Tjac 1 de 1

Mediterranean Institute for Advanced Studies, IMEDEA (CSIC-UIB)
Department of Ecology and Marine Resources

Fish Ecology Group
Selection drivers of life-history traits

in marine coastal fishes
PhD Thesis
A Thesis submitted for the degree of Doctor of Philosophy

Department of Biology
University of the Balearic Islands
By:
Josep Alós Crespí
Advisors:
Dr. Miquel Palmer Vidal & Dra. Beatriz Morales Nin
University Advisor:
Dr. Gabriel Moyà Niell
University of the Balearic Islands, April 2013

[Blank page]
[… There has been surprisingly little consideration of
human-induced selection in the wild until recently. Even
more surprising perhaps is the absence of any detailed
consideration of this effect by Darwin because he had
such a passion for hunting as a young man …]
Fred W. Allendorf, 2009

Selection drivers of life-history traits in marine coastal fishes is a PhD thesis
submitted for the degree of Doctor of Philosophy by Josep Alós Crespí
Josep Alós Crespí

(PhD Candidate)
I certify that I have read this dissertation and that, in my opinion, it is fully
adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy
Dr. Miquel Palmer Vidal

(Principal Advisor)
Dra. Beatriz Morales Nin

(Co-Advisor)
Dr. Gabriel Moyà Niell

(University Advisor)
Approved for the University Committee on graduate studies

Abstract
Humans have non-randomly harvested animals since the origin of our species, resulting
in both ecological and evolutionary changes of wild populations. Considering this fact,
this Thesis tested the hypothesis that recreational fishing is a non-random selection
process for some life-history traits of costal sedentary fishes. This Thesis provides both
empirical and theoretical evidence demonstrating how the variability existing in wild
populations could be the outcome of fishing-induced selection in marine coastal
fisheries.
Selection operates at the individual level. Therefore, the existence of inter-individual
variability is a necessary condition. As a first step, two life-history traits were selected
that may be a priori good candidates for being fishery-selected: growth and
reproduction. The wild variability in these traits of Serranus scriba were estimated using
novel methodological and theoretical approaches based on the use of calcareous
structures of fishes (otoliths), robust statistical tools and histological microscopy
techniques.
This Thesis demonstrates how recreational fishing can cause local depletion and is
size-selective per se. Fishing mortality is not necessarily linked to harvesting because
some individuals may be released after being fished for voluntary (catch-and-release) or
mandatory (minimum legal size) reasons. The potential survival rate of individuals
released by anglers was estimated for the first time in a Mediterranean marine coastal
fish, with special emphasis on Serranus scriba. Post-release mortality rates are generally
low but can be critical depending on the species’ environmental context and fishing
gear. This Thesis also provides management options for sustainable exploitation by
offering tools to decrease the post-mortality levels.
The major achievement of this Thesis is the empirical demonstration of a fishing-
induced selection process against individuals with larger growth ability after maturation
and lower reproductive investment, in accordance with life-history theory. The
theoretical and empirical expectations of this selection have been studied in wild
populations subjected to different exploitation pressures and different levels of
population connectivity and gene flow. Individuals from highly exploited populations
are smaller, have lower growth rates after maturation and invest more energy in
reproduction than fish from no-take marine protected areas, where natural selection
drives these life-history traits in the opposite direction. Elsewhere, it has been
demonstrated that most of those life-history traits are heritable and that consistent
fishing-induced selection over time can induce evolution. The plausibility of this fact is
discussed under the perspective of yield sustainability for the first time in the
Mediterranean.
This Thesis concludes by hypothesising on the outcomes of fishing-induced selection
from a behavioural perspective. A simulation analysis based on a spatially explicit
individual-based model in which fish and fishers move within an empirically supported
scenario was performed. The results show a consistent selection toward individuals that
show less exploratory behaviour. It is suggested that fishing-induced selection is most
likely related to a wide range of potentially correlated traits, including life history,
behavioural and physiological traits, thus exposing a wide range of fascinating
hypotheses to be tested in further research.
Overall, this Thesis establishes the basis of the fisheries-induced selection in marine
coastal fishes in the Mediterranean and discusses its implications for a proper resources
management, aimed to enhance sustainable fisheries and therefore mitigate the plausible
negative ecological and evolutionary effects of fishing.

Resumen
Los humanos hemos recolectado animales de manera no estocástica (azar) desde el
origen de nuestra especie causando diferentes efectos ecológicos y evolutivos en las
poblaciones salvajes. Teniendo en cuenta este hecho, esta Tesis testa la hipótesis de que
la pesca recreativa no es un proceso de azar para algunos rasgos de la historia vital de los
peces. Esta Tesis proporciona evidencias, tanto empíricas como teóricas, que
demuestran que la variabilidad existente en estos rasgos de las poblaciones salvajes puede
ser fuente de selección por pesca en pesquerías marinas costeras.
Los procesos de selección operan a nivel individual. Por lo tanto, es una condición
necesaria la existencia de variabilidad individual dentro de las poblaciones. Como primer
paso, se seleccionaron un número de rasgos de historia vital candidatos para ser fuente
de selección: crecimiento y reproducción. La variabilidad salvaje de estos rasgos se
determinó utilizando novedosas aproximaciones basadas en el uso de estructuras
calcáreas (otolitos), herramientas estadísticas robustas y técnicas de histología
microscópica en Serranus scriba.
Esta Tesis demuestra como la pesca recreativa puede generar una disminución en la
biomasa además de ser talla-selectiva. La mortalidad por pesca no tiene porque estar
necesariamente ligada a recolección debido a que los peces pueden ser voluntariamente
(captura y suelta deportiva) o obligatoriamente (tallas mínimas legales) devueltos al agua
una vez pescados. La tasa potencial de supervivencia de individuos retornados al medio
se ha estimado por primera vez en peces marinos costeros que habitan el Mediterráneo,
con un especial énfasis en S. scriba. Las tasas de mortalidad son en general bajas, pero
pueden llegar a ser críticas dependiendo del contexto ambiental y el arte de pesca. Esta
Tesis también proporciona recomendaciones para la explotación sostenible a través de
herramientas para maximizar las tasas de supervivencia.
El mayor logro de esta Tesis es la demostración empírica de un proceso de selección
por la pesca que actúa en contra de fenotipos con mayor capacidad de crecer después de l
maduración sexual y una menor inversión en reproducción que sigue las reglas generales
de la teoría clásica de historia vital. La teórica y empírica respuesta de esta selección ha
sido estudiada en diferentes poblaciones salvajes expuestas a diferentes niveles de
explotación y conectividad genética entre poblaciones. Los individuos que habitan
poblaciones altamente explotadas son de menor talla, tienen una limitada capacidad de
crecer y tiene una mayor inversión en reproducción que los individuos que habitan
reservas marinas totales, donde la selección natural juega el papel opuesto. Existen
evidencies claras que los rasgos de historia vital estudiados tienen un alto componente
heredable que expuesto a un selección constante puede desencadenar un proceso
evolutivo. La posibilidad de este hecho se discute en el marco de la explotación
sostenible de los recursos pesqueros.
La presente Tesis termina considerando una perspectiva de la selección inducida por
pesca en peces marinos costeros mas relacionada con el comportamiento. Se ha realizado
un ejercicio de simulación usando un modelo basado en el individuo espacialmente

explicito donde peces y pescadores se movían siguiendo diferentes pautas de
comportamiento con base empírica. Los resultados mostraron una consistente selección
a favor de fenotipos con un grado de exploración y de actividad menor. Esto sugiere que
el proceso de selección por pesca esta probablemente causada por un rango de rasgos
potencialmente correlacionados que incluyen rasgos de historia vital, comportamiento y
fisiología generando multitud de hipótesis para ser testadas en investigaciones futuras.
En resumen, esta Tesis establece las bases de los procesos de selección por pesca en
peces marinos costeros del Mediterráneo, y discute sus implicaciones para una correcta
gestión para potenciar la pesca sostenible y mitigar así los posibles efectos negativos de
base ecológica y evolutiva causados por la pesca.

Acknowledgements – Agradecimientos - Agraïments
Although my name is printed on the cover of this Thesis, the singular form
of the first person does not appear within its chapters. I do this to be
grateful to my advisors and collaborators, and to the support of my family,
colleagues and friends.
Miquel i Beatriu (IMEDEA), Gràcies per donar-me la oportunitat de realitzar
aquesta Tesis Doctoral amb vosaltres. Puc afirmar que he estat molt afortunat de
treballar i aprendre al vostre costat. La vostre supervisió, tant personal com humana, la
vostre confiança, i els vostres consells, han donat a la meva jove carrera investigadora
una excel·lent oportunitat per millorar la meva iniciativa, el meu pensament
independent, les meves qualitats de recerca i humanes y el treball en un grup. He gaudit
molt amb vosaltres i sempre estaré agraït per el suport incondicional que he rebut per la
vostre part. Esper que el futur ens depari multitud de bons moments, tant científics com
personals, com els que hem viscut durant aquest quatre anys.
Aquesta Tesis ha estat recolzada econòmicament per una Beca de Formació del
Personal Investigador (FPI) (ref: BES-2009-013407) emmarcada en els projecte
d’investigació CONFLICT (ref: CGL2008-00958). Diferents parts de la Tesis han
estat recalçades també per els projectes d’investigació ROQUER (ref: CTM2005-
00283), i REC2 (ref: CTM-2011-23835). Tots ells finançats per el Ministeri
d’Economia i Competitivitat del Govern Espanyol.
Ignasi, Miquel, David, Carlos, Itzi, Sílvia, i tot el personal del IMEDEA. Sense la
vostre ajuda, ànims i suport, aquesta Tesis no hagues estat possible. Moltes gràcies per
les sortides de camp, les brain-storming sessions, les ajudes amb els mostreigs, així com
el vostre temps explicant-me tècniques pròpies de la vostre línia d’investigació. Ignasi,
per tu gràcies especials per la multitud de coffee-breaks fets, quantes idees bones han
sortit d’aquests moment. Gràcies també per introduir-me en el mon dels primers estadis
de desenvolupament vital del peixos, és un món apassionant. A tots, gràcies i desitjo
poder desfruitar de moltes més ensaïmades amb tots vosaltres els divendres a mig dia.
Gotzon i Toni (IMEDEA) gràcies per la vostre ajuda en la pare de conectivitat i donar
un caire de multi-disciplinitat a la meva Tesis.
Biel (UIB), Salud (IEO) i Toni (UIB), moltes gràcies per introduir-me en el món de
la recerca quan només era un estudiant de Biologia a la UIB. Biel, gràcies especials per
ser el Ponent de la Univeristat d’aquesta Tesis. Part del meu instint de recerca és gràcies
a vosaltres i a tots els que varem compartir aquella època al laboratori de Biologia
Marina. A tota la gent que m’ha ajudat i ha recolzat aquesta Tesis de la Direcció
General de Pesca (Govern Balear), moltes gràcies. Toni, Xisco, Amàlia, Elena, M. del
Mar, Juanita, Imma, gràcies per la vostre ajuda i el vostre suport en algun moment
durant el transcurs d’aquesta Tesis. Pedro i Irine de la Reserva Marina de Palma, moltes
gràcies per la vostre col·laboració, podeu estar orgullosos de la feina que esteu fent a la
reserva.
No puc oblidar la multitud de pescadors recreatius anònims, coneguts, amics i
familiars que han col·laborat en el marc d’aquest estudi. Moltes gràcies a tots i cada un
de vosaltres. Una bona part del treball es per vosaltres, perquè pugem seguir gaudint la
nostre activitat d’oci durant molts d’anys.
A mis amigos y colegas del IIM, Fran, Alex, Gonzalo y David, sois un equipo
increíble, tanto como investigadores como personas. Gracias por los ánimos y por la
multitud de comentarios positivos hechos en este trabajo. Una parte de este trabajo es
vuestra también, y recordad, nosotros somos jóvenes, hay que luchar por el futuro.
To my international collaborators, thanks a lot. Robert (IGB), viel von dieser Arbeit
ist auch deins, Danke. Thomas (IGB), vielen Dank für deine Kommentare und deine
Hilfe während meines Aufenthaltes in Berlin, Danke für all die erstaunlichen Tage
Fischen. Sue, Luiz, Gabe and Elena (FWC), thanks for your hospitality and your help.
Joel and Sara (FWC), thanks, it was awesome enjoy with you a redfish sampling season.
Especial thanks for the fishing trips Joel, the fishing weekend down to the Florida Keys
was awesome and I will never forget it.
Finalment, Àngela, tu has esta la meva inspiració. Quan he estat disgustat, cansat o
desmoralitzat, tu sempre m’has donat ànims per seguir. Necessitaria més d’una Tesis per
enumerar totes les vegades que m’has ajudat amb paraules positives i d’ànims. Gràcies
per tot, realment soc molt afortunat de tenir-te devora. Jimmy, gràcies per aquesta
portada, i gràcies per tots els sopars i bon moment s amb la teva dona, la teva filla i na
petita, podem dir que una part d’aquesta Tesis és també per sa paina. Marta i Tomàs,
aquest treball també està dedicat a vosaltres.

