MASARYKOVA UNIVERZITA PŘÍRODOVĚDECKÁ FAKULTA ÚSTAV BOTANIKY A ZOOLOGIE

AKADEMIE VĚD ČR ÚSTAV BIOLOGIE OBRATLOVCŮ, V.V.I.

Personality, reprodukční strategie a pohlavní výběr u vybraných taxonů ryb

Disertační práce

Radomil Řežucha

ŠKOLITEL: doc. RNDr. MARTIN REICHARD, Ph.D. BRNO 2014

Bibliografický záznam

Autor: Mgr. Radomil Řežucha Přírodovědecká fakulta, Masarykova univerzita Ústav botaniky a zoologie

Název práce: Personality, reprodukční strategie a pohlavní výběr u vybraných taxonů ryb

Studijní program: Biologie

Studijní obor: Zoologie

Školitel: doc. RNDr. Martin Reichard, Ph.D. Akademie věd ČR Ústav biologie obratlovců, v.v.i.

Akademický rok: 2013/2014

Počet stran: 139

Klíčová slova: Pohlavní výběr, alternativní rozmnožovací takti- ky, osobnostní znaky, sociální prostředí, zkuše- nost, Rhodeus amarus, Poecilia wingei

Bibliographic Entry

Author: Mgr. Radomil Řežucha Faculty of Science, Masaryk University Department of Botany and Zoology

Title of Dissertation: Personalities, reproductive tactics and sexual selection in fishes

Degree Programme: Biology

Field of Study: Zoology

Supervisor doc. RNDr. Martin Reichard, Ph.D. Academy of Sciences of the Czech Republic Institute of Vertebrate Biology, v.v.i.

Academic Year: 2013/2014

Number of pages: 139

Keywords: Sexual selection, alternative mating tactics, per- sonality traits, social environment, experience, Rhodeus amarus, Poecilia wingei

Abstrakt

Vliv osobnostních znaků na alternativní reprodukční taktiky (charakteris- tické typy reprodukčního chování) patří mezi zanedbávané oblasti studia po- hlavního výběru. Současně bývá opomíjen i vliv sociálního prostředí a zkuše- nosti na tyto taktiky, a studium schopnosti jedinců v průběhu námluv mas- kovat své morfologické nedostatky. Jako studovaný systém alternativních rozmnožovacích taktik byl zvolen v přírodě nejběžnější komplex – sneaker × guarder (courter) komplex, popisující teritoriální a neteritoriální role samců. U hořavky duhové (Rhodeus amarus) byly potvrzeny osobnostní rozdíly mezi jedinci. Preference využívání konkrétní alternativní reprodukční taktiky jed- notlivých samců však u tohoto druhu nebyla ovlivněna žádným ze zkouma- ných osobnostních znaků (odvahou či agresivitou), ale vznikala zřejmě jako důsledek působení vlivu okolních environmentálních a sociálních faktorů s ohledem na morfologické znaky jedince. Nepotvrdila se ani přítomnost beha- viorálního syndromu (souboru korelovaných rysů chování). V experimentech s živorodkou Endlerovou (Poecilia wingei) se ukázalo, že i mladí nezkušení je- dinci preferují jednu z taktik a již při prvním setkání se samicí se přednostně pokouší buď o násilné páření (sneaking) nebo se samicím dvoří (courting). Re- produkční chování těchto jedinců bylo následně silně ovlivněno jejich sociál- ním prostředím. Své reprodukční chování, modifikované sociálním prostředím, si navíc přenášeli dále do budoucích setkání s potenciálními partnerkami. Je- dinci, kteří žili v prostředí bez samčího rivala, navíc vykazovali celkový pokles sexuální aktivity a relativně zvýšenou míru dvoření v kontrastu s pokusy o nedobrovolné páření. Naopak jedinci vystaveni konkurenci jiného samce se snažili o kopulaci pomocí sneakingu. Vliv sociálního prostředí na výběr alter- nativní taktiky byl velmi silný a potvrdilo se, že časově méně náročný sneaking bývá výhodný v konkurenčních prostředích, kde musí samec soupeřit často i se zdatnějším konkurentem. Dále se potvrdila schopnost samců P. wingei identifikovat v průběhu námluv svou pro samici atraktivnější stranu, a tu jí následně přednostně předvádět v průběhu dvoření. Tento efekt však trval pou- ze v konkurenčním prostředí. V konkurenceprostém prostředí tento trend vy- mizel. To vše ukazuje na silný vliv sociálního prostředí, konkurence a indivi- duální zkušenosti na epigamní chování jedinců.

Abstract

Research on the roles that personality characteristics play in the choice of alternative mating tactic (discrete alternative ways how individuals maxim- ize their reproductive success) are largely neglected area of sexual selection research. The influence of social environment and individual experience on these tactics are equally overlooked. Finally, the ability of individuals to com- pensate their morphologic deficiencies during courtship also remains poorly understood. The study system of alternative mating tactics was sneaker × guarder (courter) complex – the most common complex in nature consisting of territorial and non-territorial males. Distinct behavioural differences be- tween males of the European bitterling (Rhodeus amarus), my first model spe- cies, were confirmed. However, individual choice of alternative mating tactics was not influenced by any personality trait studied (boldness, aggressiveness). Instead, surrounding environmental and social factors coupled with morpho- logic characteristics of an individual affected the preferential use of the alter- native mating tactic. I also did not confirm presence of correlated suite of be- haviours (behavioural syndrome). In the next study, I investigated that even sexually naïve males of the Endler’s guppy (Poecilia wingei) had their preferred mating tactic (sneaking × courting) during their first encounter with a female. Behaviour of these males was later influenced by their current social envi- ronment, with consequences for their mating behaviour in future encounters with females. Individuals kept in the environment without a male rival (fe- male–biased, non–competitive) decreased their overall sexual interest in fe- males, but tended to use courtship displays rather than sneaking. On the other hand, individuals housed with a male rival (male–biased, competitive environment) used sneaking preferentially. The effect of social environment on the choice of alternative mating tactic was therefore very strong. I suggest that sneaking, being less time consuming, is more effective in competitive envi- ronment where an individual has to compete with a rival. Finally, I confirmed individual’s ability to identify more colourful side and use it preferentially dur- ing courtship. Notably, this strategic exploitation of lateral asymmetry was apparent only in male–biased (competitive) environment but not in female– biased environment. These results highlight the influence of current social en- vironment, competition and past experience on male sexual behaviour.

© Radomil Řežucha, Masarykova univerzita, 2014

7 Obsah

PODĚKOVÁNÍ ...... 8

CÍLE PRÁCE...... 9

ÚVOD ...... 10

POHLAVNÍ VÝBĚR A MAXIMALIZACE REPRODUKČNÍ ÚSPĚŠNOSTI ...... 11

PERSONALITY A BEHAVIORÁLNÍ SYNDROMY ...... 13

ALTERNATIVNÍ REPRODUKČNÍ TAKTIKY ...... 14

ZKUŠENOST A SOCIÁLNÍ PROSTŘEDÍ ...... 15

POUŢITÉ DRUHY A METODIKA ...... 17

HOŘAVKA DUHOVÁ (RHODEUS AMARUS) ...... 17 Metodika ...... 19

ŢIVORODKA ENDLEROVA (POECILIA WINGEI) ...... 20 Metodika ...... 22

IMPAKTOVANÉ PUBLIKACE ...... 25

ČLÁNEK 1 ...... 25

ČLÁNEK 2 ...... 49

ČLÁNEK 3 ...... 58

ZÁVĚR ...... 89

IDENTIFIKACE POTENCIÁLNĚ ZAJÍMAVÝCH TÉMAT PRO DALŠÍ VÝZKUM ...... 91

POUŢITÉ ZDROJE ...... 97 NAVŠTÍVENÉ KONFERENCE S VLASTNÍM PŘÍSPĚVKEM ...... 106 PŘÍLOHA – PUBLIKACE NESOUVISEJÍCÍ S DISERTACÍ ...... 109

ČLÁNEK A ...... 109

ČLÁNEK B ...... 122

ÚVOD

8

Poděkování

Na tomto místě bych chtěl poděkovat všem, kteří mě po celou dobu stu- dia podporovali a nějakým způsobem přispěli ke vzniku této práce. Přede- vším a hlavně chci poděkovat Martinu Reichardovi za jeho skvělé vedení, které trvalo přes 7 let. Děkuji mu za to, ţe to celou tu dobu vydrţel, i kdyţ to se mnou neměl jednoduché a nezlomil nade mnou hůl, ani kdyţ ze mě vypa- dla nějaká opravdu velká hloupost. Dále mu chci poděkovat za podporu při mých prvních krůčcích ve světě velké vědy a také za to, ţe jsem díky němu mohl své studium nazývat i svým zaměstnáním. Toho jsem si celou dobu ne- smírně váţil. Dále chci poděkovat Milanu Vrtílkovi za pomoc při černé práci, kritické komentáře, spoustu podnětných myšlenek a především jeho těţko popsatelnou objektivitu, která někdy ani nebyla na místě. Nikdy by mě ne- napadlo, ţe můţe existovat někdo aţ tak objektivní. Děkuji také Mateji Pola- čikovi za jeho pomoc, cenné rady zkušenějšího kolegy, jeho skepsi, realismus jdoucí ruku v ruce s pesimismem, a také za upřímné diskuze o podstatě a pozadí věcí. Především diskuze o pohlavním výběru u ţivorodek byla velmi podnětná. Jeho názory a pohled na ţivot pro mě byly oporou. Radimu Blaţ- kovi děkuji za pomoc při experimentech, rady, a také neskutečnou spoustu informací kterými mě celou tu dobu pravidelně zásoboval, stejně jako i za fotky gupek. Radku Michalkovi děkuji za četné kuloární diskuze o vědě a je- jím pozadí, a také za jeho názory a pochybovačnost. Jiřímu Farkačovi děkuji za pomoc s péčí o ryby a za akvarijní ţivočichy a informace které jsem od něj získal. Dále děkuji všem ostatním spolupracovníkům a kolegům z ÚBO za vytvoření přátelské a kolegiální atmosféry. Josefu Hufovi děkuji za revizi mých jazykových nedostatků. Nakonec bych chtěl poděkovat Lubomíru Hláskovi za laskavý souhlas s pouţitím jeho fotografií v této práci. Speciální dík patří všem těm asi padesáti anonymním oponentům, recenzentům a edi- torům, kteří strávili čas nad našimi články, a svými pozornými a detailními komentáři a starostí o spokojenost ryb velkou měrou přispěli k jejich kvalitě. Pokud jsem zde někoho důleţitého opomenul, tak se tak stalo zcela neú- myslně a omlouvám se mu tímto za to. Práce na této disertaci byla průběţně finančně podporována prostřednictvím projektů GA ČR 206/09/1163, GA

ÚVOD

9

ČR P506/11/0112 a GA ČR P505/12/G112. Dále také GA ČR 526/09/H025 a MUNI/A/0888/2013.

Cíle práce

 zkoumat souvislost mezi osobnostními charakteristikami a alternativ- ním reprodukčním chováním, které jedinec upřednostňuje při maxi- malizaci své reprodukční úspěšnosti

 testovat vliv demografických parametrů, zkušenosti a variability znaků na směr a sílu selekce znaků ovlivňovaných pohlavním výběrem

 sledovat roli sociálního prostředí a individuální zkušenosti ve stabilní koexistenci alternativního reprodukčního chování samců

ÚVOD

10

Úvod

Studium osobnostních rysů u zvířat se dostává do popředí zájmu beha- viorálních ekologů. Tento trend můţeme pozorovat jiţ několik let. Pohlavní výběr je naproti tomu studován jiţ od samého vzniku behaviorální ekologie jako vědecké displíny. Současná popularita a zájem o zvířecí osobnosti re- centně vyvrcholily i vydáním obsáhlé knihy Animal personalities: behavior, physiology, and evolution (Carere & Maestripieri 2013), která shrnuje a uce- luje vše podstatné, co je o zvířecích personalitách prozatím známo. Spojitost osobnostních rysů jedinců, tedy to, jak se většinou chovají v určitých situa- cích, a pohlavního výběru, tedy jejich vliv na volbu partnera, je však aţ na výjimky (Godin & Dugatkin 1996; Schuett et al. 2011; Ariyomo & Watt 2013; Kelley et al. 2013; Dzieweczynski et al. 2014) studována spíše zřídka (Schu- ett et al. 2010). Dokládá to i fakt, ţe i ve výše uvedené knize není problemati- ce pohlavního výběru a personalit věnována ţádná ucelená kapitola a celá problematika je probírána jen na několika málo stranách této publikace. Na- víc zde autoři nabízejí spíše více otázek neţ odpovědí. Jeden z autorů knihy a také nejvýznamnější výzkumník zabývající se zvířecími osobnostmi, An- drew Sih, poukazuje také na závaţný nedostatek prací, které by se zabývaly spojitostí konkrétních behaviorálních typů a alternativních párovacích tak- tik, nebo tím, zda jsou alternativní taktiky součástí nějakého behaviorálního syndromu (souboru navzájem korelovaných chování) (Sih 2013). Ve své di- sertační práci jsem se také pro výše zmíněný nedostatek zaměřil na obě tyto oblasti a pokusil se studovat jejich vzájemný vztah. Téměř ţádné chování jedince se neděje v izolaci, zejména u silně spole- čenských druhů, ale je ve velké míře ovlivněno i ostatními jedinci – sociálním prostředím. Jedinci se nemohou chovat jen podle svých potřeb nebo schop- ností, ale jsou nuceni se do značné míry chovat tak, jak jim dovolí okolí, kte- ré je obklopuje (Bergmüller & Taborsky 2007). Proto jsem se také zaměřil na vliv sociálního prostředí, které obklopuje jedince v průběhu jeho ontogeneze, na alternativní párovací taktiky a epigamní chování obecně. Vliv sociálního prostředí a s ním související sbírání sociálních a sexuálních zkušeností v so- bě neodmyslitelně zahrnuje zpětnou vazbu v podobě samičí reakce na sam-

ÚVOD

11 covo chování. Proto byl za účelem komplexního pochopení systému osob- nostních znaků a rozmnoţovacího chování rovněţ zkoumán vliv samic na utváření a modulaci alternativního reprodukčního chování samců. Důraz byl kladen jak na znaky behaviorální, tak i morfologické a jejich vzájemnou in- terakci, jelikoţ oba druhy znaků jedinec uplatňuje v průběhu námluv sou- časně. V případě morfologických znaků mě především zajímalo, zda dokáţe jedinec kompenzovat svou nedokonalost v podobě různě zbarvených stran svého těla (vysoká fluktuační asymetrie) a zda dokáţe na tento stav nějak re- agovat v průběhu dvoření.

Pohlavní výběr a maximalizace reprodukční úspěšnosti

Po dlouhou dobu, která uplynula od vyslovení základů teorie pohlavní- ho výběru jejím autorem Charlesem Darwinem v 19. století (Darwin 1859), se vědci zaměřovali na studium obrovského mnoţství aspektů, které pohlav- ní výběr jako teorie pojímá. Základní rozdělení pohlavního výběru na soupe- ření o partnery s jedinci stejného pohlaví (intrasexuální výběr) a snahu o za- ujetí partnerů opačného pohlaví (intersexuální výběr) dále určuje dvě hlavní skupiny zájmu. V rámci intrasexuálního výběru se pozornost výzkumníků zaměřovala především na struktury či charakteristiky určené buď k boji s protivníky (kly, parohy), nebo na znaky slouţící pouze k zastrašování kon- kurentů (zbarvení, pach, akustické projevy). Podobné znaky se ale také uplatňují i při výběru intersexuálním (Andersson 1994). Existuje nespočet prací zabývajících se problematikou, které znaky jsou samicemi preferovány a z jakého důvodu. O vysvětlení evoluce těchto znaků a s nimi spojených samičích preferencí se snaţí dokonce hned několik hypotéz (např. Fisher 1930; Zahavi 1975; Ryan 1998). Kromě zjevných vnějších znaků mohou samice také volit samce na zá- kladě znaků, které nejsou na první pohled patrné, např. na základě genetic- ké kompatibility. Nemusí tedy pro ně být nejvýhodnější páření se samcem, který je díky svým vlastnostem preferován většinou samic, ale se samcem, který je nejvýhodnější pro ně samotné díky tzv. kompatibilním genům (Neff & Pitcher 2005). Ty mohou v kombinaci s geny samice zajistit např. rychlejší

ÚVOD

12 růst potomstva a jeho vyšší ţivotaschopnost. Uvádí se, ţe do této skupiny patří tzv. MHC geny (major histocompatibility complex, soubor genů zodpo- vědných za adaptivní imunitní reakci) a samice na základě olfaktorických signálů (Agbali et al. 2010) přednostně volí samce, který má tyto geny odliš- né od těch jejích (Reichard et al. 2012). Hlavním důvodem výběru partnera jsou u samic v kaţdém případě vý- hody pro samici samotnou či pro její potomstvo. Tyto výhody mohou být pro samici přímé, které ovlivní ji samotnou, nikoliv další generaci, např. nutriční dar, lepší kvalita teritoria pro rozmnoţování či větší počet oplodněných vají- ček (Møller & Jennions 2001), nebo mohou být nepřímé, rovnající se větší genetické kvalitě jejího potomstva, tedy další generace (Jennions & Petrie 2000). Ukazatelů, kterými se samičky při volbě partnera řídí, je velké mnoţ- ství. Jejich studium se aţ donedávna zaměřovalo především na znaky akus- tické, pachové a vizuální (Andersson 1994). Tyto ukazatele jsou také silně specifické pro daný taxon. Například u ryb se studium těchto ukazatelů sou- středí především na vizuální signály (Amundsen 2003). V současné době se však začíná přikládat význam i znakům více abstraktním, jako jsou kogni- tivní schopnosti (Boogert et al. 2011) či osobnostní znaky (Schuett et al. 2010). Jedním z nejlépe prostudovaných modelových organizmů, co se týče funkce vizuálních signálů v epigamním chování, je dozajista ţivorodka duho- vá (Poecilia reticulata). O důleţitosti různých komponentů samčího zbarvení pro samičí volbu toho jiţ bylo napsáno mnoho (Houde 1997). Nejvíce se asi z hlediska samičí volby u P. reticulata diskutuje důleţitost oranţového zbar- vení na tělech a ploutvích samců. Jeho důleţitost byla potvrzena napříč po- pulacemi a má se všeobecně za to, ţe oranţovější samci jsou samicemi prefe- rováni. Nutno ale říci, ţe tato samičí preference nemusí být přítomna vţdy a jednotlivé populace se ve svých preferencích a jejich síle mohou značně li- šit (Endler & Houde 1995). Existují tak i populace, které preferují např. čer- nou, nebo dokonce modrou barvu (Sathyan & Couldridge 2013). Pokud je tedy samec více oranţový v populaci, kde je oranţová barva preferována, dá- vá mu to v intersexuálním výběru výhodu a bude samicemi pravděpodobně vybírán častěji. Mimo jiné také proto, ţe karotenoidy (jeţ jsou základem

ÚVOD

13 oranţového zbarvení samců těchto ţivorodek) slouţí jako indikátory samcovy kondice a schopnosti získávat potravu (Olson & Owens 1998), jelikoţ jinak neţ v potravě nemohou být u obratlovců získány (Latscha 1990). Bylo také dokázáno, ţe rozsah oranţového zbarvení můţe přímo indikovat samcovu plodnost (Locatello et al. 2006). Zůstává ale otázkou, jaké mají moţnosti je- dinci, kteří jsou zbarveni méně pestře. Dokáţí nějak operativně maskovat své nedostatky, a tím je efektivně kompenzovat? Téma behaviorálního masková- ní morfologických nedostatků však bývá studováno jen zřídka, i kdyţ vý- zkum laterality (předností vyuţívání jedné části těla před druhou) pokrývá velké mnoţství taxonů včetně ryb (Bisazza et al. 1998).

