www.nature.com/scientificreports
OPEN
Born small, die young: Intrinsic,
size-selective mortality in marine
larval fish
received: 30 May 2015
accepted: 22 October 2015
Published: 24 November 2015
s. Garrido1,2, R. Ben-Hamadou3, A.M.p. santos1,4, s. Ferreira1, M.A. teodósio4, U. Cotano5,
X. Irigoien6, M.A. peck7, e. saiz8 & p. Ré2
Mortality during the early stages is a major cause of the natural variations in the size and
recruitment strength of marine fish populations. In this study, the relation between the size-athatch and early survival was assessed using laboratory experiments and on field-caught larvae of
the European sardine (Sardina pilchardus). Larval size-at-hatch was not related to the egg size but
was significantly, positively related to the diameter of the otolith-at-hatch. Otolith diameter-athatch was also significantly correlated with survival-at-age in fed and unfed larvae in the laboratory.
For sardine larvae collected in the Bay of Biscay during the spring of 2008, otolith radius-at-hatch
was also significantly related to viability. Larval mortality has frequently been related to adverse
environmental conditions and intrinsic factors affecting feeding ability and vulnerability to predators.
Our study offers evidence indicating that a significant portion of fish mortality occurs during the
endogenous (yolk) and mixed (yolk /prey) feeding period in the absence of predators, revealing that
marine fish with high fecundity, such as small pelagics, can spawn a relatively large amount of eggs
resulting in small larvae with no chances to survive. Our findings help to better understand the mass
mortalities occurring at early stages of marine fish.
It is commonly accepted that mortality occurring during the early stages of development of marine fish
(eggs through to post-larvae) is a major cause of the natural variation in the abundance of marine fish
populations1–3. Recent findings suggest that the great majority of the mortality occurring at these early
phases is selective, and that an individual’s phenotype influences its probability to survive (review in 4).
Moreover, selection greatly varies amongst cohorts, implying that selective mortality is an important
source of variability in total mortality, and thus in the variability in recruitment strength4. Therefore,
understanding the phenotypic variability in selected traits is fundamental to understanding the dynamics
of fish populations and anticipating which species/populations may be most affected by environmental
change.
The high mortality experienced during the larval phase of fishes is widely regarded to be
size-dependent5. In the growth-dependent mortality hypothesis, faster-growing individuals within a
population have lower risks of mortality due to predation since i) larger larvae have stronger swimming
1
Instituto Português do Mar e da Atmosfera, Av. Brasília s/n, 1449-006 Lisboa, Portugal. 2Marine and environmental
Sciences Centre, Faculdade de Ciências, Universidade de Lisboa, Campo Grande, 1749-016 Lisboa, Portugal.
3
Department of Biological and Environmental Sciences, College of Arts and Sciences, Qatar University, PO Box
2713, Doha, Qatar. 4Centro de Ciências do Mar do Algarve, Universidade do Algarve. Campus de Gambelas, 8005139 Faro, Portugal. 5Marine Research Unit – AZTI Foundation, Herrera Kaia, Portualdea z/g, 20110 Pasaia, Spain.
6
Red Sea Research Center, King Abdullah University for Science and Technology, 23955-6900 huwal, Saudi Arabia.
7
Institute for Hydrobiology and Fisheries Science, Hamburg University, Olbersweg 24, 22767 Hamburg, Germany.
8
Institut de Ciències del Mar - CSIC, Ps. Marítim de la Barceloneta 37–49, 08003 Barcelona, Spain. Correspondence
and requests for materials should be addressed to G.S. (email: susana.garrido@ipma.pt) or B.-H.R. (email:
benhamadou@qu.edu.qa)
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
1
www.nature.com/scientificreports/
capabilities than smaller ones, and ii) fast-growing larvae experience a shorter duration of the larval
period, decreasing the period of time when they are most vulnerable to predators. Although there is a fair
amount of field and laboratory research showing that the mortality of fish larvae is growth-dependent
(e.g.6–8), most previous studies have focused on how extrinsic factors such as adverse environmental
conditions impact survival by reducing larval growth (size-at-age). However, other intrinsic factors such
as inter-individual differences in metabolic rates and maternal energy provisioning may also impact on
the size, growth and survival of larvae, even when individuals are faced with optimal environmental
conditions. Thus intrinsic and extrinsic factors combine to affect growth and survival10.
Generally, larger size-at-hatch has been associated with higher survival probabilities in fish (e.g.11–13)
especially under adverse environmental conditions14. Given that even small differences in mortality rates
experienced during the larval period can drive large fluctuations in the recruitment of marine fishes15,16,
variation in early life-history traits such as size-at-hatch can potentially explain a large portion of that
variability17. Size-at-hatch depends on extrinsic and intrinsic factors such as temperature, additive genetic
variance and maternal effects17 as well as their interaction18. Temperature influences yolk utilization efficiency which affects the size obtained by larvae at the end of endogenous feeding (e.g.19,10). After fertilization, not all embryos synchronously hatch and those hatching later are normally larger but also may have
fewer yolk reserves than earlier-hatching smaller larvae20,21. In several batch-spawning marine fishes,
maternal effects have been reported to influence larval size and viability, e.g. larger females spawn larger
eggs than smaller females (e.g.22–24) and larger eggs lead to larger larvae or larvae with higher nutritional
quality (e.g.25,26). The opposite has also been reported for the damselfish (Pomacentrus amboinensis), for
which smaller females produced larvae of longer length and larger energy reserves at hatching27. Paternal
(i.e. genetic) effects are also important in determining larval size, e.g. significantly influence larval length
and yolk-sac volume in the Atlantic herring (Clupea harengus)28,29.