Mamà, tu ets la més important, tot aquest treball és per i gràcies a tu. Tu vares ser la
persona que ens vares encoratjar i espitjar a estudiar a jo i al meus germans quan érem
uns nins, inclús en situacions molt complicades. No hi va haver excuses per donar-nos
tot el teu esforç i temps, sacrificant-te tu mateixa com a conseqüència. Aquí estan els
resultats del teu esforç. Mamà, ara seré Doctor en Ecologia Marina fent i desfruitant del
que mes m’agrada, na Marta és una eminent Advocada i en Tomàs és un reconegut
cuiner. Tot el teu esforç ha estat recompensat, i sempre estarem agraïts. No te podem
tornar el teu temps, però si que podem fer-te protagonista dels nostre èxits, per que
realment són teus també. Mamà, aquest treball està dedicat a Tu.
List of manuscripts published
Alós, J., Palmer, M., Balle, S., Grau, A. M., and Morales-Nin, B. 2010. Individual
growth pattern and variability in Serranus scriba: a Bayesian analysis. ICES
Journal of Marine Science, 67: 502-512. (Chapter 2)
Alonso-Fernández, A., Alós, J., Grau, A., Dominguez-Petit, R., and Saborido-Rey, F.
2011. The Use of Histological Techniques to Study the Reproductive Biology of
the Hermaphroditic Mediterranean Fishes Coris julis, Serranus scriba, and
Diplodus annularis. Marine and Coastal Fisheries, 3: 145-159. (Chapter 3)
Alós, J., Alonso-Fernández, A., Catalàn, I., Palmer, M., and Lowerre-Barbieri, S. (in
press). Individual variability in the factors that affect reproductive potential in
Serranus scriba. Scientia Marina. (Chapter 3)
Alós, J., and Arlinghaus, R. 2013. Impacts of partial marine protected areas on coastal
fish communities exploited by recreational angling. Fisheries Research, 137: 88-
96. (Chapter 4)
Alós, J. 2008. Influence of anatomical hooking depth, capture depth, and venting on
mortality of painted comber (Serranus scriba) released by recreational anglers.
ICES Journal of Marine Science, 65: 1620-1625. (Chapter 5)
Alós, J., Palmer, M., Grau, A. M., and Deudero, S. 2008. Effects of hook size and
barbless hooks on hooking injury, catch per unit effort, and fish size in a mixed-
species recreational fishery in the western Mediterranean Sea. ICES Journal of
Marine Science, 65: 899-905. (Chapter 6)
Alós, J., Arlinghaus, R., Palmer, M., March, D., and Alvarez, I. 2009. The influence of
type of natural bait on fish catches and hooking location in a mixed-species
marine recreational fishery, with implications for management. Fisheries
Research, 97: 270-277. (Chapter 7)
Alós, J., Arlinghaus, R., Palmer, M., Catalan, I., Basterretxea, G., Jordi, A., Alonso-
Fernández, A., et al. (submitted). Recreational harvesting can induce
evolutionary change of life-histories of an exploited marine fish at surprisingly
restricted geographical scales. Evolutionary Applications. (Chapter 8)
Alós, J., Palmer, M., and Arlinghaus, R. 2012. Consistent Selection towards Low
Activity Phenotypes When Catchability Depends on Encounters among
Human Predators and Fish. Plos One, 7: e48030. (Chapter 9)
List of tables
Table 8-1 Bayesian posterior distribution (MCMC) of means the life-history traits
(reproduction investment in year -1, infinite size inn mm of otolith, age of maturation in
years and immature growth rate referred to otolith scale in mm year -1) for the beam
trawl and the hook-and-line samples............................................................................ 74
Table 8-2 Results of the two Generalized Lineal Mixed effect models (GLMMs) fitted
to test the relationship between the direct measure of reproduction investment (batch
fecundity) and (A) the sampling method (beam trawl and hook-and-line), and (B) the
populations (inner of Palma Bay, average population and National Park of Cabrera). In
both models Fish size was also included as fixed factor and Fishing ID as random
factor. The GLMM developed for populations also included the temporal variable
month. The table shows the parameters include in the final model (the MCMC mean
of the parameters, the lower and upper highest 95% credibility intervals (HDP) and the
significance of the parameter. Interactions were not significant and were not included.
...................................................................................................................................... 76
Table 8-3 Posterior distribution (MCMC populations means, s.d. and Bayesian
credibility intervals) of the life-history traits (reproduction investment in year -1, infinite
size (mm of otlith), age of maturation in years and immature growth rate referred to
otolith scale in mm year -1) estimated for the three analysed populations (inner of Palma
Bay, average population and National Park of Cabrera). .............................................. 78
Table 8-4 Results of the three Generalized Lineal Mixed effect models (GLMMs)
fitted to test the relationship between the relative abundance (fish*30min*angler), fish
size (mm) and ages (years) of vulnerable Serranus scriba sampled in the inner of Palma
Bay, the average population and the National Park of Cabrera. The full model (without
reduction) includes the fixed factors population, depth (m), angler, bait and month. The
abundance GLMM have repeated measures within day and the size and age were
repeated measures of fishing sessions within day which were included as random effects.
The table shows the parameters include in the final model (the MCMC mean of the
parameters, the lower and upper highest 95% credibility intervals (HDP) and the
significance of the parameter. ....................................................................................... 81
List of figures
Figure 8.1 Picture of Serranus scriba (Family: Serranidae) (top left panel). Map locating
the Mallorca Island in the Balearic Sea (NW Mediterranean) (bottom left panel).The
spatial location of the 25 sites of eggs releasing is showed in the right panel highlighting
the high exploited (recreational fishing) area located in Palma Bay and the waters of
National Park of Cabrera where recreational fishing is not allowed (No-take MPA).
Black-white gradient represents to be the spatial distribution of the recreational fishing
effort in the Mallorca Island proyected in a raster of 250 m x 250 m of (n = 1,430).
Recreational anglers were censed by the coastal guards (data obtained in a work-package
of the Project CONFLICT)......................................................................................... 60
Figure 8.2 Self-recruitment (10 years mean and s.d.) of the 25 sites of eggs release
considered in this study (upper panel). The highest values observed in the inner part of
Palma Bay (sites 6 and 7) and the National Park of Cabrera (sites 14, 15 and 16) were
showed in dark grey and black respectively. An average, in terms of SC, population
(sites 9 and 10) is showed in light grey. The bottom panels shows the spatial-plots of a
fixed kernel probability distributions (estimated in raster of 250 m x 250 m) of the
positions at the end of the PLD duration of the virtual eggs released in each of these
three main areas. ........................................................................................................... 70
Figure 8.3 Directional flow diagram showing the paired direct and indirect connectivity
(black arrows) calculated among the inner part of Palma Bay (PB), the open population
(OP) and the National Park of Cabrera (NPC) (each number can be interpreted as a %
of eggs released in an specific sub-populations). Indirect connectivity values are the
maximum values of connectivity over 50 years. The grey arrows show the connectivity
attributed by other sub-populations. 360º black arrows are the self-recruitment values
for each specific sub-population.................................................................................... 72
Figure 8.4 Box-plot of the Bayesian individual means of the life-history parameters
(reproduction investment, infinite size, age of maturation and immature growth)
estimated for the beam trawl (BT in light grey) and hook-and-line (HL in dark grey)
individuals..................................................................................................................... 73
Figura 8.5 Bayesian Credibility Intervals (2.5%, 50%, and 97.5%) of the posterior
distribution of the population’s means estimated for each LH trait to assess with the
differences between-populations................................................................................... 79
Figura 8.6 Chronograms of ten years of the daily surface temperature (ºC) and
concentration of chlorophyll-a (mg m-3) in the inner part of Palma Bay (PB), in the
open population (OP) and in the National Park of Cabrera (NPC). ............................ 84
Contents
i. Abstract
ii. Resumen
iii. Acknowledgements – Agradecimientos – Agraïments
iv. List of tables
v. List of figures
vi. List of manuscripts published
Chapter 1 Human harvesting and fisheries-induced selection of

wild fish populations...................................................................37
1.1 Introduction..........................................................................37
1.2 Objectives of the Thesis........................................................41
1.3 Structure of the Thesis (chapter-by-chapter)........................42
Chapter 2 Describing between-individual variability I: Individual
growth pattern and variability in Serranus scriba .........................45
Chapter 3 Describing between-individual variability II:
Reproductive biology of Serranus scriba and variability of
reproductive potential .................................................................47
Chapter 4 Evidence of fishing mortality in small-bodied coastal
fish: basis for further analysis of the benefits of partial Marine
Protected Areas...........................................................................48
Chapter 5 Post-mortality levels of voluntary and/or mandatory
Serranus scriba released by recreational anglers: influence of
hooking injury, capture depth and venting .................................49
Chapter 6 Effects of hook size and barbless hooks on hooking
injury, catch per unit effort, and fish size in a mixed-species
recreational fishery in the western Mediterranean Sea................50
Chapter 7 The influence of type of natural bait on fish catches
and hooking location in a mixed-species marine recreational
fishery, with implications for management .................................51
Chapter 8 Recreational harvesting can induce evolutionary change
of life-histories of an exploited marine fish at surprisingly
restricted geographical scales ......................................................52
8.1 Introduction..........................................................................53
8.2 Material Methods .................................................................58
8.2.1 Species case-study ........................................................58
8.2.2 Harvesting selection experiment ..................................59
8.2.3 Sub-populations, connectivity and local harvesting
pressure .................................................................................60
8.2.4 Recreational-fishing induced life-history adaptation ...64
8.2.5 Controlling for non-fishing related environmental
impacts on life-histories........................................................70
8.3 Results ..................................................................................71
8.3.1 Selection pressures on life-history caused by recreational
angling ..................................................................................71
8.3.2 Population’s life-history strategies ...............................72
8.4 Discussion.............................................................................75
Chapter 9 A behavioural perspective of fisheries-induced selection
in marine coastal fisheries ...........................................................87
Chapter 10 General conclusions .................................................89
Reference list ..............................................................................91
List of Appendix.......................................................................105
Appendix I................................................................................105
Appendix II ..............................................................................107
Appendix III.............................................................................109
Appendix IV .............................................................................115
Appendix V ..............................................................................117
Chapter 1 37
Chapter 1 Human harvesting and fisheries-induced selection of
wild fish populations
1.1 Introduction
Humans have harvested fish from the wild since the origin of our species (Allendorf and
Hard, 2009; Stenseth and Dunlop, 2009). Today, marine ecosystems are experiencing
an accelerated loss of populations, and the current and future trends for global fisheries
remain controversial (Worm et al., 2006). Up to 63% of the assessed fish stocks
worldwide are overexploited and require rebuilding through reducing exploitation rates
(Worm et al., 2009). However, the solution for restoring fisheries is neither easy nor
non-controversial (Pauly et al., 2003). In all cases, preventing overexploitation usually
demands sophisticated management systems (Worm et al., 2009) and eco-evolutionary
approaches (Palkovacs, 2011).
Overexploitation may imply not only fishing-mortality rates exceeding the maximum
sustainable yield (Worm et al., 2006) but also unbalanced harvesting (Jørgensen et al.,
2007). The latter process can constrain the potential recovery rates of exploited
populations (Garcia et al., 2012). Unbalanced harvesting means that fishing is almost
never random (Heino and Godo, 2002). Individuals of, for example, a certain size, sex
38 J Alós (PhD Thesis) Selection drivers of life-history traits in marine coastal fishes
or behaviour are more likely to be removed from the population (Jørgensen et al., 2007).
This biased removal may result in genetic change if the expression of the selected
phenotype has some genetic basis (Allendorf et al., 2008). Therefore, detecting and
managing fishing-induced selection is a key issue for preventing potential negative
effects and ensuring sustainability (Kuparinen and Merilä, 2007).
The basis of natural or fishing-induced selection is the same: there must be some
inter-individual variability in specific phenotypic traits (Grimm and Railsback, 2005),
and such a phenotypic variability must be, at least in part, heritable (Allendorf et al.,
2008). Under these circumstances, trait-selective fishing mortality will produce a
selection differential (i.e., fish displaying some specific life-history traits will be more
prone to be fished), which in turn may result in fisheries-induced evolution of a wide
range of heritable traits (Law, 2000).
However, despite the relevance of this topic due to the potential risk for fishing
sustainability, only a few studies have directly analysed the existence and consequences
of the selection differentials exerted by fishing on different life-history traits (Enberg et
al., 2009). The results of those studies support the hypothesis that fishing-induced
selection favours individuals with smaller body size (Jørgensen et al., 2009). It is even
plausible that fishing-induced evolution could cause a more complex pattern favouring
smaller fish with reduced growth rates and increased reproductive investment that
mature sexually at a younger age and smaller size (Hutchings, 1993; Matsumura et al.,
2011; Rijnsdorp, 1993). While such traits may provide a benefit to individuals and
reduce their exposure to fishing gear, they may penalise fitness at the (wild) population
level in a process called maladaptive evolution (Olsen and Moland, 2011). Fishing-
Chapter 1 39
induced and natural selection interplay in this scenario, and such interplay will
ultimately determine the evolutionary trajectory of a population in response to fishing
pressure (Edeline et al., 2007; Matsumura et al., 2011).
Any fishing gear is size-selective per se (Heino and Godo, 2002). In consequence, it
is commonly accepted that fishing should favour the survival of individuals exhibiting
lower growth rates and smaller adult size (Conover and Munch, 2002). Indeed, some
studies conducted in a commercial marine fishing context have revealed negative
selection differentials on size-at-age in diverse species such as Atlantic cod, Gadus
morhua (Pérez-Rodríguez et al., 2009; Sinclair et al., 2002; Swain et al., 2007), and
plaice, Pleuronectes platessa (Rijnsdorp, 1993). Similarly, a classic laboratory experiment
in Menidia menida revealed that size-selective mortality may favour slow-growing fish
(Conover and Munch, 2002). However, most of the studies of fishing-related selection
differentials have been conducted in a commercial fishing context where size-selectivity
is well documented (Law, 2000).
In contrast, (Uusi-Heikkilä et al., 2008) argued that vulnerability to fishing may be
more complex: fishing-related selection might operate on a wide range of traits, either
positively or negatively and co-varying with vulnerability to fishing. This relation could
be the case of recreational fisheries, where in addition to the existence of a bias for
harvesting the larger individuals (Lewin et al., 2006), fish must actively interact with the
fishing gears (Uusi-Heikkilä et al., 2008). Given this fact, it should be expected that the
fish vulnerable to fishing would be those with not only faster growth but also bolder
behaviour (Uusi-Heikkilä et al., 2008). However, the links among life history,
physiological and behavioural traits can generate unexpected selection patterns, and the
direction of the selection differentials is an open (and hot) topic (Cooke et al., 2007).
Literature about this topic is still scarce and does not show a clear pattern of the
direction of fishing-induced selection on life-history traits. On the one hand, authors
such as (Saura et al., 2010) reported a decline in body length and weight per generation
(a decrease in the growth rate) because of recreational fishing pressure in the Atlantic
salmon, Salmo salar. On the other hand, a line of fish more vulnerable to fishing
exhibited a lower growth rate after only three generations of artificial selection for either
high or low vulnerability in largemouth bass, Micropterus salmoides (Redpath et al.,
2009). The theoretical framework justifying such a pattern is that vulnerable fish seem
to display a more accelerated metabolic rate, the energetic cost of which is compensated
by greater energy intake through an increased feeding rate (Redpath et al., 2009;
Redpath et al., 2010). Thus, the results reported for recreational fishing are
contradictory and cast doubt on the a priori expectation that more-vulnerable fish
should necessarily encompass the faster-growing individuals within a population.
Indeed, (Matsumura et al., 2011) showed that selection differentials on the growth
rate of a recreationally exploited population of pike (Esox lucius) might either be positive
or negative depending on the management strategy adopted. The adult body size was
predicted to either decline or increase in response to recreational fishing mortality
depending on which size classes of fish experienced the highest mortality. In the
absence of management regulations, the model proposed by (Matsumura et al., 2011)
predicts selection against individuals with faster growth rates. At the same time, fish
with a larger size of maturation and smaller reproductive investment were selected by
Chapter 1 41
fishing. The fishing outcomes may be the result of a multi-trait correlation (Mollet et
al., 2010), including a trade-off between growth and reproduction (Roff, 1992), that
determines a specific life-history strategy that maximises individual fitness.
Therefore, there is a need for more detailed studies to better understand whether
recreational fishing constitutes a non-random selection process with respect to key life-
history traits (Audzijonyte et al., 2013). In the case of non-random fisheries, the
putative fishing-induced selective drivers should be qualified and quantified, and their
potential ecological and evolutionary effects should be evaluated (Kuparinen and Merilä,
2007). Then, balanced fishing that distributes mortality across the widest possible range
of the selected traits should be incorporated into conventional management plans to
mitigate the undesirable consequences of fisheries-induced evolution (Garcia et al.,
2012). This lack of information especially in marine costal species is the strategic
objective of this Thesis.
1.2 Objectives of the Thesis
The main objective of this PhD Thesis was to analyse the plausibility and consequences
of fishing-induced selection on hereditable life histry traits related to recreational fishing
in the Mediterranean. A marine coastal species, Serranus scriba (Linnaeus, 1758), was
selected as a case study because this species is highly exploited by the recreational fishery
in the Mediterranean. As in many cases, the variability of the life-history traits of the
wild populations of this species with low interest for commercial fisheries is poorly
known. Thus, the first sub-objective was to accurately describe the variability in the
species’ growth and reproductive potential. The second sub-objective was to explore the
different sources of recreational fishing mortality of S. scriba. The study of fishing effects
was extended to estimate sources of fishing mortality, such as post-release mortality, and
to provide management tools to mitigate the negative effects of fishing. The third sub-
objective was to test the hypothesis that fishing can be selective for some specific life-
history traits independently of fish size and to identify how different wild populations
respond to these selective forces when gene low is limited. The fourth sub-objective was
to test the hypothesis that behaviour could be a fishing-induced selection driver as well.
Finally, the implications of the results for better management (i.e., aimed at enhancing
sustainable fisheries and avoiding plausible maladaptive effects of fishing) are discussed.
1.3 Structure of the Thesis (chapter-by-chapter)
This PhD thesis is structured in 10 chapters. The contents and structure of the chapters
one-by-one are detailed below for clarity.
Chapter 1 introduces the state of the art, including an outline of the topic, a general
description of the problem and its magnitude, and a review of the most important
references. This chapter also lists the main objectives and the hypothesis of this PhD
Thesis
Chapter 2 provides an accurate description of the individual growth of S. scriba in a
wild population using a novel theoretical approach.
First Chapter 3 provides a description of the reproduction biology of S. scriba using
microscopy histological techniques. Second, it shows the variability and the main factors
affecting the individual reproductive potential in wild populations of S. scriba.

