Maternal influence on timing of parturition , fecundity , and larval quality in three shelf rockfishes ( Sebastes spp . )

Differences in maternal investment and reproductive timing can have important consequences for offspring survival. Prior studies on nearshore rockfishes have shown significant effects of maternal age and size on the timing of parturition, fecundity, and larval quality, offering advantages to population persistence by maintaining age diversity in rockfish populations. In this study, reproduction in chilipepper Sebastes goodei, widow rockfish S. entomelas, and yellowtail rockfish S. flavidus was examined to determine whether ageand size-related effects on maternal investment and reproductive timing are exhibited in deeper-dwelling species of this genus. Parturition dates were derived from fine-scale staging of pre-parturition embryos from gravid females. Measurements of embryonic energy reserves (oil globule and yolk), indicators of condition, were used to estimate depletion rates and to test for maternal age and size effects on larval quality. For widow and yellowtail rockfish, larger or older rockfish gave birth earlier in the parturition season than younger, smaller fishes. Maternal factors of weight, length, or age were positively correlated to absolute and relative (weight-specific) fecundity in all species. A trade-off was observed between egg size and fecundity among species, with chilipepper displaying larger egg size and lower fecundity relative to widow and yellowtail rockfish. Embryonic reserves were weakly but significantly related to age only in chilipepper, with embryos from larger, older mothers having larger oil globules. Since the strength of maternal effects varies among Sebastes species, information on maternal influence can assist managers in identifying species most likely to benefit from the protection of age structure afforded by marine reserves or other fisheries regulations.


