP.

| Survivorship of EBP invertebrates unable to replenish energy reserves
Short-term starvation of individuals from the onset of the EBP did The most significant finding of this study was that all six species of invertebrates were highly tolerant of starvation in the short-term, that is, the first few days of the EBP. Survivorship was minimally impacted by starvation during the first 10 days of the EBP, with three species experiencing <2% mortality, and the remaining three species experiencing only 6%-12% mortality through this time period. This finding is not consistent with the starting hypothesis that a substantial proportion of individuals would die in the first few days of starvation due to depleted energy stores (Figure 1, curve B). Rather, these low rates of mortality suggest that for each of the examined species, very few individuals, if any, enter the EBP with such critically low energy reserves that they are below or only slightly above the minimum threshold needed to survive.
This finding of very low mortality of starved individuals during the first days of early benthic life is surprising. Many species of intertidal invertebrates have extremely high mortality rates in the field, especially during the first 10 days of the EBP, often ranging from 30% to 100% (Gosselin & Qian, 1997;Hunt & Scheibling, 1997).
More specifically, high morality during the first 3-10 days of early benthic life have been reported for wild populations of a variety of intertidal taxa, some of which were included in the present study: barnacles (Gosselin & Jones, 2010;Gosselin & Qian, 1996;Jenewein & Gosselin, 2013a;Shanks, 2009), gastropods (Moran & Emlet, 2001;Spight, 1975), mussels (Bownes & McQuaid, 2009;Von Der Meden et al., 2012;Phillips, 2017), and crabs (Spitzer et al., 2003). Several studies have hypothesized that the high rate of mortality during the first few days of the EBP might be due in part to a substantial portion of individuals having insufficient internal energy reserves at the beginning of the EBP (Gosselin & Qian, 1997;Jarrett & Pechenik, 1997;Phillips, 2002Phillips, , 2017Thiyagarajan et al., 2002Thiyagarajan et al., , 2007. an average of 23 days earlier than their larger counterparts. This is consistent with the idea that smaller individuals have a smaller reserve of metabolizable energy and reach a minimum energy threshold more quickly during periods of starvation than larger individuals.
However, the impact of size in determining starvation tolerance was subtle, as a difference was only observed between the largest and smallest size classes; the intermediate size class did not differ from either of the other size classes in their ability to tolerate starvation.
The advantage bestowed by the larger energy reserve of large hatchlings was also moderate; even small individuals were able to survive for extended periods of time without replenishing energy stores.
The modest long-term energetic advantage of a larger body size was not apparent on a shorter time scale: There was no significant effect of hatchling size on survivorship during the first 5 days of the EBP. This indicates that even the smallest size class of N. ostrina hatchlings have sufficient energy upon hatching to survive through the first 5 days without replenishing their energy stores, consistent with the null hypothesis that nearly all individuals enter the EBP with sufficient energy reserves (Figure 1, curve A). This is not entirely surprising for N. ostrina, as hatchlings of this species do not begin feeding until 3-10 days after hatching (Gosselin & Chia, 1994). These findings nevertheless further demonstrate that depletion of energy stores is not a significant cause of EBP mortality.
Offspring size is often a good indicator of offspring fitness; indeed, this is a central tenet of life history theory (Roff, 1992;Sinervo, 1990;Stearns, 1992). For benthic invertebrates, larger larval size often translates into larger EBP size, which in turn leads to greater performance (survival and growth) during the EBP (Emlet & Sadro, 2006;Jarrett & Pechenik, 1997;Marshall & Keough, 2004;Phillips, 2002;Thiyagarajan, Harder, Qiu, et al., 2003). While many of these studies have focused on the larval stage, very few studies have examined the relationship between offspring size and offspring performance after the onset of EBP. The present study found no significant difference in short-term survivorship among starved individuals of different initial body sizes, and only modest differences in long-term survivorship. These findings are not dissimilar to those of Moran and Emlet (2001), who found increased survivorship for larger hatchlings in longer outplant experiments (36-54 days), but no significant difference in shorter duration outplants (9 days). This indicates that the short-term response to starvation and energy depletion is independent of the size of recently hatched individuals.
In populations of many benthic species, larger offspring do tend to outperform their smaller counterparts; larger offspring grow more quickly (Marshall et al., 2003;Moran & Emlet, 2001;Thiyagarajan, Harder, Qiu, et al., 2003;Torres et al., 2016), reach maturity earlier (Marshall et al., 2003;Moran & Emlet, 2001), and often have lower rates of mortality (Emlet & Sadro, 2006;Phillips, 2002;Thiyagarajan, Harder, Qiu, et al., 2003) than smaller offspring. The results of the present study, however, indicate that the larger energy reserve of a larger body size does not lead to greater intrinsic performance during the first days of the EBP. Rather, having a larger body size may provide other benefits that allow for greater success. A larger body size may allow individuals to handle and consume prey more efficiently and thus grow at a faster rate (Moran & Emlet, 2001;Palmer, 1990). Being larger, or being able to grow quickly to reach larger body sizes, also decreases risk of predation (Bashevkin & Pechenik, 2015;Paine, 1976;Rumrill, 1990) and vulnerability to desiccation (Gosselin, 1997) by reaching a size refuge.