Chapter 1 43
Chapter 4 review the literature and provide evindences that recreational fishing is a
source of mortality on the Posidonia oceanica fish community with, especial emphasize
on S. scriba. This chapter also shows possible benefits of management trough partial
Marine Protected Areas.
Chapter 5 explores alternative sources of fishing mortality induced by the mandatory
(minimum legal size) or voluntary release of S. scriba made by recreational anglers. This
chapter identifies the most important post-release mortality factors that affect survival
of S. scriba.
Following the results obtained in Chapter 5, the Chapters 6 and 7 suggest science-
based management tools to reduce the fishing mortality induced by mandatory or
voluntary release.
Chapter 8 demonstrates how recreational fishing selects for certain phenotypes in
terms of life history traits independently of fish size. Also, this chapter demonstrated
how the matching population’s genetic connectivity and the selective forces, the optimal
life history strategy of harvested populations of S. scriba can be explained.
Chapter 9 shows the results of a simulation analysis sugesting how behaviour may be
modulated by fisheries-induced selection. Most coastal fish seems to move according to
home range behaviour and empirical data on between-fish variability of such behaviour
was used to test if fishing can be selective for some movement charcateristic. Results
were clear and the life history/behavioural relationship emerge as an excellent
opportunity for further research.
Finally, Chapter 10 provides the general conclusions of this PhD Thesis.

The references cited throughout this PhD Thesis are included at the end followed by
the appendixes including additional information.

Chapter 2 Describing between-individual variability I: Individual
growth pattern and variability in Serranus scriba
Alós, J., Palmer, M., Balle, S., Grau, A. M., and Morales-Nin, B.
2010. Individual growth pattern and variability in Serranus scriba:
a Bayesian analysis. ICES Journal of Marine Science, 67: 502-
512.
Chapter 3 Describing between-individual variability II:
Reproductive biology of Serranus scriba and variability of
reproductive potential
Alonso-Fernández, A., Alós, J., Grau, A., Dominguez-Petit, R.,

and Saborido-Rey, F. 2011. The Use of Histological
Techniques to Study the Reproductive Biology of the
Hermaphroditic Mediterranean Fishes Coris julis, Serranus
scriba, and Diplodus annularis. Marine and Coastal Fisheries,
3: 145-159.
Alós, J., Alonso-Fernández, A., Catalàn, I., Palmer, M., and
Lowerre-Barbieri, S. (in press). Individual variability in the
factors that affect reproductive potential in Serranus scriba.
Scientia Marina.
Chapter 4 Evidence of fishing mortality in small-bodied coastal
fish: basis for further analysis of the benefits of partial Marine
Protected Areas
Alós, J., and Arlinghaus, R. 2013. Impacts of partial marine
protected areas on coastal fish communities exploited by
recreational angling. Fisheries Research, 137: 88-96.

Chapter 5 Post-mortality levels of voluntary and/or mandatory
Serranus scriba released by recreational anglers: influence of
hooking injury, capture depth and venting
Alós, J. 2008. Influence of anatomical hooking depth, capture
depth, and venting on mortality of painted comber (Serranus
scriba) released by recreational anglers. ICES Journal of Marine
Science, 65: 1620-1625.

Chapter 6 Effects of hook size and barbless hooks on hooking
injury, catch per unit effort, and fish size in a mixed-species
recreational fishery in the western Mediterranean Sea
Alós, J., Palmer, M., Grau, A. M., and Deudero, S. 2008. Effects of
hook size and barbless hooks on hooking injury, catch per unit
effort, and fish size in a mixed-species recreational fishery in the
western Mediterranean Sea. ICES Journal of Marine Science, 65:
899-905.
Chapter 7 The influence of type of natural bait on fish catches
and hooking location in a mixed-species marine recreational
fishery, with implications for management
Alós, J., Arlinghaus, R., Palmer, M., March, D., and Alvarez, I.
2009. The influence of type of natural bait on fish catches and
hooking location in a mixed-species marine recreational fishery,
with implications for management. Fisheries Research, 97: 270-
277.
Chapter 8 Recreational harvesting can induce evolutionary
change of life-histories of an exploited marine fish at surprisingly
restricted geographical scales
Recreational harvesting was sometimes considered neither persistent nor efficient
enough to be of concern for management. In particular in marine environments the
impacts of recreational anglers were usually considered negligible. However, consistent
selection acting on certain phenotypes should lead to adaptation of life-history strategies
whenever gene flow is low. Following this argument one would expect characteristic
adaptations of exploited life-histories to recreational fishing pressure in coastal fish as
long as population connectivity is low and evolutionary change conceivable to
materialize. Here we provide compelling evidence of recreational fishing can induce
evolution of fast-life-histories in a small coastal fish as surprisingly small geographical
scales. We first show that recreational fishing indeed preferentially captures fish with
low reproductive investment and large adult size. Secondly, using oceanographic particle
motion models we identify subpopulations of Serranus scriba that were reproductively
isolated. We finally found fish life histories in exploited sites to be substantially different
from unexploited sites. Individuals characterized by high reproductive investment and

small adult size were much more prevalent in exploited compared to unexploited sites
suggesting evolutionary adaptations to recreational harvesting that was unlikely be
caused by site-specific environmental factors. We propose that recreational harvesting
changes life histories also in open marine systems at small geographic scales. A possible
meta-population structure is therefore important to consider when managing coastal
fisheries, even for small bodied, geographically restricted fish that are a targeted by
anglers mainly.
Keywords: Bayesian approach, bi-phasic growth model, life-history traits, fisheries-
induced evolution, meta-population structure, recreational fishing, sustainable yield
8.1 Introduction
When fishing exploits a naturally reproducing fish stock, exploitation is almost never
random. Indeed humans selectively target certain species (e.g., top predators) and
specific phenotypic traits and underlying genotypes within species (e.g., larger sized
fish) (Law, 2000; Jørgensen et al., 2007; Allendorf and Hard, 2009; Laugen et al.,
2012). Elevated and/or non-random harvesting can generate selection on certain traits
and result in fisheries-induced adaptive changes of maturation or growth, which may be
of relevance to management (Jørgensen et al., 2007). Whether adaptive changes in life-
histories are evolutionary (i.e., adaptive genetic) or merely phenotypically plastic is
controversially debated for many stocks (Browman et al., 2008; Jørgensen et al., 2008).
Available literature nevertheless supports the hypothesis that fisheries-induced evolution

(FIE) should select for “fast life-histories” characterized by small size and age at
maturation and high reproductive investment, which should collectively reduce adult
body size (Jørgensen et al., 2007; Law, 2000) and elevate adult natural mortality
(Jørgensen and Fiksen, 2009; Jørgensen and Holt, 2013).
The potential for FIE happens has been mainly discussed in the context of
commercial fisheries where positive size-selection is well documented (Law, 2000;
Sharpe and Hendry, 2009). More complex exploitation patterns with respect to traits
selective by fishing are conceivable for passive fishing gears such as trapping or
recreational angling (Uusi-Heikkilä et al., 2008). In fact, in many recreational harvesting
situations, selection on behavioural or physiological traits should be prevalent in
addition to directly or indirectly correlated life-history traits (Alós et al., 2012; Cooke et
al., 2009; Cooke et al., 2007; Redpath et al., 2009; Redpath et al., 2010; Wilson et al.,
2011; Sutter et al., in press). The resulting complexity of selection pressures on a range
of traits suggests that selection differentials are not necessarily to be expected to have
the expected sign. For example, it has often been found intuitive to think of selection on
body size in fishing context (Conover and Munch, 2002). Indeed, in recreationally
exploited Atlantic salmon and Chinook salmon, negative selection differentials on body
size were reported (Saura et al., 2010; Kendall and Quinn, 2011). By contrast, (Redpath
et al., 2009) found that largemouth bass selectively captured by anglers were smaller
than those surviving exploitation, likely because food limitation met with a correlation
between metabolism and vulnerability to fishing. Therefore, the evolutionary impact of
recreational harvesting on life-history cannot be easily foreseen and more stock-specific
analyses that represent local fishing conditions and patterns are needed. To our
knowledge, no study on recreationally exploited fish stocks exists that has attempted to
sort this question out in marine environments.
To support the hypothesis of FIE two not mutually exclusive conditions have to be
met. First, any phenotypic changes has to be due to fisheries, and second, at least part of
the phenotypic change has to have a genetic basis (Kuparinen and Merilä, 2007). A
range of approaches have been put forward to study FIE in these two dimensions. First,
theoretical modelling of FIE (e.g., (Dunlop et al., 2009; Matsumura et al., 2011) has
shown that fishing exploitation can be strong enough to cause selection differentials and
fisheries-induced adaptive change to be plausible. Naturally, inference of models to the
wild are limited, and models will also never conclusively answer whether evolution has
truly happened in nature (Audzijonyte et al., 2013). Second, many observational studies
using time series of phenotypic data from the wild have been conducted to study
evidence for FIE (e.g., (Rijnsdorp, 1993; Swain et al., 2007). The main issue of this
approach is to disentangle genetic from plastic change in life-histories and sophisticated
statistical approaches to control for some of the most important environmentally-
induced plastic responses on phenotypes have been developed (e.g., the popular
probabilistic maturation reaction norm approach, (Heino and Dieckmann, 2008).
Clearly, observational studies cannot discount conclusively the potential for non-fishing
factors to affect phenotypic change (Audzijonyte et al., 2013), and usually such studies
also lack the molecular tools needed to study genetic changes caused by fishing. There
are, however, some few observational studies that have used molecular techniques to
study historical samples from exploited stocks, which suggest genetic change has taken
place (Ãrnason et al., 2009; Jakobsdóttir et al., 2011; Pukk et al., 2013). The third and
final popular approach to the study of FIE has been experimental in the laboratory or in
experimental facilities such as ponds (e.g.,(Conover and Munch, 2002; Philipp et al.,
2009). The strength of these studies is that they can provide cause-and-effect evidence
and that phenotypic change has a genetic basis (e.g., through common-garden
experiments) (Conover and Baumann, 2009), but the inference to wild conditions is
highly difficult.
A final approach that joins many of the benefits outlined in the three approaches
above is a comparative field approach that studies phenotypes of fish striving in similar
environmental conditions but differ by historical exposure to harvesting pressure (e.g.,
(Bevacqua et al., 2012; Drake et al., 1997). Whenever there is no contemporary natural
gene flow among sub-populations, differential fishing slection should generate sub-
populations that adapt their life-histories to the local selective pressures (Hutchings,
1993; Reznick et al., 1990). Some comparative field approaches has some history in
freshwater systems, where spatial segregation and limited natural gene flow (Haugen
and Vøllestad, 2001; Drake et al., 1997). However, meta-populations with historically
connected fish species are increasingly identified in the open ocean and marine systems
(Mora and Sale, 2002), which constitutes a good opportunity to study evolutionary
consequences of fisheries in wild fish stocks. In fact, although marine coastal
environment has been traditionally assumed that are open and operate as genetically
open systems (Roberts, 1997), recent development matching genetics and oceanography
suggests that marine fish populations are not as open as expected (Palumbi and Warner,
2003) even at spatial-scales of a few km (e.g., (Carreras-Carbonell et al., 2007; Jones et

al., 2005; Taylor and Hellberg, 2003). Small-scale meta-populations in the marine
environment would render a field approach useful to provide evidence of FIE change in
the wild (Pukk et al., 2013). The Mediterranean Sea offers a suitable model system for
population’s connectivity because oceanic fronts and local hydrodynamic conditions can
generate patterns of isolation within sub-populations in many coastal species (Schunter
et al., 2011). Physical processes can induce an increase of local self-recruitment and
increase the vulnerability of local populations to human impacts such as fishing (Galarza
et al., 2009). Therefore, the local depletion of adult individuals due to fishing activity
could therefore change the genetic architecture of the sub-population due to
contemporarily restricted gene flow from nearby areas (Roberts, 1997).
Capitalizing on this idea, in this Chapter we adopted a multidisciplinary approach
linking an individual, function-based estimation approach to infer life-history traits
(annual reproduction investment, theoretical maximum size, age of maturation and
immature growth rate) and high resolution hydrodynamic modelling to test if between-
population differences in life-history traits can be adequately explained by variation in
fishing and population connectivity. Our study is based on a comprehensive analysis of
life-history traits in nearby areas at small geographical scales subjected to varying
historic fishing pressure (from intense harvesting to no-take) combined with the results
from a high resolution numerical model that allowed estimating the connectivity of the
studied fish populations. Our hypotheses was that historic recreational fishing pressure
has led to life-history adaptations towards “fast life histories” in a small marine fish with
limited gene flow among spatially close coastal areas.

8.2 Material Methods
To generate compelling evidence for fisheries to be the most parsimonies explanation
for phenotypic change in an open marine environment, at least three lines of evidence
have to be accumulated: i) one needs to show that the particular gear under
investigation induces selection differentials on heritable traits (e.g., body size or
maturation), ii) one has to provide evidence of restricted gene flow among local
populations and iii) one has to test the predicted changes in phenotypes among stocks
experiencing differential fishing pressures and exclude non-fishing related factors as
drivers of change as far as possible. We have compiled all three lines of evidence using a
small-bodied coastal fish exploited by recreational fisheries as a model species.
8.2.1 Species case-study
Serranus scriba (Figure 8.1), is a small-sized Serranid that inhabits shallow seagrass
meadows (Posidonia oceanica and Cymodocea nodosa). Its reproductive behaviour is
characterized by simultaneous hermaphroditism with relative short-life-span whcih
achieves sexual maturity during its first or second year (Alonso-Fernández et al., 2011).
It is characterized by daily-batch egg releases (with peak spawning season in May and
June in the Mediterranean), indeterminate fecundity and asynchronous oocyte
maturation development (Chapter 3 of this PhD Thesis). S. scriba is of no importance
to commercial fisheries, but the species is highly appreciated by recreational anglers. In
fact, the species is among the most intensively harvested (both in terms of numerical
abundance and biomass yield) by the small-scale recreational boat anglers in the NW
Mediterranean (Morales-Nin et al., 2005).