The European sardine Sardina pilchardus is one of the most abundant small pelagic fish species in
the Northeast Atlantic and the Mediterranean30,31, dominating the ichthyoplankton community in the
Western Iberian upwelling ecosystem, particularly during the colder months of the year32. In recent
years, the Atlanto-Iberian sardine stock has displayed large fluctuations in size and spatial distribution
and has now reached historically minimum values of population abundance and recruitment strength33.
Population fluctuations of short-lived small pelagic fishes are known to result from recruitment variability34, and changes in prey availability are considered to be a main cause of such variability, either
by affecting the intensity of reproduction or by directly impacting the growth and survival of early life
stages35. Trophodynamically-mediated successful reproduction and/or larval survival promote strong year
classes, preventing the collapse of small pelagic fish stocks, which often has dramatic socio-economic
consequences36.
The main objective of this work was to assess the relation between size-at-hatch and early viability of
the larvae of European sardine. Our analysis included both wild individuals collected from the field and
sardine larvae reared in the laboratory under different feeding conditions. The second objective was to
study whether egg size influenced the size-at-hatch of sardine larvae. Our main tool was the microstructure analysis of larval otoliths, a technique previously used to examine the growth-selective mortality
hypothesis in marine fish larvae (e.g.37,38,9). Unveiling the relationship between size-at-hatch and survival
potential in the laboratory and field will help clarify whether this intrinsic factor plays a direct role in
the mortality experienced by sardine larvae and the potential for year-class success.
Results
Egg size and sardine larvae size-at-hatching. The size of the sardine eggs incubated individually
(Experiment 1), estimated from the area of the egg, ranged from ~5.4 to 6.1 mm2 (mean: 5.7 ± 0.17 SD mm2).
The average yolk-sac area of the eggs was 0.9 ± 0.03 SD mm2. Larval total length-at-hatch varied from
2.57 to 4.18 mm, with a mean of 3.6 ± 0.37 SD mm (n = 47). No statistically significant relationship was
found between either egg size or yolk-sac area and the total length-at-hatch of sardine larvae (simple
linear regression analysis, P-values = 0.16 and 0.52, respectively; Fig. 1).
Sardine larvae development and mortality.
Approximately 72 h after spawning, sardine eggs had
already hatched and the mean (± SD) larval length-at-hatch was 3.6 ± 0.34 mm (n = 73). Individual larvae ranged from 2.8 to 4.2 mm at hatch. Larvae hatched with no eye pigmentation and a closed mouth.
Larvae started exogenous feeding at an age of 4 days post-hatch (dph).
When reared under optimal feeding conditions (Experiment 2), larval size (total length) increased
linearly throughout ontogeny (Fig. 2A; Supplementary Table S1 online) following the linear regression
equation (n = 213):
Size (mm) = 3.862 + 0.25 × Age (days)
(R 2 = 0.85 , P − value = 0.00007).
When reared with no prey (Experiment 3), sardine larvae were able to grow in length only during the
endogenous feeding period (< 5 dph) and then the mean size-at-age declined by 9.7% from 6 dph to
11 dph when all larvae died (Fig. 2B).
Total larval mortality during the experimental period for larvae reared with optimal feeding conditions was 61.3%, and mostly occurred during the first 10 dph (57.3%). In particular, larvae mortality
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
2
www.nature.com/scientificreports/
Figure 1. Relationship between sardine larvae size-at-hatching and egg size (panel A) and egg yolk-sac
area (panel B).
Figure 2. Growth of sardine larvae reared in the laboratory under starvation (panel A) and under
optimal feeding conditions (panel B).
Figure 3. Cumulative mortality (%) through ontogeny of sardine larvae reared in laboratory conditions
at 15 °C and salinity of 35 under defined optimal feeding conditions.
concentrated during the first 5 days after hatch, comprising 42.8% of the total (cumulative) larval mortality experienced during the whole experimental period (Fig. 3). These first 5 dph correspond to an
exclusively endogenous feeding period until 2 dph, when total mortality was 34.5% and a period of mixed
endogenous and onset of the exogenous feeding period (3 to 5 dph), when mortality was 8.3% of the total
cumulative mortality during the experimental period.
Hatch check diameter and size-at-hatch for larvae reared in the laboratory under optimal
feeding conditions. For each of the 820 sardine larvae reared under optimal feeding conditions
(Experiment 2), we extracted one or (when possible) both sagittal otoliths for analysis. From the 184
individuals sampled at the day of hatching, a clear hatch check was present in 161 larvae, with a diameter of 10.4 ± 1.11 µ m (mean ± SD). The relationship between larval size at hatch (L0) and otolith hatch
check diameter (HCD) was best described by a simple linear regression model (Fig. 4; Supplementary
Table S1 online):
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
3
www.nature.com/scientificreports/
Figure 4. Relationship between larvae size-at-hatch (total length, mm) and otolith hatch check diameter
(µm) for sardine larvae reared in the laboratory under optimal feeding conditions.
Figure 5. Relation between sardine larvae age (days post-hatch) and otolith first increment check
diameter (µm) for European sardine (Sardina pilchardus) larvae reared in the laboratory under optimal
feeding conditions.