| Ability to recover after delayed feeding for EBP invertebrates
Although EBP individuals could survive extended periods of starvation, their ability to recover once access to food was regained was negatively affected by the duration of starvation to which they were exposed at the onset of the EBP. The effect, however, was expressed differently among species. In C. dalli and N. ostrina, the proportion of individuals able to survive and begin to grow once food was introduced decreased with increasing duration of starvation. Conversely, in B. glandula the proportion of individuals able to recover was not impacted by the duration of starvation, but the rate of growth throughout the 30-day feeding period decreased significantly as the duration of starvation increased.
In C. dalli and N. ostrina, recovery was impacted by starvation, and in both species, the duration of starvation that resulted in only 50% of individuals able to recover was considerably shorter than the starvation LD 50 , found in the previous experiment. Thus, although individuals are able to survive on internal energy stores for extended periods of time, they may still be destined to die if unable to replenish those energy stores much sooner than the LD 50 . One possible explanation for the difference between the starvation LD 50 and the starvation period resulting in 50% recovery is that extended periods of starvation may cause EBP individuals to lose the capacity to feed by inducing a reabsorption or atrophy of digestive structures (Espinoza et al., 2017;Takami et al., 2000). From that point on, individuals would then continue to deplete all remaining available internal energy reserves but be incapable of efficiently feeding and thus replenishing those energy stores if food were to become available again. We did note that C. dalli exhibited behaviors associated with feeding, such as extending and beating their cirri, throughout the 30-day feeding period even in individuals that did not grow; cirral beating in those individuals most likely served primarily for gas exchange. Similarly, each N. ostrina, including those that did not grow, attacked at least one juvenile mussel during the 30-day feeding period, as evidenced by drill marks observed on the valves of dead mussels. This suggests that although EBP individuals are still capable of exhibiting behaviors associated with feeding, extended periods of starvation may have reduced their ability to digest or absorb nutrients, even if they are able to access food items. Another possible explanation for the inability of individuals to recover from prolonged periods of starvation is that they were unable to extract sufficient energy from consumed food to both replenish the energy they had spent during the starvation period and also begin to grow. In this case, more food, or a longer feeding period, may be required before these animals can fully recover.
The recovery of starved B. glandula after regaining access to food differed relative to the other species in this study. Regardless of the duration of starvation experienced in this study, virtually all B. glandula individuals showed some ability to recover. The rate of growth throughout the 30-day feeding period, however, was significantly impacted by the duration of starvation. The 0-day group, which did not experience any starvation, increased in size five times more during the 30-day feeding period than the group starved for 50 days.
This considerable reduction in growth rate following starvation may have further implications for survivorship, as slower growing individuals will take longer to reach a size refuge, and thus remain susceptible to factors such as competition (Connell, 1961;Pechenik, 1990), predation (Gosselin & Rehak, 2007;Osman & Whitlach, 1995), and desiccation (Hamilton and Gosselin 2020) for a longer period of time.
B. glandula also differed from the other species in terms of growth of shell material while being starved (i.e., prior to feeding).
During the first 20 days of starvation, some growth was apparent in all groups of EBP B. glandula. While growth throughout the starvation period was modest, it indicates that EBP B. glandula are capable of depositing new shell material and thus increasing their shell diameter for the first few weeks of the EBP even without feeding and thus without replenishing their energy stores. This suggests that increasing shell size during the EBP is not entirely reliant on obtaining food, nor is it exclusively stimulated by the growth of body tissues.
However, the increase in shell diameter in newly settled B. glandula that have been without food for any duration was greatly reduced when compared to individuals that had fed.
Finally, in all three species studied in this experiment, an extended period of starvation following the onset of the EBP eventually had a negative impact on the ability to recover, either by directly causing mortality or by impacting growth. Although depletion of energy was not found to be a direct, major source of early mortality among EBP individuals, these findings suggest that EBP individuals with reduced energy reserves would still be at a marked disadvantage for longer-term survival in the intertidal zone.

| Ecological implications
Cohorts of benthic invertebrates in the natural environment suffer extremely high mortality rates during the first days and weeks of the EBP. This study revealed that insufficient energy reserves are not likely a direct cause of this EBP mortality, as had previously been hypothesized (Gosselin & Qian, 1997;Hunt & Scheibling, 1997;Jarrett & Pechenik, 1997;Phillips, 2017). For those individuals that survive through the larval phase and make it to the start of the EBP, energy reserves obtained through larval feeding or maternal provisioning thus appear to be sufficient to sustain the metabolism of the individual through the critical first days of early benthic life. This further indicates that low food availability during the first 5-10 days of the EBP would not be a direct cause of mortality during that time. EBP survivorship in wild populations might nevertheless be impacted by initial energy reserves through indirect effects. Low energy reserves may cause EBP individuals to have low tolerance thresholds to environmental stressors, such as extreme temperatures (Hamilton and Gosselin 2020;Jenewein & Gosselin, 2013b;Miller et al., 2009), desiccation (Foster, 1971;Gosselin & Chia, 1995;Jenewein & Gosselin, 2013a, 2013bMiller et al., 2009), and low salinity (Qiu & Qian, 1999;Thiyagarajan et al., 2007). If so, this could explain previously reported associations between low energy reserves and high EBP mortality rates in some species (Emlet & Sadro, 2006;Phillips, 2002;Thiyagarajan, Harder, Qiu, et al., 2003). The role of energy levels on stress tolerance thresholds and the possible indirect effects of initial energy reserves on EBP mortality are not well understood and require further investigation.

ACK N OWLED G EM ENTS
We would like to thank A. Skurikhina, B. Iwabuchi, and R. Gaudet for assistance with field and laboratory work. A big thank you as well to the directors and staff of the Bamfield Marine Sciences Centre for providing the facilities, equipment, and support necessary to make this study possible and to the Huu-ay-aht First Nation for granting us access to their ancestral territory. This research was supported by NSERC Discovery Grants (RGPIN-2014-04779 and RGPIN-2020-04935) to LAG and by a BC Graduate Scholarship, a TRU ENVS Fellowship Award, and a Ken Lepin Graduate Student Award to SRM.

CO N FLI C T O F I NTE R E S T
None declared. Writing-review & editing (equal).