Personality a behaviorální syndromy

V současné době se u ryb (a nejen u nich) začíná zdůrazňovat také stu- dium osobnostních rysů, které utvářejí personalitu jedinců. Z výsledků zá- kladního výzkumu rybích personalit nyní čerpají dokonce i publikace zabý- vající se lidskou behaviorální psychologií (např. Cain 2012). Vlastní persona- litou jedince se rozumí konzistence jeho chování skrze čas a kontext. Ne- znamená to však, ţe se jeho chování nemůţe měnit, ale pouze to, ţe budou existovat inter- individuální rozdíly (Réale et al. 2007). Např. určitý jedinec bude v určitém časovém úseku průměrně více agresivní nebo odváţný neţ jedinec jiný. To, jak se jedinec chová v určitých situacích, má vliv např. na jeho přeţívání (Réale et al. 2009), prozkoumávání nových území a šíření (Co- te et al. 2010), mnoţství získané potravy (Harfmann Short & Petren 2008) a v neposlední řadě i na jeho reprodukční úspěšnost (Dingemanse et al. 2005; Reaney & Backwell 2007; Patterson & Schulte-Hostedde 2011; Colléter & Brown 2011). Nejčastěji zkoumaným osobnostním znakem bývá u zvířat, a především u ryb, jejich odvaha nebo naopak plachost (Budaev & Brown 2011). Jinými slovy se zkoumá jejich umístění na ose plachost – od- vaha (angl. shy – bold axis). Odvahou se pak rozumí ochota podstoupit riziko či prozkoumávat nové prostředí (Wilson et al. 1993, 1994). Tento hlavní osobnostní rys byl zdokumentován u široké škály druhů a má přímý vliv na fitness jedinců. U ryb se většinou stanovuje tzv. standard emergence testem,

ÚVOD

14 kdy se měří doba potřebná do opuštění bezpečného úkrytu (Brown et al. 2005), či méně častěji reakcí na neznámý předmět (novel object test) nebo sledováním chování jedince v neznámém prostředí (open field test) (Burns 2008). Výsledky emergence testu se zdají být dostatečně vypovídající o stup- ni skutečné odvahy jedince, a to dokonce i napříč různými taxony (Wilson & Godin 2009). Současně se však začíná prosazovat místo zkoumání osobnostních znaků odděleně jejich zkoumání coby souborů znaků, tzv. behaviorálních syndro- mů (Sih et al. 2004). Behaviorální syndromy jsou navzájem korelovaná cho- vání skrze čas a kontext. Prvním takovým zkoumaným syndromem byl sou- bor odvaha – agresivita. Agresivnější jedinci koljušky tříostné (Gasterosteus aculeatus) byli zároveň i odváţnější vůči predátorovi (Huntingford 1976). Kromě agresivity bývá ještě často s odvahou korelována aktivita (Sih et al. 2004). Spojení takových zdánlivě nesouvisejících chování můţe v počátku působit nelogicky, či dokonce maladaptivně, ale v důsledku můţe mít dale- kosáhlé ekologické či evoluční dopady (Sih et al. 2012; Wolf & Weissing 2012). Není ovšem zcela jasné, zda nějaké osobnostní znaky nebo soubory znaků, tzv. behaviorální syndromy, mohou mít přímý vliv na důleţitý pro- středek pohlavního výběru s nezanedbatelnými ekologickými a evolučními důsledky – alternativní reprodukční taktiky (ART) (Carere & Maestripieri 2013).

Alternativní reprodukční taktiky

ART jsou flexibilní rozdíly v chování jedinců, kterými se samci uvnitř druhu či populace snaţí maximalizovat svůj reprodukční úspěch během pá- ření či oplodňování gamet u druhů s vnějším oplozením (Oliveira et al. 2008). ART slouţí především kompetičně méně schopným jedincům k tomu, aby zvýšili svůj reprodukční úspěch v probíhající kompetici s více zdatnými kon- kurenty. Mohou být fixní, kdy jedinec zastává celý ţivot jednu taktiku, nebo naopak flexibilní, coţ v praxi znamená, ţe je jedinec můţe libovolně střídat v reakci na okolní podmínky (Candolin 2004). Pokud se v populaci vyskytuje více takovýchto taktik a všechny poskytují svým nositelům aspoň podobný

ÚVOD

15 reprodukční úspěch, pak mohou dohromady tvořit evolučně stabilní strategii (Andersson 1994). Tyto taktiky jsou dány morfologií jedinců (Shuster & Wa- de 1991) či fyzickými předpoklady a schopnostmi jedince samotného (Tho- rnhill 1981) i jeho konkurentů v sociální hierarchii (Chan & Ribbink 1990), nebo kondicí a věkem (Isvaran 2005). Zároveň jsou ovlivněny i velkým po- čtem environmentálních, ekologických a demografických faktorů, které je- dince obklopují a ovlivňují. Mohou jimi být například populační hustota (Kokko & Rankin 2006), intenzita osvětlení (Denoël & Doellen 2010), rychlost proudění vody (Magellan & Magurran 2006), úroveň predace (Godin 1995) nebo dostupnost zdrojů pro rozmnoţování (Przybylski et al. 2007). Taktiky mohou být také získané zkušeností a naučené (Guevara-Fiore 2012) či gene- ticky podmíněné (Lank et al. 1995). Zatímco všechny výše zmíněné vlivy byly a jsou intenzivně zkoumány, vliv osobnostních znaků jedince na vyuţívání alternativních reprodukčních taktik je studován jen zřídka. V přírodě je asi nejznámějším systémem ART tzv. guarder (courter) vs. non-guarder (sneaker) komplex (Oliveira et al. 2008). Jedná se o dichotomii mezi teritoriální a neteritoriální strategií. Guarder (teritoriální samec) je cha- rakterizován snahou o monopolizaci nějakého území, popř. zdrojů, jako jsou třecí substrát či samice. Alternativně také investuje čas a energii do předvá- dění svých kvalit nebo do dvoření se samicím. Sneaker (neteritoriální samec) se naproti tomu snaţí oplodnit vajíčka nepozorovaně, ze zálohy, často i proti vůli samice a obejít tak její volbu. Sneaking je také většinou znatelně časově úspornější neţ dvoření a jedinec tak můţe ušetřený čas věnovat hledání dal- ších potenciálních partnerek (Parker 1974). Největší variabilitu v ART vyka- zují mezi obratlovci ryby (Oliveira et al. 2008).

Zkušenost a sociální prostředí

U společenských druhů zvířat se ţádné chování neděje v izolaci. Proto je důleţité zakomponovat do úvah o konkrétním problému i vliv sociálního pro- středí na alternativní reprodukční taktiky a nevytrhávat tak toto chování z kontextu. Vliv demografických činitelů na vyuţívání ART byl jiţ studován (např. Taborsky 2001) a je známo, ţe různé taktiky mají v různých prostře-

ÚVOD

16 dích různou úspěšnost. Většinou platí, ţe při vysoké konkurenci by mělo být aspoň pro část jedinců výhodnější zastávat pozici sneakerů a snaţit se získat kopulace nepozorovaně či násilně, jelikoţ tato taktika není tak časově nároč- ná a je výhodná v boji se schopnějšími či vybarvenějšími konkurenty. Nao- pak v prostředí, kde konkurence není tak silná a je zde větší dostupnost sa- mic, by mělo být pro samce výhodné se samicím dvořit a pokusit se je pře- svědčit ke kooperativnímu páření. To je důleţité např. u ţivorodých ryb, kde jsou náznaky, ţe samec dokáţe takto přenést do samice více spermií neţ při sneakingu (Pilastro & Bisazza 1999). Některá tato chování mohou být získaná pozorováním nebo díky vlastní zkušenosti (Freeberg 2000). Ovšem to, jestli jedinec okamţitě reaguje na ak- tuální podmínky, nebo je jeho chování ovlivněno vlastními minulými zkuše- nostmi, nebo do jaké míry je jeho chování ovlivněno sociálním prostředím, zůstává často opomíjeno. Z demografických charakteristik sociálního pro- středí jsem se při hledání odpovědi na tuto otázku zaměřil především na zá- sadní aspekty jako počet a dostupnost samic a také na jejich následný vliv na samčí sexuální chování.

POUŽITÉ DRUHY A METODIKA

17

Použité druhy a metodika

Hořavka duhová (Rhodeus amarus)

Jako jeden z modelo- vých druhů jsem zvolil ho- řavku duhovou (Rhodeus amarus). Je to drobná kap- rovitá ryba z podčeledi Acheilognathinae a vyznaču- je se nevšedním způsobem rozmnoţování. Je rozšířená na většině území dnešní Ev- Obrázek 1 Samec Rhodeus amarus s detailem duhov- ropy (Van Damme et al. ky. Fotografie samce Roman Hlásek, detail duhovky 2007). V České republice se Martin Reichard. vyskytuje v povodí všech velkých řek, především ve slepých ramenech a po- malu tekoucích vodách. Dorůstá obvykle délky přibliţně 7 cm (Smith et al. 2004). Způsob její reprodukce umoţňuje nezávislé porovnání významu pří- mých a nepřímých zisků v samičí volbě (Smith et al. 2002). Chování hořavky duhové je jiţ velmi dobře prostudováno. Samci vyzná- vají alternativní reprodukční taktiky – mohou se stát dominantními jedinci, kteří brání své teritorium obsahující jednoho či více mlţů. Nebo mohou být neteritoriálními jedinci, tzv. sneakery, a mohou se snaţit oplodnit jikry na- kladené samičkou do mlţe nepozorovaně (Smith et al. 2004). Více neţ na fy- zickou kompetici se tito jedinci zaměřují na kompetici spermií. Sneakeři mo- hou být u hořavek jak obligatorní, tak i fakultativní (Kanoh 1996). Obě tak- tiky také nejsou fixní a jedinec je můţe často střídat (Candolin & Reynolds 2001; Smith et al. 2002). To, jakou taktiku zvolí, záleţí na dostupnosti mlţů (Przybylski et al. 2007) a samic, denzitě jedinců (Reichard et al. 2004) či na samcově velikosti (Smith et al. 2002; Reichard et al. 2008) nebo jeho kompe- tičních schopnostech (Candolin & Reynolds 2002).

POUŽITÉ DRUHY A METODIKA

18

Vlastní párovací systém hořavky je promiskuitní a zaloţený na zdrojích (Reichard et al. 2008). Teritoriální samci (Obrázek 1) se v době rozmnoţová- ní, které u nás začíná v dubnu, vrcholí v květnu a pokračuje aţ do července, pestře zbarví na ventrální části těla a řitní ploutvi a lá- kají samičky s prodlouţeným kladélkem (Obrázek 2) ke tření do ţivých mlţů čeledi Unionidae, na jejichţ výskyt je hořavka vázána. U nás to jsou především druhy Ano- donta cygnea, A. anatina, Obrázek 2 Samička Rhodeus amarus s prodlouženým Unio tumidus a U. pictorum kladélkem. Fotografie Roman Hlásek. (Smith et al. 2004). Samička za pomoci proudu moči umístí kladélkem do ţaberní dutiny mlţe 1 – 6 jiker (průměrně 3) a můţe je takto klást i několikrát denně. Sameček je pak oplodní vypuštěním spermatu nad inhalačním sifonem mlţe. Do mlţe však ejakuluje i před nakladením jiker z důvodu zvýšení šance na úspěšné oplo- zení. Mladé ryby svůj úkryt opouští asi po jednom měsíci (Smith et al. 2004). Na počáteční rozhodnutí, zda bude samička samečka k mlţi následovat, má vliv intenzita jeho červeného zbarvení a také intenzita námluv (Candolin & Reynolds 2001). Ty překvapivě neslouţí jako ukazatel nepřímých zisků pro samičku, jak by se dalo tušit, kdeţto spíše zisků přímých v podobě větší šance na oplození jejích jiker, protoţe červenější samci mívají větší zásobu spermií (Smith et al. 2014). Své partnery ke tření si vybírají samičky hořavek také za pomoci olfaktorických signálů, které značí genetickou kompatibilitu (Agbali et al. 2010; Reichard et al. 2012). Na finální rozhodnutí, do kterého mlţe samička jikry naklade, má vliv především obsah kyslíku vycházející z jeho exhalačního sifonu (Smith et al. 2001). Samičky ovšem také za účelem maximalizace svých zisků překvapivě vyhledávají tření v blízkosti sneakerů (Smith & Reichard 2005), které jim také zajišťuje zvýšení jejich reprodukční úspěšnosti.

POUŽITÉ DRUHY A METODIKA

19

To, jakou alternativní taktiku pro oplození jiker sameček hořavky zvolí, ovlivňuje, jak jiţ bylo zmíněno výše, velké mnoţství faktorů. Avšak aţ do současnosti se nikdo nezaměřil na to, zda zastávání jedné z taktik popř. je- jich střídání mohou nějak ovlivňovat osobnostní rysy jedince a zda tyto rysy nemohou společně tvořit nějaký behaviorální syndrom (soubor navzájem ko- relovaných chování).

Metodika

Pokusní jedinci byli pro studii o personalitách (Řeţucha et al. 2012), která vychází jiţ z mé diplomové práce (Řeţucha 2010), odchyceni za pomoci elektrolovu v říčce Kyjovce u soutoku Moravy a Dyje na jiţní Moravě. Poté byli převezeni na Ústav biologie obratlovců v Brně. Kaţdý pokusný samec byl individuálně označen barevnou elastomerovou značkou injektovanou pod kůţi. Ke stanovení základních a také nejčastěji studovaných osobnostních rysů – odvahy a plachosti, jsem pouţil zavedenou metodiku, tzv. emergence test, který spočívá v měření času, který jedinec potřebuje k opuštění bezpeč- ného úkrytu a vyplavání na volné prostranství. Dále se u jedinců stanovovala míra jejich dominance – schopnost uhájit teritorium. To bylo prováděno v la- boratorních podmínkách a pak dále konfrontováno s podmínkami ve velké venkovní nádrţi na úrovni mezokosmu v polopřírodních podmínkách. Tam měli jedinci k dispozici časově variabilní počet teritorií reprezentovaných jed- notlivými mlţi a byla kvantifikována jejich schopnost teritoria obsadit a udr- ţet. Dále byla softwarem Genepop 4.0.10 (Rousset 2008) zjišťována hetero- zygotnost jedinců na základě třinácti mikrosatelitových lokusů. Rovněţ se stanovoval gonadosomatický index (GSI, poměr hmotnosti gonád ke zbytku těla) a také velikost a hmotnost jedinců spolu s plochou červené barvy v du- hovce coby relativním ukazatelem dominance (Reichard et al. 2009).

POUŽITÉ DRUHY A METODIKA

20

Živorodka Endlerova (Poecilia wingei)

Jako další modelový druh jsem si pro svou další práci (Řeţucha & Reichard 2014, Řeţucha & Reichard, submitováno) vybral ţivo- rodku Endlerovu (Poecilia wingei). Je to endemický druh jihoamerické ţivorodé ryby z čeledi Poeciliidae,

Obrázek 3 Samec Poecilia wingei. Fotografie Radim Bla- který je rozšířen pouze žek. v nevelké oblasti kolem se- verovenezuelského města Cumaná (Poeser et al. 2005; Evans et al. 2011). Je blízce příbuzná velmi dobře známé a prozkoumané ţivorodce duhové (Poecilia reticulata), od níţ se odlišuje především svým barevným vzorem (zejména větší plochou oranţové a černé barvy), lehce rozdílnou morfologií gonopodia (pářícího orgánu samců vzniklého přeměnou řitní ploutve) a také několika aspekty chování (Poeser et al. 2005). Od ţivorodky duhové byla odlišena také na základě molekulárních dat (Schories et al. 2009) a druhový status byl ověřen i v recentní studii potvrzující postavení Poecilia wingei coby sester- ského druhu Poecilia reticulata (Meredith et al. 2010). U obou druhů existuje silný pohlavní dimorfismus, kdy samci jsou pestře zbarvení (Obrázek 3) a velcí zhruba 2,5 cm. Naopak šedozelené samice (Obrázek 4) dorůstají při- bliţně 4 cm. V přírodě ţijí ve stojatých vodách i menších řekách a tvoří smí- šená hejna čítající přibliţně 60 jedinců (Poeser et al. 2005). Párovací systém ţivorodky Endlerovy je čistě promiskuitní, bez zaloţení na jakýchkoliv zdro- jích. Samci P. wingei jsou charakterističtí přítomností alternativních repro- dukčních taktik, které jsou za účelem maximalizace individuálního repro- dukčního úspěchu samců flexibilní, a jedinec je tak můţe podle potřeby stří- dat. Samec se můţe samici dvořit formou tzv. sigmoidního předvádění, kdy se před ní rychle natřásá a ukazuje své pestře zbarvené, doširoka roztaţené

POUŽITÉ DRUHY A METODIKA

21 ploutve a tělo má zkroucené v pozici připomínající písmeno S. Tímto dvoře- ním se snaţí samici přesvědčit ke kooperativnímu páření. Samec se také můţe pokusit oplodnit samici nepozorovaně za pomoci sneakingu, kdy potají zezadu zasune vztyčené gonopodium do samičina gonoporu (pohlavní otvor) a vypustí malé mnoţství spermatu uskladněného ve váčcích (spermatozeug- matech). Celá akce je oproti dvoření velmi rychlá a většinou trvá od zlomku sekundy po pár vteřin. Objem přesunutého spermatu během sneakingu je obyčejně menší neţ během koo- perativního páření (Pilastro & Bisazza 1999) a je při něm také menší šance, ţe samička nako- nec porodí nějaká mláďata (Sato Obrázek 4 Samice P. wingei. Fotografie Radim Bla- et al. 2011). Samička je recep- žek. tivní pouze pár dní po porodu a v této době také samečkům dovoluje se s ní pářit. Mimo toto období však můţe být oplodněna nedobrovolně za pomoci sneakingu, jelikoţ spermie přenesené do samičích vejcovodů (gonoduktů) v nich dokáţí zůstat ţivotaschopné dlouhou dobu (Houde 1997). Často se stává, ţe spousta samečků příbuzné Poecilia reticulata tak zplodí potomstvo dokonce i po své smrti (López-Sepulcre et al. 2013). U těchto ryb platí priori- ta posledního samce, kdy poslední samec, se kterým se samička páří, zplodí většinu potomstva (Evans & Magurran 2001). Samci věnují snaze o páření a pronásledování samiček většinu svého času (Magurran & Seghers 1994). Jak pouţití alternativních taktik samec vybalancuje, závisí u těchto ţi- vorodek jak na environmentálních faktorech, jako jsou intenzita okolního světla (Reynolds et al. 1993; Chapman et al. 2009) či rychlost proudění vody (Magellan & Magurran 2006), tak i na faktorech ekologických – např. do- stupnost potravy (Kolluru & Grether 2004; Kolluru et al. 2007), stupeň pre- dace (Godin 1995), intrasexuální kompetice (Evans & Magurran 1999), po- pulační hustota (Farr & Herrnkind 1974), poměr pohlaví (Magurran & Magellan 2007), stupeň parazitace (Kolluru et al. 2009) či faktorech fyzio-

POUŽITÉ DRUHY A METODIKA

22 logických, morfologických a genetických – březost samice (Ojanguren & Ma- gurran 2004), zbarvení (Jirotkul 2000; Kiritome et al. 2012) či délka ocasu (Karino & Kobayashi 2005; Karino & Kamada 2009). Pouţívání těchto taktik můţe být zároveň vrozené (Rodd & Sokolowski 1995) nebo získané vlastní zkušeností či pozorováním ostatních jedinců (Farr 1980; Jordan & Brooks 2012; Guevara-Fiore et al. 2012). V mém experimentu jsem se zaměřil na jeden z posledních jmenovaných faktorů, a to na vliv zkušenosti v kombinaci s vlivem sociálního prostředí (Řeţucha & Reichard 2014). Vliv sociálního prostředí a zkušenosti byl jiţ zkoumán u Poecilia reticulata (např. Guevara-Fiore 2012; Guevara-Fiore et al. 2012), nicméně vliv dlouhodobého působení sociálního prostředí (pomě- rově vychýleného na stranu samců nebo samic) na osvojení alternativní tak- tiky u dospělých jedinců a její budoucí vyuţívání, byl hodně opomíjen. Zamě- řil jsem se na to, jak bude u samců budoucí vyuţití alternativních taktik modifikováno jejich současným sociálním prostředím a zda dokáţí rychle re- agovat na změnu v tomto prostředí a promptně přizpůsobí své chování aktu- ální sociální situaci.