INTRODUCTION
The commercial, recreational, and ecological importance of rockfishes Sebastes spp.merit additional early life history research to understand factors determining their recruitment success or failure.Knowledge of species-specific reproductive phenologies and factors determining both the quality and quantity of offspring is integral to evaluating reproductive success.All nearshore Sebastes species researched to date have exhibited some significant effect of maternal size or age on reproductive traits (Berkeley et al. 2004a, Bobko & Berkeley 2004, Fisher et al. 2007, Sogard et al. 2008, Rodgveller et al. 2012).This study extends our understanding of maternal effects to deep-water rockfishes, examining the timing of parturition, fecundity, and larval quality.
Rockfishes are among the longest-lived marine fishes along the Pacific coast (Cailliet et al. 2001) and do not display reproductive senescence (de Bruin et al. 2004).All Sebastes spp.are live-bearers and can mate several months prior to egg fertilization.Females have the ability to store sperm, control fertilization, and ultimately determine the timing of parturition (Love et al. 2002).Offspring develop and hatch within the ovaries synchronously.Species north of central California typically extrude 1 brood per year, whereas some southern populations may have an additional brood (MacGregor 1970, Wyllie Echeverria 1987, Love et al. 2002).Gestation period ranges from 1 to 2 mo, with a longer duration in colder temperatures.At birth, larvae are free-swimming and able to begin exogenous feeding in their pelagic environment.High fecundity and typically long reproductive lifespan in rockfishes serve as a bethedging strategy to buffer against high variability in recruitment success in the physically dynamic environment of the Northeast Pacific (Leaman & Beamish 1984, Longhurst 2002).
An age/size-related maternal effect on the timing of parturition potentially provides a stabilizing influence on recruitment during periods of poor environmental conditions.In some rockfishes, older females extrude larvae earlier in the season than younger females, effectively extending the parturition season (Nichol & Pikitch 1994, Bobko & Berkeley 2004, Sogard et al. 2008, Rodgveller et al. 2012).This protracted parturition season results in a greater likelihood of larval production during times of peak food availability, as in the match-mismatch hypothesis (Cushing 1990).Sogard et al. (2008) suggest that a progression of earlier parturition dates with increasing female age represents a diversified bet-hedging strategy, effectively spreading annual reproductive effort over time within a maternal lineage despite the constraint of each individual spawning only once each year.
In some nearshore rockfishes, older mothers provide greater nutrient reserves to their offspring compared to younger mothers (Berkeley et al. 2004a, Sogard et al. 2008, Rodgveller et al. 2012).During embryonic development, each larva is provisioned with endogenous energy in the form of yolk containing an oil globule (see Fig. 1) comprised of triacylglycerol lipids (Norton & MacFarlane 1999).At parturition, larvae have absorbed their yolk, but a portion of their oil globule remains, sustaining the larvae after extrusion until they are able to locate prey and feed (Fisher et al. 2007).In laboratory experiments, rockfish larvae from older, larger mothers had larger oil globules and exhibited increased resistance to starvation compared to larvae from younger, smaller mothers (Berkeley et al. 2004a, Sogard et al. 2008).
In many teleosts, absolute fecundity (total number of eggs produced), often used as a proxy for reproductive success, typically increases as maternal size increases (Blaxter 1969, Love et al. 2002).In rockfishes, older, larger females also exhibit increased relative fecundity (number of eggs produced per gram of female body weight), thereby contributing more larvae per unit biomass than younger females (Boehlert et al. 1982, Bobko & Berkeley 2004, Sogard et al. 2008).The strength of this maternal effect on fecundity varies among species of Sebastes (Dick 2009).In long-term studies, interannual differences have been found in the fecundities of young (<15 yr old) yellowtail rockfish (Eldridge & Jarvis 1995).Older yellowtail, however, were less affected by environmental influences, thus further demonstrating the role older fishes play in stabilizing populations during poor environmental conditions.
Stock assessment models have assumed constant relative fecundity among mature females such that annual egg production is directly proportional to the population's spawning stock biomass regardless of age or size structure (Berkeley 2006, Dick 2009).For species displaying age-and size-modulated differences in fecundity and larval quality, the assumption of equivalent relative fecundity can result in a misestimation of population productivity.Increasing recognition of the disproportionate contribution of larger, older females to fecundity has resulted in the incorporation of maternal effects in assessments for some but not all Sebastes species.Larval quality differences among mothers have received less recognition, although modeling studies suggest that their inclusion may improve estimates of recovery time for overfished populations (Lucero 2008(Lucero , 2009) ) and stock productivity (Spencer & Dorn 2013).
This study investigates whether age-and sizerelated effects in maternal investment and reproductive timing are exhibited in deeper-dwelling species of this genus (chilipepper Sebastes goodei and widow rockfish S. entomelas).In prior studies (Berkeley et al. 2004a, Sogard et al. 2008), gravid females of nearshore species were held in laboratory tanks until parturition, allowing the precise measurement of larval traits upon completion of development.However, many species that live in deeper habitats are more susceptible to barotrauma-related injuries when brought to the surface, making laboratoryrearing studies impractical.Our approach here was to use fine-scale developmental staging to determine maternal effects on the oil globule depletion rate during embryogenesis and to predict parturition date.
Yellowtail rockfish S. flavidus are included in this study to compare methodologies for the prediction of parturition timing, larval quality, and fecundity.Although they are a shelf species, yellowtail rockfish migrate vertically to feed (Pereyra et al. 1969, Pearcy 1992) and, in contrast to most Sebastes species, can purge their swim bladders quickly upon ascent (Love et al. 2002), reducing barotrauma-related injuries.Adult females have successfully been held in the laboratory until parturition, allowing the assessment of larval quality, timing of parturition, and fecundity (Eldridge et al. 2002, Fisher et al. 2007, Sogard et al. 2008).Results from laboratory studies of yellowtail rock fish are compared with predictions made from preparturition yellowtail rockfish embryos in this study.

Sampling
Gravid female chilipepper, widow, and yellowtail rockfishes were collected via hook and line fishing at 150 to 400 m depth from Cordell Bank, California (approximate location 38.0666°N, 123.3678°W) from 2005 through 2008 (Table 1).All 3 species give birth in the winter (Table 2, Wyllie Echeverria 1987, Love et al. 2002); thus, sampling began in December and continued bi-weekly, weather permitting, through March or until fish with developing embryos were no longer found.Morphometric data collected from each adult included fork length, total weight, ovary weight, and somatic weight (total weight minus ovary weight).
Sagittal otoliths were extracted from each adult to obtain maternal age data.Age validation has been conducted by marginal increment analysis on all 3 species: yellowtail rockfish (Leaman & Nagtegaal 1987), widow rockfish (Lenarz 1987), and chili pepper (J.Mello pers.comm.).It was therefore assumed that all study species deposit annual growth band pairs, composed of 1 opaque and 1 trans lucent band (Cailliet et al. 2006), which were counted for age estimates.Otoliths were prepared using the break and burn method (Beamish & Chilton 1982, Laidig et al. 2003), which is routinely used for rockfish stock assessments.