L 0 = 1.96 + 0.156 × HCD
2
(R = 0.25 , P−value = 0.0007) .
Hatch check diameter and survival for larvae reared in the laboratory under optimal feeding
conditions. The hatch check was clearly visible in the sagittal otolith of 608 individuals (70% of all
sardine larvae reared under optimal feeding conditions). The hatch check diameter of the otoliths of
larvae sampled throughout the laboratory rearing period was 11.4 ± 1.4 µ m (mean ± SD).
Changes in the distribution of the hatch check diameters versus larval age (Fig. 5) indicated that larvae with smaller hatch check diameters disappeared from the population significantly faster than larvae
having larger hatch check diameters. Quantile regression analysis confirmed that the slope of the regression using the lower, 10th percentile of diameters was positive and statistically different from 0 (+ 0.142,
P-value = 0.00003; Supplementary Table S1 online), indicating that larvae having relatively small hatch
check diameters disappeared from the population through time. In contrast, the slope of the regression for the upper, 90th percentile was not significantly different from zero (+ 0.035, P-value = 0.013;
Supplementary Table S1 online), indicating that larvae with larger hatch check diameter were not
removed from the population through time.
Hatching check diameter and viability for larvae reared under starvation. Regarding the larvae
reared in starving conditions, it was possible to extract the otoliths of 391 out of the 430 larvae sampled.
From the otoliths corresponding to the larvae sampled at the day of hatching (35 individuals), 28 had
a clear hatch check. Average hatch check diameter of larvae reared under starvation was 11.9 ± 1.33 µ m
(mean ± SD). Larvae with smaller hatch check diameters disappeared from the population significantly
faster than larvae having larger hatch check diameters, particularly from age 3 onwards (Fig. 6). Quantile
regression analysis confirmed that the slope of the lower, 10th percentile fit was positive and statistically
different from 0 (+ 0.327, P-value = 0.000002; Supplementary Table S1 online), indicating that larvae
provided with small hatch check diameters were disappearing from the population through time. In
contrast, the regression slope for the 90th percentile was not statistically different than zero (+ 0.087,
P-value = 0.13; Supplementary Table S1 online), indicating that larvae with larger hatch check diameters
remained within the population.
Otolith hatch check diameter and age of wild sardine larvae from the Bay of Biscay. The
otoliths of 492 sardine larvae collected in the Bay of Biscay were analysed. Sardine larvae ranged from 2.6
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
4
www.nature.com/scientificreports/
Figure 6. Relation between sardine larvae age (days post-hatching) and otolith first increment check
diameter (µm) for European sardine (Sardina pilchardus) larvae reared in the laboratory under
starvation.
Figure 7. Relation between sardine larvae age (days post-hatching) and otolith first increment check
diameter (µm) for larvae captured in the Bay of Biscay.
to 24.8 mm in length and had between 0 and 28 otolith increments. Hatch check radius was measured for
all larvae, ranging from 3.9 to 9.9 µ m. Similarly to what was observed in the reared sardine larvae, in the
wild sardine larvae there appeared to be a positive linear relationship between hatch check size (radius)
and age of the larvae (Fig. 7). Quantile regression analysis shows that this pattern result from the gradual
loss of relatively small size-at-hatch larvae from the population (10th quantile regression slope + 0.12,
P-value = 0.039; Supplementary Table S1 online) and a relatively stable contribution of larvae with larger
hatch check diameters in the population (i.e, the upper, 90th quantile regression fit shows no statistically
significant slope (+ 0.017, P-value = 0.045; Supplementary Table S1 online).
Discussion
Our study provides solid evidence that the survival probability of fish larvae, both for larvae reared in
the laboratory and for wild specimens, is closely linked to the size of the larvae at hatch. The first otolith
check, formed at hatching, was linearly related to the larval size-at-hatch, and larvae surviving past the
first-feeding stage had higher mean hatch-check diameters when compared to those analysed at the day
of hatching. This fact means that larvae that are larger at hatch have higher chances to survive.
This size-dependent mortality cannot simply be driven by extrinsic factor such as predation, since
our laboratory experiments did not include predators. One could argue that the size-selective mortality
observed in the laboratory experiments when larvae were provided prey is a consequence of larger larvae
being superior feeders and utilizing a broader prey size spectrum and, thus, tending to grow faster than
smaller larvae. However, a considerable amount (34.5%) of the total larval mortality occurred before
the age of 3 dph, which falls within the exclusively endogenous feeding period while an additional 8.3%
occurred at 3 and 4 dph, a period of mixed endogenous and exogenous feeding (at 15 °C, sardine larvae
start exogenously feeding at 3 to 4 dph). In comparison, only 15% of the total mortality occurred at
first-feeding and only 2% occurred for larvae ≥ 10 dph. Correspondingly, the selective disappearance of
relatively small larvae at hatch was noticed during the endogenous feeding period and early start of the
exogenous feeding (< 5 dph) and after that period no larvae with hatch-check diameters smaller than
10 µ m were found. In the absence of prey, larvae did not live more than 12 dph and larvae with relatively
small hatch size died earlier than larvae that were relatively large at hatch.