Metodika

Původně získaní jedinci byli potomky ryb z přírody (oblast kolem seve- rovenezuelského města Cumaná) a byli na Ústavu biologie obratlovců AV ČR, v.v.i. v Brně drţeni po několik generací ve velké společenské nádrţi. Na za- čátku experimentů bylo vybráno 65 juvenilních samců hned, jak je bylo moţné rozlišit od samic, a ti byli dále drţeni odděleně od zbytku ryb po celou dobu jejich adolescence. Nepřišli do kontaktu ani se samicemi či jinými sam- ci. Vlastní pokus začal jejich rozdělením do 2l individuálních nádrţí, které od sebe byly vizuálně odděleny. Po době aklimatizace následovalo 1. měření epi- gamního chování ve standardizovaných podmínkách. Probíhalo v malé nádr- ţi s jednou samicí. Zaznamenáván byl počet a délka jednotlivých dvoření (sigmoidních předvádění, levá × pravá strana), počet pokusů o sneaking, po- hyby gonopodia do vzpřímené pozice (napumpování gonopodia spermatem (Liley 1966)) a počet kousnutí do gonoporu samice (zkoumání olfaktorických

POUŽITÉ DRUHY A METODIKA

23 informací vylučovaných samicí (Herdman et al. 2004)). Kvantifikace byla prováděna za pomoci programu pro hodnocení chování – JWatcher (Blum- stein et al. 2006). Díky němu bylo moţné zaznamenávat zároveň více druhů chování, jak jejich počet, tak i délku. Tyto údaje pak software umoţňuje na konci také analyzovat a vyhodnocovat, ale protoţe tento program ukládá průběh vlastního pozorování do časové osy, je moţné se k výsledkům měření i zpětně vrátit a analýzu opravit či zpřesnit. Po skončení prvního testování byla do poloviny nádrţí s pokusnými samci přidána samice a další samec (ri- val) a do druhé poloviny dvě samice. Tím vnikla dvě sociální prostředí – jedno s více samci (male – biased) a jedno s více samicemi (female – biased). Po pěti týdnech ţivota v těchto prostředích následovalo 2. testování rozmnoţovacího chování pokusných samců. Paralelně byla za účelem srovnání otestována a vyfocena skupina 32 náhodně vybraných samců P. wingei ze společenské nádrţe. Ti vyrůstali a ţili společně s velkým počtem rivalů a samic, a měli tak od nich ke svému chování zpětnou vazbu. Výsledky experimentu jsou disku- továny v rámci pohlavního výběru a vlivu zkušenosti, sociálního prostředí a zbarvení na rozmnoţovací chování samců. V druhé části experimentu jsem cílil na to, zda má sociální prostředí vliv nejen čistě na behaviorální aspekty samčího epigamního chování, ale také na aspekty, které souvisí se vzhledem jedince. Konkrétně bylo zkoumáno, zda pokusní jedinci dokáţí nějak účinně behaviorálně maskovat své fyzické ne- dostatky a zda se tato jejich případná schopnost bude lišit svou intenzitou v různých sociálních prostředích (Řeţucha & Reichard, submitováno). Gross et al. (2007) zjistil, ţe nezkušení samečci Poecilia reticulata předváděli samič- kám během dvoření především svou více oranţovou stranu. Nezkoumali však vznik a vývoj tohoto fenoménu a také jaký vliv mohou mít na tento trend okolní faktory, především zkušenost jedinců a sociální prostředí, jehoţ vliv je u této skupiny společenských ryb rozhodující. Můţe být např. důleţitější snaţit se vypadat barevněji v prostředí, kde jsou přítomni konkurenti, a samice tak můţe samce navzájem porovnávat, coţ je důleţitý proces v samičí volbě (Gasparini et al. 2013), neţ v prostředí bez konkurentů, kde tlak na kvalitu dvoření nebude tak silný.

POUŽITÉ DRUHY A METODIKA

24

Za účelem pochopení těchto vlivů byli jedinci v našem pokusu vyfoto- grafováni fotoaparátem Canon EOS Rebel XTi z kaţdé strany, dvakrát v prů- běhu experimentu a v návaznosti na behaviorální pozorování. Analýza jejich zbarvení byla provedena za pomoci programu pro analýzu obrazu Fiji (Schindelin et al. 2012). Sledována byla především plocha nejdůleţitějších barevných odstínů – černé, oranţové a iridescentní. Ta byla stanovena po- mocí funkce programu Fiji Color Threshold, která oproti jiným dostupným a většinou pouţívaným řešením umoţňuje precizní, přesnou a především opakovatelnou kvantifikaci dané plochy zájmu. Plochu iridescentních barev by bez pouţití mnou zvoleného postupu nebylo moţné adekvátně změřit. A pokud ano, tak jen velmi nepřesně. Počet skvrn byl pak určen na základě jejich vizuálního součtu. Zaznamenávány dále byly i další proměnné, jako např. délka gonopodia, celková délka, délka ocasu a hřbetní ploutve. Tyto údaje pak byly dány do souvislosti s behaviorálními daty a byl zkoumán vliv rozloţení těchto barevných elementů na těle samce na jeho chování během dvoření. Konkrétně na frekvenci předvádění jeho barevnější strany během námluv v různých sociálních prostředích (male biased × female biased) a s rozdílnými zkušenostmi (naivní × zkušený). Výsledkem mělo být zjištění, zda dokáţe samec operativně kompenzovat své nedostatky ohledně zbarvení a zda dokáţe přednostně předvádět během sigmoidních předvádění samicím svou atraktivnější stranu, a tím u nich budit dojem, ţe je barevnější neţ ve skutečnosti. Výsledky jsou diskutovány v kontextu pohlavního výběru a s ním související samičí volby.

IMPAKTOVANÉ PUBLIKACE

25

Impaktované publikace

Článek 1

Řeţucha, R., Smith, C. & Reichard, M. 2012. Personality traits, reprodu- ctive behaviour and alternative mating tactics in male European bitter- ling, Rhodeus amarus. Behaviour, 149, 531-553.

V této práci jsem provedl všechny experimenty a měření (kromě analýzy mikrosatelitových lokusů) a napsal první verzi rukopisu. Dále jsem přispíval k vytvoření finální verze.

Behaviour 149 (2012) 531–553 brill.nl/beh

Personality traits, reproductive behaviour and alternative mating tactics in male European bitterling, Rhodeus amarus

Radomil Režuchaˇ a, Carl Smith b and Martin Reichard a,∗ a Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Kvetnᡠ8, 603 65 Brno, Czech Republic b School of Biology, University of St Andrews, St Andrews, Fife KY16 8LB, UK *Corresponding author’s e-mail address: [email protected]

Accepted 23 April 2012

Abstract Individual differences in behavioural traits may play a role in reproductive behaviour and it is likely that different personality types have different reproductive success across specific social environ- ments. This suggests a role for sexual selection for personality types, including a link between behavioural traits and alternative reproductive tactics. While research on morphological differ- ences between guarders (bourgeois males defending resources) and sneakers (males parasitizing the territories of bourgeois males) is well characterized, the role of personality in the adoption of alternative mating tactics has hitherto been largely ignored. We investigated individual behavioural differences in a sneaker–guarder mating system using size/age-matched males of a small freshwa- ter fish, the European bitterling (Rhodeus amarus). We predicted distinct behavioural responses by guarder and sneaker males that were consistent in different contexts, indicating the existence of behavioural syndromes associated with male mating tactic. No behavioural syndromes were detected in male R. amarus, despite ability of individual male bitterling to establish dominance, boldness and investment in sperm competition being relatively repeatable across three consecu- tive trials. Male aggression, though not repeatable, was negatively correlated with the number of ejaculations, indicating a trade-off between aggression and sperm loading. No association between the tendency of males to guard a territory and behavioural traits was found, despite a significant association between the tendency to guard and morphological and physiological traits, with higher relative testis size and more breeding tubercles in guarder males. Our data suggest male bitterling mating tactics are largely unconstrained by innate factors and likely to be the product of prevailing environmental and social conditions. Keywords alternative reproductive behaviour, animal personality, boldness, behavioural syndrome, mat- ing system, sperm competition, territorial males.

© 2012 Koninklijke Brill NV, Leiden DOI:10.1163/156853912X643908 532 Personality and alternative male mating behaviour

1. Introduction Research on animal behaviour increasingly recognizes that behavioural traits can be consistent through time and contexts, resulting in the emergence of distinct behavioural types (Stamps & Groothuis, 2009). Such types have been termed ‘personalities’ (defined as a set of personality traits consis- tent through time). Additionally, ‘behavioural syndromes’ are recognized as an inter-correlated set of individual personality traits (Sih & Bell, 2008). For example, ‘personality’ can be recognized when an individual is boldest among a group of conspecifics as a juvenile and expresses a relatively high level of boldness later in life, when compared to similarly-aged conspecifics. A ‘behavioural syndrome’ is observed when a positive or negative corre- lation is found between traits, such as aggression directed at conspecifics and boldness towards predators (Huntingford, 1976). Such correlated suites of behaviours may be favourable in certain environments or situations (Re- aney & Backwell, 2007), while maladaptive in others (Johnson & Sih, 2005). Consequently the existence of distinct behavioural types has important impli- cations for the way in which selection acts on behavioural traits, and thereby on the maintenance of behavioural variation among and within populations. To investigate personality, a group of individuals can be observed through a suite of contexts and their behavioural traits compared. A typical approach is to expose animals to new or threatening stimuli (Weinstein et al., 2008) and score their behaviour with respect to a shy–bold axis, with the relative position of an individual along this continuum representing its willingness to take risks (Wilson et al., 1994). Bold individuals are more prone to engage in risky behaviour, such as predator inspection, more readily switch to unfamil- iar diets, interact readily with novel objects or individuals, and are generally more explorative (Budaev, 1997; Godin & Clark, 1997; Wilson, 1998). In contrast, shy individuals engage in behaviours yielding lower profits but also less risk (Ward et al., 2004). The position of individuals on the shy-bold axis can predict their behaviour in the presence of predators, under competition, and their ability to monopolize resources (Weinstein et al., 2008). Behavioural traits can play a role in reproductive behaviour, either directly or by their covariance with other traits via pleiotropic effects. For exam- ple, Colléter & Brown (2011) have shown that the most dominant males of the crimson-spotted rainbowfish (Melanotaenia duboulayi) were also more aggressive, bold and more active than subordinates. Males at the top of rain- bowfish social hierarchies have better access to females and are able to R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 533 monopolize them, which significantly enhances their reproductive success (Young et al., 2010). Personality also plays a role in individual reproductive success in the Trinidadian guppy (Poecilia reticulata). Godin & Dugatkin (1996) demonstrated that females preferred mating with bolder males who were more likely to inspect a predator in a female’s presence, and apparently were better informed about imminent attack from predators. Under certain demographic and environmental conditions, females may, thus, increase their fitness by mating with bolder males in the case there is heritable component to boldness (Brown et al., 2007). While it appears that different personality types may have different re- productive success across specific social environments, the impact of sexual selection on personality traits has only rarely been studied (Schuett et al., 2010), including the link between personality traits and alternative repro- ductive tactics (ARTs) (Brockmann et al., 2008; Bergmüller & Taborsky, 2010). ARTs represent mating polymorphisms, often maintained through negative frequency-dependent selection (Shuster, 2010). ARTs are a general phenomenon among competing males, particularly in taxa with external fer- tilization reflecting a functional distinction between ‘producers’ (bourgeois males) and ‘scroungers’ (parasitic males) (Taborsky et al., 2008). One of the most common forms of ART is the sneaker–guarder system, in which guarders monopolize females or critical breeding resources, while sneakers attempt to mate deceptively and engage in sperm competition (Parker et al., 1997; Taborsky, 1998). The relative success of an ART depends on indi- vidual competitive capability, environmental and population parameters, and the level at which reproductive competition occurs (Reichard et al., 2004b; Taborsky, 2008; Konecnᡠet al., 2010; Shuster, 2010). While research on morphological differences between guarders and sneakers has received some attention, the role of behavioural syndromes and personality in the adoption of alternative tactics has so far been largely ignored (Brockmann et al., 2008; Bergmüller & Taborsky, 2010). Here, we investigated individual behavioural differences in bourgeois- parasitic tactics using males of a small freshwater fish, the European bit- terling (Rhodeus amarus). Because ARTs are associated with contrasting roles, we predicted that guarder (bourgeois) and sneaker (parasitic) males would differ in personality traits across contexts. We anticipated that the expression of discrete behavioural syndromes would predispose individual males to largely act in one mating role, with guarder (bourgeois) males being 534 Personality and alternative male mating behaviour bolder, more aggressive and dominant. We further tested the association be- tween particular behavioural traits and morphological (body size, abundance of breeding tubercles, fluctuating asymmetry), physiological (testis size, red colour in eye iris, growth rate) and genetic (heterozygosity at microsatellite loci) characters. We tested a group of bitterling males in a series of aquarium and semi-field experiments. We controlled for the effect of body size, since this trait is known to affect dominance and territorial status in R. amarus.

2. Methods 2.1. Study species The European bitterling is a cyprinid fish (subfamily Acheilognathinae) that uses freshwater mussels for oviposition (Smith et al., 2004). Bitterling have a promiscuous resource-based mating system (Reichard et al., 2008), with adult body size ranging from 35 to 65 mm (Smith et al., 2004). Females deposit their eggs in the gills of mussels using an unusually long oviposi- tor, while males release sperm over the inhalant siphon of mussels, so that water filtered by a mussel carries the sperm to the eggs. During the reproduc- tive season males defend small territories around living freshwater mussels. Guarder males actively court females and attempt to lead them to mussels in their territory. Males can also act as sneakers, interrupting spawning pairs by darting next to a spawning pair from a shelter and fertilizing eggs by re- leasing sperm into the mussel defended by another male. Larger males are more likely to become guarders than smaller males (Reichard et al., 2008), but there are no apparent morphological differences between guarders and sneakers. Male tactics are sensitive to prevailing environmental and social conditions (Przybylski et al., 2007) and it appears that all males are capable of both tactics, though disproportionally perform one of the roles (Smith et al., 2004). For further details on bitterling mating systems see Smith et al. (2004). 2.2. Subjects Experimental fish were collected in the River Kyjovka in the southeast of the Czech Republic using a portable electrofishing backpack, modified to capture small fish efficiently (Lena, www.r-bednar.cz), and transported in aerated river water to the Institute of Vertebrate Biology (IVB) in Brno. Prior to experiments, a stock of wild fish was held in a large pond in the garden of R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 535 the IVB under natural conditions. In March (prior to the onset of the bitter- ling spawning season), a sample of males was measured for Standard Length (SL, body size excluding caudal fin) and 24 selected on the basis of their sim- ilar body size (mean ± SE = 53.8 ± 0.29 mm, range 51.5–57.0 mm). All the males were uniquely marked with coloured visible implant elastomer tags (VIE, Northwest Marine Technology Company). After marking, fish were housed in a large aquarium (126 × 41 × 40 cm) (henceforth referred to as the social tank) with a sand substrate and artificial plants as refuges. Fish experienced natural daylight (approx. 13 h of light) and water temperature fluctuation (mean 18°C, range 16–22°C) and were fed once each day with frozen chironomid larvae. The start of experiments coincided with the on- set of the bitterling reproductive season (early April). Experimental mussels were collected from an oxbow lake adjacent to the River Kyjovka prior to the start of bitterling spawning season. 2.3. Experimental procedures Individual boldness was measured as the time taken for a male to emerge from a shelter after introduction to a novel aquarium. The test aquarium was 75 × 40 × 38 cm and contained a 2 cm layer of river sand. Three sides of the aquarium were covered with opaque barriers to minimize external dis- turbance. Two large artificial plants were provided as shelters and situated in both of the rear corners. Before the start of the experiment all the experi- mental males were carefully removed from the social tank to six plastic tubs to standardize the effects of capture in fish that were tested. Trials were con- ducted on 7, 8 and 10 April 2009. After a short period of habituation in the tubs (15 min), an experimental fish was gently captured and released into the right front corner of the experimental tank. All fish immediately hid in the fronds of the artificial plants. The time taken to emerge from the shelter and swim in open water (i.e., to start tank exploration) was measured as index of boldness. Some individuals did not leave the shelter within 1200 s and were assigned a score of 20 min (maximum shyness). After emerging from shelter, or after 1200 s, whichever was sooner, the test male was captured and gen- tly returned to the social tank. Note that our measure of boldness could also reflect a measure of stress resistance, because it measures the recovery time after being caught, handled and moved. The test consisted of three replicates for each male. Each male was tested once each day of the experiment. The order of testing males followed a predetermined random sequence. 536 Personality and alternative male mating behaviour