Embryonic staging/timing of parturition
Upon dissection, each fish was assigned a macroscopic maturity stage (Table 3 Bobko & Berkeley 2004).Frequency histograms of stage by collection month were created to provide a coarse-scale assessment of differences in the timing of ovarian development during the reproductive season based on maternal size (in 20 mm bins).For finer scale analyses, samples of fertilized eggs and developing larvae were photographed using a Nikon dissection microscope equipped with a high-resolution calibrated camera.Larvae were as signed one of 33 developmental stages, each corresponding to an estimated number of days to birth, according to the classification scheme for kurosoi Sebastes schlegeli (Yamada & Kusakari 1991, Table 4, Fig. 1).Gestation time was modified using a conversion formula developed by Eldridge et al. (2002) (r 2 = 0.94) for yellowtail rockfish also captured on Cordell Bank: Days to parturition = -0.022x 2 − 0.097x + 28.9 where x = developmental stage following the Yamada & Kusakari (1991) scheme.Under laboratory conditions at 12.0°C, the yellowtail rockfish in Eldridge et al.'s (2002) study had a mean gestation period of 29.2 d (range 27 to 33 d).
Chilipepper and widow rockfish larvae were assumed to follow a similar time course of development under the same temperature regime.Mean sea surface temperatures in the Cordell Bank area ranged from 10.5 to 13.0°C during our collections, suggesting that 12°C provided a reasonable approximation of the temperatures experienced by female rockfish during the well-mixed, pre-upwelling conditions of the winter spawning period.
To calculate parturition date, the estimated number of days to parturition was added to the capture date for each mother.Regression analyses (ordinary least squares, OLS) were used to determine if there were age-or size-related maternal influences on the timing of parturition in each species.Results for yellowtail rockfish were compared with those previously observed for fish held in laboratory tanks until larval release (Sogard et al. 2008).

Fecundity
The gravimetric method (Bagenal & Braum 1978) was used for fecundity analyses.Fecundity estimates  1991), with converted developmental timing (Eldridge et al. 2002) for yellowtail rockfish.Refer to Fig. 1 for a visual representation of embryonic stages were obtained by counting all of the eggs/larvae in each of 2 weighed subsamples (0.5 to 1 g) per mother and scaling by total ovary weight.The 2 fecundity estimates were averaged to determine total fecundity for each fish.
Relative fecundity was determined by dividing the total fecundity by somatic weight, providing a standardized measure to compare fish of different sizes.Regression analyses (OLS) were used to determine if there were age-or size-related maternal effects on relative fecundity in each species.Data from previous laboratory studies on yellowtail rockfish relative fecundity at parturition (Sogard et al. 2008) were compared with predicted values from pre-parturition samples.

Egg size comparison among species
A calibrated pen tablet (Wacom Technology) was used to measure the circumference of fertilized eggs in stages < 22 (lens formation) from embryos of each species.These stages were chosen to confine measurements to early development, when eggs remain spherical.ANCOVA was used to determine if there were differences in egg size among species, with stage and maternal size (fork length) as covariates.The relationship be tween egg size and relative fecundity was compared among species at a standardized size of 450 mm fork length.