8.2.2 Harvesting selection experiment
The first question to be answered was whether recreational angling is selective for life-
history traits such as annual reproduction investment, theoretical maximum size, age of
maturation and immature growth rate. Our empirical approach was based in exploring
the potential direction of selection by comparing the life-history of individuals sampled
using an experimental active method (random sample of the population using
experimental beam trawl) and the dominant fishing gear used in the area (passively
fished hook-and-line fishing) similar to those performed by (Wilson et al., 2011) to
explore selection of fish personality. We selected an experimental area of 1.1 km2
(average of the movement area used by S. scriba, (March et al., 2010)) located at the
south of Mallorca Island (Figure 8.1). During May (peak of the spawning season),
individuals of S. scriba from this population were sampled using either the experimental
beam trawl (active method) or hook-and-line (passive method) at the identical within-
day (mid-day) and spatial-scale.
We conducted three different fishing hauls using the experimental beam trawl (see
details of the method in (Deudero et al., 2008) to obtain a large enough sample (n = 50).
We also conducted a number of experimental angling sessions done at the experimental
area to obtain a hook-and-line sample (n = 165). The experimental angling sessions
were done following protocols described in Chapter 7 of this PhD Thesis. All
individuals were processed fresh, measured (total length, mm), weighted (total weight,
g), and their sagitta otoliths were dissected for estimation of the individual life-history
traits using a function-based assessment approach that infers the traits from an
individual growth curve (see for details below). Additionally, the gonads of a sub-sample
for each sampling method (n = 54) were used for direct measurement of reproduction
investment (batch-fecundity) following the method based on the automated image
analysis described by (Alonso-Fernández et al., 2009).
Figure 8.1 Picture of Serranus scriba (Family: Serranidae) (top left panel). Map locating the Mallorca
Island in the Balearic Sea (NW Mediterranean) (bottom left panel).The spatial location of the 25 sites of
eggs releasing is showed in the right panel highlighting the high exploited (recreational fishing) area
located in Palma Bay and the waters of National Park of Cabrera where recreational fishing is not allowed
(No-take MPA). Black-white gradient represents to be the spatial distribution of the recreational fishing
effort in the Mallorca Island proyected in a raster of 250 m x 250 m of (n = 1,430). Recreational anglers
were censed by the coastal guards (data obtained in a work-package of the Project CONFLICT).
8.2.3 Sub-populations, connectivity and local harvesting pressure
After identifying the direction of possible selection operating on life-history traits, sub-
populations of S. scriba were identified that should show adaptations in such traits due
to historic harvesting pressure. The ideal populations to study should experience
different historic fishing intensity and have limited gene flow among each other. It is
under these conditions that distinct life history adaptations should be expected. We thus
selected the sub-populations of S. scriba in the present study following two criteria: i)
limited gene flow and ii) contrasting historical and contemporary harvesting pressures.
We first explored the gene flow approximated by the population’s connectivity and
subsequently we selected reproductively isolated sub-populations exposed to contrasting
harvesting pressure (old no-takes marine protected areas, MPA, and a highly exploited
area, see below) to test whether life-history changes in life-histories have taken place.
Based on life-history theory and the results of the potential selection pressures caused by
angling from section 8.2.2, we predicted fish subjected to high harvesting pressure to
have evolved smaller maturation age and size, higher reproductive investment and
smaller adult size, i.e., a faster life-history (Matsumura et al., 2011).
Similar to many other coastal fish inhabiting seagrass meadows, early life-stages of S.
scriba are pelagic with dispersal characteristics and gene flow linked to the dominant
coastal flow patterns during the spawning season. To estimate the degree of the sub-
population’s connectivity, we selected 25 areas (see spatial distribution in Figure 8.1)
potentially inhabited by adults of S. scriba (any seagrass or rocky bottoms with < 30 m
depth) along the inner-shelf region of the SE coast of Mallorca (Figure 8.1). The self-
recruitment and the direct and indirect connectivity of each sub-population of S. scriba
were calculated following the procedures of (Basterretxea et al., 2012). Briefly, the
Lagrangian trajectories (Watson et al., 2011) of weekly releases of passive tracers (virtual
eggs) during the peak reproductive season were numerically simulated using a high
resolution (200 m) three-dimensional density-resolving model of the hydrodynamics of
the southern Mallorca through a previously-validated Princeton Ocean Model (POM).

The flow fields produced by surface wind forcing were obtained from the 12-hourly
wind maps from an historical data-base of ten years (2000-2009, data from the coastal
meteorological station of the Palma de Mallorca airport, accessed February 2012:
http://www.aemet.es/).
The pelagic larval duration (PLD) was assumed to be 28 days (Macpherson and
Raventos, 2006) and the floatability of the eggs was assumed to be positive following
the results obtained in the experiment provided in Appendix III. The virtual eggs were
homogeneously distributed in each area, and a reflective boundary condition was used to
prevent particles from moving onto land. We assumed individual larvae to have been
recruited to a particular area if they were found in the inner-shore habitat (defined by
each area) at the end of the PLD simulation. Degree of self-recruitment (mean and s.d.)
was defined as the number of particles remaining in area j at the end of the prescribed
PLD period divided by the numbers of particles initially released in area j per each
releasing period (Basterretxea et al., 2012).
The probability that a particle is transported from one area to another one by the end
of the PLD (28 days) was considered the direct connectivity. Direct connectivity
estimates provide a good indication of the relationship between self-recruitment and
dispersal from and to other areas during one hatching cycle. In this sense, it is indicative
of organism dispersal (Basterretxea et al., 2012). Hereditable components of phenotypic
variation can also be exchanged through indirect transport involving several generations.
The indirect connectivity reflects the probability of this transport across generations
(Casabianca et al., 2011) and is dependent on the existence of suitable habitat corridors
that allow the survival of intermediate populations between the source and the receiving
population. Indirect connectivity (IC) between paired populations was calculated as:
N
IC   DC n
n 1 (1),
where N is the number of hatching cycles, DC identifies the direct routes and DCn
includes all the routes requiring n-1 steps (or n direct routes). It is important to note
that indirect connectivity conveys a temporal term that provides information on the
velocity of the exchange. In the case of phenotypic variations this may be relevant
because longer routes of exchange, involving a higher number of generations, may imply
loss of the hereditable signal.
Patterns of population’s connectivity showed how the local self-recruitment processes
of the early life-stages of S. scriba were dictated by the coastal morphology at the spatial-
scale studied. Two main areas, the inner part of Palma Bay (IPB) and the National Park
of Cabrera (NPC) were markedly different from the others showing the highest
percentages of self-recruitment (Figure 8.2 and 8.3). In comparison to an open
population (OP, Figure 8.2), the final position of the virtual eggs released suggested a
notorious high degree of retention of eggs in these two main areas (Figure 8.2). Direct
connectivity between these two populations (IPB and NPC) areas was low (Figure 8.3),
indicating low potential gene flow. These two areas also showed low degree of direct
connectivity with the OP (Figure 8.3). When longer term dispersal was analyzed
(indirect connectivity based in 50 years) the effect of along shelf connections between
the IPB and NPC remained low (Figure 8.3). Therefore, while the populations in OP
were considered to have certain degree of heritable exchange, IPC and NPC were
considered to relatively mutually isolated, which constitutes a good opportunity to study
the potential for life-history change due to fishing in IPC with NPC serving as a
control.
Apart of the limited gene flow of both NPC and IPB, these sub-populations were
unique because they were the only ones exposed to the minimum and maximum
harvesting selection pressure, rendering them an ideal and pretty unique case-study
candidates to test our hypothesis. The NPC was declared a sanctuary (no-take MPA) in
1991. It is formed by 19 small Islands and is one of the largest marine protected areas of
the Mediterranean Sea (Francour et al., 2001). Since recreational fisheries was
abandoned in NPC, top-predator species and potential predators of small individuals of
S. scriba notably increased its abundance (e.g., Epinephelus marginatus, (Reñones et al.,
1999). By contrast, the second area IPB (South of Mallorca Island) is the most
important recreational fishery in Mallorca Island receiving substantial recreational
fishing effort (Figure 8.1).
8.2.4 Recreational-fishing induced life-history adaptation
Studies on FIE depend on assessment of individual traits. Many life-history traits such
as maturation decisions or amount of resources invested into reproduction cannot be
directly observed or measured on captured fish. Therefore, many studies on FIE have
relied on inferring life-history traits from hard structures (e.g., (Swain et al., 2007). The
back-calculation of length-at-age using growth marks contained in the otoliths can offer
a reliable method to obtain repeated measures over the life-span of the same individual
when there is a strong relationship between fish and otolith size as is often the case
(Mollet et al., 2010; Pilling et al., 2002; Brunel et al., 2012). We chose a function-valued
approach (Stinchcombe and Kirkpatrick, 2012) to infer key life-history traits from
individual growth curves. In this approach, parameters (traits) estimated from individual
growth curves were used as data characterizing individual fish. As far as we are aware,
this is the first time this method is used in studies on FIE.
We estimated individual life-history strategies in the NPC, IPB and the OP sub-
populations to test our prediction that fish in the NPC should be characterized by
“slow” life-histories compared to fish in IPB (the exploited area). Samples of adult
individuals of S. scriba were obtained from each study area: IPB (n = 201), OP (n = 215)
and NPC (n = 353). Fish sampling was carried out with the same gears (conventional
hook-and-line) at similar habitat characteristics (seagrass meadows of P. ocenaica) and
depths. All the individuals were processed as described in section 8.2.2 and the gonads
from sub-sampled individuals from May and June (n = 53 from the IPB, n = 41 from
the OP and n = 79 from the NPC) were used for direct measurement of the
reproductive investment (batch-fecundity). Moreover, direct measures of relative
abundance (catch per unit effort as index) and the size and age of vulnerable fish were
taken using highly standardized fishing sessions with conventional hook-and-line gear.
Fishing sessions (IPB n = 54, AP n = 24 and NPC n = 64) were realized using the
standardized recreational angling gear described in (Alós et al., 2009) visiting each sub-
population on randomly selected days during a year period. Briefly, two anglers
(accompanied by a researcher) fished the same time with the same gear (hook size)
using the same effort (30 min) using two different baits (worm and shrimp). At the end
of each fishing session the fish were identified and measured (fish sizes: IPB n = 61, OP
n =40 and NPC n = 202). The otoliths of a random-sample of the individuals were
processed as described below for aging purposes (IPB n = 61, OP n = 27 and NPC n =
179).
We fitted each individual back-calculated length-at-age data from otoliths to the bi-
phasic growth model proposed by (Lester et al., 2004) to provide indirect measures of
the annual reproduction investment (g), individual maximum theoretical size (L∞), age
of initiation of maturation (T) and immature growth rate (h) (Arlinghaus et al., 2009;
Matsumura et al., 2011). The model assumes linear growth in the juvenile phase and a
trade-off between the energy allocation to somatic growth and reproduction in adult
ages (Lester et al., 2004). The model by (Lester et al., 2004) is a special case of a more
general version published by (Quince et al., 2008a; Quince et al., 2008b) and is valid for
a length-weight relationship with an exponent b = 3, which is the case in S. scriba (see
Chapter 3 of this Thesis) and bi-phasic growth pattern (as detected in S. scriba in
Chapter 2). Somatic growth rate before the age of the onset of maturation (T, years) in
the model by (Lester et al., 2004) is given by
L t  h (t  t 1 )   when t ≤ T (2),
where Lt is the length at age t (mm), h is the immature growth (mm year-1), t1 is the
theoretical age at size 0 mm and ε is the Gaussian error.

When the individual takes the “decision” to start investing into reproduction at age
T, the somatic growth of the individual is represented by the more classical von
Bertalanffy (VB) growth equation with parameters that provide explicit information on
the energy allocated to reproduction (g, annual reproduction investment, year-1) as
 
Lt  L  1  e  k  t  t 0    when t ≥ T (3),
where L∞ = 3h/g, k = ln(1 + g/3) and to = T + (ln(1 - g(T - t1)/3)/ln(1+g/3), and ε is the
Gaussian error (Lester et al., 2004).
These two equations provide a biologically meaningful interpretation of the VB
growth parameters traditionally used in fisheries science but representing simple
mathematical descriptors (Lester et al., 2004). Accordingly, the growth factor k is
proportional to reproduction investment (k ≈ g/3) and reflects the relative investment
averaged over the entire reproductive cycle (g) (see also (Charnov, 2008), L∞ (maximum
theoretical body size) is the ratio of net production to reproductive investment (L∞ =
3h/g) and t0 in the VB growth model is another simple function of T and g (Lester et al.,
2004).
The growth model was fitted using a Bayesian approach for each individual. This
approach has several advantages for inferring parameters of complex models (Ellison,
2004) and has become popular to estimate individual growth parameters (Helser and
Lai, 2004). Two different models were fitted. The first model was applied to the data of
either beam or hook-captured fish to estimate the type of selection on life-history traits
exerted by hook-and-line captured fish relative to trawled individuals. The second
model was fit to the individual data of the three study populations to infer life-history
traits of fish in the three sites (IPB vs. OP vs. NPC). The model was implemented and
run using the runjags library of the R package (http://www.r-project.org/), which opens
JAGS (http://mathstat.helsinki.fi/openbugs/). The model was fitted for individuals of
age 5 or older at capture, which improves the accuracy of the method substantially (Alós
et al., 2010). Moreover, due the over-dispersed representation of older ages in the
unexploited NPC site in the second model, 50 individuals were randomly selected. By
this approach, the ages and the number of observations per individual included in the
analysis were not different among sites in fitting of the first (Chi-squared independence
test, p = 0.46) and second model (Chi-squared independence test, p = 0.1). Therefore,
any significant variation in life-history traits among gears or sites was independent of
the fish size and the number of samples (ages) per fish including in the fitting. A
hierarchical (individuals within a population) structure of minimal prior and hyper-prior
knowledge was used and is fully provided in Appendix IV. Three Monte Carlo Markov
Chains (MCMC) were run using randomly chosen initial values for the life-history
traits and After accepting convergence of the three chains, assessed using Gelman-
Rubin diagnostic (Plummer et al., 2006), the number of valid iterations after burning
and thinning was 1,000. The feasibility of the Bayesian framework adopted here was
assessed through a simulation analysis (Appendix V).
We estimated the potential direction of selection of recreational fishing by
comparing the distribution of the individual Bayesian means of each life-history trait for
the two sampling methods beam trawl and hook-and-line to test whether angling would
be expected to exert selection on S. scriba life-history traits. Subsequently, between-
population comparisons in average life-history traits were carried out by comparing the
Bayesian population’s means of each life-history traits using the Bayesian Credibility
Intervals (BCI) (2.5%, 50%, and 97.5% overlapping). Unlike the frequentist’s p value,
the Bayesian posterior distributions are interpreted correctly as one’s belief that there is
a p% probability that the parameter of interest lies within the interval (Ellison, 2004).
To make robust conclusions, we considered differences in the life-history traits when
the BCI did not overlap at all. L∞ and h were expressed in mm as the distance from the
otolith centre to the border following the growth path. However, in order to facilitate
interpretation, in some cases the scale was translated to fish size (total length) using the
regression model provided in Chapter 2.
Between-sampling methods and between-population differences in direct measures
of reproduction investment (batch-fecundity), and populations differences in the relative
abundance, age and size of vulnerable fish sampled by experimental angling were
analysed using frequentist statistics using Generalized Lineal Mixed Effects Models
(GLMM, (Zuur et al., 2009). Minimum adequate GLMMs were fitted following a
forward step-by-step (by comparing the model with and without the factor through
ANOVA) until maximum explanatory power was reached (the AIC criterion) (Zuur et
al., 2009). The p values for each parameter were estimated through MCMC simulations
using the function pvals.fnc from the library languageR of the R package.
Figure 8.2 Self-recruitment (10 years mean and s.d.) of the 25 sites of eggs release considered in this
study (upper panel). The highest values observed in the inner part of Palma Bay (sites 6 and 7) and the
National Park of Cabrera (sites 14, 15 and 16) were showed in dark grey and black respectively. An
average, in terms of SC, population (sites 9 and 10) is showed in light grey. The bottom panels shows the
spatial-plots of a fixed kernel probability distributions (estimated in raster of 250 m x 250 m) of the
positions at the end of the PLD duration of the virtual eggs released in each of these three main areas.
8.2.5 Controlling for non-fishing related environmental impacts on life-histories
To help sorting out whether FIE would be the most parsimonious explanation of the
observed life-history differences among the three sites investigated in the present study,
we also included some environmental data in our study. A 10 year record of satellite
(MODIS-Aqua) derived monthly averages of sea surface temperature (SST) and
chlorophyll (Chl-a) as measure of productivity was obtained for each of the three sites
included in the analyses and compared among them. In all cases (temperature and
productivity) chronograms were used to visual inspection for differences.