Male dominance was measured as the ability of a male to establish a ter- ritory under competition with other males (conducted 14, 15 and 17 April 2009). All males likely interfered with each other prior to the test in the social tank, though their potential interactions before the test were standard- ized by using a single social tank for all males. Pre-experimental and tank conditions were identical to the test for boldness, except that a freshwater mussel, placed in a sand-filled plastic box (10 × 6 × 6 cm), was located in the centre of the tank. Three males were captured from a plastic box and simultaneously released into the experimental tank. The establishment of a dominance hierarchy was checked each 15 min. Dominant males were identified through mussel defence and aggressive displays directed towards other males (Reichard et al., 2005). Physical contact between males was oc- casionally recorded, but no injuries were observed. Territorial disputes are a common feature of male reproductive behaviour and are typically settled prior to physical contact (Smith, 2011). Subordinate males were provided with artificial vegetation and were able to readily escape and hide from dom- inant individuals. Once one male had unambiguously established dominance (chased the rivals from the territory and patrolled over the mussel), he was removed. This process was repeated with the two remaining males to enable the second and third ranked males to be identified. The trials took between 30 and 75 min before dominance was established. This period corresponds with the rapid establishment of territoriality observed in the field (Smith et al., 2003). Dominance was scored on a 3-point ordinal scale from 2 (the most dominant) to 0 (no dominance). Winner/loser effects were not considered. Each male was tested three times and male combinations were random ex- cept to ensure that no individual males were ever tested together more than once. Each male was tested once each day of the experiment. Male investment in aggressive behaviour and sperm competition was tested as a response by each male to a rival (conducted 21–28 April 2009). Pre-experimental and aquarium conditions were identical to the previous ex- periment, but with the addition of rival male in a glass jar (10 × 10 × 28 cm) positioned in the centre of the aquarium, 15 cm from the mussel, with a mesh top to permit ventilation. Small bitterling males (SL ± SE = 35 ± 0.7 mm) were used as rivals. Test males could not attack the bottled rival males directly, but the presence of a rival elicited the focal male to respond to the risk of sperm competition (Smith et al., 2003, 2009). The response of focal males was recorded 30 s after the first response to the rival male R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 537 and lasted 10 min. Head butting and parallel swimming (Smith et al., 2004) were recorded as aggressive bouts and their frequency was used in analyses. No other aggressive behaviour was observed. At the same time, investment in sperm competition was quantified as the number of ejaculations over the inhalant siphon of the mussel, which is a good approximation of sperm in- vestment by male bitterling (Smith et al., 2009). Male bitterling frequently ejaculate over mussels in the absence of females. This behaviour probably relates to the importance of sperm precedence in fertilization success, with preoviposition ejaculation rate (often in the absence of females) being the only significant predictor of paternity success (Reichard et al., 2004b). Rival males and mussels were replaced in each trial (to minimize their stress when constrained) and each male was tested on three separate occasions. In a final, large-scale trial, the ability of each male to defend a terri- tory under semi-natural conditions, their adoption of a specific mating tactic (guarder, sneaker), and a measure of their reproductive success were mea- sured in an outdoor concrete pool measuring 12.6 × 6 m filled to a water depth of 0.6 m (approximately 45 000 l). Large (1 m long) artificial plants were placed around the perimeter of the pool as shelters. On 30 April, 23 experimental males and 135 females were released into the pool and allowed to settle for one week. Fish fed on naturally occurring food items (detri- tus, algae, aquatic insect larvae) supplemented with a daily ration of frozen chironomids. Note that only 2–20% of female bitterling are in reproductive condition on any given day (Konecnᡠ& Reichard, 2011) yielding a local op- erational sex ratio (Seger & Stubblefield, 2002) typically encountered under natural conditions, despite the population sex ratio being significantly female biased. Both reproductively active and non-reproducing females swim in large shoals during the reproductive season, forage and visit male territories (Smith et al., 2004), and this behaviour was also observed in the experimen- tal pool (Reichard et al., 2004a; Konecnᡠet al., 2010). After one week, live mussels were placed in small sand-filled plastic boxes (18 × 14 cm) to rep- resent varying number of territories. A total of 18 plastic boxes were placed along three walls of the pool (one side was left empty to allow a diver to enter and leave the pool without disturbing fish) 1.7 m apart. A mussel was placed in a box to enable an active territory to be formed. The number of potential territories was varied between 1 and 18, with 1, 2, 3, 4, 6, 8, 10, 14 or 18 territories active on a given day. The number of active territories was ran- domized among days. The position of active territories within the pool was 538 Personality and alternative male mating behaviour selected to maximize distances between adjacent territories. On each day of the experiment, a snorkeler observed every mussel for 15 min and recorded male behaviour. For each marked male, the number of defended territories, number of sneaking attempts, leading of females to a mussel, and number of spawnings were recorded. After completion of behavioural observations, all mussels were removed and replaced with new set of mussels that were positioned in accordance with the next experimental treatment. Behavioural observations took place between 9:00 and 17:00 Central European Summer Time; i.e., >3.5 h after sunrise and <3 h before sunset. A single treatment was completed each day. At the end of the experiment, the males were recaptured using baited net traps. Nine males were not recovered (and not observed during snorkelling after the end of the experiment) and those males appeared to have died over the course of the experiment. Intrinsic male mortality is naturally high dur- ing the reproductive season of the bitterling (Smith et al., 2000), with only 4–5% of males surviving until the end of reproductive season under natural conditions in the study population (Konecnᡠ& Reichard, 2011). Only recov- ered males were used in data analysis. Captured males were photographed (Olympus Mju 1030 SW) in a glass cell under identical light conditions from a distance of 25 mm to estimate the extent of red area in their iris, serving as carotenoid-based nuptial signal (Reichard et al., 2009). The male was killed with an overdose of anaesthetic (clove oil), a finclip taken for DNA extrac- tion and fixed in ethanol, and the body preserved in an 8% formaldehyde solution. After preservation, each male was measured for SL to the nearest mm and weighted to the nearest 0.001 g (wet weight). A count was made of the number of ventral fin rays, scales in the longest row along the body, and breeding tubercles on each side of the body and head. Males were dissected and their testes were weighted to the nearest 0.001 g. Gonad mass measured at the end of reproductive season denotes residual gonad mass and represents a valid relative index of investment in gonad tissue in the foregoing breeding season (Reichard et al., 2009). This measure overcomes the problem of tem- porary sperm depletion which commonly affects ejaculate size and sperm density (traits that can be measured during the reproductive season) in bitter- ling (Smith et al., 2009). Fish were individually marked using elastomer tags, which are widely used in experimental studies and have proven harmless (Halls & Azim, 1998; R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 539

Weston & Johnson, 2008), which was also the case in the present study. Mor- tality of a single male during the first phase of experimental work (aquarium), and a further nine males during the second phase was recorded. These mor- talities were not associated with experimental treatments. Our target sample size was estimated based on previous comparable experiments (Smith et al., 2003; Reichard et al., 2004b, 2005, 2006; Smith & Reichard, 2005) to achieve a trade off between maximizing experimental power while minimiz- ing the number of experimental fish used. 2.4. Data analysis The repeatability of behavioural traits was measured as intraclass correlation sensu Lessells & Boag (1987) (ICC1). However, our experimental design (constrained by the duration of the bitterling reproductive season) necessi- tated identical trials to be repeated within a relatively short period, with a risk that subjects habituated to experimental conditions (Chervet et al., 2011), which was indeed the case for some measures of personality traits (see Re- sults). To account for this, we calculated alternative intraclass correlations according to Shrout & Fleiss (1979) and implemented in package ‘psych’ for R 2.9.1 (R Development Core Team, 2009). Those intraclass correlations remove mean differences in response variables between trials (McGraw & Wong, 1996) and we present the outcomes of analyses where subjects were modelled as either random (ICC2) or fixed (ICC3) effects. We acknowledge, however, that our experimental design did not control for potential carry-over effects (Dochtermann, 2010). A fluctuating asymmetry index (FA) was calculated according to formula FA11 of Palmer (1994), whereby FA = |R − L|,whereR is the value for a trait on right side of the fish and L the value for the same trait on left side. Meristic data for ventral fins, breeding tubercles and the number of scales along the body axis were summed. The extent of eye redness was estimated as the proportion of the total iris that was red using ImageJ 1.43n (Abramoff et al., 2004; Reichard et al., 2009). The gonadosomatic index (GSI) was calculated as GSI = (WG/WE) × 100, where WG is gonad weight and WE is the weight of the eviscerated fish. The use of WE (weight of internal organs not included) minimized the effect of gut fullness. Growth rate was calculated as a difference in the natural log of body size measured on 1 April and 15 June, which coincided with the reproductive period and can be considered as growth rate over the breeding season (Reichard et al., 2009). 540 Personality and alternative male mating behaviour

Individual heterozygosity was estimated on the basis of 10 microsatellite loci (Rser01, Rser03, Rser05, Rser06, Rser08, Rser09, Rser10, Rser11, Rser12 and Rser13) (Dawson et al., 2003; Reichard et al., 2008) using Genepop 4.0.10 (Rousset, 2008). DNA extraction and microsatellite analysis followed established protocols given in Bryja et al. (2010). Prior to statistical analysis, all data were checked for assumptions of parametric tests (normality, homoscedasticity of variances) and transformed if necessary. Proportional data were square-root arcsine transformed. Final sample size was 23 males in aquarium experiments and a subset of 14 males in the outdoor experiments due to mortalities (see below). Given that some mortalities are common among males during the reproductive season, we anticipated that final sample size (at the end of reproductive season) to be lower than the number of males used at the beginning of experiments (at the start of reproductive season). Consequently, with a target final sample size of 18 males, we initially used 24 males. Therefore, our final sample size in aquarium experiments (N = 23) was slightly higher than planned, while the sample size for outdoor experiments was slightly lower (N = 14). Associations between personality traits and morphological traits were tested using summed values across all three trials. To test whether fish habit- uated to experimental conditions and changed their response among consec- utive trials, we fitted linear mixed models (nlme package in R 2.9.1, LMM) with particular personality trait as response variable and fish identity as ran- dom factor. The relationships between male territoriality in the outdoor pool, personality traits, and morphological characters were tested using a Spear- man correlation. A measure of territoriality (tendency of the males to guard a territory) was expressed as the number of cases when a particular male was observed to guard a territory.

3. Results Individual differences in the capacity to establish dominance, boldness, and investment in sperm competition were repeatable, though male aggression was not (Table 1); the significance of repeatability increased when habitua- tion was accounted for (ICC2 and ICC3 measures in Table 1). There was a clear increase in boldness score across trials, indicating that fish decreased their time to emerge from a refuge with increasing experience with experi- mental conditions (LMM, trial effect F2,44 = 8.75, p<0.001). Investment R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 541

Table 1. Results of repeatability analysis (N = 23 males, aquarium).

ICC1 ICC2 ICC3

rF22,46 pr rF22,44 p

Dominance 0.385* 2.88 0.001 0.379* 0.369* 2.76 0.002 Boldness 0.160 1.27 0.098 0.210* 0.267* 2.09 0.017 Sperm competition 0.178 1.65 0.076 0.207* 0.230* 1.90 0.035 Aggression −0.195 0.51 0.955 −0.119 −0.147 0.62 0.889

Intraclass correlations were calculated sensu Lessells & Boag (1987) (ICC1) and with mean differences between trials taken into account, with subjects modelled as random (ICC2)or fixed (ICC3) effects. * Statistically significant result. in sperm competition (LMM, trial effect F2,44 = 4.45, p = 0.017) and ag- gression rate varied among trials (LMM, trial effect F2,44 = 6.62, p = 0.003) (Figure 1). These personality traits were significantly repeatable when differ- ences in mean values among trials were accounted for (Table 1), indicating that despite overall numerical changes in the response variable, rank order of boldness and sperm competition score for individual fish was consistent across trials. Dominance was measured on a rank scale, resulting in no dif- ferences in mean values across trials and significant repeatability using all three intraclass correlation coefficients (Table 1). There was a significant negative correlation between male aggression and investment in sperm competition (Pearson correlation, r =−0.544, N = 23, p = 0.004; Figure 2). The statistical significance was retained after Bonfer- roni correction (critical p = 0.008). Other associations between personality traits were not significant (all p>0.229), giving little support for the exis- tence of behavioural syndromes, at least under our experimental conditions. Male body size was experimentally constrained in the study, and as antici- pated did not correlate with any behavioural traits (boldness: r = 0.237, p = 0.276; investment in sperm competition: r =−0.111, p = 0.615; aggres- sion: r =−0.150, p = 0.494; dominance: r = 0.367, p = 0.085). Within individual dominance trials, male body size still had some importance. It was not associated with the rank order among males during the first (Kendall coefficient of concordance W = 0.204, p = 0.195) and second (W = 0.094, p = 0.472) trial, though it did affect the outcome of the third trial, with the relatively largest males being most dominant (W = 0.717, p = 0.003). 542 Personality and alternative male mating behaviour

Figure 1. Mean values of personality traits across individual trials in aquarium experiments (N = 23 males). Whiskers represent one standard error.

There was no association between the tendency to act as a guarder in the outdoor experiment and personality traits (dominance: rS = 0.401, p = 0.156, investment in sperm competition: rS = 0.460, p = 0.098, boldness: rS = 0.325, p = 0.257, aggression: rS =−0.077, p = 0.793). For morpho- logical and physiological traits, a greater tendency to act as a guarder was positively associated with larger GSI and more breeding tubercles (Table 2). The number of spawning acts a male participated in (either as guarder or sneaker) was positively related to the ability of a male to establish a territory (measured as the overall number of territories a male defended) (Spearman correlation, rS = 0.740, N = 14, p = 0.002). The ability of males to estab- R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 543

Figure 2. Relationship between aggression and number of ejaculations during simultaneous assessment in male R. amarus in aquarium experiments (N = 23 males). The larger point identifies an overlap between two individuals. lish dominance in aquarium tests correlated positively with the number of breeding tubercles, but no other relationship between morphological (and physiological) traits and personality traits proved significant, including the effect of body size (Table 2).

Table 2. Association between territoriality (tendency to act as a guarder in an outdoor study) and dom- inance (in an aquarium study) and morphological and physiological traits in male European bitterling (N = 14 males).

Territoriality Dominance

rS prS p ∗ ∗ GSI 0.697 0.006 0.399 0.158 ∗ ∗ ∗ ∗ Breeding tubercles 0.610 0.021 0.633 0.015 Growth rate 0.407 0.149 0.117 0.692 Eye redness 0.191 0.514 0.045 0.879 Body size 0.093 0.753 0.230 0.430 Body mass 0.038 0.898 0.173 0.555 Heterozygosity −0.137 0.641 0.197 0.499 Fluctuating asymmetry −0.153 0.601 −0.294 0.308

Territoriality was expressed as the number of cases when a particular male was observed to defend a territory. Results of a Spearman correlation and its statistical significance are reported. * Statistically significant result. 544 Personality and alternative male mating behaviour

4. Discussion We investigated individual behavioural differences in a sneaker–guarder sys- tem using males of the European bitterling. We predicted distinct behavioural responses by guarder and sneaker males that were consistent in different contexts, indicating the existence of behavioural syndromes. We demon- strated the ability of individual male bitterling to establish dominance was highly repeatable, while boldness and investment in sperm competition were repeatable only at relative scales. Individuals retained their relative ranks, but numerical measures of their personality traits varied. Male aggression was not repeatable. Overall, the existence of a behavioural syndrome was not supported, because personality traits were not correlated. The exception was the association between male aggression and sperm competition which were negatively correlated, indicating that individual males varied in their response to the presence of a rival along the bourgeois-parasitic continuum. Despite significant association between the tendency to act as a guarder and morphological and physiological traits (with greater relative testis size and more breeding tubercles in guarders), no difference between males along the guarder–sneaker continuum was found in personality traits. High repeatabil- ity in dominance was detected even though male size was constrained within a narrow (5 mm) range. This suggests that an individual’s ability to establish and successfully defend a territory is consistent even among similarly-sized rivals. However, even within this size-constrained cohort, slight body size differences explained variance in dominance in one experimental trial. Male body size is the major predictor of dominance hierarchy in most animals, including fishes (Andersson, 1994). Male boldness gradually increased across trials, with the time to emerge from a shelter declining across successive trials (Figure 1). This indicates that the measure of boldness might also reflect stress resistance or ‘train- ing effect’. Chervet et al. (2011) report a ‘training effect’ in a Tanganyikan cichlid (Neolamprologus pulcher). If two behavioural tests were conducted only a short time apart in their study, individuals were usually bolder and more explorative in the second test. Wilson & Godin (2009) also found that individual bluegill sunfish varied in their response to predators across trials due to individuals habituating to the novel stimuli at different rates. This im- plies that different capacities of individual males to learn and habituate to experimental conditions can limit repeatability of behavioural traits. In our study, however, repeatability of male boldness was high after accounting for R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 545 habituation in the calculation of intraclass correlation coefficient, consistent with a situation of relatively minor differences to habituate to experimental conditions among individual fish. Behavioural syndromes in males were not demonstrated by the study. This result may be an artifact of our experimental design (using size-matched 1-year-old males tested during the reproductive season or the choice of be- havioural measures and their order). Alternatively it may be a feature of our study population or, more generally, of the European bitterling. Study sub- jects may have come from an environment where strong correlations between behaviours across contexts were not adaptive. For example, Bell (2005) re- ported a behavioural syndrome in a population of three-spined sticklebacks (Gasterosteus aculeatus) that experienced strong predator pressure. In an adjacent population subjected to limited predation pressure no behavioural syndrome was detected. However, the results may have been confounded by testing different populations at different phases of reproduction. Onto- genetic patterns may also affect expression of behavioural syndromes. Bell & Stamps (2004) found a weaker correlation across behaviours in subadult sticklebacks compared with adults, probably linked to hormonal changes as- sociated with sexual maturation. Male aggression and investment into sperm competition were measured in the same experiment and a negative correlation between aggression and number of ejaculations supports the observation that male bitterling traded- off between the two behaviours. These two activities are time and energy expensive and males may be constrained in their ability to express both si- multaneously (Candolin & Reynolds, 2002). A trade-off between aggression and sperm competition has been observed in other fish species, including the Mediterranean wrasse (Symphodus ocellatus) (Alonzo & Warner, 2000), three-spined stickleback (Le Comber et al., 2003) and zebrafish (Danio re- rio) (Spence & Smith, 2005), and may be a general feature of animal mating systems. Dominant male bitterling were shown to have more breeding tubercles; epidermal keratin-based structures found on the head of reproductively ac- tive males. The development of breeding tubercles in fish is controlled by androgens (Wiley & Collette, 1970), themselves influenced by social status (Oliveira et al., 1996). Dominant individuals with elevated levels of these hormones express more breeding tubercles (Oliveira et al., 2002) and they may be perceived as an indicator of androgen levels. Tubercles have also 546 Personality and alternative male mating behaviour been hypothesized as functioning as signals of male quality affecting female mate choice (Huuskonen et al., 2009), though whether this occurs through direct or indirect selection is unclear. A large number of breeding tubercles in dominant males has also been demonstrated in the roach (Rutilus rutilus) (Kortet et al., 2004), a species of fish with a lek mating system (Wedekind, 1996). Breeding tubercles may provide protection from mechanical friction during spawning (Ahnelt & Keckeis, 1994). In the European bitterling, tuber- cles are concentrated on the tip of the snout and may function as a weapon in male–male combat. During territorial fights males frequently engage in bouts of what has been termed headbutting (Smith et al., 2004), but which actually involves males striking each other with their snouts. Hence, the number of breeding tubercles may potentially be under intrasexual selection and further research is needed to establish their function in the bitterling mating system. Male tendency to act as a guarder positively correlated with testis mass. This finding appears to contradict a key prediction of sperm competition theory (Parker, 1998). An assumption of classical sneak-guard models is that sneakers face sperm competition in all their reproductive attempts, while guarder males are confronted with sperm competition only in a subset (Parker, 1998). However, in R. amarus this assumption may be violated, with the overall level of sperm competition risk consistently high for all males in a population, irrespective of mating tactic (Candolin & Reynolds, 2001; Re- ichard et al., 2004a,b; Smith et al., 2004, 2009). Male R. amarus display specialized reproductive adaptations for high levels of sperm competition and sperm economy in comparison with other bitterling species (Pateman- Jones et al., 2011), and with no distinction in sperm quality or testis structure in relation to mating tactic (Pateman-Jones, 2008). Male bitterling release sperm both before and after a female deposits her eggs, and each spawning is accompanied by a mean ± SE of 4.4 ± 0.47 ejaculations/mating, with an excess of 250 ejaculations per guarder male on each day of the reproduc- tive season in a wild population (Smith et al., 2009). Our observations that guarder males had greater testis mass than sneakers may better reflect the pattern in resource-based mating systems than is generally recognized (see also Zbinden et al., 2001; Fitzpatrick et al., 2006; Reichard et al., 2007), with only males in sufficiently good condition being able to afford the high cost of sperm production and territorial defence simultaneously. Indeed, the ability of a male to guard a territory was positively related to the overall number of spawning acts a male participated in as guarder or sneaker and, therefore, R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 547 with reproductive success under our experimental conditions. However, the relative success of guarders and sneakers varies extensively across demo- graphic parameters (Reichard et al., 2004a,b). We predicted that specific morphological, physiological and genetic traits would be associated with personality traits and mating tactics. Contrary to predictions (Tiira et al., 2003, 2006; Vilhunen et al., 2008), we failed to de- tect a relationship between genetic heterozygosity, the level of fluctuating asymmetry (an index of developmental instability), carotenoid-based red col- oration and behaviour, especially the ability to establish dominance. While this may be partially ascribed to our relatively low sample size, no appar- ent trend was detected and any potential relationship must have had only a relatively small effect. Males that employ alternative mating tactics are sometimes morphologi- cally dissimilar and their behaviour may be, at least temporarily, dictated by morphological and physiological conditions related to ontogenetic stage or phenotype. However, the majority of known cases of alternative mating tac- tics reflect conditional responses to transitory social conditions (Oliveira et al., 2008), including in R. amarus (Smith et al., 2004). Social conflict can se- lect for the stable coexistence of different behavioural types and Bergmüller & Taborsky (2010) hypothesized that consistent individual differences in personality traits may lead to a diversification into discrete social roles, including guarder and sneaker males as alternative avenues to maximize fitness. Our data do not support Bergmüller & Taborsky’s (2010) hypoth- esis; there was no difference in any personality trait between guarder and sneaker males, nor was there consistency in behaviour across contexts. The assumption that bolder individuals are more likely to become guarders was also not confirmed. However, we acknowledge a relatively small sample size in our analysis, and the fact that our experimental design excluded the role of predators, despite an attempt to mimic natural conditions in a large semi-natural setting. Predators can have a significant impact on individual behaviour (Bell, 2005), and guarder males may be more susceptible to pre- dation when courting or displaying to females (Magurran, 2005; Konecnᡠ& Reichard, 2011). In conclusion, we did not detect evidence for behavioral syndromes cor- responding with male mating tactics in male R. amarus. We further failed to find strong associations between individual behavioural, morphological, physiological and genetic traits, despite repeatability of some personality 548 Personality and alternative male mating behaviour traits. These data suggest male bitterling mating tactics are largely uncon- strained by innate or ontogenetic factors, implying male bitterling mating tactics to be the product of the prevailing environmental and social conditions rather than being evolutionarily constrained (Duckworth, 2006), making the bitterling an especially attractive model for research on conditional male re- productive behaviour.

