Rodgveller et at: Effects of maternal age and size on embryonic energy reserves and developmental timing of Sebastes maliger 
43 
that were held in the laboratory, we found that many 
females began resorbing embryos. Therefore, for this 
species progressive sampling may not be possible. Our 
samples included a wide range of sizes and ages and 
even though we could not track embryonic development 
of each female, it is likely that we captured various life 
stages of female quillback rockfish that likely contrib- 
uted to the significant relationships we found between 
development stage and age and size. 
Maternal length and weight were statistically re- 
lated to OGD, but the correlations were not strong. 
Weight and OGD had the strongest relationship and 
therefore it is possible that energy stores have more 
of an effect on OGD than age (e.g., Larson, 1991). In 
the literature, oil globule size and maternal factors 
are only sometimes related for rockfish. Sogard et al. 
(2008) found correlations between oil globule volume 
and weight, age, and length for one out of five species 
(a different species for all three factors). However, 
sample sizes were relatively small (n=28-40 females) 
and the significant relationship with age was heavily 
influenced by one sample. Berkeley et al. (2004a) also 
found that black rockfish oil globule volume was related 
to age (n= 20). 
In our study, weight-specific fecundity did not in- 
crease with size and age. Besides a decrease in fe- 
cundity, significant absence of energy stores due to a 
decreased weight can cause females to skip spawning, 
where mature fish do not spawn during a spawning 
season (e.g., Hannah and Parker, 2007). Practically, 
a skip in spawning would cause a decrease in weight 
specific fecundity if it was negatively related to weight. 
Because only gravid females were collected, we did not 
examine skip spawning. 
Because rockfish are matrotrophic (e.g., MacFarlane 
and Bowers, 1995), effects on larvae of holding gravid 
females in captivity may be more pronounced than hold- 
ing other species of fishes. In previous studies pregnant 
rockfish were held in captivity for 1-14 weeks until 
parturition or until larvae were fully developed (Berke- 
ley, 2004a; Fisher et al., 2007; Sogard et al., 2008). If 
embryonic development is minimally compromised in 
a laboratory setting, there are benefits that cannot be 
accomplished in a field-based study, such as following 
embryonic development within a female and measure- 
ments of larvae after parturition. However, results of 
maternal age effects on larval energy stores and growth 
in the laboratory may not be an accurate predictor 
of success in the natural environment (Marshall et 
al., 2010). The laboratory favors the survival of larger 
larvae. Depending on the environment, bigger may not 
always be better (Marshall and Keough, 2008). Multiple 
environmental factors play a role in determining the 
ideal larval size and parturition date. Larger size does 
not alleviate pressure from many factors in the envi- 
ronment, such as intra- and interspecific competition, 
predation, water temperature and chemistry, and food 
availability (Marshall et al., 2010). 
This is the first study where field measurements of 
OGD were used to avoid effects that laboratory rearing 
could have on the energy use and health of developing 
embryos. The use of field data increased the number 
of covariates we had to consider in our analyses com- 
pared to those based on laboratory studies. However, 
staging may also be necessary for laboratory studies. 
It is a likely assumption that larvae from all females 
that parturiated in the laboratory are at the same de- 
velopmental stage. This may not be accurate; we found 
that embryos at stages 11-13 often broke out of their 
egg envelope easily and appeared to have been hatched 
internally and this activity would imply that they 
were all at stage 13. With closer investigation of the 
head and pigment, we classified these hatched larvae 
into three different developmental stages. Therefore, 
it is possible that developmental stage may need to be 
considered even for laboratory experiments. 
Staging embryos as accurately as possible and con- 
sidering developmental stage in analyses excludes vari- 
ability in OGD that can be attributed to developmental 
use of the oil globule. We divided late developmental 
stages, when embryos and preparturition larvae are 
mobile, into more stages than were used in previous 
studies of rockfish embryonic development (Yamada 
and Kusakari, 1991; Eldridge et al, 2002). The poly- 
nomial shape of the relationship between OGD and 
stage, which is typical for rockfish (e.g., Eldridge et 
al., 2002), illustrates that the oil globule is used more 
rapidly at later stages of development. Therefore, late 
stages may require further division to properly account 
for variation in OGD that is attributable to stage. We 
also saw increased within-stage variability in OGD be- 
tween females at later stages of development, possibly 
because the preparturition larvae and embryos were at 
different stages of development that we were not able 
to detect. However, measurements from late-stage em- 
bryos are likely a more accurate portrayal of OGD at 
parturition and the most obvious maternal effects may 
be detectable at late stages. Therefore, we would expect 
there to be more variability at later stages. Eldridge 
et al. (2002) found similar variability at early and 
late stages; however, samples size was small (n = 21), 
especially at middle stages, and therefore their results 
are inconclusive. Morphometric measurements may be 
useful for further dividing development because we 
noticed differences in body and cephalic morphometries 
through development. 
Studies that examined maternal effects on rockfish 
larvae have focused on pelagic, relatively shallow-water 
species sampled off of Oregon and California (Berkeley 
et al., 2004a; Fisher et al., 2007; Sogard et al., 2008). 
We studied a relatively long-lived, demersal rockfish in 
Alaska waters (up to 95 years, depths to 274 m, Love 
et al., 2002) and had some results that differed from 
studies of rockfish from southern latitudes. Distribu- 
tional ranges of many rockfishes extend north to Alaska 
where the temperature, food availability, and other 
environmental variables differ substantially. These en- 
vironmental differences may affect parturition timing 
and strategies for optimizing larval survival compared 
to species at lower latitudes. Owing to these environ- 