8.3 Results
8.3.1 Selection pressures on life-history caused by recreational angling
The individual values of maximum theoretical size (L∞), annual reproduction
investment (g), immature growth (h) and age of maturation (T) were successfully
estimated for the experimental beam trawl (active gear) and hook-and-line samples
(passive gear) (Table 8-1). Mean L∞ was significantly different between the fish sampled
with the two gears (Table 8-1). Similarly, the average g of fish vulnerable to hook-and-
line were smaller than those captured in the trawl suggesting that removal of vulnerable
fish to angling would exert selection for high reproductive investment (Table 8-1).
Results in immature growth rate h and onset of maturation T suggested on average
older age of maturation and smaller immature growth in the hook-and-line individuals
compared to fish vulnerable to trawls, but in both cases the Bayesian credibility intervals
overlapped notably suggesting smaller selection pressures on these traits (Table 8-1).
Figure 8.4 shows a box-plot of the individual values of life-history traits estimated for
the two fishing treatments. Overall, the results suggested a specific direction of
recreational-fishing induced selection, which should favour fish with higher
reproductive investment and smaller adult growth capacity under conditions of intensive
recreational harvesting.
The results of the direct measures of reproduction investment (batch-fecundity) were
in the line of those obtained for the indirect measures of g just described. Accordingly,
there were significant differences between the beam trawl and hook-and-line (GLMM;
p <0.05, Table 8-2). Individual’s sampled using hook-and-line invested less into
reproduction compared to trawl-captured fish (Table 8-2). The relationship between
batch fecundity and g estimated from the growth curve was highly significant (GLM, p
< 0.01).
Figure 8.3 Directional flow diagram showing the paired direct and indirect connectivity (black arrows)
calculated among the inner part of Palma Bay (PB), the open population (OP) and the National Park of
Cabrera (NPC) (each number can be interpreted as a % of eggs released in an specific sub-populations).
Indirect connectivity values are the maximum values of connectivity over 50 years. The grey arrows show
the connectivity attributed by other sub-populations. 360º black arrows are the self-recruitment values for
each specific sub-population.
8.3.2 Population’s life-history strategies
The individual values of g, L∞, T and h were also successfully estimated for the
individuals sampled in the three sub-populations studied (IPB, NPC and OP). There
were non-overlapping posterior distributions of g between the IPB (Bayesian mean =

0.75 years-1) and the NPC (Bayesian mean = 0.61 years-1). Thus, we accepted the
hypothesis of higher reproductive investment in the exploited site IPB relative to the
unexploited site NPC (Table 8-3 and Figure 8.5). In contrast, the posterior distribution
of the open population OP had intermediate values (Bayesian mean = 0.70 years-1)
notably overlapping with the distributions of the other two populations. Results were
similar for L∞, and differences between the Bayesian means of IPB (Bayesian mean =
228 mm of fish total length) and NPC (Bayesian mean = 244.7 mm of fish total length)
were significant, indicating lower final size in fish living in the exploited site IPB (Table
8-3 and Figure 8.5). In the case of T and h, the posterior distributions of the three
populations were highly overlapped discarding the plausibility of any significant
differences (Table 8-3 and Figure 8.5).
4.5 0.9
g (years‐1)
4.0 0.8
L∞ (mm)
3.5 0.7
3.0 0.6
1.7 1.2
1.6
T (years)
1.1
h (mm)
1.5
1.4 1.0
1.3 0.9
1.2
0.8
BT HL BT HL
Figure 8.4 Box-plot of the Bayesian individual means of the life-history parameters (reproduction
investment, infinite size, age of maturation and immature growth) estimated for the beam trawl (BT in
light grey) and hook-and-line (HL in dark grey) individuals.
Individuals sampled in the IPB (high exploited population) also had a significant
higher batch-fecundity than those the individuals sampled in the NPC (GLMM; p
<0.05, Table 8-2). A summary of the GLMMs performed to explore differences in the
population means of relative abundance, fish sizes and ages of vulnerable fish is shown
in Table 8-4, which revealed significant effects in all the response variables. The
population of IPB was characterized by a significantly lower catch rate as indicator of
abundance, and the fish were on average smaller and younger as is typical in exploited
sites. By contrast, vulnerable individuals sampled in the NPC were more abundant,
larger and older, and the fish in the OP site showed average values (Table 8-4). Daily
temperature oscillated in the same way both at a daily and seasonal basis in the IPB and
NPC sites suggesting no differences in temperature regimes as to be expected due the
close proximity between the two areas (Figure 8.6). By contrast, chlorophyll-a
concentrations estimated from the sea-colour images available for the IPB and NPC
showed in both cases a clear seasonal periodicity, and the IPB had higher values
suggesting being a more productive site than the NPC (Figure 8.6). Therefore, one
would have expected larger growth rates in IPB fish, but the opposite was found
Table 8-1 Bayesian posterior distribution (MCMC) of means the life-history traits (reproduction
investment in year -1, infinite size inn mm of otolith, age of maturation in years and immature growth rate
referred to otolith scale in mm year -1) for the beam trawl and the hook-and-line samples.
Beam trawl Hook-and-line
Parameter Mean s.d. 2.5% Median 97.5% Mean s.d. 2.5% Median 97.5%
Reproduction investment 0.88 0.12 0.75 0.86 1.11 0.74 0.06 0.68 0.73 0.81
Infinite size 3.50 0.11 3.28 3.50 3.70 3.84 0.07 3.72 3.84 3.95
Age of maturation 1.33 0.08 1.17 1.33 1.45 1.43 0.05 1.35 1.43 1.49
Immature growth 1.03 0.11 0.90 1.01 1.24 0.94 0.05 0.89 0.93 1.01
8.4 Discussion
Recreational fisheries-induced evolution of “fast life-histories” characterized by high
reproductive investment and small adult size in the small coastal fish S. scriba
constituted the most parsimonious explanation of the observed phenotypic data. Four
lines of evidence support this conclusion. We have first shown that recreational angling
selects for certain life-histories, i.e., those with high reproductive investment and small
terminal size. We have secondly shown that oceanographic currents and local
hydrodynamics prevented gene flow among spatially closed sites. Therefore, two
ingredients of FIE of life histories in the marine environment were met. Accordingly,
we found average life-histories in exploited and unexploited sites to mirror the
directional changes expected from the theoretical selection pressures exerted by anglers.
Finally, the smaller adult growth and larger reproductive investment of fish surviving
the exploitation in high intensity sites could not be explained by envrioenmatlel relates
factors. Our comparative findings therefore joins previous work in freshwater
ecosystems (Drake et al., 1997; Haugen and Vøllestad, 2001; Magnan et al., 2005) that
harvesting can result in FIE of fast life-histories characterized by early maturation, high
investment into reproduction and small adult growth when gene flow is limited at small
spatial scales in marine environments.

Table 8-2 Results of the two Generalized Lineal Mixed effect models (GLMMs) fitted to test the
relationship between the direct measure of reproduction investment (batch fecundity) and (A) the
sampling method (beam trawl and hook-and-line), and (B) the populations (inner of Palma Bay, average
population and National Park of Cabrera). In both models Fish size was also included as fixed factor and
Fishing ID as random factor. The GLMM developed for populations also included the temporal variable
month. The table shows the parameters include in the final model (the MCMC mean of the parameters,
the lower and upper highest 95% credibility intervals (HDP) and the significance of the parameter.
Interactions were not significant and were not included.
Log (batch fecundity) MCMC HPD 95%
Parameter mean lower upper P- value
Selection experiment
Intercept 5.83 5.00 6.63 <0.001
Fish size 0.02 0.01 0.02 <0.001
Method (hook-and-line) -1.31 -1.93 -0.68 <0.05
Populations comparisons
Intercept 5.21 4.67 5.75 <0.001
Fish size 0.02 0.02 0.02 <0.001
Population (inner of Palma Bay) 0.74 0.42 1.08 <0.05
Population (average population) -0.08 -0.38 0.22 0.87
Month (may) -1.04 -1.30 -0.77 <0.001
Our findings agreed with theoretical models on recreational-fishing induced
selection that proposed selection for high reproductive investment and declines in adult
body size of Northen pike (Arlinghaus et al., 2009), and with the empirical findings on
recreationally exploited Atlantic salmon (Saura et al., 2010) and Chinook salmon,
(Kendall and Quinn, 2011). Because it is likely that the life-history of a fish is correlated
to underlying behavioural and physiological traits (Uusi-Heikkilä et al., 2008), it is very
likely that the fish surviving exploitation are also characterized by higher shyness and
less aggressive foraging (Biro and Post, 2008; Sutter et al., 2012). Because the selection
found in our work is expected to drive populations of S. scriba to evolve slower capacity
to growth after maturation due to higher reproductive investment and because the
adapted individuals are expected to also become harder to catch (Philipp et al., 2009) the
resulting outcome is likely to affect fishing communities negatively by affecting
reproductive fitness and catchability negatively (Sutter et al., 2012; Johnson et al.,
2011). In direct contrast to such recreational harvest-based selection, natural selection in
natural environments such as the National Park of Cabrera will drive population in the
opposite direction favouring individuals with large expected life-span to avoid gape-
limited predation and with low reproductive investment due to the trade off with post
maturation growth and to postpone the survival cost of reproduction (Hutchings, 1993;
Reznick, 1982; Aikio et al., 2013). In particular, in environments where the abundance
of predators of S. scriba is high (for example, presence of Ephinephelus marginatus),
directional selection on fast (juvenile and adult) growth is to be expected (Edeline et al.,
2007). Therefore, the optimal life-history strategy should favour a version of “bigger-is-
better” which is characterized by fast early growth, late maturation and moderate to
small annual reproductive investment (Edeline et al., 2007; Hutchings, 1993; Olsen and
Moland, 2011; Reznick, 1982). This is because the (surviving) individual theoretically
would maximize the fitness over a larger the life-span under natural conditions
(Charnov, 2005; Roff, 1984).

Table 8-3 Posterior distribution (MCMC populations means, s.d. and Bayesian credibility intervals) of
the life-history traits (reproduction investment in year -1, infinite size (mm of otlith), age of maturation in
years and immature growth rate referred to otolith scale in mm year -1) estimated for the three analysed
populations (inner of Palma Bay, average population and National Park of Cabrera).
Population Bayesian credibility intervals

LH Parameter Mean s.d. 2.5% Median 97.5%
Inner of Palma Bay
Reproduction investment 0.75 0.04 0.67 0.75 0.83
Infinite size 3.59 0.07 3.45 3.60 3.74
Age of maturation 1.50 0.05 1.40 1.51 1.61
Immature growth 0.89 0.04 0.81 0.89 0.98
Average population
Infinite size 3.85 0.07 3.70 3.85 3.99
Immature growth 0.89 0.04 0.82 0.89 0.98
National Park of Cabrera
Infinite size 3.88 0.06 3.75 3.87 4.00
Immature growth 0.79 0.03 0.72 0.79 0.85
By contrast and in line with our work, heavy exploitation by anglers would expect to
lead to evolution of fast life-histories characterized by early reproduction and large
reproductive investment, which results in penalties on adult growth. While these patters
were supported in our work in relation to reproductive investment and adult growth, no
change in maturation timing was found. However, given that maturation happens in S.
criba usually at age 1 (Alonso-Fernández et al., 2011) there was no room for maturation
evolution in this species. This constraint also likely explains the tendency for exploited
fish to growth fast early in life because it pays off to be as large as possible when first
reproduction starts and investment into the future (by growing to larger sizes) does not
pay off. The same was found in a theoretical model by (Matsumura et al., 2011) in
recreationally exploited northern pike, Exos lucius. When size selectivity favoured heavy
cropping of very young fish already, selection pressures on fast early immature growth
was predicted. The direction of selection changed to negative selection on immature
growth when size-selectivity only targeted large mature fishes, e.g., due to a large
minimum-size limit (Matsumura et al., 2011).
4.4 0.9
4.2 0.8
g (years‐
4
L∞ (mm)
0.7
3.8
3.6 0.6
3.4 0.5
3.2 0.4
1.8 1.1
1.7 1
T (years)
h (mm)
1.6 0.9
1.5
1.4 0.8
1.3 0.7
1.2 0.6
NPC I IPB NPC I IPB
Figura 8.5 Bayesian Credibility Intervals (2.5%, 50%, and 97.5%) of the posterior distribution of the
population’s means estimated for each LH trait to assess with the differences between-populations
One caveat of any comparative study such as the one performed here is to discount
any non-fishing related environmental drivers that might also explain the among-site
differentiation in life-history (Kendall and Quinn, 2011). In fact, all differences
observed in the life-history in our sub-populations could be related to unaccounted
environmental characteristics directly or indirectly related to harvest pressure (Dunlop et
al., 2005; McDermid et al., 2010). While we cannot discount that the phenotypic
change may be entirely plastic, the bulk of evidences presented here makes this
explanation unlikely. First of all, all life-history traits have some modest heritability and
the selection caused by recreational fishing was found pronounced, which should
generate selection differentials on heritable traits (Law, 2000). Secondly, the variation in
fishing pressure among sites was undisputed and the gene flow among the protected
area site and the exploited site was found very small. Thirdly, our control sites (OP)
showed intermediate phenotypes to the high exploitation and non-exploitation site.
Finally and although cold waters may slower pre-maturation growth and result in older
age of maturation and larger longevity in some species (McDermid et al., 2010; Wright
et al., 2011), we could discount temperature to vary substantially among our sites. The
only environmental driver that did vary among the high and no exploitation sites was
the productivity of the system, which was approximated by a 10-years time-series of
chlorophyll-a concentration. Although differences were small, there was a consistently
higher productivity during the peak of annual productivity in spring in the exploited
IPB. Greater resource availabilities should on the one hand increase relative fecundity
due to greater resource intake (e.g., (Craig and Kipling, 1983), which might explain the
greater annual reproduction investment found in the IPB sites. At the same time,
however, growth rate of both immature and mature should profit from higher
productivity in the high exploitation, and maximum size should then increase
(McDermid et al., 2010). Similarly, the reduced density of fish should additionally
elevate growth due to relaxed competition (Lorenzen and Enberg, 2002). By contrast,
our life-history traits in relation to growth directly showed the exact opposite response.
First, the immature growth rates (h) did not differ statistically among sites. Secondly,
the theoretical maximum size was significantly smaller in the high exploitation sites
although the environmental plasticity hypothesis should have predicted larger maximum
sizes inter alia because relaxed intra-specific competition for prey should have increased
the food availability generating fast growth and small reproductive investment (Dunlop
et al., 2005). Therefore, we have no environmental factor that may explain the life-
history difference and thus have limited supports for a plasticity hypothesis, rendering
FIE the more parsimonious mechanism of life-history differentiation in our study. In
particular, the high heritability of the life-history traits studied here (Law, 2000), and
the limited genetic exchange among populations of very likely common ancestors makes
genetic divergence through fisheries-induced adaptive change the most likely
explanation.
Table 8-4 Results of the three Generalized Lineal Mixed effect models (GLMMs) fitted to test the
relationship between the relative abundance (fish*30min*angler), fish size (mm) and ages (years) of
vulnerable Serranus scriba sampled in the inner of Palma Bay, the average population and the National
Park of Cabrera. The full model (without reduction) includes the fixed factors population, depth (m),
angler, bait and month. The abundance GLMM have repeated measures within day and the size and age
were repeated measures of fishing sessions within day which were included as random effects. The table
shows the parameters include in the final model (the MCMC mean of the parameters, the lower and
upper highest 95% credibility intervals (HDP) and the significance of the parameter.
MCMC HPD 95%
Parameter mean lower upper P- value
Log (Abundance + 1)
(Intercept) 1.35 0.86 1.87 <0.001
Population (inner of Palma Bay) -0.64 -1.00 -0.27 <0.01
Population (average population) -0.30 -0.77 0.16 0.200
Depth (m) -0.04 -0.06 -0.01 <0.05
Month (Jan) 0.43 0.03 0.81 0.082
Month (Feb) 0.12 -0.26 0.53 0.588
Month (Mar) 0.16 -0.30 0.61 0.570
Month (May) 0.31 -0.12 0.75 0.170
Month (Jun) 0.52 -0.02 1.03 0.076
Month (Jul) 1.16 0.49 1.84 <0.01
Month (Aug) 0.99 0.53 1.43 <0.001
Month (Oct) 0.75 0.00 1.57 0.073
Month (Nov) 0.91 0.15 1.69 <0.05
Log (Fish size)
(Intercept) 4.83 4.73 4.93 <0.001
Population (average population) -0.10 -0.17 -0.03 <0.01
Bait (Shrimp) 0.08 0.04 0.13 <0.01
Depth (m) 0.01 0.00 0.02 <0.05
Log (Fish age + 1)
(Intercept) 1.33 1.10 1.54 <0.001
Population (average population) -0.46 -0.62 -0.29 <0.001
Bait (Shrimp) 0.14 0.04 0.24 <0.05
Depth (m) 0.03 0.01 0.04 <0.01
Cause-and-effect evidence for evolutionary changes (i.e., genetic) in the wild can
only unequivocally determined by commen garden experiments (Conover and
Baumann, 2009; Enberg et al., 2012). Molecular ecological tools are rapidly emerging
but the link of genotypes to complex phenotypes is not straightforward (Ãrnason et al.,
2009; Jakobsdóttir et al., 2011; Pukk et al., 2013). In this context, comparative field
studies such as ours may provide a useful compromise to study whether fishing
differentially targets certain phenotypes and whether populations with common ancestry
and low current gene flow differentiate in their average phenotype in directions expected
from theory and in line with selection differentials estimated in the wild. Often, the
large effective population size of marine fish has been proposed to constitute barriers to
FIE (Cuveliers et al., 2011; Pukk et al., 2013). We have shown, however, that meta-
populatoins of the small coastal species S. scriba inhabiting Mallorca Island develop
reproductively isolated sub-populations on small spatial scales, which are characterized
by high levels of local self-recruitment and low direct and indirect connectivity among
each other. Those findings are in line with recent literature indicating that contrary to
traditional wisdom that considered marine populations to constitute panmictic systems
(Roberts, 1997), hydrodynamics and life-history may actually also result in fine meta-
population structures with limited gene flow in many marine areas (Cowen et al., 2006;
Mora and Sale, 2002; Schunter et al., 2011). This provides conditions for fisheries-
induced evolution to happen also in coastal species targeted by recreational fisheries on
very small geographical scales.