Acknowledgements We are grateful to Matej Polacik,ˇ Hanka Patzenhauerová, Markéta On- dracková,ˇ Markéta Konecná,ˇ Jiríˇ Huml, Josef Bryja, Milan Vrtílek, Pavel Šebek and Klára Morongová for help with experiments and M. Taborsky, R. Earley and two anonymous referees for comments on an earlier draft of the manuscript. RRˇ and MR hold a license for conducting experimental work on vertebrates. Experimental procedures were approved by the ethical com- mittees of the IVB and Ministry of Agriculture (CZ 62760203) and are in accordance with Czech legal requirements. All authors conceived and de- signed the research. R.R.ˇ collected the data, R.R.ˇ and M.R. analyzed data and drafted the paper. All authors contributed to writing. Financial support came from Czech Science Foundation GACR 206/09/1163.

References Abramoff, M.D., Magelhaes, P.J. & Ram, S.J. (2004). Image processing with ImageJ. — Biophoton. Int. 11: 36-42. Ahnelt, H. & Keckeis, H. (1994). Breeding tubercles and spawning behaviour in Chondros- toma nasus (Teleostei: Cyprinidea): a correlation? — Ichtyol. Explor. Freshw. 5: 321-330. Alonzo, S.H. & Warner, R.R. (2000). Allocation to mate guarding or increased sperm pro- duction in a Mediterranean wrasse. — Am. Nat. 156: 266-275. Andersson, M. (1994). Sexual selection. — Princeton University Press, London. Bell, A.M. (2005). Behavioural differences between individuals and two populations of stick- leback (Gasterosteus aculeatus). — J. Evol. Biol. 18: 464-473. Bell, A.M. & Stamps, J.A. (2004). Development of behavioural differences between indi- viduals and populations of sticklebacks, Gasterosteus aculeatus.—Anim.Behav.68: 1339-1348. Bergmüller, R. & Taborsky, M. (2010). Animal personality due to social niche specialisation. — Trends Ecol. Evol. 25: 504-511. Brockmann, H.J., Oliveira, R.F. & Taborsky, M. (2008). Integrating mechanisms and func- tion: prospects for future research. — In: Alternative reproductive tactics: an integrative approach (Oliveira, R.F., Taborsky, M. & Brockmann, H.J., eds). Cambridge University Press, New York, NY, p. 471-489. R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 549

Brown, C., Braithwaite, V.A. & Burgess, F. (2007). Heritable and experiential effects on boldness in a tropical poeciliid. — Behav. Ecol. Sociobiol. 62: 237-243. Bryja, J., Smith, C., Konecný,ˇ A. & Reichard, M. (2010). Range-wide population genetic structure of the European bitterling (Rhodeus amarus) based on microsatellite and mito- chondrial DNA analysis. — Mol. Ecol. 19: 4708-4722. Budaev, S.V. (1997). Alternative styles in the European wrasse, Symphodus ocellatus — boldness-related schooling tendency. — Environ. Biol. Fish. 49: 71-78. Candolin, U. & Reynolds, J.D. (2001). Sexual signaling in the European bitterling: females learn the truth by direct inspection of the resource. — Behav. Ecol. 12: 407-411. Candolin, U. & Reynolds, J.D. (2002). Adjustments of ejaculation rates in response to risk of sperm competition in a fish, the bitterling (Rhodeus sericeus). — Proc. Roy. Soc. Lond. B: Biol. 269: 1549-1553. Chervet, N., Zöttl, M., Schürch, R., Taborsky, M. & Heg, D. (2011). Repeatability and heri- tability of behavioural types in a social cichlid. — Int. J. Evol. Biol. 2011: 1-15. Colléter, M. & Brown, C. (2011). Personality traits predict hierarchy rank in male rainbowfish social groups. — Anim. Behav. 81: 1231-1237. Dawson, D.A., Burland, T.M., Douglas, A., Le Comber, S.C. & Bradshaw, M. (2003). Isola- tion of microsatellite loci in the freshwater fish, the bitterling Rhodeus sericeus (Teleostei: Cyprinidae). — Mol. Ecol. Notes 3: 199-202. Dochtermann, N. (2010). Behavioral syndromes: carryover effects, false discovery rates, and a priori hypotheses. — Behav. Ecol. 21: 437-439. Duckworth, R.A. (2006). Behavioral correlations across breeding contexts provide a mecha- nism for a cost of aggression. — Behav. Ecol. 17: 1011-1019. Fitzpatrick, J.L., Desjardins, J.K., Stiver, K.A., Montgomerie, R. & Balshine, S. (2006). Male reproductive suppression in the cooperatively breeding fish Neolamprologus pulcher.— Behav. Ecol. 17: 25-33. Godin, J.G.J. & Clark, K.A.V. (1997). Risk-taking in stickleback fishes faced with different predatory threats. — Ecoscience 4: 246-251. Godin, J.G.J. & Dugatkin, L.A. (1996). Female mating preference for bold males in the guppy, Poecilia reticulata. — Proc. Natl. Acad. Sci. USA 93: 10262-10267. Halls, A.S. & Azim, M.E. (1998). The utility of visible implant (VI) tags for marking tropical river fish. — Fish. Manag. Ecol. 5: 71-80. Huntingford, F.A. (1976). The relationship between anti-predator behaviour and aggression among conspecifics in the three-spined stickleback, Gasterosteus aculeatus.—Anim. Behav. 24: 245-260. Huuskonen, H., Haakana, H. & Kekäläinen, J. (2009). Offspring performance is linked to parental identity and male breeding ornamentation in whitefish. — Biol. J. Linn. Soc. 98: 532-539. Johnson, J.C. & Sih, A. (2005). Precopulatory sexual cannibalism in fishing spiders (Dolomedes triton): a role for behavioral syndromes. — Behav. Ecol. Sociobiol. 58: 390- 396. 550 Personality and alternative male mating behaviour

Konecná,ˇ M. & Reichard, M. (2011). Seasonal dynamics in population characteristics of European bitterling Rhodeus amarus in a small lowland river. — J. Fish. Biol. 78: 227- 239. Konecná,ˇ M., Smith, C. & Reichard, M. (2010). Population and individual consequences of breeding resource availability in the European bitterling (Rhodeus amarus). — Behav. Ecol. Sociobiol. 64: 1069-1079. Kortet, R., Taskinen, J., Vainikka, A. & Ylonen, H. (2004). Breeding tubercles, papillomatosis and dominance behaviour of male roach (Rutilus rutilus) during the spawning period. — Ethology 110: 591-601. Le Comber, S.C., Faulkes, C.G., Formosinho, J. & Smith, C. (2003). Response of territorial males to the threat of sneaking in the three-spined stickleback (Gasterosteus aculeatus): a field study. — J. Zool. 261: 15-20. Lessells, C.M. & Boag, P.T. (1987). Unrepeatable repeatabilities: a common mistake. — Auk 104: 116-121. Magurran, A.E. (2005). Evolutionary ecology: the trinidadian guppy. — Oxford University Press, New York, NY. McGraw, K.O. & Wong, S.P. (1996). Forming inferences about some intraclass correlation coefficients. — Psychol. Methods 1: 30-46. Oliveira, R.F., Almada, V.C. & Canario, A.V.M. (1996). Social modulation of sex steroid concentrations in the urine of male cichlid fish Oreochromis mossambicus.—Horm. Behav. 30: 2-12. Oliveira, R.F., Hirschenhauser, K., Carneiro, L.A. & Canario, A.V.M. (2002). Social modula- tion of androgen levels in male teleost fish. — Comp. Biochem. Physiol. B 132: 203-215. Oliveira, R.F., Taborsky, M. & Brockmann, H.J. (2008). Alternative reproductive tactics: an integrative approach. — Cambridge University Press, New York, NY. Palmer, A.R. (1994). Fluctuating asymmetry analyses: a primer. — In: Developmental insta- bility: its origins and evolutionary implications (Markow, T.A., ed.). Kluwer, Dordrecht, p. 335-364. Parker, G.A. (1998). Sperm competition and the evolution of ejaculates: towards a theory base. — In: Sperm competition and sexual selection (Birkhead, T.R. & Møller, A.P., eds). Cambridge University Press, Cambridge, p. 3-54. Parker, G.A., Ball, M.A., Stockley, P. & Gage, M.J.G. (1997). Sperm competition games: a prospective analysis of risk assessment. — Proc. Roy. Soc. Lond. B: Biol. 264: 1793- 1802. Pateman-Jones, C. (2008). Sperm competition and male mating tactics in the bitterling fishes. — PhD thesis, Department of Biology, University of Leicester, Leicester. Pateman-Jones, C., Rasotto, M., Reichard, M., Liao, C., Liu, H., Zieba, G. & Smith, C. (2011). Variation in male reproductive traits among three bitterling fishes (Acheilognathi- nae: Cyprinidae) in relation to mating system. — Biol. J. Linn. Soc. 103: 622-632. Przybylski, M., Reichard, M., Spence, R. & Smith, C. (2007). Spatial distribution of ovipo- sition sites determines variance the reproductive rate of European bitterling (Rhodeus amarus). — Behaviour 144: 1403-1417. R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 551

R Development Core Team (2009). R: a language and environment for statistical computing. — R Foundation for Statistical Computing, Vienna, Austria. Available online at http:// www.R-project.org Reaney, L.T. & Backwell, P.R.Y. (2007). Risk-taking behavior predicts aggression and mating success in a fiddler crab. — Behav. Ecol. 18: 521-525. Reichard, M., Bryja, J., Ondracková,ˇ M., Dávidová, M., Kaniewska, P. & Smith, C. (2005). Sexual selection for male dominance reduces opportunities for female mate choice in the European bitterling (Rhodeus sericeus). — Mol. Ecol. 14: 1533-1542. Reichard, M., Jurajda, P. & Smith, C. (2004a). Male–male interference competition decreases spawning rate in the European bitterling (Rhodeus sericeus). — Behav. Ecol. Sociobiol. 56: 34-41. Reichard, M., Le Comber, S.C. & Smith, C. (2007). Sneaking from a female perspective. — Anim. Behav. 74: 679-688. Reichard, M., Ondracková,ˇ M., Przybylski, M., Liu, H. & Smith, C. (2006). The costs and benefits in an unusual symbiosis: experimental evidence that bitterling fish (Rhodeus sericeus) are parasites of unionid mussels in Europe. — J. Evol. Biol. 19: 788-796. Reichard, M., Ondracková,ˇ M., Bryjová, A., Smith, C. & Bryja, J. (2009). Breeding resource distribution affects selection gradients on male phenotypic traits: experimental study on lifetime reproductive success in the bitterling fish (Rhodeus amarus). — Evolution 63: 377-390. Reichard, M., Smith, C. & Jordan, W.C. (2004b). Genetic evidence reveals density-dependent mediated success of alternative mating behaviours in the European bitterling (Rhodeus sericeus). — Mol. Ecol. 13: 1569-1578. Reichard, M., Smith, C. & Bryja, J. (2008). Seasonal change in the opportunity for sexual selection. — Mol. Ecol. 17: 642-651. Rousset, F. (2008). Genepop’007: a complete re-implementation of the Genepop software for Windows and Linux. — Mol. Ecol. Resour. 8: 103-106. Schuett, W., Tregenza, T. & Dall, S.R.X. (2010). Sexual selection and animal personality. — Biol. Rev. 85: 217-246. Seger, J. & Stubblefield, J.W. (2002). Models of sex ratio evolution. — In: Sex ratios: con- cepts and research methods (Hardy, I.C.W., ed.). Cambridge University Press, New York, NY, p. 2-25. Shrout, P.E. & Fleiss, J.L. (1979). Intraclass correlations: uses in assessing rater reliability. — Psychol. Bull. 86: 420-3428. Shuster, S.M. (2010). Alternative mating strategies. — In: Evolutionary behavioral ecology (Fox, C. & Westneat, D.F., eds). Oxford University Press, New York, NY, p. 434-450. Sih, A. & Bell, A.M. (2008). Insights for behavioral ecology from behavioral syndromes. — Adv. Stud. Behav. 38: 227-281. Smith, C. (2011). Good fences make good neighbours: the role of landmarks in territory partitioning in the rose bitterling (Rhodeus ocellatus). — Behaviour 148: 233-246. Smith, C., Pateman-Jones, C., Zi˛eba, G., Przybylski, M. & Reichard, M. (2009). Sperm depletion as a consequence of increased sperm competition risk in the European bitterling, Rhodeus amarus. — Anim. Behav. 77: 1227-1233. 552 Personality and alternative male mating behaviour

Smith, C. & Reichard, M. (2005). Females solicit sneakers to improve fertilization success in the bitterling fish (Rhodeus sericeus). — Proc. Roy. Soc. Lond. B: Biol. 272: 1683-1688. Smith, C., Reichard, M. & Jurajda, P. (2003). Assessment of sperm competition by European bitterling, Rhodeus sericeus. — Behav. Ecol. Sociobiol. 53: 206-213. Smith, C., Reichard, M., Jurajda, P. & Przybylski, M. (2004). The reproductive ecology of the European bitterling (Rhodeus sericeus). — J. Zool. 262: 107-124. Smith, C., Reynolds, J.D. & Sutherland, W.J. (2000). The population consequences of repro- ductive decisions. — Proc. Roy. Soc. Lond. B: Biol. 267: 1327-1334. Spence, R. & Smith, C. (2005). Male territoriality mediates density and sex ratio effects on oviposition in the zebrafish. — Anim. Behav. 69: 1317-1323. Stamps, J. & Groothuis, T.G.G. (2009). The development of animal personality: relevance, concepts and perspectives. — Biol. Rev. 85: 301-325. Taborsky, M. (1998). Sperm competition in fish: “bourgeois” males and parasitic spawning. — Trends Ecol. Evol. 13: 222-227. Taborsky, M. (2008). Alternative reproductive tactics in fish. — In: Alternative reproductive tactics: an integrative approach (Oliveira, R.F., Taborsky, M. & Brockmann, H.J., eds). Cambridge University Press, New York, NY, p. 251-299. Taborsky, M., Oliveira, R.F. & Brockmann, H.J. (2008). The evolution of alternative re- productive tactics: concepts and questions. — In: Alternative reproductive tactics: an integrative approach (Oliveira, R.F., Taborsky, M. & Brockmann, H.J., eds). Cambridge University Press, New York, NY, p. 1-22. Tiira, K., Laurila, A., Enberg, K., Piironen, J., Aikio, S., Ranta, E. & Primmer, C.R. (2006). Do dominants have higher heterozygosity? Social status and genetic variation in brown trout, Salmo trutta. — Behav. Ecol. Sociobiol. 59: 657-665. Tiira, K., Laurila, A., Peuhkuri, N., Piironen, J., Ranta, E. & Primmer, C.R. (2003). Aggres- siveness is associated with genetic diversity in landlocked salmon (Salmo salar). — Mol. Ecol. 12: 2399-2407. Vilhunen, S., Tiira, K., Laurila, A. & Hirvonen, H. (2008). The bold and the variable: fish with high heterozygosity act recklessly in the vicinity of predators. — Ethology 114: 7-15. Ward, A.J.W., Thomas, P., Hart, P.J.B. & Krause, J. (2004). Correlates of boldness in three- spined sticklebacks (Gasterosteus aculeatus). — Behav. Ecol. Sociobiol. 55: 561-568. Wedekind, C. (1996). Lek-like spawning behaviour and different female mate preferences in roach (Rutilus rutilus). — Behaviour 133: 681-695. Weinstein, T.A.R., Capitanio, J.P. & Gosling, S.D. (2008). Personality in animals. — In: Handbook of personality theory and research (John, O.P., Robins, R.W. & Pervin, L.A., eds). The Guilford Press, New York, NY, p. 328-348. Weston, M.R. & Johnson, R.L. (2008). Visible implant elastomer as a tool for marking etheostomine darters (Actinopterygii: Percidae). — Southeast. Nat. 7: 159-164. Wiley, M.L. & Collette, B.B. (1970). Breeding tubercles and contact organs in fishes: their occurrence, structure, and significance. — Bull. Am. Mus. Nat. Hist. 143: 145-216. Wilson, A.D.M. & Godin, J.G.J. (2009). Boldness and behavioral syndromes in the bluegill sunfish, Lepomis macrochirus. — Behav. Ecol. 20: 231-237. R. Režuchaˇ et al. / Behaviour 149 (2012) 531–553 553

Wilson, D.S. (1998). Adaptive individual differences within single populations. — Philos. Trans. Roy. Soc. Lond. B: Biol. Sci. 353: 199-205. Wilson, D.S., Clark, A.B., Coleman, K. & Dearstyne, T. (1994). Shyness and boldness in humans and other animals. — Trends Ecol. Evol. 9: 442-446. Young, M., Simmons, L. & Evans, J. (2010). Pre- and post-mating sexual selection both favor large males in a rainbowfish. — Behav. Ecol. Sociobiol. 64: 915-925. Zbinden, M., Largiader, C.R. & Bakker, T.C.M. (2001). Sperm allocation in the three-spined stickleback. — J. Fish Biol. 59: 1287-1297.