Size-at-hatch can be the result of maternal or intrinsic (metabolic) factors and otolith growth is
directly related to the metabolic rate and is only indirectly correlated with somatic growth; therefore the
hatch-check area (measured in this study) is likely a proxy for embryonic metabolic rate. Larger larvae
at hatch correspond to individuals with a relatively high metabolic rate39 while larvae that are smaller
at hatch have lower metabolic rate during the embryonic stage than larger larvae40. We speculate that
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
5
www.nature.com/scientificreports/
size-selective mortality during the endogenous feeding period most likely stems from metabolic differences arising during early embryonic development, meaning that sardines with small size-at-hatch are
unable to grow and survive due to metabolic limits and the high costs of development in young larvae.
Another hypothesis to explain the pattern found in this study would be that larger larvae correspond to
those with more yolk reserves, as shown e.g. for capelin, Mallotus villosus41, and the small larvae with
higher mortalities are below a given threshold of reserves where embryonic or early life development
fail. In fact, it is generally accepted that females producing eggs with higher yolk reserves would have
progeny which were larger-at-hatch and with an increased chance of survival. In our laboratory experiments, we found that egg size and egg yolk content were not significantly related to the size-at-hatch of
the sardine larvae. However, maternal effects have indeed been shown to influence egg quality in wild
sardine populations42,43. Moreover, in wild populations, the viability of larval sardines has been shown
to depend on the quality of the eggs produced, particularly their protein content44. On the other hand,
size-selective mortality that occurred during the mixed feeding period might have been related to the
lower feeding abilities of the smaller larvae. In fact, it has been demonstrated that size-selective mortality of first-feeding larvae is related to the inefficiency of small larvae to capture prey, even if present
in high concentrations, due to limited feeding performance at low Reynolds numbers (“hydrodynamic
starvation”45).
A large number of field studies has demonstrated that larger larvae at a given age have a survival
advantage in terms of feeding success and predator avoidance (e.g.46,47) because they swim faster and
have a broader prey size spectrum. Particularly for sardines, back-calculated otolith radius-at-age suggests the existence of growth-selective mortality9. On the other hand, it has been demonstrated for the
radiated shanny (Ulvaria subbifurcata), that larvae with small otolith size-at-hatch and low metabolic
rates survived longer than larger larvae that consumed metabolic fuel at higher rates40, but such mortality pattern was only apparent when food was limiting. According to the authors, the advantage of
producing small larvae was related to those species facing variable feeding conditions, where relatively
slow-growing, small larvae would consume their yolk reserves more slowly, extending the endogenous
period and increasing their chances to encounter adequate food patches. Therefore, the higher the unpredictability of the environment, the higher the variability of hatch-check size (e.g. intrinsic metabolic
rates). Sardines are mostly abundant in productive upwelling ecosystems, where plankton productivity is
highly variable, both spatially and temporally. European (and other) sardine larvae are fast-growing and
require high prey concentrations for survival48. Finding these prey patches by active foraging is enhanced
starting at 20 dph, when larvae start to actively swim49. Furthermore, swimming ability was positively
correlated to size-at-age49, which can explain why larger sardines are favoured in the wild. In fact, the
high genetic correlation between larval size and swimming performance suggests that direct selection
on swimming performance likely induces an appreciable, correlated response in larval size-at-hatch17.
A key question is: if larger and fast-growing larvae are always favoured in relation to smaller and
slow-growing larvae, irrespective of feeding conditions, then why should the small progeny be produced?
According to50, selection operates on overall fitness and other traits influencing overall life-time mortality
(but are associated with smaller size-at-hatch) must also be considered. For example, it has been shown
that mean offspring size evolves to a value that maximizes maternal fitness (i.e. selection of maternal
fecundity) even though mean offspring survival might be much less than optimal10. To our knowledge,
there are no studies demonstrating a trade-off between the number and size of offspring for sardines.
However, according to10, evolutionary responses of larvae would likely be balanced by reproductive selection favouring mothers that produce more, smaller offspring. Therefore, female sardines may trade-off
smaller larval size for higher fecundity, even if it is done at the expense of decreasing the probability of
survival for a proportion of her offspring.
The existence of selective mortality can be an important source of bias on the estimation of growth51.
For a tropical clupeid, the silver-stripe round herring Spratelloides gracilis52, selective mortality only acted
on one out of six cohorts examined. In that study, selective mortality occurred from 0 to 15 dph in a
cohort of particularly small and slow-growing larvae, which increased the mean size-at-age in older
larvae to values above that of cohorts that had not undergone this process. It is important to note that
the growth rate bias (overestimation) caused by size-selective mortality is likely greatest when cohorts
are growing under sub-optimal environmental conditions (e.g. too cold or warm temperatures) which
cause higher mortalities of small larvae. Therefore, in some cases it will be essential to parametrize
size-selective mortality if one hopes to generate robust growth functions for fish larvae.
Fish experience massive mortalities during the first phases of development, which represents a key
time for phenotypical selection to take place. Although not all mortality is selective, recent findings
have shown that the majority of the mortality occurring at the early stages of marine fish larvae, either
during the embryonic or during the onset of the exogenous feeding, is selective4. The distribution of
phenotypes that affect early life survival vary substantially among cohorts, implying that the strength
and direction of the selection also vary significantly4. This explains the wide range relationships reported
between larval size-at-hatch and selective mortality, ranging from undetected size-selective mortality
for a specific cohort (Atlantic herring, Clupea harengus53) to strong selection of larger larvae-at-hatch
(Atlantic cod, Gadus morhua13,14). Size selective mortality seems to be particularly intense under adverse
environment14, revealing the importance of food availability for early larvae, when small larvae with
lower feeding abilities45 are forced to forage intensely to sustain high growth rates, at the cost of increased
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
6
www.nature.com/scientificreports/
predation risk54 other factors can also influence larval size-at-hatch, and therefore influence the direction
and intensity of selective mortality, and these include temperature55 and size-selective fishery pressure,
affecting both maternal and larvae size56. Given the importance of larval survival to fish population
dynamics, it is essential to understand the different causes of larval mortality and the combined influence of intrinsic and external factors in shaping selective mortality, which are responsible for the natural
variations of recruitment strength for marine fish populations.