Figura 8.6 Chronograms of ten years of the daily surface temperature (ºC) and concentration of
chlorophyll-a (mg m-3) in the inner part of Palma Bay (PB), in the open population (OP) and in the
National Park of Cabrera (NPC).
Although we consider our evidence for angling-induced evolutionary change in S.
scriba to be indirect, we cannot disentangle unequivocally whether the phenotypic
change has a genotypic basis. Empirical evidence of FIE in our study species could be
generated using common-garden experiments of fish collected of different sites and
raised and tested in the laboratory. For example, the offspring from the different
populations reared under identical environmental conditions could explain the origin of
the differences of S. scriba in the wild (Conover and Baumann, 2009). Until this
research becomes available, this Thesis proposes a new paradigm of the potential for
widespread FIE of small bodied at small scales in marine coastal fishes. Due the
potential negative effects of fisheries-induced evolution, which may affect recovery,
productivity and catch per unit effort negatively (Sutter et al., 2012; Enberg et al., 2009),
fisheries scientist and mangers may adjust their management polices to prevent or
decrease the evolutionary impacts of harvesting for successful fish stock management.
To do so requires understanding the meta-population structure and avoid sub-
population specific harvesting as much as possible if avoiding FIE is a management
goal.
Chapter 9 A behavioural perspective of fisheries-induced selection
in marine coastal fisheries
Alós, J., Palmer, M., and Arlinghaus, R. 2012. Consistent
Selection towards Low Activity Phenotypes When Catchability
Depends on Encounters among Human Predators and Fish. Plos
One, 7: e48030.
Chapter 10 89
Chapter 10 General conclusions
1. The individual is the optimal level of study understanding life-history traits such as
growth and/or reproduction, since differential mortality (either natural or fishing-
related) happens at such level.
2. The study of the individual growth and reproduction revealed both population-level
patterns such as the mating behaviour, and individual-level patterns such as the broad
between-fish variability on most of the life-history traits or the bi-phasic growth
pattern.
3. Recreational fishing constitutes a significant source of mortality (additional to the
natural mortality). Specifically, it may cause local depletion in highly sedentary
species, such those that inhabit Posidonia oceanica meadows.
4. Fishing mortality results not only from the direct extraction of biomass but also from
indirect morality related with fish captured and released but that fails to survive.
Management measures such as gear modifications or the stipulation of partial Marine
Protected Areas should be useful to reduce recreational fishing mortality.
5. The fishing mortality induced by recreational fishing is not a random but a biased
process in relation to some hereditable life-history traits. Biased harvesting generates
selection differentials against phenotypes with higher growth capability and lower
J Alós (PhD Thesis) Selection drivers of life-history traits in marine coastal fishes 90
reproduction investment in line with the theoretical expectations of optimal life
history.
6. The pelagic nature of the early life stages and its hydrodynamic-dependent dispersal
characteristics can generate sub-populations, even at a small spatial scale (few
kilometres). This process can generate reproductively isolated populations with limited
gen flow.
7. Harvesting should able to induce fast evolution of life-history traits. Selected
phenotypes in isolated sub-populations of coastal fish exposed to the recreational
fishing seems to be characterized by high reproductive investment and small adult size
8. Selection of behavioural traits constitutes an overlooked component of the emerging
concept of human-induced evolution, and it can play a key role in sedentary marine
species determining the odds of encounters between fish and fishers.
9. Empirical evidence of fisheries-induced evolution can be only unequivocally detected
by common-garden experiments. This constitutes an exciting opportunity for further
research. Until this research becomes available, this PhD Thesis proposes a new
paradigm: the evidence, even at small spatial-temporal scales, that fisheries-induced
evolution on coastal sedentary species exploited by recreational fisheries is plausible.
10. Due the potential negative effects of fisheries-induced evolution, which may affect
stock recovery and productivity, fisheries scientist and mangers should adjust
management polices to prevent or buffer the evolutionary impacts of harvesting. This
goal requires better understanding the meta-population structure and identifying
fishing syndromes or potentially correlated selective traits.

Reference list 91
Reference list
Aikio, S., Herczeg, G., Kuparinen, A., Alós, J., Arlinghaus, R., Palmer, M.,
and Merilä, J. 2013. Optimal March, D., and Alvarez, I. 2009.
growth strategies under The influence of type of natural
divergent predation pressure. bait on fish catches and hooking
Journal of Fish Biology, 82: location in a mixed-species
318-331. marine recreational fishery, with
Alonso-Fernández, A., Alós, J., Grau, implications for management.
A., Dominguez-Petit, R., and Fisheries Research, 97: 270-277.
Saborido-Rey, F. 2011. The Use Alós, J., Palmer, M., and Arlinghaus, R.
of Histological Techniques to 2012. Consistent selection
Study the Reproductive Biology towards low activity phenotypes
of the Hermaphroditic when catchability depends on
Mediterranean Fishes Coris julis, encounters among human
Serranus scriba, and Diplodus predators and fish. PLoS ONE,
annularis. Marine and Coastal 7: e48030.
Fisheries, 3: 145-159. Alós, J., Palmer, M., Balle, S., Grau, A.
Alonso-Fernández, A., Vallejo, A. C., M., and Morales-Nin, B. 2010.
Saborido-Rey, F., Murua, H., Individual growth pattern and
and Trippel, E. A. 2009. variability in Serranus scriba: a
Fecundity estimation of Atlantic Bayesian analysis. ICES Journal
cod (Gadus morhua) and of Marine Science, 67: 502-512.
haddock (Melanogrammus Allendorf, F. W., England, P. R.,
aeglefinus) of Georges Bank: Luikart, G., Ritchie, P. A., and
Application of the autodiametric Ryman, N. 2008. Genetic effects
method. Fisheries Research, 99: of harvest on wild animal
47-54.
populations. Trends in Ecology Evolutionary Applications: n/a-

& Evolution, 23: 327-337. n/a.
Allendorf, F. W., and Hard, J. J. 2009. Basterretxea, G., Jordi, A., Catalán, I.
Human-induced evolution A., and Sabatés, A. 2012.
caused by unnatural selection Model-based assessment of
through harvest of wild animals. local-scale fish larval
Proceedings of the National connectivity in a network of
Academy of Sciences of the marine protected areas. Fisheries
United States of America, 106: Oceanography, 21: 291-306.
9987-9994. Bevacqua, D., Capoccioni, F., MeliÃ ,
Arlinghaus, R., Matsumura, S., and P., Vincenzi, S., Pujolar, J. M.,
Dieckmann, U. 2009. De Leo, G. A., and Ciccotti, E.
Quantifying selection 2012. Fishery-Induced Selection
differentials caused by for Slow Somatic Growth in
recreational fishing: European Eel. PLoS ONE, 7:
development of modeling e37622.
framework and application to Biro, P. A., and Post, J. R. 2008. Rapid
reproductive investment in pike depletion of genotypes with fast
(Esox lucius). Evolutionary growth and bold personality
Applications, 2: 335-355. traits from harvested fish
Ãrnason, E., Hernandez, U. B., and populations. Proceedings of the
Kristinsson, K. 2009. Intense National Academy of Sciences
habitat-specific fisheries- of the United States of America,
induced selection at the 105: 2919-2922.
molecular Pan I Locus predicts Browman, H. I., Law, R., and
imminent collapse of a major Marshall, C. T. 2008. The role
Cod fishery. PLoS ONE, 4: of fisheries-induced evolution.
e5529. Science, 320: 47-50.
Audzijonyte, A., Kuparinen, A., and Brunel, T., Ernande, B., Mollet, F. M.,
Fulton, E. A. 2013. How fast is and Rijnsdorp, A. D. 2012.
fisheries-induced evolution? Estimating age at maturation
Quantitative analysis of and energy-based life-history
modelling and empirical studies. traits from individual growth
Reference list 93
trajectories with nonlinear energetic characteristics in

mixed-effects models. migratory adult sockeye salmon?
Oecologia: 1-13. Evolutionary Applications, 2:
Carreras-Carbonell, J., Macpherson, E., 299-311.
and Pascual, M. 2007. High Cooke, S. J., Suski, C. D., Ostrand, K.
self-recruitment levels in a G., Wahl, D. H., and Philipp,
Mediterranean littoral fish D. P. 2007. Physiological and
population revealed by behavioral consequences of
microsatellite markers. Marine long-term artificial selection for
Biology, 151: 719-727. vulnerability to recreational
Casabianca, S., Penna, A., Pecchioli, angling in a teleost fish.
E., Jordi, A., Basterretxea, G., Physiological and Biochemical
and Vernesi, C. 2011. Zoology, 80: 480-490.
Population genetic structure and Cowen, R. K., Paris, C. B., and
connectivity of the harmful Srinivasan, A. 2006. Scaling of
dinoflagellate Alexandrium Connectivity in Marine
minutum in the Mediterranean Populations. Science, 311: 522-
Sea. Proceedings of the Royal 527.
Society B: Biological Sciences, Craig, J. F., and Kipling, C. 1983.
279: 129-138. Reproduction effort versus the
Conover, D. O., and Baumann, H. environment; case histories of
2009. The role of experiments in Windermere perch, Perca
understanding fishery-induced fluviatilis L., and pike, Esox
evolution. Evolutionary lucius L. Journal of Fish Biology,
Applications, 2: 276-290. 22: 713-727.
Conover, D. O., and Munch, S. B. Cuveliers, E. L., Volckaert, F. A. M.,
2002. Sustaining fisheries yields Rijnsdorp, A. D., Larmuseau,
over evolutionary time scales. M. H. D., and Maes, G. E.
Science, 297: 94-96. 2011. Temporal genetic stability
Cooke, S. J., Donaldson, M. R., Hinch, and high effective population
S. G., Crossin, G. T., Patterson, size despite fisheries-induced
D. A., Hanson, K. C., English, life-history trait evolution in the
K. K., et al. 2009. Is fishing North Sea sole. Molecular
selective for physiological and Ecology, 20: 3555-3568.
Charnov, E. L. 2005. Reproductive Dunlop, E. S., Orendorff, J. A., Shuter,

effort is inversely proportional to B. J., Rodd, F. H., and Ridgway,
average adult life span. M. S. 2005. Diet and divergence
Evolutionary Ecology Research, of introduced smallmouth bass
7: 1221-1222. (Micropterus dolomieu)
Charnov, E. L. 2008. Fish growth: populations. Canadian Journal
Bertalanffy k is proportional to of Fisheries and Aquatic
reproductive effort. Sciences, 62: 1720-1732.
Environmental Biology of Edeline, E., Carlson, S. M., Stige, L.
Fishes, 83: 185-187. C., Winfield, I. J., Fletcher, J.
Deudero, S., Morey, G., Frau, A., M., Ben James, J., Haugen, T.
Moranta, J., and Moreno, I. O., et al. 2007. Trait changes in
2008. Temporal trends of littoral a harvested population are
fishes at deep Posidonia driven by a dynamic tug-of-war
oceanica seagrass meadows in a between natural and harvest
temperate coastal zone. Journal selection. Proceedings of the
of Marine Systems, 70: 182-195. National Academy of Sciences
Drake, M. T., Claussen, J. E., Philipp, of the United States of America,
D. P., and Pereira, D. L. 1997. 104: 15799-15804.
A Comparison of Bluegill Ellison, A. M. 2004. Bayesian inference
reproductive strategies and in ecology. Ecology Letters, 7:
growth among lakes with 509-520.
different fishing intensities. Enberg, K., Jørgensen, C., Dunlop, E.
North American Journal of S., Heino, M., and Dieckmann,
Fisheries Management, 17: 496- U. 2009. Implications of
507. fisheries-induced evolution for
Dunlop, E. S., Baskett, M. L., Heino, stock rebuilding and recovery.
M., and Dieckmann, U. 2009. Evolutionary Applications, 2:
Propensity of marine reserves to 394-414.
reduce the evolutionary effects of Enberg, K., Jørgensen, C., Dunlop, E.
fishing in a migratory species. S., Varpe, Ø., Boukal, D. S.,
Evolutionary Applications, 2: Baulier, L., Eliassen, S., et al.
371-393. 2012. Fishing-induced evolution
Reference list 95
of growth: concepts, Ecology, Princeton University

mechanisms and the empirical Press, Princeton and Oxford,.
evidence. Marine Ecology, 33: Haugen, T. O., and Vøllestad, L. A.
1-25. 2001. A century of life-history
Francour, P., Harmelin, J. G., Pollard, evolution in grayling. Genetica,
D., and Sartoretto, S. 2001. A 112-113: 475-491.
review of marine protected areas Heino, M., and Dieckmann, U. 2008.
in the northwestern Detecting Fisheries-Induced
Mediterranean region: siting, Life-History Evolution: An
usage, zonation and Overview of the Reaction-Norm
management. Aquatic Approach. Bulletin of Marine
Conservation: Marine and Science, 83: 69-93.
Freshwater Ecosystems, 11: Heino, M., and Godo, O. R. 2002.
155-188. Fisheries-induced selection
Galarza, J. A., Carreras-Carbonell, J., pressures in the context of
Macpherson, E., Pascual, M., sustainable fisheries. Bulletin of
Roques, S., Turner, G. F., and Marine Science, 70: 639-656.
Rico, C. 2009. The influence of Helser, T. E., and Lai, H. L. 2004. A
oceanographic fronts and early- Bayesian hierarchical meta-
life-history traits on connectivity analysis of fish growth: with an
among littoral fish species. example for North American
Proceedings of the National largemouth bass, Micropterus
Academy of Sciences, 106: salmoides. Ecological
1473-1478. Modelling, 178: 399-416.
Garcia, S. M., Kolding, J., Rice, J., Hutchings, J. A. 1993. Adaptive life
Rochet, M. J., Zhou, S., histories effected by age-specific
Arimoto, T., Beyer, J. E., et al. survival and growth rate.
2012. Reconsidering the Ecology, 74: 673-684.
Consequences of Selective Jakobsdóttir, K. B., Pardoe, H.,
Fisheries. Science, 335: 1045- Magnússon, Á., Björnsson, H.,
1047. Pampoulie, C., Ruzzante, D. E.,
Grimm, V., and Railsback, S. F. 2005. and Marteinsdóttir, G. 2011.
Individual-based Modeling and Historical changes in genotypic
frequencies at the Pantophysin
locus in Atlantic cod (Gadus Evolutionary Applications, 2:

morhua) in Icelandic waters: 356-370.
evidence of fisheries-induced Jørgensen, C., and Fiksen, O. 2009.
selection? Evolutionary Modelling fishing-induced
Applications, 4: 562-573. adaptations and consequences
Johnson, D. W., Christie, M. R., Moye, for natural mortality. Canadian
J., and Hixon, M. A. 2011. Journal of Fisheries and Aquatic
Genetic correlations between Sciences, 67: 1086-1097.
adults and larvae in a marine Jørgensen, C., and Holt, R. E. 2013.
fish: potential effects of fishery Natural mortality: Its ecology,
selection on population how it shapes fish life histories,
replenishment. Evolutionary and why it may be increased by
Applications, 4: 621-633. fishing. Journal of Sea Research,
Jones, G. P., Planes, S., and Thorrold, 75: 8-18.
S. R. 2005. Coral reef fish larvae Kendall, N. W., and Quinn, T. P. 2011.
settle close to home. Current Length and age trends of
Biology, 15: 1314-1318. Chinook salmon in the
Jørgensen, C., Enberg, K., Dunlop, E. Nushagak River, Alaska, related
S., Arlinghaus, R., Boukal, D. to commercial and recreational
S., Brander, K., Ernande, B., et fishery selection and
al. 2007. Managing Evolving exploitation. Transactions of the
Fish Stocks. Science, 318: 1247- American Fisheries Society, 140:
1248. 611-622.
Jørgensen, C., Enberg, K., Dunlop, E. Kuparinen, A., and Merilä, J. 2007.
S., Arlinghaus, R., Boukal, D. Detecting and managing
S., Brander, K., Ernande, B., et fisheries-induced evolution.
al. 2008. Response. Science, Trends in Ecology & Evolution,
320: 48-50. 22: 652-659.
Jørgensen, C., Ernande, B., and Fiksen, Laugen, A. T., Engelhard, G. H.,
Ø. 2009. Size-selective fishing Whitlock, R., Arlinghaus, R.,
gear and life history evolution in Dankel, D. J., Dunlop, E. S.,
the Northeast Arctic cod. Eikeset, A. M., et al. 2012.
Evolutionary impact assessment:
Reference list 97
accounting for evolutionary Macpherson, E., and Raventos, N.

consequences of fishing in an 2006. Relationship between
ecosystem approach to fisheries pelagic larval duration and
management. Fish and geographic distribution of
Fisheries: n/a-n/a. Mediterranean littoral fishes.
Law, R. 2000. Fishing, selection, and Marine Ecology Progress Series,
phenotypic evolution. ICES 327: 257-265.
Journal of Marine Science, 57: Magnan, P., Proulx, R., and Plante, M.
659-668. 2005. Integrating the effects of
Lester, N. P., Shuter, B. J., and fish exploitation and
Abrams, P. A. 2004. interspecific competition into
Interpreting the von Bertalanffy current life history theories: an
model of somatic growth in example with lacustrine brook
fishes: the cost of reproduction. trout (Salvelinus fontinalis)
Proceedings of the Royal Society populations. Canadian Journal
of London Series B-Biological of Fisheries and Aquatic
Sciences, 271: 1625-1631. Sciences, 62: 747-757.
Lewin, W. C., Arlinghaus, R., and March, D., Palmer, M., Alós, J., Grau,
Mehner, T. 2006. Documented A., and Cardona, F. 2010.
and potential biological impacts Short-term residence, home
of recreational fishing: Insights range size and diel patterns of
for management and the painted comber Serranus
conservation. Reviews in scriba in a temperate marine
Fisheries Science, 14: 305-367. reserve. Marine Ecology
Lorenzen, K., and Enberg, K. 2002. Progress Series, 400: 195-206.
Density-dependent growth as a Matsumura, S., Arlinghaus, R., and
key mechanism in the regulation Dieckmann, U. 2011. Assessing
of fish populations: evidence evolutionary consequences of
from among-population size-selective recreational fishing
comparisons. Proceedings of the on multiple life-history traits,
Royal Society of London Series with an application to northern
B-Biological Sciences, 269: 49- pike (Esox lucius). Evolutionary
54. Ecology, 25: 711-735.
McDermid, J. L., Shuter, B. J., and Evolutionary Ecology, 25: 695-

Lester, N. P. 2010. Life history 710.
differences parallel Palkovacs, E. P. 2011. The overfishing
environmental differences debate: an eco-evolutionary
among North American lake perspective. Trends in ecology &
trout (Salvelinus namaycush) evolution (Personal edition), 26:
populations. Canadian Journal 616-617.
of Fisheries and Aquatic Palumbi, S. R., and Warner, R. R.
Sciences, 67: 314-325. 2003. Why Gobies are like
Mollet, F. M., Ernande, B., Brunel, T., hobbits. Science, 299: 51-52.
and Rijnsdorp, A. D. 2010. Pauly, D., Alder, J., Bennett, E.,
Multiple growth-correlated life Christensen, V., Tyedmers, P.,
history traits estimated and Watson, R. 2003. The
simultaneously in individuals. future for fisheries. Science, 302:
Oikos, 119: 10-26. 1359-1361.
Mora, C., and Sale, P. F. 2002. Are Pérez-Rodríguez, A., Morgan, M. J.,
populations of coral reef fish and Saborido-Rey, F. 2009.
open or closed? Trends in Comparison of demographic
Ecology & Evolution, 17: 422- and direct methods to calculate
428. probabilistic maturation reaction
Morales-Nin, B., Moranta, J., Garcia, norms for Flemish Cap cod
C., Tugores, M. P., Grau, A. (Gadus morhua). Evolutionary
M., Riera, F., and Cerda, M. Applications, 2: 291-298.
2005. The recreational fishery Philipp, D. P., Cooke, S. J., Claussen, J.
off Majorca Island (western E., Koppelman, J. B., Suski, C.
Mediterranean): some D., and Burkett, D. P. 2009.
implications for coastal resource Selection for vulnerability to
management. ICES Journal of angling in Largemouth Bass.
Marine Science, 62: 727-739. Transactions of the American
Olsen, E., and Moland, E. 2011. Fisheries Society, 138: 189-199.
Fitness landscape of Atlantic Pilling, G. M., Kirkwood, G. P., and
cod shaped by harvest selection Walker, S. G. 2002. An
and natural selection. improved method for estimating
Reference list 99
individual growth variability in in relation to vulnerability to

fish, and the correlation between angling in an experimentally
von Bertalanffy growth selected teleost fish.
parameters. Canadian Journal of Evolutionary Applications, 2:
Fisheries and Aquatic Sciences, 312-323.
59: 424-432. Redpath, T. D., Cooke, S. J., Suski, C.
Plummer, M., Best, N., Cowles, K., D., Arlinghaus, R., Couture, P.,
and Vines, K. 2006. CODA: Wahl, D. H., and Philipp, D. P.
convergence diagnosis and 2010. The metabolic and
output analysis for MCMC. R biochemical basis of
News, 6: 5. vulnerability to recreational
Pukk, L., Kuparinen, A., Järv, L., angling after three generations
Gross, R., and Vasemägi, A. of angling-induced selection in a
2013. Genetic and life-history teleost fish. Canadian Journal of
changes associated with Fisheries and Aquatic Sciences,
fisheries-induced population 67: 1983-1992.
collapse. Evolutionary Reñones, O., Goñi, R., Pozo, M.,
Applications: n/a-n/a. Deudero, S., and Moranta, J.
Quince, C., Abrams, P. A., Shuter, B. 1999. Effects of protection on
J., and Lester, N. P. 2008a. the demographic structure and
Biphasic growth in fish I: abundance of Epinephelus
Theoretical foundations. Journal marginatus (Lowe, 1834).
of Theoretical Biology, 254: Evidence from Cabrera
197-206. Archipelago National Park
Quince, C., Shuter, B. J., Abrams, P. (West-central Mediterranean).
A., and Lester, N. P. 2008b. Marine Life, 9: 2.
Biphasic growth in fish II: Reznick, D. 1982. The impact of
Empirical assessment. Journal of predation on life history
Theoretical Biology, 254: 207- evolution in Trinidadian
214. Guppies: genetic basis of
Redpath, T. D., Cooke, S. J., observed life history patterns.
Arlinghaus, R., Wahl, D. H., Evolution, 36: 1236-1250.
and Philipp, D. P. 2009. Life- Reznick, D. A., Bryga, H., and Endler,
history traits and energetic status J. A. 1990. Experimentally
induced life-history evolution in Vidal-Vijande, E., Pascual, A.,

a natural population. Nature, Guidetti, P., et al. 2011.
346: 357-359. Matching genetics with
Rijnsdorp, A. 1993. Fisheries as a large- oceanography: directional gene
scale experiment on life-history flow in a Mediterranean fish
evolution: disentangling species. Molecular Ecology, 20:
phenotypic and genetic effects in 5167-5181.
changes in maturation and Sharpe, D. M. T., and Hendry, A. P.
reproduction of North Sea 2009. Life history change in
plaice, (Pleuronectes platessa). commercially exploited fish
Oecologia, 96: 391-401. stocks: an analysis of trends
Roberts, C. M. 1997. Connectivity and across studies. Evolutionary
management of Caribbean coral Applications, 2: 260-275.
reefs. Science, 278: 1454-1457. Sinclair, A. F., Swain, D. P., and
Roff, D. A. 1984. The Evolution of Hanson, J. M. 2002. Measuring
Life History Parameters in changes in the direction and
Teleosts. Canadian Journal of magnitude of size-selective
Fisheries and Aquatic Sciences, mortality in a commercial fish
41: 989-1000. population. Canadian Journal of
Roff, D. A. 1992. The evolution of life Fisheries and Aquatic Sciences,
histories: theory and analysis, 59: 361-371.
Chapman and Hall, New York, Stenseth, N. C., and Dunlop, E. S.
US. 2009. Unnatural selection.
Saura, M., Moran, P., Brotherstone, S., Nature, 457: 803-804.
Caballero, A., Alvarez, J., and Stinchcombe, J. R., and Kirkpatrick, M.
Villanueva, B. 2010. Predictions 2012. Genetics and evolution of
of response to selection caused function-valued traits:
by angling in a wild population understanding environmentally
of Atlantic salmon (Salmo salar). responsive phenotypes. Trends
Freshwater Biology, 55: 923- in Ecology & Evolution, 27:
930. 637-647.
Schunter, C., Carreras-Carbonell, J., Sutter, D. A. H., Suski, C. D., Philipp,
Macpherson, E., Tintoré, J., D. P., Klefoth, T., Wahl, D. H.,
Reference list 101
Kersten, P., Cooke, S. J., et al. McWilliams, J. C., Blanchette,

2012. Recreational fishing C. A., et al. 2011. Currents
selectively captures individuals connecting communities:
with the highest fitness nearshore community similarity
potential. Proceedings of the and ocean circulation. Ecology,
National Academy of Sciences. 92: 1193-1200.
Sutter, D. A. H., Suski, C. D., Philipp, Wilson, A. D. M., Binder, T. R.,
D. P., Klefoth, T., Wahl, D. H., McGrath, K. P., Cooke, S. J.,
Kersten, P., Cooke, S. J., et al. and Godin, J.-G. J. 2011.
in press. Recreational fishing Capture technique and fish
selectively captures individuals personality: angling targets timid
with the highest fitness bluegill sunfish, Lepomis
potential. Proceedings of the macrochirus. Canadian Journal of
National Academy of Sciences. Fisheries and Aquatic Sciences,
Swain, D. P., Sinclair, A. F., and Mark 68: 749-757.
Hanson, J. 2007. Evolutionary Worm, B., Barbier, E. B., Beaumont,
response to size-selective N., Duffy, J. E., Folke, C.,
mortality in an exploited fish Halpern, B. S., Jackson, J. B. C.,
population. Proceedings of the et al. 2006. Impacts of
Royal Society B: Biological biodiversity loss on ocean
Sciences, 274: 1015-1022. ecosystem services. Science, 314:
Taylor, M. S., and Hellberg, M. E. 787-790.
2003. Genetic Evidence for Worm, B., Hilborn, R., Baum, J. K.,
Local Retention of Pelagic Branch, T. A., Collie, J. S.,
Larvae in a Caribbean Reef Costello, C., Fogarty, M. J., et
Fish. Science, 299: 107-109. al. 2009. Rebuilding Global
Uusi-Heikkilä, S., Wolter, C., Klefoth, Fisheries. Science, 325: 578-
T., and Arlinghaus, R. 2008. A 585.
behavioral perspective on Wright, P. J., Millar, C. P., and Gibb,
fishing-induced evolution. F. M. 2011. Intrastock
Trends in Ecology & Evolution, differences in maturation
23: 419-421. schedules of Atlantic cod, Gadus
Watson, J. R., Hays, C. G., Raimondi, morhua. ICES Journal of Marine
P. T., Mitarai, S., Dong, C.,
Science: Journal du Conseil, 68: M. 2009. Mixed Effects Models

1918-1927. and Extensions in Ecology with
Zuur, A. F., Leno, E. N., Walker, N. J., R, New York, USA.
Saveliev, A. A., and Smith, G.
List of Appendix 103
List of Appendix
Appendix I
The conventional, deterministic form of the von Bertalanffy model for the size y of a
fish at age t reads:
yt= L∞* (1- exp(-k* (t- t0)),
where L∞ is the asymptotic size at infinite age, k is the growth parameter, and t0 is a
correction term allowing that the fish has some specific size y0 = L∞* (1-exp(k * t0)) at
time t = 0. Alternatively, t = t0 might be seen as the time when the fish has zero size.
This model corresponds to the solution of the differential equation
dty= k * (L∞- y), subject to y(t0)= 0.
The model can be easily extended to the case of non-constant growth parameter
k(t), allowing for growth changes. In the general case, the size of the fish at age t is
given by
yt= L∞* (1- exp(-Q(t))),
t
where Q(t )   k ( s ) ds . If we consider a single spurt occurring at time t1 > t0, such that
t0
the growth parameter changes from k = k0 for t < t1 to k = k1 for t > t1, then
Q(t)= k0* (t- t0), if t < t1
Q(t)= k0* (t1- t0)+ k1* (t- t1), if t > t1
Hence, the fish size at age t is given by
yt= L∞* (1- exp(-k0* (t- t0))), if t < t1
yt= L∞* (1- exp(-k0* (t1- t0)- k1* (t- t1))), . if t > t1
In addition we have fitted a generalization of this model that allows for a change in the
growth at a specific moment, t1, of the life span:
yij= L∞i* (1- exp(-k0i* (tj- t0i)))+ εij, for tj < t1
yij= L∞i* (1- exp(-k0i *(t1i- t0i)- k1i* (tj- t1i)))+ εij, for tj > t1
where k0i and k1i are the growth parameters before and after the moment of the
change (t1i).
Appendix II
Bayesian implementation of conventional Von Bertalanffy growth model proposed by
Pilling et al. (2002), Helser and Lai (2004), and Helser et al. 2007 imposes structure of
the variance covariance matrix of the model parameters. This means that covariance
between any pair of parameters is explicitly estimated. It is claimed that this procedure
ameliorates parameter estimations.
Here we use a simpler model and covariance is left free (i.e., not explicitly
estimated by the model). Obviously, in the case that the individual values (i.e. for any
single fish) of a given pair of parameters (e.g., L∞i and t1i) are well estimated, covariance
between them will be well estimated also a posteriori. Demonstration of this was carried
out through a simulation experiment. We defined a variance-covariance matrix using
the between-parameter correlation (i.e., parameters obtained using the observed data).
Then, we generated n sets of growth parameters using the mvrnorm() function of the
MASS library of the R-package. This n*5 (i.e., in the case of 5 parameter model)
reflects the known variance-covariance matrix. Then, we generate n growth curves with
theses parameters (adding a proper error term, i.e., the observed one) and with the
observed length-at-age (i.e., the currently observed age). After that, these n curves, for
which we know the individual values of the growth parameters, were submitted to the
Bayesian approach described in the text (see Materials and methods). Finally, the
estimated growth parameters were compared with the known (simulated) values and
they were found to be very similar (r2 ≈ 0.9) and unbiased (slope ≈ 1 and intercept ≈ 0).
This is shown in the next figure:

Appendix III
Field experiment to estimate the buoyancy, or specific gravity, of the eggs of Serranus
scriba during development for dispersal potential
Six adult S. scriba individuals of similar sizes (mean and s.d. 145 ± 15 mm) were
captured on June 2010 in Palma Bay (Fig. 1 of the manuscript) using conventional
hook-and-line recreational fishing gear at depths <20 m and surface water temperatures
of approximately 22°C. Individuals were caught early in the morning and transported
live in tanks for approximately 30 min to aquaculture facilities (LIMIA, Government of
the Balearic Islands). The fish were measured, weighed and placed in a grey cylinder-
conical 400 l polyvinyl chloride tank mounted with flow-through UV-sterilised
seawater. The temperature and salinity of the system was monitored continuously.
In a previous experiment, the fish did not spawn under similar conditions, likely due
the combination of captivity-induced stress and the lack of an appropriate natural
spawning environment (J Alós, unpublished data). Therefore, spawning was induced
through administration of synthetic agonists of gonadotropin-releasing hormone
(GnRHa), which has been successfully used in other species of the Serranidae family,
such as Ephinephelus marginatus (Marino et al., 2003). GnRHa stimulates the pituitary
synthesis and release of LH and induces ovarian steroidogenesis and oocyte maturation
(Mylonas and Zohar, 2000; Zohar and Mylonas, 2001).
Individuals were slightly anesthetised with MS222, and a dose of GnRHa (mean and
s.d. of 0.0018 ± 0.0009 ml per g of fish) was inserted in the peritoneal cavity using a
hypodermic needle. The fish were allowed to spawn undisturbed in the tank, which was
examined for the presence of eggs every hour. After spawning, the tanks were checked
for the presence of fertilised eggs (eggs floating in an intermediate position). Adult
individuals were removed from the holding tank, and the water flow and aeration was
reduced to prevent damage to the eggs. The temperature and salinity remained constant
throughout the experiment (T = 22 ± 0.2°C, salinity = 37.25 ± 0.150).
A sample of approximately 10 eggs was collected every three hours until hatching to
estimate changes in buoyancy through development. Egg buoyancy (specific gravity)
was measured using the density-gradient column method (Coombs, 1981). Briefly, the
method consisted of a 1 m column of water with a stable, constant gradient of seawater
salt. The eggs were introduced at the top of the column and settled to a known density
level (i.e., the location where the egg is in hydrostatic equilibrium with the surrounding
fluid). Glass spheres of a known density were used to calibrate the system (Coombs,
1981; Coombs et al., 1985; Coombs et al., 1990). A temperature-controlled water jacket
maintained the apparatus at a stable temperature, similar to the conditions at the time
of capture.
Each sub-sample of eggs was introduced into the gradient column. To select a stable
density measurement, we recorded the changes in egg density at increasing
observational time points (10, 30, and 60 min). The position of the glass spheres in the
density column (cm) was recorded at each observational time point and converted to a
temperature-corrected specific gravity (g cm-3) (Coombs, 1981). Only fertilized eggs
were selected for the experiment. Dead or moribund eggs, which were characterized by
an opaque appearance or rapid sinking to the bottom, were discarded. The sinking or
ascending velocity (w) at each measured time point was calculated based on Stokes’ law:
gd 2 
w
18
where g is the gravitational force, d is the diameter of the egg, Δρ is the difference
between the density of the environment and the density of the egg and ν is the
molecular viscosity (Sundby, 1997). We were confident that the properties of the egg
remained in agreement with Stokes’ law because the diameter of the egg and differences
in density yielded Reynolds numbers < 0.5 (see Sundby, 1997).
Induced spawning occurred between 21:00 and 21:30 on the first night after the
morning hormone treatment (sunset was at 21:21). We successfully measured the
buoyancy of the eggs over five different time periods prior to hatching. The last
measurement was performed after 29.5 h, and 100% of the eggs in the rearing tank had
hatched 32.5 h after fertilisation, which was the morning of the second day after
fertilisation (Fig. 1 of this appendix). The observed mean and s.d. of the egg diameter
was 799 μm ± 32.7 (n=90). The eggs had a single oil globule and a relatively small
perivitelline space (Figure 1 of this appendix).
The mean specific gravity was at neutral buoyancy at the time of the first
measurement, although the eggs remained lighter than the surrounding water for the
remaining duration of the experiment (Figure 1 of this appendix). At 200 min after
fertilisation, the average egg density was close to a neutral buoyancy (5.47x10-4 g cm-
3), and some of the eggs were becoming heavier at an ambient temperature, which
corresponded mainly to the gastrula stages (Figure 1 of this appendix). The average
excess positive buoyancy at the end of the experiment was 2.07 x 10-3 g cm-3, which
corresponds to an average ascent velocity of 4.19 m h-1 under homogeneous
experimental conditions.
Figure 1 Specific gravity (g cm-3) of Serranus scriba eggs through development at 22°C
water temperature. A local smoother (95% CI, in grey) has been added to aid in the
visualization (function “loess”=local polynomial regression in the library ggplot2 from
the R package). Dashed line depicts the specific gravity of the local seawater at the
culture temperature. Figure also shows three images of the eggs development at the
stages of morula (A, 100 min after fertilization), early (B, 897 min) and late embryo (C,
1,845 min).
A B C
1.027
38.0
1.026 Ambient salinity=37.25 37.5
37.0
T‐corrected specific gravity
1.025 36.5
1.024
1.023
1.022
Min 500 1000 1500

h:m 0:50 10:00 14:11 20:11 02:45
Time after fertilization
Appendix references
Coombs, S. H. 1981. A density-gradient column for determining the specific gravity of
fish eggs, with particular reference to eggs of the mackerel (Scomber scombrus). Marine
Biology, 63: 101-106.
Coombs, S. H., Fosh, C. A., and Keen, M. A. 1985. The Buoyancy and Vertical
Distribution of Eggs of Sprat (Sprattus Sprattus) and Pilchard (Sardina Pilchardus).
Journal of the Marine Biological Association of the United Kingdom, 65: 461-474.
Coombs, S. H., Nichols, J. H., and Fosh, C. A. 1990. Plaice eggs (Pleuronectes platessa
L.) in the southern North Sea: abundance, spawning area, vertical distribution, and
buoyancy. Journal du Conseil: ICES Journal of Marine Science, 47: 133-139.

Marino, G., Panini, E., Longobardi, A., Mandich, A., Finoia, M. G., Zohar, Y., and
Mylonas, C. C. 2003. Induction of ovulation in captive-reared dusky grouper,
Epinephelus marginatus (Lowe, 1834), with a sustained-release GnRHa implant.
Aquaculture, 219: 841-858.
Mylonas, C. C., and Zohar, Y. 2000. Use of GnRHa-delivery systems for the control of
reproduction in fish. Reviews in Fish Biology and Fisheries, 10: 463-491.
Sundby, S. 1997. Turbulence and ichthyoplankton: influence on vertical distributions
and encounter rates. Scientia Marina, 61: 159-176.
Zohar, Y., and Mylonas, C. C. 2001. Endocrine manipulations of spawning in cultured
fish: from hormones to genes. Aquaculture, 197: 99-136.

Appendix IV
Prior distributions used for individuals (i) and populations (j) used for Bayesian
estimations of the individuals life-history estimations. Normal distributions are denoted
as dnorm (mean and variance), uniform distribution as dunif (minimum and maximum
value) and gamma distribution as dgamma (shape and scale). Lmax is referred to otolith
scale in mm, h is in mm year-1, g and k in year -1 and, T, t0 and t1 are in years.
Priors for individuals

for ( i in 1 : n fish)
T[i] dnorm(mT[group[j]],1/sqrt(taut))
k[i] dnorm(mk[group[j]],1/sqrt(tauk))
Lmax[i] dnorm(mL∞[group[j]],1/sqrt(tauL∞))
t0[i] dnorm(mt0[group[j]],1/sqrt(taut0))
g[i] 3*(exp(k[i]) - 1)
h[i] L∞[i]*(exp(k[i]) - 1)
t1[i] T[i] - (1 - exp(-k[i]*(T[i] - t0[i])))/(exp(k[i]) - 1)
Prior for groups (sampling method)
for( j in 1 : n groups)
mT[j] dnorm(2,1/sqrt(0.5))
mk[j] dunif(0,1)
mLmax[j] dnorm(5, 1/sqrt(0.001))
mt0[j] dunif(-2,1)
mg[j] 3*(exp(mk[j]) - 1)
mh[j] mL∞[j]*(exp(mk[j]) - 1)
mt1[j] mT[j] - (1 - exp(-mk[j]*(mT[j] - mt0[j])))/(exp(mk[j]) - 1)
Prior for variances
tau dgamma (0.001, 0.001)
tauT dgamma (0.001, 0.001)
tauk dgamma (0.001, 0.001)
tauLmax dgamma (0.001, 0.001)
taut0 dgamma (0.001, 0.001)
Appendix V
The feasibility of the Bayesian approach estimating individual values of life-history
traits (LH) was tested via simulation. We generated simulated data (length-at-age data)
for one individual with known LHs (Lmax , g, T and h). A Gaussian error (reasonable
otolith reading error) was included to the simulated data using a distribution of mean 0
and standard deviation 0.1 (see Figure 1 of this Appendix). Then the individual values
of LH traits were estimated using the model parameterization described in the text (see
M&M). The distributions of the Bayesian means estimated were compared with the
true values. 5
Figure 2 Individual trajectory of one individual growing

Fish size in otolith scale (mm)
following the somatic growth derived by Lester el al.

4
(2004) (solid line). Grey dashed line shows the timing of

maturation (in that case T = 1.8). Dashed black lines
3
after maturation (T) is the projection considering

immature lineal growth, while before T is the projection
2
of the VB growth curve exclusively fitted with the

mature length‐at‐age data. Solid circle are the back‐
1
calculated length‐at‐age data at otolith scale (mm).

0
0 5 10 15
Age (years)
Simulation exercise demonstrated high feasibility of the Bayesian framework
estimating individual life-history traits Lmax , g, h and T (Table 1 of this appendix).
Values estimated trough MCMC from the simulated individual (with known
parameters) were very close to the true values (Table 1 of this Appendix). Credibility
intervals are relatively small and unbiased for all the parameters (Table 1 of this
Appendix).
Table 1 True and estimated (MCMC) values obtained in the simulation exercise done to check
the performance of the Bayesian framework used in this study (Fig 2). Estimated values are
the Bayesian mean and standard deviation as well as Bayesian credibility intervals. Estimated
values are very close to the true values showing the robustness of the approach.
True values Estimated values (MCMC)
Parameter Bayesian credibility intervals
value Mean SD 2.5% Median 97.5%
Lmax 5 5.01 0.09 4.82 5.01 5.19
T 1.8 1.79 0.20 1.35 1.81 2.14
g 0.6 0.62 0.06 0.53 0.61 0.77
h 1 1.04 0.09 0.92 1.02 1.26
[Blank page]

Tjac 1 de 1

Uploaded by

Copyright:

Available Formats

Tjac 1 de 1

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tjac 1 de 1

Uploaded by

Copyright:

Available Formats

Mediterranean Institute for Advanced Studies, IMEDEA (CSIC-UIB)

Department of Ecology and Marine Resources

Selection drivers of life-history traits

A Thesis submitted for the degree of Doctor of Philosophy

University of the Balearic Islands, April 2013

Fred W. Allendorf, 2009

Josep Alós Crespí

Dr. Miquel Palmer Vidal

Dra. Beatriz Morales Nin

Dr. Gabriel Moyà Niell

Approved for the University Committee on graduate studies

populations could be the outcome of fishing-induced selection in marine coastal

Selection operates at the individual level. Therefore, the existence of inter-individual

novel methodological and theoretical approaches based on the use of calcareous

structures of fishes (otoliths), robust statistical tools and histological microscopy

offering tools to decrease the post-mortality levels.

The major achievement of this Thesis is the empirical demonstration of a fishing-

and lower reproductive investment, in accordance with life-history theory. The

populations subjected to different exploitation pressures and different levels of

This Thesis concludes by hypothesising on the outcomes of fishing-induced selection

from a behavioural perspective. A simulation analysis based on a spatially explicit

behavioural and physiological traits, thus exposing a wide range of fascinating

hypotheses to be tested in further research.

negative ecological and evolutionary effects of fishing.

Los humanos hemos recolectado animales de manera no estocástica (azar) desde el

origen de nuestra especie causando diferentes efectos ecológicos y evolutivos en las

ser fuente de selección por pesca en pesquerías marinas costeras.

necesaria la existencia de variabilidad individual dentro de las poblaciones. Como primer

de selección: crecimiento y reproducción. La variabilidad salvaje de estos rasgos se

determinó utilizando novedosas aproximaciones basadas en el uso de estructuras

calcáreas (otolitos), herramientas estadísticas robustas y técnicas de histología

microscópica en Serranus scriba.

una vez pescados. La tasa potencial de supervivencia de individuos retornados al medio

Tesis también proporciona recomendaciones para la explotación sostenible a través de

herramientas para maximizar las tasas de supervivencia.

El mayor logro de esta Tesis es la demostración empírica de un proceso de selección

de la teoría clásica de historia vital. La teórica y empírica respuesta de esta selección ha

sido estudiada en diferentes poblaciones salvajes expuestas a diferentes niveles de

explotación y conectividad genética entre poblaciones. Los individuos que habitan

heredable que expuesto a un selección constante puede desencadenar un proceso

evolutivo. La posibilidad de este hecho se discute en el marco de la explotación

sostenible de los recursos pesqueros.

La presente Tesis termina considerando una perspectiva de la selección inducida por

pesca en peces marinos costeros mas relacionada con el comportamiento. Se ha realizado

un ejercicio de simulación usando un modelo basado en el individuo espacialmente

potencialmente correlacionados que incluyen rasgos de historia vital, comportamiento y

fisiología generando multitud de hipótesis para ser testadas en investigaciones futuras.

base ecológica y evolutiva causados por la pesca.

Miquel i Beatriu (IMEDEA), Gràcies per donar-me la oportunitat de realitzar

una excel·lent oportunitat per millorar la meva iniciativa, el meu pensament

independent, les meves qualitats de recerca i humanes y el treball en un grup. He gaudit

Personal Investigador (FPI) (ref: BES-2009-013407) emmarcada en els projecte

d’investigació CONFLICT (ref: CGL2008-00958). Diferents parts de la Tesis han

d’Economia i Competitivitat del Govern Espanyol.

el vostre temps explicant-me tècniques pròpies de la vostre línia d’investigació. Ignasi,

de desenvolupament vital del peixos, és un món apassionant. A tots, gràcies i desitjo

un caire de multi-disciplinitat a la meva Tesis.

nostre activitat d’oci durant molts d’anys.

was awesome and I will never forget it.