IMPAKTOVANÉ PUBLIKACE

49

Článek 2

Řeţucha, R. & Reichard, M. 2014. The effect of social environment on al- ternative mating tactics in male Endler's guppy, Poecilia wingei. Animal Behaviour, 88, 195-202.

V této práci jsem navrhl její design včetně řešení chovných nádrţí. Vy- choval jsem pokusné jedince a staral se o ně v průběhu celého experimentu. Provedl jsem také všechny experimenty a vymyslel vlastní způsob jejich rea- lizace. Sesbíral jsem všechna data, vyhodnotil je a provedl jejich primární statistickou analýzu. Napsal jsem první verzi článku a později přispíval k je- ho finální verzi.

Animal Behaviour 88 (2014) 195e202

Contents lists available at ScienceDirect

Animal Behaviour

journal homepage: www.elsevier.com/locate/anbehav

The effect of social environment on alternative mating tactics in male Endler’s guppy, Poecilia wingei

, , Radomil Rezucha a b, Martin Reichard a * a Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Brno, Czech Republic b Department of Botany and Zoology, Masaryk University, Brno, Czech Republic article info Variation in sexual behaviour has both genetic (no experience or social contact needed) and environ- fi Article history: mental (based on individual experience) components. Within species, males can maximize their tness Received 9 July 2013 during inter- and intrasexual selection in a variety of ways. This may lead to evolution of alternative male Initial acceptance 20 August 2013 mating behaviours resulting in a dichotomy between courting/guarding and sneaking tactics. We used Final acceptance 21 November 2013 Endler’s guppies to investigate (1) whether individual sexually naïve males have a preferred mating Published online tactic (courting or sneaking), and (2) how male mating behaviour is affected by recent social history MS. number: 13-00569R (male- or female-biased social environment). We found that individual males approached females by either courting or sneaking, even with no previous sexual experience. We further demonstrated that Keywords: male sexual behaviour in standardized mating trials was strongly affected following exposure to a sex- alternative mating tactics biased social environment for 5 weeks. Males from female-biased social environments showed decreased courtship mating effort overall compared to the baseline level (measured prior to assignment to a social envi- experience Poecilia wingei ronment treatment) but were more likely to court females rather than attempt sneak copulations. Males sneaking from male-biased social environments maintained a high level of mating effort and increased the rate of social environment sneaking attempts compared to courtship displays. Our study highlights that both genetic and envi- ronmental sources of variation contribute to individual sexual behaviour. Current sexual behaviour was strongly modulated by recent social history, underpinning the importance of previous experience in the expression of sexual behaviour. Ó 2013 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

Males adopt a variety of ways to maximize their fitness during mating behaviour and mating tactics may have profound conse- inter- and intrasexual selection, leading to high interindividual quences for reproductive success (Reichard, Smith, & Jordan, 2004). variability in sexual behaviour. Some aspects of sexual behaviour Despite generally continuous variation, discrete alternative are genetically underpinned and are performed by an individual mating tactics are often readily recognized (Taborsky, 1994). The without any previous experience or social contact (Andersson, relative fitness benefit of a particular mating tactic is potentially 1994). Other components of sexual behaviour and mate prefer- affected by a variety of environmental, ecological and demographic ence, however, may be affected by the local conditions an indi- factors. It has been demonstrated, for example, that predation level vidual experiences. Some behaviour patterns may be learnt (Godin, 1995; Luyten & Liley, 1985), food availability and parasite through observation of conspecifics or through an individual’sown load (Kolluru, Grether, Dunlop, & South, 2009), water velocity experience (Freeberg, 2000), while others may be influenced by the (Magellan & Magurran, 2006), population density (Jirotkul, 1999a), immediate social environment (Rodríguez, Rebar, & Fowler-Finn, adult sex ratio (Magurran & Magellan, 2007), operational sex ratio 2013). Individuals vary in their capacity to respond to changes in (Jirotkul, 1999b; Mills & Reynolds, 2003) and breeding resource their social environment, however, and in their ability to modify abundance (Konecná, Smith, & Reichard, 2010) may all influence behaviour (Bretman, Cage, & Chapman, 2011). A degree of such the outcome of maleemale competition for fertilization and hence plasticity determines an individual’s ability to adapt rapidly to the the success of individual males and their tactics. prevailing social conditions. Indeed, the ability to respond quickly The most common system of alternative male mating tactic is to changes in the social environment through modulation of the sneakereguarder complex (Oliveira, Taborsky, & Brockmann, 2008). Under this system, guarder males maximize their repro- ductive success through investment in guarding a limiting resource, such as a breeding substrate or females, and often in * Correspondence: M. Reichard, Institute of Vertebrate Biology, Academy of Sci- performing courtship behaviour. Alternatively, males can mate as ences of the Czech Republic, Kvetná 8, 603 65 Brno, Czech Republic. E-mail address: [email protected] (M. Reichard). sneakers, that is, fertilizing females or ova through concealment or

0003-3472/$38.00 Ó 2013 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.anbehav.2013.12.010 196 R. Rezucha, M. Reichard / Animal Behaviour 88 (2014) 195e202 coercion, and frequently engage in sperm competition (Oliveira environment were predicted to mate more often through gon- et al., 2008; Parker, Ball, Stockley, & Gage, 1997; Taborsky, 1998). opodial thrusting (sneaking), as sneaking is more successful when In many species, the choice of tactic is flexible, with males maleemale competition for copulation increases and the risk of switching between the two in response to local conditions sperm competition is high (Evans & Magurran, 1999; Mills & (Candolin, 2004; Chapman, Morrell, & Krause, 2009; Reichard et al., Reynolds, 2003). In contrast, males exposed to a female-biased 2004). However, it is less clear whether males respond directly to environment were predicted to court females using sigmoid dis- their current state, or whether the expression of mating behaviour plays, since they had no previous experience of mating competition. is influenced by experience from previous mating attempts. The costs and benefits of mating tactics differ and can vary across METHODS social environments. When competition is greatest, for example, it should be advantageous for at least a proportion of males to act as Study Species and its Mating Behaviour sneakers and force copulations. Sneaking can circumvent female choice when a particular male is unattractive, unsuccessful or not Poecilia wingei is closely related to the common (Trinidadian) preferred (Head, Lindholm, & Brooks, 2007). Sneaking can also be guppy P. reticulata (Schories et al., 2009) and endemic to the Cumaná less time consuming, increasing opportunities for further mate Region of Venezuela. Despite being recognized as distinct from searching (Parker, 1974). Attempting to mate through cooperative, P. reticulata since 1975, it has only recently been formally described courtship-based mating will be more favourable when a male is (Poeser, Kempkes, & Insbrücker, 2005). The two species can be attractive and courtship is expected to lead to successful copulation. readily distinguished by morphological traits (e.g. colour pattern), In some mating systems at least, males transfer more sperm during structure of the gonopodium and some features of their behaviour cooperative mating (Pilastro & Bisazza, 1999) and there is a greater (Houde, 1997; Poeser et al., 2005), but share most behavioural likelihood of successful female parturition after cooperative rather characteristics. After taking into account the behavioural differences than forced mating (Sato, Shimoichi, & Karino, 2011). described by Poeser et al. (2005), which reflect our own compre- We used Poecilia wingei, a small live-bearing poecilid fish with hensive observations, it is clear that data on P. wingei may be internal fertilization, to examine whether male mating behaviour is compared with studies on P. reticulata with confidence. affected by social environment. This guppy is a sister species of the Poecilia wingei are sexually dimorphic (Poeser et al., 2005), Trinidadian guppy, Poecilia reticulata, from coastal northern males being brightly coloured and females dull. Males develop a Venezuela (Schories, Meyer, & Schartl, 2009). The two species share gonopodium, a modified anal fin that serves to transfer sperm many aspects of reproductive behaviour, which has been well during insemination. Their mating system is promiscuous and studied and documented for P. reticulata (Houde, 1997), making nonresource-based (Poeser et al., 2005). Females are courted via a quantification convenient. Male guppies may either court females sigmoid display (Poeser et al., 2005), whereby the male arches his through sigmoid displays and mate cooperatively or sneak copulate body into a sigmoid shape then quickly undulates his body and through gonopodial thrusts (Houde, 1997). exposes his bright coloration and spread fins (Houde, 1997). If the In P. reticulata, there is evidence for a genetic basis to male sexual male is successful, the female allows him to copulate cooperatively behaviour (Evans, 2010), but also for an effect of different rearing and enables him to insert his gonopodium into her gonopore. The environments (Evans & Magurran, 1999; Guevara-Fiore, 2012; but pair then rotate around each other to complete insemination see Magellan & Magurran, 2009), and ecological and environmental (Houde, 1997). Males also use sneaking tactics, whereby they chase factors (Devigili, Kelley, Pilastro, & Evans, 2012; Rodd & Sokolowski, females with their gonopodium in the upright position and try to 1995) on the adoption of particular mating tactics or mating copulate coercively, with no prior courtship (termed gonopodial behaviour patterns. The effect of social environment on the adoption thrusting). Male tactics are flexible and there is no indication of of male mating tactics, however, remains largely untested. There is genetic determination to these mating tactics (Houde, 1997). Fe- some evidence that males are able to modify mating tactics in males can be fertilized by gonopodial thrusts at any stage of their response to their social environment (Barbosa, Ojanguren, & reproductive cycle (Farr, 1980) as sperm viability in the female Magurran, 2013; Jirotkul, 1999b), but how they respond to recent gonoducts is extended (Houde, 1997; López-Sepulcre, Gordon, social history is not clear (Evans & Magurran, 1999). The first sexual Paterson, Bentzen, & Reznick, 2013). Behavioural elements typi- experience (Guevara-Fiore, Svensson, & Endler, 2012), contact with cally preceding copulation attempts include gonopodial swings different numbers and quality of females (Jordan & Brooks, 2012) (male moves his gonopodium repeatedly up and down), apparently and comparative evaluation of potential mates (Bateson & Healy, serving to load the gonopodium with sperm (Liley, 1966), and 2005) are known to modify male sexual behaviour. Our study is gonopore nipping, where the male nips the female’s gonopore with novel in testing the effect of long-term exposure to biased sex ratios his mouth. The function of gonopore nipping is unclear but it has on male mating behaviour. been suggested that it allows males to assess female olfactory cues We investigated the effect of a varying social environment on (Herdman, Kelly, & Godin, 2004). Male guppies spend most of their adult, sexually naïve males. First, we measured whether individual, time following and courting females and trying to copulate either sexually naïve males had a preferred tactic, as suggested by cooperatively or coercively (Magurran & Seghers, 1994). Magurran and Magellan (2007) for P. reticulata and, if so, whether this tactic was modified by social experience. We examined how Subjects and Housing male mating behaviour was affected by recent social history, manipulated by housing adult males at a particular sex ratio (male- Wild P. wingei were collected in 2007 from the Laguna de Los or female-biased) for a period of 5 weeks. Our specific aims were (1) Patos (northern Venezuela), the species’ original site of occurrence. to investigate whether individual sexually naïve males have a The breeding stock was kept as a large outbred population (100 preferred mating tactic (courting or sneaking), and (2) how male adult fish) in a 120-litre tank (75 40 cm and 40 cm deep) in the mating behaviour is affected by recent social history through aquarium facility at the Institute of Vertebrate Biology, Brno, Czech exposure to a male- or female-biased social environment. We pre- Republic, and fed twice each day with commercial flake food and dicted that individual sexually naïve males would already have a frozen chironomid larvae. The aquarium was exposed to natural preferred mating tactic and that this would be modified by social daylight for 10e14 h each day from a glass ceiling and water tem- experience. Specifically, males experiencing a male-biased social perature was maintained at 25 2 C. Aquarium water was aerated R. Rezucha, M. Reichard / Animal Behaviour 88 (2014) 195e202 197 continuously and its quality maintained through an air-driven foam The fish were gently immobilized using the sponge and two pho- filter and regular water changes. Live plants (Taxiphyllum barbieri tographs were quickly taken. The fish were then returned to their and Najas guadalupensis) were provided as shelter. Over 37 days (7 original tanks. No distress was observed during photography. March to 14 April 2011), juveniles were collected from the stock Upon completion of all first trials, two recently mated females aquaria and isolated from adults in a separate 72-litre aquarium from the stock population were randomly assigned to 33 tanks (60 30 cm and 40 cm deep). The juveniles were frequently containing a focal male to create a female-biased social environ- checked and females (recognized by the development of a black spot ment consisting of one male and two females. In the remaining 32 around the gonopore) were moved to a separate tank from the focal male tanks, a single male and a single female from the stock males. population were added to create a male-biased social environment of two males and one female. Mortalities reduced the sample size to Experimental Set-up and Behavioural Assays 32 males in female-biased environment and 27 males in male- biased environment. At the age of 3e4 months, 65 males, readily distinguished from The second male trials were completed after 5 weeks of expo- females by the presence of a gonopodium and nuptial coloration, sure to the male- and female-biased social environments. Males were placed in individual 2-litre plastic tanks (16 10 cm and had unrestricted access to females (and rivals), with full visual, 12 cm deep). The tanks were visually isolated using opaque barriers physical and olfactory contact, and gained mating experience and exposed to a natural light regime, with additional light pro- within their particular sex-biased social environment. While some vided by a 40 W daylight spectrum Sun Glo fluorescent tube to focal male mortality was recorded during the treatment phase, this maintain 12 h (0800e2000 hours) of light exposure. Water tem- was not associated with the experimental treatments and simply peratures were maintained at 22e25 C and each tank was equip- resulted in a smaller sample size during the second trials. Trial 2 ped with an artificial plant as a shelter. Males were fed once each testing (after exposure to divergent sex ratios for 5 weeks) followed day on commercial flake food and the water was changed every 2 the same protocol, except that focal males were isolated from all weeks. Males were housed under these conditions for 2 months. conspecifics for 1.5 days prior to testing in order to enable the males During this period no data on their behaviour were recorded. replenish their sperm reserves and standardize mating effort. The first tests of male mating behaviour were performed after 2 Previous work on P. reticulata suggests that 1 day should be suffi- months. Fully sexually mature males were tested to limit any cient to restore male sperm reserves (Kuckuck & Greven, 1997 in confounding effects of hormonal changes during adolescence, as Evans, Pierotti, & Pilastro, 2003). Behavioural observations took adolescent individuals have been shown to be less sensitive to place when males were 7e8 months old. All observations (Trials 1 hormonal sex pheromones (Irvine & Sorensen, 1993). Males have and 2) were completed between 0900 and 1700 hours. been shown to modify their mating behaviour in response to fe- Upon completion of the second trials, males were returned to male olfactory cues (Guevara-Fiore, Stapley, Krause, Ramnarine, & their original tanks until natural death, their survival time being Watt, 2010). Sexually naïve males, with no previous experience or recorded. Social environment treatment conditions were main- contact with receptive adult females, were tested to determine tained by replacing any dead females or nonexperimental males. their baseline sexual behaviour. This procedure was imposed because there is evidence that court- Mating behaviour of sexually naïve males (Trial 1) was scored in ship effort affects longevity (Miller & Brooks, 2005) and different a 6-litre tank (25 25 cm and 10 cm deep) with the back and sides mating tactics have different energetic demands (Cummings & covered with black plastic sheets to minimize disturbance. Illumi- Gelineau-Kattner, 2009). nation was provided by a 25 W daylight spectrum Sun Glo lamp positioned above the aquarium, ensuring even lighting throughout Scoring Behaviour the tank. Water was aerated between trials, but not during the behavioural observations. Trial 1 observations were conducted A female was allowed to settle for 5 min in the test tank. A when males were 5e6 months old. This observation allowed the randomly chosen focal male was then gently captured, added to the comparison of male post-treatment behaviour using a paired test tank and left for 10 min to settle. Male mating behaviour was design (male identity treated as random factor). then scored for 15 min by a single observer using JWatcher 1.0 Experimental females were taken from a tank of virgin females software for behavioural scoring (Blumstein & Daniel, 2007). The and left for 1 day prior to experimental trials with a group of three following male behaviour was scored: the number of sigmoid dis- adult nonexperimental males from the stock aquarium in order to plays (courtship); gonopodial thrusts (sneaking); gonopodial standardize female sexual receptivity and ensure that females were swings (loading gonopodium with sperm); and gonopore nipping nonvirgin and in the same reproductive state. We did not use virgin (assessment of female olfactory cues). We further quantified gen- females, as encountering already mated females is a more realistic eral male interest in the female as the time (s) devoted to interest in scenario for males under natural conditions. Male P. reticulata can the female. We defined male interest as occurring when a male’s readily distinguish between virgin and even recently mated females head was oriented towards the female and there was a maximum and modulate their mating tactics accordingly (Guevara-Fiore, distance of 10 cm between the male and female. This behaviour Skinner, & Watt, 2009), and we anticipated that male P. wingei had covered the time taken for the male to inspect the female, assess the same capacity. Each female was tested with four experimental cues related to her reproductive state and her willingness to males. Tank water was not exchanged between replicates and olfac- copulate, and the time taken preparing for a courtship or copulation tory cues were not considered. Focal males were not fed for 1 day attempt. After 15 min, the focal male was gently recaptured and prior to testing to standardize their metabolic level and sexual returned to his original tank. Fish showed no signs of stress and no motivation. aggressive behaviour or injury was observed during the trials. All After completion of each trial, focal males were photographed individuals were fed ad libitum after each trial. from both sides under standardized lighting conditions using a Canon EOS Rebel XTi camera fitted with a Sigma Macro 100 mm Data Analysis lens to enable identification through their unique colour patterns. During photography, fish were released into the photographic tank We tested all behavioural measurements separately as they and gently moved to the front part of the tank using a soft sponge. describe different aspects of the male mating tactic. Given 198 R. Rezucha, M. Reichard / Animal Behaviour 88 (2014) 195e202 significant correlations between several behaviours, we addition- Research. Experimental procedures were approved by the ethical ally used principal component analysis (PCA) to construct com- committees of the IVB and the Ministry of Agriculture (CZ posite variables. For the PCA, we did not include male interest as it 62760203) and are in accordance with Czech legal requirements. is already a composite variable summing all male behaviours directed towards a female and would be redundant in the PCA. Data RESULTS from both pre- and post-treatment sets were pooled in a PCA analysis as a single matrix to allow comparison of temporal changes Associations between Mating Behaviours in male behaviour. Axes with eigenvalues >1 were retained for further analysis. The analysis was performed using STATISTICA 10 The PCA on male behaviour data produced two axes with ei- software (StatSoft, Inc., Tulsa, OK, U.S.A.). genvalues >1(Table 1). The first (PCA 1) explained 43% of variation The relationship between treatment and temporal change in and was strongly negatively associated with gonopodial thrusts each male behaviour was analysed using the general linear mixed (sneaking attempts) and gonopodial swings (loading sperm re- model from the nlme package (normal distribution: general male serves into the gonopodium). The second (PCA 2) explained 27% of interest, exp-transformed; PCA1; PCA2), and the generalized linear variation and was strongly negatively associated with sigmoid mixed model in the lme4 package (Poisson distribution with log- displays (cooperative mating). link function: number of sigmoid displays, gonopodial thrusts, Univariate analysis of sexually naïve males (Trial 1) indicated a gonopodial swings and gonopore nipping per 15 min) from the R negative correlation between number of sigmoid displays and environment (R Development Core Team, 2009). This approach gonopodial thrusts (Spearman correlation: rS ¼0.28, N ¼ 59, enabled us to account for the paired design (male identity treated P ¼ 0.035), suggesting that, in the absence of any sexual experience, as a random factor) and repeated use of the same females. The individual males tended to approach females by either courting or independent variables were Trial (pre- and post-treatment), sneaking, although the variation was not discrete. Males with a Treatment (male- and female-biased environment) and their stronger general interest in females performed more gonopore nips interaction. The significance of the interaction would indicate that (Spearman correlation: rS ¼ 0.34, N ¼ 59, P ¼ 0.008) and gon- the two treatments influenced males differently. To reveal temporal opodial thrusts (rS ¼ 0.37, N ¼ 59, P ¼ 0.004). Sigmoid displays changes in behaviour within treatments, we re-ran the analyses were positively associated with gonopodial swings (rS ¼ 0.59, from each treatment separately (within male- and female-biased N ¼ 59, P < 0.001). environments) and compared behaviours before the treatment After treatments were imposed (Trial 2), male interest remained was imposed (standard social environment) against the same positively associated with the number of gonopore nips in female- behaviour after exposure in female- or male-biased treatments. biased environments and with gonopodial thrusts in male-biased Finally, we compared male survival between the two treatments environments (Table 2). Additionally, a positive correlation be- using the log-rank test. tween male interest and sigmoid displays was observed in female- biased environments and, in both treatments, gonopodial swings Ethical Note were positively associated with gonopodial thrusts rather than sigmoid displays (Table 2). Fish were housed according to approved institutional and governmental experimental procedures and all authors hold a licence for conducting experimental work on vertebrates. No Treatment Effects on Male Mating Behaviour distress or aggression was observed over the course of the study; all behaviour concurred with natural behaviour of the study species. The effect of social environment on male sexual behaviour was Mortalities were observed during the progress of the experiment tested using general (composite variables) and generalized mixed but were not associated with experimental procedures; tracking models, where interaction between Trial and Treatment factor the survival of experimental fish formed part of the experimental indicated whether social environment affected male behaviour protocol. Life span in the experiment was considerably higher than differently. As significant interactions indicated that social envi- in natural populations (López-Sepulcre et al., 2013). We minimized ronment did indeed affect male sexual behaviour (Table 3, Figs 1, 2, the number of fish used in the study through a trade-off between Supplementary Material S1), we then analysed the temporal sample size and statistical power, our sample size being chosen pattern of change in all behaviours separately for each social according to comparable studies on Trinidadian guppies. The sub- environment (Table 3). Males from male-biased treatments showed jects came from a captive population. The treatment tank envi- increased attention to females, largely resulting from a high rate of ronment was enriched by the use of live and artificial plants and all gonopodial thrusting (Table 3, Fig. 1). In contrast, males from fish were exposed to natural daylight and an optimum water female-biased treatments showed more limited attention to fe- temperature. males and expressed lower copulation-oriented behaviour The original stock was collected by a conservation breeder of compared to the level prior to the imposition of the experimental wild populations of poecilid fishes. Fish for our breeding stock were treatment (Table 3, Fig. 1). As a result, males from male-biased obtained noncommercially, directly from the collector. The fish were collected by hand net, thereby minimizing any impact on Table 1 fi other native organisms. The sh were imported in large plastic bags Proportion of variance explained (in %), eigenvalues (in parentheses) and factor at a density of no more than four fish per bag. No mortalities were loadings for the first two principal components from behavioural data (baseline and recorded during import. In captivity, fish were always kept in post-treatment data pooled) environmentally enriched tanks (e.g. gravel, live aquatic plants) and PC 1 PC 2 fed ad libitum. The sexes were not separated. When moved be- fi Variance (eigenvalues) 42.8 (1.71) 26.9 (1.08) tween tanks, sh were gently captured using an aquarium hand net Gonopodial swings L0.760 0.231 and placed in a plastic jar containing aquarium water and released Gonopore nipping 0.568 0.001 into the new tank. Fish were observed until their natural death, Sigmoid displays 0.131 L0.953 which was part of the experimental protocol. The research protocol Gonopodial thrusts L0.806 0.372 adheres to the ASAB/ABS Guidelines for the Use of Animals in Two other components had eigenvalues <1. Significant factors are shown in bold. R. Rezucha, M. Reichard / Animal Behaviour 88 (2014) 195e202 199