Material and Methods
Sardine rearing and experimentation. The experimental procedures used in this study were
approved by the Animal Welfare and Protection of the General Direction of Veterinary (DGV) of
the Ministry of Agriculture, Rural Development and Fisheries, approval number 0420/000/000/2011,
Lisbon, Portugal, and were in accordance with the guidelines of the European Directive 2010/63/EU of
September, 22 and the Portuguese legislation that implements such directive nationally.
Sardine larvae were hatched from eggs spawned by captive sardines (Sardina pilchardus), captured
by purse-seiners in western Portugal (Peniche) during 2009 and 2010, and kept in a 15000-L cylindrical
tank at the Lisbon Aquarium (Oceanário de Lisboa). Adult fish started spawning naturally by adjusting
the photoperiod and temperature to natural conditions. The eggs for the experiments were collected from
the captive brood stock by a 500-µ m mesh bag placed in the skimmers of the tank.
We conducted 3 experiments designed for different aims. In Experiment 1, we aimed to assess whether
or not there was a relationship between egg size and larval size-at-hatch. To do that, 82 sardine eggs were
reared individually in 100-mL containers submerged in a water bath, maintaining the same temperature,
salinity and light regime conditions as described below. Egg size was measured immediately after collection from the adult tank, and larvae size (total length) was measured at hatching.
To investigate the relationship between size-at-hatch and larval viability under optimal feeding conditions (Experiment 2), ~2000 sardine eggs were transferred to 30-L cylindrical tanks filled with seawater
of salinity 35 and kept at 15 °C; these temperature and salinity conditions are the ones experienced during the peak spawning season for sardines off Portugal57. The light regime was kept at 16:8 light:dark,
and gentle aeration was used to maintain steady water circulation and high oxygen concentrations. The
sardine larvae were reared with two types of feeding conditions: a) optimal, which corresponded to those
previously shown to translate into feeding and growth rates at saturated feeding levels48,49, and b) under
starvation. The diet of sardine larvae in the rearing tanks with optimal feeding conditions comprised
the dinoflagellate Gymnodinium sp., the rotifer Brachionus sp. and nauplii, copepodites and adults of the
copepod Acartia grani. Details of the feeding regime are provided in48. The concentration of prey in the
larval tanks was monitored daily by counting the prey left from the previous day and adding new prey
to maintain the same concentration throughout the experimental period. Every day, the bottom of the
tank was syphoned and dead larvae were counted to determine age-specific mortality rates.
In Experiment 3, the set-up was similar to the one in Experiment 2, but the larvae were kept only
under starvation conditions, in order to assess the relationship between the size-at-hatch and the time
until death of starved larvae.
Egg and larvae size determinations, and otolith analysis. Sardine eggs from experiment 1 were
measured individually after collection from the rearing tanks under a stereoscopic microscope (Zeiss
STEMI 2000, magnification of 80× ). Eggs were first inspected to be at an advanced stage of development,
with the embryo with a well formed tail (Stage X58). Digital pictures of eggs were taken using the software
Visilog Expert 6.300 and two types of measurements were done for each egg: egg total area and yolk sac
area. At hatch, total length of all larvae was determined under the stereoscopic microscope.
In experiments 2 and 3, groups of larvae were randomly sampled from tanks to determine their size
and conduct otolith analysis. Sardine larvae reared under optimal feeding conditions (Experiment 2) were
sampled from 0 to 75 days post hatching (dph), whereas sardines reared under starvation (Experiment 3)
were sampled until 10 dph (longest survival time was 11 dph). Sampling included 20 larvae at hatching,
≥ 10 individuals from 3 to 15 dph, and from 15 dph until the end of the experiment, 5–10 individuals
were collected every 5 days. Each sampling day, larvae were immediately measured (total length) and
preserved in formaldehyde (4%) for posterior otolith extraction.
Sardine otoliths were extracted using fine needles and both sagittae were removed and mounted on
glass slides, using DPX. Otoliths were examined with an Olympus BX60 microscope. Examination and
reading of the otoliths was carried out using 2000× for small otoliths and 1000× for large otoliths; resolution was 0.24 µ m. A DP50 colour photo-camera was coupled to the microscope, and the Viewfinder
Lite image-analyzer software was used to capture digital images of the otoliths. Otolith increments were
measured with the ImageJ software. Increments were counted, beginning with the first clearly visible
increment outside the core. Otoliths measurements included the total diameter and the diameter of the
first check.
Study of wild sardine larvae. To examine whether otolith size-at-hatch was related to larval survival
in wild conditions, the otoliths of field-caught sardine larvae were analysed. Sardine larvae were collected
on board the R/V Investigador from 6th to 26th May 2008 throughout 124 sampling stations distributed in
the eastern region of the Bay of Biscay (Fig. 8). At each station, Bongo (336 µ m) and MIK (1000 µ m) nets
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
7
www.nature.com/scientificreports/
Figure 8. Sampling sites of European sardine (Sardina pilchardus) larvae in the Bay of Biscay. The map
outline was generated using the software Surfer 8 (Golden Software, LLC).