Larval quality
Mean oil globule volumes from each developing brood of larvae were used to determine the oil utilization rate during gestation for each species.Optimas image analysis software (version 6.51) was used to measure the oil globule volume (mm 3 ) from 5 to 20 randomly subsampled embryos from each fe male.The average coefficient of variation for measurements within a female's brood was less than 5%.Since meristics from photographs only offer 2-dimensional viewing, oil globule volumes, assumed to be spherical, were obtained by averaging 2 dia meter measurements and calculating the volume for a sphere, V = (4/3)πr 3 .Embryonic yolk utilization rates were also determined using photomicrographs.For the non-spherical yolk sac, a proxy for size was obtained by tracing the circumference of the yolk using a calibrated pen tablet and subtracting the enclosed oil globule circumference.
A regression (OLS) of oil globule volume versus embryonic developmental time to parturition (derived from fine-scale staging as noted above) was used to examine the rate of oil depletion for each species.Models were selected by choosing the regression with the highest coefficient of determination.The residuals of measured oil volumes from this regression were examined to determine if there was a relationship with maternal age and size.Data from previously published laboratory studies on yellowtail rockfish oil globule volume at parturition (Sogard et al. 2008) were compared with the observed relationship in this study.IBM SPSS Statistics (v.19) software was used for all statistical analyses in this study.ANOVAs and linear regressions were performed following veri fication of the assumptions of normality and homo scedasticity of variance.Preliminary analyses found no interannual differences for any of the response variables; therefore, collection years were pooled.

Collection summary
A broad age and size range of each species was collected (Table 1).Based on female ages derived from otoliths, the proportion of fish in each year class (year of birth) was determined.Chilipepper ranged in age from 5 to 26 yr but were clearly dominated by the 1999 year class (Fig. 2).Yellowtail rockfish ranged from 6 to 27 yr of age.The 1994 year class was most abundant but did not strongly dominate the population, as observed for the 1999 year class of chilipepper.Widow rockfish ranged in age from 5 to 27 yr, with the 2000 year class most abundant but, as with yellowtail, not clearly dominant within the population.

Macroscopic staging
Based on macroscopic staging, larger widow rockfish females appeared to fertilize their eggs and initiate embryo development earlier in the parturition season than smaller fish (Fig. 3a).In December, the majority of widow rockfish had unfertilized eggs, whereas a few larger fish (460 mm bin) were already spent.In January, larger fish (420 to 480 mm bins) were further along in ovarian development, with some spent individuals, but none of the smaller size classes (< 420 mm) were spent.By February, the largest fish in the collection were spent, but some smaller fish (360 to 400 mm) were still developing larvae.
Macroscopic staging of yellowtail rockfish suggested that larger females extruded larvae earlier in the parturition season (Fig. 3b).In January and February, smaller fish (340 to 440 mm) in the collection had unfertilized eggs or newly fertilized larvae, whereas the majority of the largest fish (> 440 mm) were spent.
In chilipepper, macroscopic staging suggested a weak trend of the reverse pattern, with larger fish somewhat delayed in parturition timing compared to smaller fish (Fig. 3c).In December, both small and large chilipepper were represented by all stages.In January, a higher proportion of larger fish (> 400 mm) were earlier in development, at stages 2 and 3, compared to smaller fish (< 420 mm).
Chilipepper displayed unusual re sidual patterns from the linear regression between estimated parturition date and age or size; thus, parametric statistical analyses were not possible.Younger fish released larvae throughout the season, whereas older fish showed a bimodal trend in the timing of parturition (Fig. 4).
There was a positive relationship between maternal size and relative fecundity in yellowtail rockfish (Fig. 5e, linear regression with weight: F 1, 92 = 11.012,p = 0.001, r 2 = 0.11, linear regression with length: F 1, 92 = 12.440, p = 0.001, r 2 = 0.12).There was no significant relationship between relative fecundity and age (F 1, 91 = 0.419, p = 0.519) in yellowtail rockfish.In comparing yellowtail rockfish pre-parturition relative fecundity with previously published data on fish held in laboratory tanks until parturition (Sogard et al. 2008), both groups displayed a positive relationship between relative fecundity and maternal fork length; however, the coefficient of determination was greater for the laboratory study (r 2 = 0.34, Fig. 5e).The slope of the relationship did not differ between the 2 studies (fixed-effects ANCOVA: F 1,106 = 0.558, p = 0.457), suggesting a similar maternal effect, but the intercept was lower in the laboratory study, suggesting a lower overall fecundity estimate compared to that calculated from pre-parturition fish (fixedeffects ANCOVA: F 1,107 = 5.616, p = 0.02).