Table 2 Spearman rank correlation associations between individual behaviours following five weeks of social environment treatment

Gonopodial swings Gonopore nipping Sigmoid displays Gonopodial thrusts

Bias in environment Female biased Male biased Female biased Male biased Female biased Male biased Female biased Male biased

Male interest 0.317 0.380 0.553 0.022 0.428 0.098 0.373 0.661 Gonopodial swings 0.244 0.366 0.216 0.260 0.670 0.547 Gonopore nipping 0.259 0.209 0.058 0.042 Sigmoid displays 0.191 0.130

Female-biased environment (N ¼ 24 males); male-biased environment (N ¼ 17 males). Significant correlations (P < 0.05) are shown in bold. treatments showed higher rates in all copulation-oriented behav- was likely to be innate, rather than learned through interactions iours (except gonopore nipping), related to either cooperative or with other males. forced copulations (Table 3, Figs 1, 2). For composite variables (PC1 We found that males exhibiting greater general interest in fe- and PC2), significant variables were negatively loaded to the PCs; males also performed more gonopore nipping. As gonopore hence in Fig. 2 an increase denotes a decrease in the particular nipping appears to be used for detecting olfactory cues released by behaviour and vice versa. females (Herdman et al., 2004), males spent an extended period assessing female reproductive state. As males can discriminate between virgin and recently mated females (Guevara-Fiore et al., Treatment Effect on Male Longevity 2009), they may alter their reproductive effort and mating tactic (sneaking or courting) on the basis of these female cues (Guevara- fi There was no signi cant difference in longevity between males Fiore, Stapley, & Watt, 2010). Males that invested in courtship from male-biased and female-biased treatments (log-rank test: (sigmoid displays) prior to copulation frequently performed gon- ¼ P 0.836). Median survival was 274 days and 90% survival was 451 opodial swings, perhaps to load the gonopodium with sperm (Liley, days (Fig. 3). 1966). The adoption of gonopodial swings by males may increase the amount of sperm immediately available prior to a cooperative copulation with females, enabling them to transfer a larger volume DISCUSSION of sperm (Pilastro & Bisazza, 1999). We experimentally demonstrated the effect of social environ- We studied the effect of social environment on mating tactics of ment on the reproductive behaviour of male P. wingei. Males male Endler’s guppies. We demonstrated that even sexually naïve experiencing female-biased environments, with no competition for males, with no prior mating experience, showed a tendency to mating from other males, decreased their mating effort in terms of approach females by either sneaking or courting behaviour, all four premating behaviours measured, despite maintaining the although variation was continuous rather than discrete. This same level of general interest in terms of time spent near the focal finding corroborates previous evidence that alternative mating female. Males from the female-biased environments preferentially behaviour in guppy males has a genetic component (Evans, 2010). used sigmoid displays to court females and their attempts to sneak Although guppy males court each other in the absence of females copulations by gonopodial thrusts were relatively rare. In contrast, (Field & Waite, 2004), interaction with an adult female, along with males from male-biased environments used opportunities to mate the female’s response, is necessary for the adoption of the most by attempting sneak copulations; however, they also maintained efficient mating tactic leading to successful copulation. Indeed, relatively high levels of sigmoid displays. inexperienced males often do not copulate successfully during their Low interest in copulation and associated mating behaviour of first sexual encounter (Japanese quail, Coturnix japonica: Cornil & males from female-biased environments was particularly strong. Ball, 2010) and males reach their maximum sexual performance This finding supports theoretical predictions that males should only after several copulations with females, irrespective of poten- learn to become choosy if encounter rates with females are high tial previous homosexual attempts (goat, Capra hircus: Imwalle & (Dukas, Clark, & Abbott, 2006). These males were not sperm Katz, 2004). These findings demonstrate the crucial role of female limited, as the level of sperm competition risk has no effect on behavioural cues on effective male mating behaviour in some taxa. sperm production in poeciliids (Evans et al., 2003) and males had The behaviour of sexually naïve males in the first trial, therefore,

Table 3 The effect of social treatment on male mating behaviour, analysed as the interaction between treatment level and trial (Comparison), and temporal changes in behaviour between baseline and post-treatment trials (Male-biased, Female-biased), analysed as temporal pattern within each treatment

Comparison* Male-biased Female-biased

Statistics P Diff Statistics P Trend Statistics P Trend

Composite variables Male interest 2.27F 0.140 3.08t 0.007 Increase 0.76t 0.457 PCA1 9.39F 0.004 M>F 0.35t 0.730 4.06t 0.001 Increase PCA2 2.94F 0.094 (M>F) 1.28t 0.220 1.21t 0.24 Individual behaviours Gonopodial swings 4.81z <0.001 M>F 1.71z 0.083 (Increase) 5.05z <0.001 Decrease Gonopore nipping 1.27z 0.204 0.28z 0.778 2.38z 0.017 Decrease Sigmoid displays 4.77z <0.001 M>F 1.43z 0.152 8.72z <0.001 Decrease Gonopodial thrusts 7.15z <0.001 M>F 2.26z 0.024 Increase 6.99z <0.001 Decrease

Test statistic (F, t or z value, as in superscript); Trend ¼ temporal trend (increase or decrease in behaviour rate between baseline Trial 1 and post-treatment Trial 2); Diff ¼ direction of difference in behaviour between male- and female-biased social environment treatments (M: male-biased, F: female-biased). * For Comparison, only the interaction term is reported. For full outcomes of the statistical test, see Supplementary Material S1. 200 R. Rezucha, M. Reichard / Animal Behaviour 88 (2014) 195e202

(a) (b) 14 2.5

12 2 10

8 1.5

6 1 4 0.5 2 Gonopore nipping (rate per 15 min) Gonopore Gonopodial swings (rate per 15 min) swings Gonopodial 0 0 Male Female Male Female

(c) (d) 14 8

12 6 10

8 4 6

4 2

Sigmoid displays (rate per 15 min) displays Sigmoid 2 Gonopodial thrusts (rate per 15 min) Gonopodial 0 0 Male Female Male Female Bias in environment

Figure 1. Temporal changes and differences between treatments in sexual behaviour trials (baseline: open circles; post-treatment: closed circles) in male- and female-biased environments. Error bars represent 1 SE. See Table 3 for the statistical significance of comparisons. Data on ‘male interest’ were exp-transformed prior to analysis to normalize residuals, the transformed data being presented in the figure. sufficient time to replenish their sperm reserves prior to trials. & Birkhead, 2003), including fish (Kelley, Graves, & Magurran, 1999; Decreased mating effort was detected despite the use of a new, Spence, Reichard, & Smith, 2013). unfamiliar female, running counter to the predictions for a Coolidge Unfortunately, such a strong decrease in mating effort was not effect (Dewsbury, 1981). The Coolidge effect describes a decrease in anticipated and our experimental design did not allow for deter- sexual interest in one female with sexual interest being revived by a mination of potential partner compatibility related to, for example, new female, a process that has been demonstrated across many MHC genes, which are known to play an important role in mate taxa (Koene & Ter Maat, 2007; Pizzari, Cornwallis, Løvlie, Jakobsson, choice in fishes (Agbali, Reichard, Bryjová, Bryja, & Smith, 2010;

(a) 1 (b)1 (c) 2.5

2 0.5 0.5

1.5 PC 1 0 PC 2 0 1 –0.5 –0.5 0.5

0 –1 –1 Male interest (rate per 15 min) Male Female Male Female Male Female Sigmoid displays Gonopodial swings Gonopodial thrusts Bias in environment

Figure 2. Temporal changes and differences between treatments in sexual behaviour trials for composite variables (baseline: open circles; post-treatment: closed circles) in male- and female-biased environments. See Table 3 for the statistical significance of comparisons. Text in arrows indicates significantly loaded variables and the direction of their as- sociation with the PC axis. R. Rezucha, M. Reichard / Animal Behaviour 88 (2014) 195e202 201

1 ratio. In another poecilid, Gambusia holbrooki, males increased their mating effort in the presence of male competitors, and the effect 0.9 Male-biased held even after removal of competitors (Evans et al., 2003). Our 0.8 Female-biased results corroborate the importance of previous experience in 0.7 modulation of sexual behaviour. We hypothesize that the use of previous experience can benefit males by saving time from 0.6 repeated assessments of social condition, but can become costly in 0.5 a rapidly changing social environment. 0.4 Courtship effort may affect longevity (Miller & Brooks, 2005). It is energetically costly (Cady, Delaney, & Uetz, 2011; Kotiaho, 2000) 0.3 and males are known to have a higher mortality during the mating 0.2 season (Hunt et al., 2004; McKellar, Turcotte, & Hendry, 2009). 0.1 Furthermore, males performing different mating tactics pay Cumulative proportion surviving 0 different energetic costs for interactions with females (Cummings & 0 50 100 150 200 250 300 350 400 450 500 550 Gelineau-Kattner, 2009). We observed no effect of sex-biased Days environment and associated mating effort on longevity or mortal- ity patterns. It is possible, however, that survival costs of mating Figure 3. Survival of focal males for each experimental treatment, visualized by effort would become detectable in more challenging (predation e Kaplan Meier estimates. Triangles: male-biased treatment; circles: female-biased risk, food limitation) environments (Magnhagen, 1991), rather than treatment. under benign laboratory conditions. In conclusion, we demonstrated that individual males displayed Reusch, Häberli, Aeschlimann, & Milinski, 2001). Individual gup- either sneaking or courting behaviour, even during their first pies, however, do not discriminate between siblings and unrelated mating attempts. We further established that previous experience, individuals during mate choice or courtship (Pitcher, Rodd, & Rowe, in the form of a sex-biased social environment, strongly affected 2008; Viken, Fleming, & Rosenqvist, 2006), reducing the possibility male sexual behaviour in standardized mating trials. Males with of a strong mate incompatibility effect in our experiment. Males experience of female-biased social environments decreased their from male-biased environments exposed to strong maleemale mating effort considerably, but were more likely to court females competition for access to a female used every opportunity to mate, rather than sneak copulations compared to males from male-biased resulting in an increased rate of sneak attempts (gonopodial social environments which maintained a high level of mating effort thrusts), which was in accordance with our predictions. by performing frequent sneak mating attempts through gonopodial Our study was primarily designed to compare the effects of thrusts. These results underline the importance of previous expe- different sex ratios on male behaviour. However, our experimental rience in the expression of sexual behaviour. design inevitably included variation in other factors and these factors may have contributed to the results. For example, the male- Acknowledgments biased treatment excluded all femaleefemale interactions while the female-biased treatment excluded all maleemale interactions We are grateful to Matej Polacik, Milan Vrtílek and Radim Blazek during the 5-week period prior to testing. Furthermore, the quan- for their help with the experiments, practical advice and com- titative and qualitative characteristics of maleefemale interactions ments, Carl Smith, Culum Brown and two anonymous referees for probably varied between treatments. While this was a natural constructive comments on the manuscript and Jirí Farkac for help outcome of our experimental design, we did not control for their with fish maintenance. This work was supported by the Czech behavioural effects, for example the effect of modulation of female Science Foundation (206/09/1163). behaviour and its feedback to male mating behaviour (Kahn, Dolstra, Jennions, & Backwell, 2013). Males frequently modify their mating effort in accordance with female responsiveness (fish: Supplementary Material Casalini, Reichard, Phillips, & Smith, 2013; fiddler crabs: Kahn et al., 2013). Thus, although we tested males with standardized, unfa- Supplementary material associated with this article is available, miliar females, we acknowledge that male mating behaviour during in the online version, at http://dx.doi.org/10.1016/j.anbehav.2013. the second assay may have been affected by all these unmeasured 12.010. variables and their potential interactions. ’ There is clear evidence that a male s mating behaviour is References influenced by its environment and previous experience (Bretman et al., 2011; Freeberg, 2000; Rodríguez et al., 2013). Our results Agbali, M., Reichard, M., Bryjová, A., Bryja, J., & Smith, C. (2010). Mate choice for corroborate previous findings of Jordan and Brooks (2012), who nonadditive genetic benefits correlate with MHC dissimilarity in the rose bit- terling (Rhodeus ocellatus). Evolution, 64, 1683e1696. also found that guppy males modify their sexual behaviour ac- Andersson, M. (1994). Sexual selection. Princeton, New Jersey: Princeton University cording to previous mating experience. They found that guppy Press. males exposed to multiple females during the conditioning period Barbosa, M., Ojanguren, A. F., & Magurran, A. E. (2013). Courtship display persists e subsequently invested less effort in courting a new female than despite early social deprivation. Ethology, 119, 496 502. Bateson, M., & Healy, S. D. (2005). Comparative evaluation and its implications for males housed with only one female at a time. Guppy males housed mate choice. Trends in Ecology & Evolution, 20, 659e664. in male-only environments, on the other hand, have been recorded Blumstein, D. T., & Daniel, J. C. (2007). Quantifying behavior the JWatcher way. performing homosexual behaviour (courting other males and Sunderland, Massachusetts: Sinauer Associates, Inc. Bretman, A., Gage, M. J. G., & Chapman, T. (2011). Quick-change artists: male plastic attempting to sneak copulate) that persisted even in the presence behavioural responses to rivals. Trends in Ecology & Evolution, 26,467e473. of females (Field & Waite, 2004). Similarly, Evans and Magurran Cady, A. B., Delaney, K. J., & Uetz, G. W. (2011). Contrasting energetic costs of (1999) found that guppy males reared in a male-biased social courtship signaling in two wolf spiders having divergent courtship behaviors. Journal of Arachnology, 39,161e165. environment were more inclined to perform gonopodial thrusts, Candolin, U. (2004). Effects of algae cover on egg acquisition in male three-spined despite being briefly housed in an environment with an equal sex stickleback. Behaviour, 141, 1389e1399. 202 R. Rezucha, M. Reichard / Animal Behaviour 88 (2014) 195e202