®
were deployed and oblique tows from 5 m above the bottom to surface conducted to capture the fish larvae. Fish larvae samples were preserved in 80% Ethanol for posterior identification. Sardine larvae were
identified in the samples, and then 504 well preserved and intact specimens were randomly selected and
measured with 0.1 mm precision. Otoliths were extracted and analysed as described above except for the
reading of the otoliths, which was done with 500× magnification and the size measurement of the first
check (hatch check), for which the radius was determined instead of the diameter. Since the temperature
and feeding history of field-caught larvae was not known, larval size-at-hatch was compared with larval
total length instead of otolith-derived determinations of larval age.
statistical analyses. A linear regression model was applied to describe the relationship between
laboratory-reared sardine larval length (total length) and otolith hatching check radius.
To infer the differential disappearance of larvae as a function of their initial size, a quantile regression
analysis was carried out at respectively quantiles 0.10 and 0.90 of larvae’s hatching check diameter vs.
their actual length. This analysis was carried on laboratory reared in optimal feeding conditions, under
starvation and for field sardine larvae separately. 0.10 and 0.90 quantile regression coefficients are used
to infer the differential mortality (disappearance from the batch) between small and large sized larvae at
hatch, respectively. Quantile regression analysis was limited to the age interval where a sufficient number of larvae (minimum of 6 individuals) was sampled per length class, corresponding to all ages for
starving larvae (0–10 dph) and from 0 to 20 dph for fed larvae in the laboratory, and from 6 to 22 mm
for wild-caught larvae.
Daily mortality rates of sardine larvae hatched in the laboratory were calculated based on observed
mortality of experiments conducted at 15 °C and a salinity of 35 psu. Cumulative mortalities were calculated by adding up daily rates from day 0 to day 55.
All statistical analyses and graphics were performed using Matlab 8.1 (R2013a) from Mathworks.
References
1. Houde, E. D. & Schekter, R. C. Feeding by marine fish larvae: developmental and functional responses. Environ. Biol. Fishes 5,
315–334 (1980).
2. Hjort, J. Fluctuations in the year classes of important food fishes. J. Conseil Int. Explor. Mer 1, 5–38 (1926).
3. Houde, E. D. Emerging from Hjort’s Shadow. J. Northw. Atl. Fish. Sci. 41, 53–70 (2008).
4. Johnson D. W., Grorud-Colvert K., Sponaugle S. & Semmens B. X. Phenotypic variation and selective mortality as major drivers
of recruitment variability in fishes. Ecology Letters 17, 743–755 (2014)
5. Anderson, J. T. A review of size dependent survival during prerecruit stages of fishes in relation to recruiment. J. Northw.t Atl.
Fish Sci. 8, 55–66 (1988).
6. Houde, E. D. Comparative growth, mortality, and energetics of marine fish larvae - temperature and implied latitudinal effects.
Fish. Bull. 87, 471–495 (1989).
7. Pepin, P. & Myers, R. A. Significance of egg and larval size to recruitment variability of temperate marine fish. Can. J. Fish. Aquat.
Sci. 48, 1820–1828 (1991).
8. Houde, E. D. Patterns and trends in larval-stage growth and mortality of teleost fish. J. Fish Biol. 51, 52–83 (1997).
9. Ramírez ,T., Cortés, D. & García, A. Growth of North Alboran Sea sardine larvae estimated by otolith microstructure, nucleic
acids and protein content. J. Fish Biol. 59, 403–415 (2001).
10. Peck, M. A., Huebert, K. B. & Llopiz, J. K. Intrinsic and extrinsic factors driving match-mismatch dynamics during the early life
history of marine fishes. Adv. Ecol. Res. 47, 177–302 (2012).
11. Vigliola, L. & Meekan, M. G. Size at hatching and planktonic growth determine post-settlement survivorship of a coral reef fish.
Oecologia 131, 89–93 (2002).
12. Raventos, N. & Macpherson, E. Effect of pelagic larval duration and size-at-hatching on post-settlement survivorship in two
temperate labrid fish of genus Symphodus. Mar. Ecol. Progr. Ser. 285, 205–211 (2005).
13. Grønkjær, P. & Schytte, M. Non-random mortality on Baltic cod larvae inferred from otolith hatch-check sizes. Mar. Ecol. Progr.
Ser. 181, 53–59 (1999).
14. Meekan, M. G. & Fortier, L. Selection for fast growth during the larval life of Atlantic cod Gadus morhua on the Scotian Shelf.
Mar. Ecol. Progr. Ser. 137, 25–37 (1996).
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
8
www.nature.com/scientificreports/
15. Houde, E. D. Fish early life dynamics and recruitment variability. Am. Fish. Soc. Symp. 2, 17–29 (1987).
16. Johnson, D. W. Habitat complexity modifies post-settlement mortality and recruitment dynamics of a marine fish. Ecology 88,
1716–1725 (2007).
17. Johnson, D. W., Christie, M. R. & Moye, J. Quantifying evolutionary potential of marine fish larvae: heritability, selection and
evolutionary constraints. Evolution 64, 2614–2628 (2010).
18. Politis, S. N. et al. Temperature, paternity and asynchronous hatching influence early developmental characteristics of larval
Atlantic cod, Gadus morhua. J. Exp. Mar. Biol. Ecol. 459, 70–79 (2014).