Larval quality
Oil and yolk depletion rates The 2 forms of energy reserves, yolk and oil globule, were utilized differently during embryogenesis (Fig. 6).Early in development, yolk was utilized at a  and (c,f) chilipepper (linear function).Dashed line in (e) presents yellowtail rockfish data from laboratory studies for comparison (Sogard et al. 2008) faster rate than oil, followed by the rapid depletion of both resources beginning around 15 d prior to parturition.Just prior to parturition, yolk depletion accelerated whereas oil depletion slowed.In all 3 species, yolk depletion curves followed a logarithmic trend with time, whereas oil depletion displayed a sigmoidal trend.Matching their larger egg sizes, chilipepper embryos were provisioned with more oil and yolk than widow and yellowtail rockfishes.
Relationship between maternal age/size and oil globule volume Oil globule volume was not related to maternal age, length, or weight in widow or yellowtail rockfishes but was weakly correlated in chilipepper, which had a positive relationship between oil globule volume residuals from the depletion curve and maternal weight (linear regression: F 1, 304 = 5.523, p = 0.019, r 2 = 0.02, Fig. 7).

Timing of parturition
Both methods of analyzing maternal effects on the timing of parturition -macroscopic staging and finescale staging -showed analogous patterns for each species, with a clear trend of older, larger fish releasing larvae earlier than younger, smaller fish in widow and yellowtail rockfish and a bimodal pattern for older chilipepper.Macroscopic staging analysis provided a larger sample size that included stage 2 (unfertilized) and stage 5 (spent) fish and showed the general course of gonadal development for different size classes through time.Fine-scale staging allowed increased precision in estimating the date of parturition but with a reduced sample size.Our estimates of parturition dates based on fine-scale staging relied on the assumption that widow rockfish and chilipepper have similar gestation times to those established for yellowtail rockfish by Eldridge et al. (2002).Few studies have characterized rockfish embryonic devel opment from fertilization through parturition; thus, knowledge of accurate gestation times is limited.Developmental sequence, however, appears to be com mon in the genus, as observed in studies of yellowtail rockfish (Eldridge et al. 2002), black rockfish Sebastes melanops (Boehlert & Yoklavich 1984), and kurosoi S. schlegeli (Yamada & Kusakari 1991), although there were variations in the rate of development because of differences in water temperatures.Despite the much greater resource provisioning of chilipepper embryos compared to yellowtail and widow, the 3 species exhibited a strikingly similar pattern of yolk and oil depletion by developmental stage (Fig. 6, where time to birth is a function of stage).Stage definition generally depends on physical features, such as the appearance of otoliths, and is therefore independent of yolk and oil quantities.This similarity in embryo development among species suggests that the assumption of a common, temperature-dependent gestation time is reasonable, but we were not able to test this assumption with our Fig. 6.Energy reserve depletion.Each of the 3 species displayed a sigmoidal trend in the depletion of oil reserves during embryogenesis: chilipepper, y = 0.0312/[1 + e −(x − 6.9512)/2.5956]; widow rockfish, y = 0.0147/[1 + e −(x − 7.4240)/2.3289]; yellowtail rockfish, y = 0.0125/[1 + e −(x − 8.9894)/2.3041].Yolk reserves depleted in a logarithmic fashion in the 3 species: chilipepper, y = 0.6489 + 0.4724 × ln(x); widow rockfish, y = 0.1084 + 0.5054 × ln(x); yellowtail rockfish, y = 0.0813 + 0.4810 × ln(x) data.If the gestation period is longer or shorter than predicted, the estimated dates of parturition would be comparably expanded or compressed, but any maternal effects would remain.Individual variability in gestation period, documented to range between 27 and 33 d in yellowtail rockfish (Eldridge et al. 2002), can also impact precision in estimates of extrusion date.The extent of individual variability as well as species differences in gestation time need to be addressed in future research.By giving birth earlier in the season, older, larger yellowtail and widow rockfishes effectively extended parturition at the population level, thus offering an advantage to the population in contending with seasonal environmental variability encountered by newly released larvae in a match-mismatch scenario.At the individual level, this pattern is consistent with the diversified bet-hedging strategy for maternal lineages proposed by Sogard et al. (2008).A pattern of earlier spawning by older, larger females is evident in a phylogenetically diverse range of species (review in Hixon et al. 2013).Age truncation in fish populations, with a consequent shift of reproduction to younger females, has been associated with increased variability in recruitment, potentially a consequence of the restricted period of spawning and increased mismatch with environmental conditions (Hixon et al. 2013).
In contrast to yellowtail and widow rockfish, older chilipepper did not contribute to an extension of the parturition season; the 1999 year class alone spanned the range of observed parturition dates.The bimodal pattern observed in the parturition dates of older chilipepper (Fig. 3) visually suggests multiple parturition within a season; however, we saw no evidence of multiple broods in ovary examination.In a more recent study (Beyer et al. 2014), multiple broods were common in chilipepper collected in a southern population (Santa Barbara Channel) and occurred occasionally in central California populations.Our detection of multiple broods depends on the presence of residual larvae that were not released from the first brood.It is possible that multiple broods occurred in chilipepper during this study but were not detected.

Fecundity
Older, larger fishes produced more offspring in both absolute and relative terms; thus, the assumption that larval production is a function of female biomass regardless of size structure is inappropriate for these species.In general, fecundity estimates derived in this study were comparable to estimates for the 3 species in other studies (Boehlert et al. 1982, Haldorson & Love 1991, Eldridge & Jarvis 1995, Beyer et al. 2014).Likewise, the differences among species, with the reduced absolute fecundity of chilipepper com-  6.The maternal factor with the highest coefficient of determination for each species is shown; the relationship was significant only for chilipepper pared to widow or yellowtail rockfish and the much weaker maternal effect on relative fecundity in chilipepper, were consistent with Haldorson & Love's (1991) comparisons of relative fecundity at 50% maturity and at maximum fish size.
Although a significant maternal effect on relative fecundity was observed in all 3 species, there was substantial variability among individual females, suggesting that female age/size only partly determines the allocation of resources to egg production.The success of females in storing energy reserves during the pre-reproductive period potentially plays a major role.For example, Eldridge & Jarvis (1995) found that yellowtail rockfish collected in Washington had greater mesenteric fat reserves and higher relative fecundity than fish collected in central California.Beyer et al. (2014) found several maternal condition indices correlated with relative fecundity in yellowtail, chilipepper, blackgill Sebastes melanostomus, and speckled rockfish S. ovalis.In particular, liver size, as indexed by the hepato-somatic index, appeared to be influential.
Pre-parturition yellowtail rockfish in this study had higher relative fecundity than laboratory-held yellowtail rockfish collected from the same location in the same year but a similar slope for the maternal effect of increasing relative fecundity with increasing female size (Fig. 5c, Sogard et al. 2008).This difference may reflect the different methods used to calculate fecundity.A gravimetric method was used for pre-parturition fish in this study, whereas fecundity of the laboratory fish was estimated by the collection of all released larvae followed by the use of a plankton splitter to create countable subsamples.

Larval quality
No clear age-or size-related maternal effects were evident on the quantity of lipid provisioning and thus larval quality in the 3 study species.A statistically significant (p = 0.019) effect of increased oil globule volume with increasing maternal weight, observed in chilipepper, was weak (r 2 = 0.02) and perhaps biologically insignificant (Fig. 7b).
The lack of a maternal effect on larval quality in yellowtail rockfish in this study was contrary to findings for yellowtail rockfish in laboratory studies, which showed a positive relationship between maternal weight and oil globule size at parturition (r 2 = 0.23, Sogard et al. 2008).Laboratory studies have also found maternal effects on larval quality in black, blue (S.mystinus), and gopher (S. carnatus) rock-fishes (Berkeley et al. 2004a, Sogard et al. 2008).Variations in methodology be tween laboratory rearing until parturition versus sampling pre-term gravid females could account for this difference.Measurement of larvae at the time of parturition allows more precise estimates of available energy reserves in the oil globule.Examining pre-parturition larvae allows for greater sample sizes; however, drawbacks include possible misestimation in projected time intervals between fine-scale developmental stages and in accuracies in depletion rate estimations, including lack of accounting for matro trophic effects.
Small errors in staging have the potential to greatly influence the oil depletion relationship, particularly during the period of rapid catabolism of oil reserves.During the later stages, depletion rates change over a short time period and are difficult to estimate with small sample sizes.A greater sample size of latestage embryos (stages 30 to 33, < 7 d to birth) would be beneficial to critically analyze yolk and oil utilization in stages just prior to parturition.
Although laboratory-rearing studies cannot duplicate the natural environment completely, they provide standardization by reducing the variability from factors associated with field-caught specimens.These include differences in maternal nutrition, activity levels, and water temperature, which can affect developmental rates of embryos.By allowing females to complete gestation, more precise estimates of parturition timing and larval quality can be obtained.