19. Pepin, P., Orr, D. C. & Anderson, J. T. Time to hatch and larval size in relation to temperature and egg size in Atlantic cod (Gadus
morhua). Can. J. Fish. Aquat. Sci. 54 (S1), 2–10 (1997).
20. Laurel, B. J. & Blood, D. M. The effects of temperature on hatching and survival of northern rock sole larvae (Lepidopsetta
polyxystra). Fish B-NOAA 109(3), 282–291 (2011).
21. Bian, X. et al. Temperature-mediated survival, development and hatching variation of Pacific cod Gadus macrocephalus eggs. J.
Fish Biol. 84, 85–105 (2014).
22. Beacham, T. D. & Murray, C. B. Effect of female size, egg size, and water temperature on development of chum salmon
(Oncorhynchus keta) from the Nitinat River, British Columbia. Can. J. Fish. Aquat. Sci. 42, 1755–1765 (1985).
23. Kjesbu, O. S., Solemdal, P., Bratland, P. & Fonn, M. Variation in annual egg production in individual captive Atlantic cod (Gadus
morhua). Can. J. Fish. Aquat. Sci. 53, 610–620 (1996).
24. Beldade, R. et al. Larger female fish contribute disproportionately more to self-replenishment. Proc. R. Soc. B 279, 2116–2124
(2012).
25. Marteinsdottir, G. & Steinarsson, A. Maternal influence on the size and viability of Iceland cod Gadus morhua eggs and larvae.
J. Fish Biol. 52, 1241–1258 (1998).
26. Sogard, S. M., Berkeley, S. A. & Fisher,R. Maternal effects in rockfishes Sebastes spp.: a comparison among species. Mar. Ecol.
Progr. Ser. 360, 227–236 (2008).
27. Maddams, J. C. & McCormick, M. I. Not All Offspring Are Created Equal: Variation in Larval Characteristics in a Serially
Spawning Damselfish. PLoS ONE 7(11), e48525 (2012).
28. Evans, J. P. & Geffen, A. J. Male characteristics, sperm traits, and reproductive success in winter-spawning Celtic Sea Atlantic
herring. Clupea harengus. Mar. Biol. 132, 179–186 (1998).
29. Bang, A., Grønkjær, P., Clemmesen, C. & Høie, H. Parental effects on early life history traits of Atlantic herring (Clupea harengus
L.) larvae. J. Exp. Mar. Biol. Ecol. 334, 51–63 (2006).
30. Suau, P. Contribution a la connaissance des stocks de sardines de la Méditerranée espagnole. Proc. Gen. Fish. Coun. Medit. 5,
453–458 (1959).
31. Andreu, B. Las branquispinas en la caracterización de las poblaciones de Sardina pilchardus (Walb.). Inv. Pesq. 33, 425–607
(1969).
32. Garrido, S., Santos, A. M. P., Santos, A. & Ré, P. Spatial distribution and vertical migrations of fish larvae communities off
Northwestern Iberia sampled with LHPR and Bongo nets. Estuar. Coast. Shelf S. 84, 463–475 (2009).
33. ICES. Management plan evaluation for sardine in Divisions VIIIc and IXa. In Report of the ICES Advisory Committee, 2013.
ICES Advice, 2013. Book 7, Section 7.3.5.1 (2013).
34. Peck, M. A., Reglero, P., Takahashi, M. & Catalán, I. A. Life cycle ecophysiology of small pelagic fish and climate-driven changes
in populations. Progr. Oceanogr. 116, 220–245 (2013).
35. Blaxter, J. H. S. & Hunter, J. R. The biology of the clupeoid fishes. Adv. Mar. Biol. 20, 3–223 (1982).
36. Pita, C., Silva, A., Prellezo, R., Andrés, M. & Uriarte, A. Socioeconomics and Management. Ch 9 in Biology and Ecology of
Sardines and Anchovies. (Ed. Ganias, K.) 335–366 (CRC Press/ Taylor & Francis – Science Publishers (2014).
37. Hovenkamp, F. Growth-dependent mortality of larval plaice Pleuronectes platessa in the North Sea. Mar. Ecol. Progr. Ser. 82,
95–101 (1992).
38. Plaza, G. & Ishida, M. The growth–mortality relationship in larval cohorts of Sardinops melanostictus, revealed by using two new
approaches to analyse longitudinal data from otoliths. J. Fish Biol. 73, 1531–1553 (2008).
39. Yamamoto, T., Ueda, H. & Higashi, S. Correlation among dominance status, metabolic rate and otolith size in masu salmon. J.
Fish Biol. 52, 281–290 (1998).
40. Bochdansky, A. B., Grønkjær, P., Herra, T. P. & Leggett, W. C. Experimental evidence for selection against larvae with high
metabolic rates in a food limited environment. Mar. Biol. 147, 1413–1417 (2005).
41. Chambers, R. C., Leggett, W. C. & Brown, J. A. Egg Size, Female Effects, and the Correlations Between Early Life History Traits
of Capelin, Mallotus villosus: An Appraisal at the Individual Level. Fish B-NOAA 87, 515—523 (1989).
42. Garrido, S. et al. Effect of maternal fat reserves on the fatty acid composition of sardine (Sardina pilchardus) oocytes. Comp.
Biochem. Phys. B 148(4), 398–409 (2007).