Species differences
In this study, variability in reproductive strategies among 3 shelf rockfish species was observed, with an apparent trade-off between egg size and fecundity.Chilipepper had the largest egg size and yolk and oil reserves and the lowest absolute and relative fecundity, whereas the reverse pattern was observed for both widow and yellowtail rockfishes.Beyer et al. (2014) also found the mass of chilipepper eggs/larvae to be greater than that of yellowtail at all embryonic stages.Larger larval size at parturition may aid in survivorship because of increased motility for es caping prey and capturing food, while the greater energy reserves in the larger oil globule provide re sistance to starvation (Fisher et al. 2007).Several life history traits distinguish chilipepper from widow and yellowtail rockfish (Table 2).Chilipepper tend to be more demersal, their parturition season begins earlier in the winter, they have a shorter lifespan, and they are capable of producing multiple broods.Interestingly, speckled rockfish Sebastes ovalis also have low fecundity but large egg size and, like chilipepper, are multiple brooders, begin parturition early, and have a comparable life span of about 34 yr (Beyer et al. 2014).These differences in life history may relate to the importance of quality over quantity for chilipepper larvae.

Conclusions
Widow rockfish, yellowtail rockfish, and chilipepper each exhibited a positive age-or size-related maternal effect on 1 or more reproductive metrics in this study, extending this pattern previously reported in nearshore species (Berkeley et al. 2004a, Bobko & Berkeley 2004, Fisher et al. 2007, Sogard et al. 2008, Rodgveller et al. 2012) to deeper-dwelling rockfishes.The strength and pattern of such maternal effects differ among species, potentially reflecting differences in life histories.These results further support the suggested benefits of preserving age structure in rockfishes as well as accounting for differences in life histories among species.Numerous rockfish species have experienced age/size truncation because of fishing pressure on these slow-growing, late-maturing fishes (Parker et al. 2000, Berkeley et al. 2004b, Harvey et al. 2006).A recent review by Hixon et al. (2013) demonstrates that positive maternal effects on reproductive capacity are common in exploited fish species and that extensive age truncation results in widespread diminished ability to buffer against temporal environmental variability.The effects of continual size/age truncation may be phenotypic and reversible or they may induce evolutionary re sponses, with less plastic shifts toward reduced growth rate, age at maturity, body size, and productivity (Conover & Munch 2002, Law 2007).Management options that protect older fish, accounting for the phenology of parturition and oceanic regime, can enhance conditions for the successful rebuilding of species in the genus Sebastes.

Fig. 4 .
Fig.4.Maternal influence on parturition date for widow rockfish, yellowtail rockfish, and chilipepper.Each symbol shows the estimated parturition date (day of the year [DOY], 1 = January 1) for the brood of 1 female, based on fine-scale staging of embryos.The maternal factor (age, length, or weight) with the strongest relationship with extrusion date is displayed.Dashed line in (b) presents yellowtail rockfish data from laboratory studies for comparison(Sogard et al. 2008)

Fig. 5 .
Fig. 5. (a−c) Absolute (no. of larvae) and (d−f) relative (larvae g −1 somatic weight) fecundity versus the maternal factor (length or weight, or age) with the highest significant coefficient of determination for (a,d) widow rockfish (power function), (b,e) yellow tail rockfish (linear function),and (c,f) chilipepper (linear function).Dashed line in (e) presents yellowtail rockfish data from laboratory studies for comparison(Sogard et al. 2008)

Fig. 7 .
Fig. 7. Maternal influence on oil globule volume.The y-axis represents residuals from the sigmoidal depletion curves shown in Fig. 6.The maternal factor with the highest coefficient of determination for each species is shown; the relationship was significant only for chilipepper Thomas Turner, Albuquerque, New Mexico, USA Submitted: August 7, 2013; Accepted: March 13, 2014 Proofs received from author(s): May 12, 2014 ,Gunderson et al.