43. Garrido, S., et al. Spatio-temporal variability in fatty acid trophic biomarkers in stomach contents and muscle of Iberian sardine
(Sardina pilchardus) and its relationship with spawning. Mar. Biol. 154, 1053–1065 (2008).
44. Riveiro, I., Guisande, C., Lloves, M., Maneiro, I. & Cabanas, J. M. Importance of parental effects on larval survival in Sardina
pilchardus. Mar. Ecol. Progr. Ser. 205, 249–258 (2000).
45. China, V. & Holzman, R. Hydrodynamic starvation in first-feeding larval fishes. P. Natl. Acad. Sci. USA 111, 22, 8083–8088
(2014).
46. Cowan, J. H., Houde, E. D. & Rose, K. A. Size-dependent vulnerability of marine fish larvae to predation: an individual-based
numerical experiment. ICES J. Mar. Sci. 53, 23–37 (1996).
47. Dower, J. E., Pepin, P. & Kim, G. C. Covariation in feeding success, size-at-age and growth in larval radiated shanny (Ulvaria
subbifurcata): insights based on individuals. J. Plankton Res. 31(3), 235–247 (2009).
48. Caldeira, C. et al. Effect of prey concentration on the Ingestion rates of Sardine (Sardina pilchardus) larvae reared in laboratory
conditions. Mar. Ecol. Progr. Ser. 517, 217–228 (2014).
49. Silva, L., Faria, A. M., Teodósio M. A. & Garrido, S. Ontogeny of swimming behaviour in sardine Sardina pilchardus larvae and
effect of larval nutritional condition on critical speed. Mar. Ecol. Progr. Ser. 504, 287–300 (2014).
50. Johnson, D. W. & Hixon, M. A. Ontogenetic and spatial variation in size-selective mortality of a marine fish. J. Evol. Biol. 23,
724–737 (2010).
51. Methot, R. D. & Kramer D. Growth of northern anchovy, Engraulis mordax larvae in the sea. Fish B-NOAA 77, 413–423 (1979).
52. Durieux, E. H. D., Meekan, M. G., Ponton, D. & Vigliola, L. Temperature, selective mortality and early growth in the short-lived
clupeid Spratelloides gracilis. J. Fish Biol. 74(4), 921–938 (2009).
53. Feet, P. O., Ugland, K. I. & Mokness, E. Accuracy of age estimates in spring spawning herring (Clupea harengus L.) reared under
different prey densities. Fish. Res. 56, 59–67 (2002).
54. Jørgensen, C., Opdal, A. F. & Fiksen, Ø. Can behavioural ecology unite hypotheses for fish recruitment? ICES J. Mar. Sci. 71,
909–917 (2014).
55. Rankin, T. L. & Sponaugle, S. Temperature Influences Selective Mortality during the Early Life Stages of a Coral Reef Fish. PLoS
ONE 6(5), e16814 (2011).
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
9
www.nature.com/scientificreports/
56. Johnson, D. W., Christie, M. R., Moye, J. & Hixon, M. A. Genetic correlations between adults and larvae in a marine fish:
potential effects of fishery selection on population replenishment fishery selection may reduce recruitment. Evol. Appl. 4,
621–633 (2011).
57. Stratoudakis, Y. et al. Sardine (Sardina pilchardus) spawning seasonality in European waters of the northeast Atlantic. Mar. Biol.
152(1), 201–212 (2007).
58. Gamulin, T. & Hure, T. Contribution à la connaisance de l'écologie de la ponte de la sardine (Sardina pilchardus Walb.) dans
l’Adriatique. Acta Adriat. 8 (8), 1–22 (1955).
Acknowledgements
This work was financed by Project VITAL (Vital rates of pelagic fish larvae PTDC/MAR/111304/2009).
S.G. is supported by the Portuguese Foundation for Science and Technology (FCT) through a PostDoctoral Fellowship (SFRH/BPD/38332/2007). This work was also partially funded by project MODELA
(PTDC/MAR/098643/2008). The microalgae Rhodomonas baltica was provided by the Assemble
Program (grant 227799) funded under the EU (FP7). This publication was made open access through a
support from Qatar University (IDC of R.B-H). The authors thank Mestre P. Leitão and his crew (‘Mestre
Comboio’) , H. Delgado and S. Farinha (ANOPCERCO), N. Baylina and H. Batista (ODL), C. Caldeira
and A. R. Cristovão (Project VITAL).
Author Contributions
S.G., R.B., E.S. and M.A.P. designed the laboratory experiments; S.G., M.A.T., A.M.S. and P.R. conducted
the laboratory experiments; UC and XI design and coordinated the collection of field data; S.G., M.A.T.,
S.F. and P.R. performed the otolith analysis; R.B. performed the statistical analysis; R.B. and S.G. produced
the figures; S.G. wrote the first draft of the manuscript, and all authors contributed substantially to
revisions.
Additional Information
Supplementary information accompanies this paper at http://www.nature.com/srep
Competing financial interests: The authors declare no competing financial interests.
How to cite this article: Garrido, S. et al. Born small, die young: Intrinsic, size-selective mortality in
marine larval fish. Sci. Rep. 5, 17065; doi: 10.1038/srep17065 (2015).
This work is licensed under a Creative Commons Attribution 4.0 International License. The
images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the
Creative Commons license, users will need to obtain permission from the license holder to reproduce
the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
Scientific RepoRts | 5:17065 | DOI: 10.1038/srep17065
10