Casalini, M., Reichard, M., Phillips, A., & Smith, C. (2013). Male choice of mates and Kotiaho, J. S. (2000). Testing the assumptions of conditional handicap theory: costs mating resources in the rose bitterling (Rhodeus ocellatus). Behavioral Ecology, and condition dependence of a sexually selected trait. Behavioral Ecology and 24,1199e1204. Sociobiology, 48,188e194. Chapman, B. B., Morrell, L. J., & Krause, J. (2009). Plasticity in male courtship Kuckuck, C., & Greven, H. (1997). Notes on the mechanically stimulated discharge of behaviour as a function of light intensity in guppies. Behavioral Ecology and spermiozeugmata in the guppy, Poecilia reticulata: a quantitative approach. Sociobiology, 63, 1757e1763. Zeitschrift für Fischkunde, 4,73e88. Cornil, C. A., & Ball, G. F. (2010). Effects of social experience on subsequent sexual Liley, N. R. (1966). Ethological isolating mechanisms in four sympatric species of performance in naïve male Japanese quail (Coturnix japonica). Hormones and poeciliid fishes. Behaviour, 13,1e197. Behavior, 57, 515e522. López-Sepulcre, A., Gordon, S. P., Paterson, I. G., Bentzen, P., & Reznick, D. N. (2013). Cummings, M. E., & Gelineau-Kattner, R. (2009). The energetic costs of alternative Beyond lifetime reproductive success: the posthumous reproductive dynamics male reproductive strategies in Xiphophorus nigrensis. Journal of Comparative of male Trinidadian guppies. Proceedings of the Royal Society B: Biological Sci- Physiology A, 195, 935e946. ences, 280, 20131116. Devigili, A., Kelley, J. L., Pilastro, A., & Evans, J. P. (2012). Expression of pre- and Luyten, P. H., & Liley, N. R. (1985). Geographic variation in the sexual behaviour of postcopulatory traits under different dietary conditions in guppies. Behavioral the guppy, Poecilia reticulata (Peters). Behaviour, 95, 164e179. Ecology, 24,740e749. Magellan, K., & Magurran, A. E. (2006). Habitat use mediates the conflict of interest Dewsbury, D. A. (1981). Effects of novelty of copulatory behavior: the Coolidge between the sexes. Animal Behaviour, 72,75e81. effect and related phenomena. Psychological Bulletin, 89,464e482. Magellan, K., & Magurran, A. E. (2009). The effect of social environment during Dukas, R., Clark, C. W., & Abbott, K. (2006). Courtship strategies of male insects: ontogeny on life history expression in the guppy Poecilia reticulata. Journal of when is learning advantageous? Animal Behaviour, 72, 1395e1404. Fish Biology, 74, 2329e2337. Evans, J. P. (2010). Quantitative genetic evidence that males trade attractiveness for Magnhagen, C. (1991). Predation risk as a cost of reproduction. Trends in Ecology & ejaculate quality in guppies. Proceedings of the Royal Society B: Biological Sci- Evolution, 6,183e186. ences, 277,3195e3201. Magurran, A. E., & Magellan, K. (2007). Behavioural profiles: individual consistency Evans, J. P., & Magurran, A. E. (1999). Male mating behaviour and sperm production in male mating behaviour under varying sex ratios. Animal Behaviour, 74, 1545e characteristics under varying sperm competition risk in guppies. Animal 1550. Behaviour, 58,1001e1006. Magurran, A. E., & Seghers, B. H. (1994). A cost of sexual harassment in the guppy, Evans, J. P., Pierotti, M., & Pilastro, A. (2003). Male mating behavior and ejaculate Poecilia reticulata. Proceedings of the Royal Society of London. Series B: Biological expenditure under sperm competition risk in the eastern mosquitofish. Sciences, 258(1351), 89e92. Behavioral Ecology, 14, 268e273. McKellar, A. E., Turcotte, M. M., & Hendry, A. P. (2009). Environmental factors Farr, J. A. (1980). The effects of sexual experience and female receptivity on influencing adult sex ratio in Trinidadian guppies. Oecologia, 159,735e745. courtship-rape decisions in male guppies, Poecilia reticulata (Pisces: Poecilii- Miller, L. K., & Brooks, R. (2005). The effects of genotype, age, and social environ- dae). Animal Behaviour, 28,1195e1201. ment on male ornamentation, mating behavior, and attractiveness. Evolution, Field, K. L., & Waite, T. A. (2004). Absence of female conspecifics induces homo- 59,2414e2425. sexual behaviour in male guppies. Animal Behaviour, 68, 1381e1389. Mills, S. C., & Reynolds, J. D. (2003). Operational sex ratio and alternative repro- Freeberg, T. M. (2000). Culture and courtship in vertebrates: a review of social ductive behaviours in the European bitterling, Rhodeus sericeus. Behavioral learning and transmission of courtship systems and mating patterns. Behav- Ecology and Sociobiology, 54,98e104. ioural Processes, 51,177e192. Oliveira, R. F., Taborsky, M., & Brockmann, H. J. (Eds.). (2008). Alternative repro- Godin, J. G. J. (1995). Predation risk and alternative mating tactics in male Trini- ductive tactics: an integrative approach (1st ed.). New York: Cambridge Univer- dadian guppies (Poecilia reticulata). Oecologia, 103, 224e229. sity Press. Guevara-Fiore, P. (2012). Early social experience significantly affects sexual behav- Parker, G. A. (1974). Courtship persistence and female-guarding as male time in- iour in male guppies. Animal Behaviour, 84,191e195. vestment strategies. Behaviour, 48,157e184. Guevara-Fiore, P., Skinner, A., & Watt, P. J. (2009). Do male guppies distinguish Parker, G. A., Ball, M. A., Stockley, P., & Gage, M. J. G. (1997). Sperm competition virgin females from recently mated ones? Animal Behaviour, 77, 425e431. games: a prospective analysis of risk assessment. Proceedings of the Royal So- Guevara-Fiore, P., Stapley, J., Krause, J., Ramnarine, I. W., & Watt, P. J. (2010). Male ciety of London. Series B: Biological Sciences, 264, 1793e1802. mate-searching strategies and female cues: how do male guppies find receptive Pilastro, A., & Bisazza, A. (1999). Insemination efficiency of two alternative male females? Animal Behaviour, 79,1191e1197. mating tactics in the guppy Poecilia reticulata. Proceedings of the Royal Society of Guevara-Fiore, P., Stapley, J., & Watt, P. J. (2010). Mating effort and female recep- London. Series B: Biological Sciences, 266, 1887e1891. tivity: how do male guppies decide when to invest in sex? Behavioral Ecology Pitcher, T. E., Rodd, F. H., & Rowe, L. (2008). Female choice and the relatedness of and Sociobiology, 64, 1665e1672. mates in the guppy (Poecilia reticulata): mate choice and inbreeding depression. Guevara-Fiore, P., Svensson, P. A., & Endler, J. A. (2012). Sex as moderator of early life Genetica, 134,137e146. experience: interaction between rearing environment and sexual experience in Pizzari, T., Cornwallis, C. K., Løvlie, H., Jakobsson, S., & Birkhead, T. R. (2003). So- male guppies. Animal Behaviour, 84, 1023e1029. phisticated sperm allocation in male fowl. Nature, 426,70e74. Head, M. L., Lindholm, A. K., & Brooks, R. (2007). Operational sex ratio and density Poeser, F. N., Kempkes, M., & Isbrücker, I. J. H. (2005). Description of Poecilia do not affect directional selection on male sexual ornaments and behavior. (Acanthophacelus) wingei n. sp. from the Paría Peninsula, Venezuela, including Evolution, 62,135e144. notes on Acanthophacelus Eigenmann, 1907 and other subgenera of Poecilia Herdman, E. J. E., Kelly, C. D., & Godin, J. J. (2004). Male mate choice in the guppy Bloch and Schneider, 1801 (Teleostei, Cyprinodontiformes, Poeciliidae). Contri- (Poecilia reticulata): do males prefer larger females as mates? Ethology,110,97e111. butions to Zoology, 74,97e115. Houde, A. E. (1997). Sex, color, and mate choice in guppies. Princeton, NJ: Princeton R Development Core Team. (2009). R: A language and environment for statistical University Press. computing. Vienna, Austria: R Foundation for Statistical Computing. Hunt, J., Brooks, R., Jennions, M. D., Smith, M. J., Bentsen, C. L., & Bussière, L. F. Reichard, M., Smith, C., & Jordan, W. C. (2004). Genetic evidence reveals density- (2004). High-quality male field crickets invest heavily in sexual display but die dependent mediated success of alternative mating behaviours in the Euro- young. Nature, 432, 1024e1027. pean bitterling (Rhodeus sericeus). Molecular Ecology, 13, 1569e1578. Imwalle, D. B., & Katz, L. S. (2004). Development of sexual behavior over several serving Reusch, T. B., Häberli, M. A., Aeschlimann, P. B., & Milinski, M. (2001). Female capacity tests in male goats. Applied Animal Behaviour Science, 89,315e319. sticklebacks count alleles in a strategy of sexual selection explaining MHC Irvine, I. A. S., & Sorensen, P. W. (1993). Acute olfactory sensitivity of wild common polymorphism. Nature, 414,300e302. carp, Cyprinus carpio, to goldfish hormonal sex pheromones is influenced by Rodd, F. H., & Sokolowski, M. B. (1995). Complex origins of variation in the sexual gonadal maturity. Canadian Journal of Zoology, 71,2199e2210. behaviour of male Trinidadian guppies, Poecilia reticulata: interactions between Jirotkul, M. (1999a). Population density influences male-male competition in gup- social environment, heredity, body size and age. Animal Behaviour, 49,1139e pies. Animal Behaviour, 58,1169e1175. 1159. Jirotkul, M. (1999b). Operational sex ratio influences female preference and male- Rodríguez, R. L., Rebar, D., & Fowler-Finn, K. D. (2013). The evolution and evolu- male competition in guppies. Animal Behaviour, 58, 287e294. tionary consequences of social plasticity in mate preferences. Animal Behaviour, Jordan, L. A., & Brooks, R. C. (2012). Recent social history alters male courtship 85,1041e1047. preferences. Evolution, 66, 280e287. Sato, A., Shimoichi, A., & Karino, K. (2011). Copulation type affects parturition in the Kahn, A. T., Dolstra, T., Jennions, J. D., & Backwell, P. R. Y. (2013). Strategic male guppy. Zoological Science, 28,98e104. courtship effort varies in concert with adaptive shifts in female mating pref- Schories, S., Meyer, M. K., & Schartl, M. (2009). Description of Poecilia (Acantho- erences. Behavioral Ecology, 24, 906e913. phacelus) obscura n. sp., (Teleostei: Poeciliidae), a new guppy species from Kelley, J. L., Graves, J. A., & Magurran, A. E. (1999). Familiarity breeds contempt in western Trinidad, with remarks on P. wingei and the status of the ‘Endler’s guppies. Nature, 401,661e662. guppy’. Zootaxa, 2266,35e50. Koene, J. M., & Ter Maat, A. (2007). Coolidge effect in pond snails: male motivation Spence, R., Reichard, M., & Smith, C. (2013). Strategic sperm allocation and a in a simultaneous hermaphrodite. BMC Evolutionary Biology, 7, 212. Coolidge effect in an externally fertilizing species. Behavioral Ecology, 24,82e88. Kolluru, G. R., Grether, G. F., Dunlop, E., & South, S. H. (2009). Food availability and Taborsky, M. (1994). Sneakers, satellites, and helpers: parasitic and cooperative parasite infection influence mating tactics in guppies (Poecilia reticulata). behavior in fish reproduction. Advances in the Study of Behavior, 23,1e100. Behavioral Ecology, 20,131e137. Taborsky, M. (1998). Sperm competition in fish: ‘bourgeois’ males and parasitic Konecná, M., Smith, C., & Reichard, M. (2010). Population and individual conse- spawning. Trends in Ecology and Evolution, 13, 222e227. quences of breeding resource availability in the European bitterling (Rhodeus Viken, A., Fleming, I. A., & Rosenqvist, G. (2006). Premating avoidance of inbreeding amarus). Behavioral Ecology and Sociobiology, 64, 1069e1079. absent in female guppies (Poecilia reticulata). Ethology, 112,716e723.

IMPAKTOVANÉ PUBLIKACE

58

Článek 3

Řeţucha, R. & Reichard, M. Strategic exploitation of fluctuating asymmetry in male Endler’s guppy courtship displays is modulated by social envi- ronment. Journal of Evolutionary Biology, submitováno.

V této práci jsem navrhl její design včetně řešení chovných nádrţí. Vy- choval jsem pokusné jedince a staral se o ně v průběhu celého experimentu. Provedl jsem také všechny experimenty a vymyslel vlastní způsob jejich rea- lizace. Sesbíral jsem všechna data, vyhodnotil je a provedl jejich primární statistickou analýzu. Napsal jsem první verzi článku a později přispíval k je- ho finální verzi. Dále jsem pokusné jedince několikrát fotografoval a vymyslel vlastní způsob analýzy získaných fotografií. Ty jsem pak také sám vyhodno- til.

1 Title

2 Strategic exploitation of fluctuating asymmetry in male Endler’s guppy courtship displays is

3 modulated by social environment

4

5 Running title

6 Strategic exploitation of lateral asymmetry

7

8 Authors

9 Radomil Řežucha and Martin Reichard

10

11 Affiliation

12 R. Řežucha and M. Reichard

13 Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Květná 8, 603 65

14 Brno, Czech Republic

15 Address correspondence to: M. Reichard, e-mail: [email protected], telephone: +420543422522

16

17 R. Řežucha

18 Department of Botany and Zoology, Masaryk University, Kotlářská 2, 611 37 Brno, Czech Republic

19

1 20 Abstract

21 Lateral asymmetry in signalling traits enables males to strategically exploit their best side. In many

22 animals, both body colouration and fluctuating asymmetry are signals of male attractiveness. We

23 experimentally demonstrated that even sexually naïve males Poecilia wingei were able to identify

24 their most attractive side (i.e. that with a higher proportion of carotenoid pigmentation) and use it

25 preferentially during courtship. Notably, males retained their strategic signalling in a male-biased

26 social environment while they ceased to signal strategically in a female-biased environment. The

27 degree of asymmetry in colouration did not affect overall courtship activity. Strategic lateralization

28 in courtship displays was strongest and most repeatable in the male-biased social environment

29 where males competed with rivals for matings. Individual asymmetry in coloration changed

30 considerably over a period of 3 months. This suggests that colouration is a dynamic feature during

31 adulthood and that males are capable of tracking and strategically exploiting their lateral asymmetry

32 in accordance with their social environment.

33

34 Keywords

35 Poecilia wingei, sigmoid display, sidedness, social environment, experience, behavioural laterality

2 36 Introduction

37

38 Females use male signalling to make mating decisions. In response, males adopt strategies to

39 increase the efficiency of their signalling (Rodriguez et al., 2012; Kahn et al., 2013). This is

40 particularly important in mating systems where the male contribution to reproduction is limited to

41 sperm transfer. Despite growing evidence that females can attend to complex traits such as

42 cognitive skills (Boogert et al., 2011) or particular personality traits (Schuett et al., 2010),

43 ornamental traits are still considered to be the major targets of female choice across taxa

44 (Amundsen, 2003; Kuijper et al., 2012). There is also abundant evidence that females base their

45 choice on relative variation among males (Hughes et al., 2013), highlighting the role of social

46 environment. A potential link between abstract, cognitive traits and male ornamentation is strategic

47 modification of male signalling in response to feedback from females. This allows males to

48 effectively emphasize their signalling effort (Patricelli et al., 2002; Sullivan-Beckers & Hebets,

49 2014) or mask their particular deficits (Gross et al., 2007).

50

51 Fluctuating asymmetry (FA) is a deviation from bilateral symmetry, most likely resulting from

52 developmental errors or instability during ontogeny (Palmer, 1996). FA has been demonstrated to

53 play a significant role in mate choice across taxa (humans: Koehler et al., 2002; birds: Møller, 1992;

54 invertebrates: Harvey & Walsh, 1993) including fish (Sheridan & Pomiankowski, 1997; Morris &

55 Casey, 1998; Schlüter et al., 1998, but see Brooks & Caithness, 1995; Gross et al., 2007). The FA

56 can signal lower male quality in general (Møller, 1993; Watson & Thornhill, 1994), but also enables

57 asymmetric males to exploit their superior side by its preferential lateral display (Gross et al., 2007;

58 Amcoff et al., 2009).

59

60 Most teleost fishes have laterally compressed bodies and can therefore benefit from strategic

61 behavioural laterality during courtship. Even sexually naïve males of the common guppy Poecilia

3 62 reticulata Peters adjusted their courtship to exploit their more colourful side (Gross et al., 2007) and

63 male swordtail characins Corynopoma riisei Gill responded to experimental manipulation of their

64 enlarged pectoral fin by strategically biasing their lateral displays (Amcoff et al., 2009). It remains

65 unclear, however, how such strategic signalling is affected by individual experience and how such

66 experience is modified in various social environments. Laterality is widespread in fish (Bisazza et

67 al., 1998), but our understanding of whether it may be adaptive by being utilized strategically in

68 response to FA is surprisingly weak.

69

70 Social environment has a major effect on individual behaviour, including mating behaviour (Pruett-

71 Jones, 1992; Plath & Bierbach, 2011; Barbosa et al., 2013). Indeed, male courtship and signalling is

72 strongly influenced by experience and current social environment (Miller & Brooks, 2005;

73 Guevara-Fiore et al., 2012). It is therefore critical to relate the behaviour of an individual to its

74 current and past social experience. Surprisingly, the question of how individual experience and

75 social environment influence lateralized behaviour remains unexplored.

76

77 Carotenoid pigments are reliable and taxonomically widespread indicators of individual condition

78 and social status (Sefc et al., 2014). Carotenoids cannot be synthetized by animals de novo and must

79 be obtained in the diet (Latscha, 1990), making them potential indicators of male foraging ability

80 (Kodric-Brown, 1989). Females can evaluate several critical features of male phenotypic and

81 genetic quality, including his condition (Olson & Owens, 1998), genome-wide heterozygosity

82 (Herdegen et al., 2014), and fertility (Locatello et al., 2006; Pike et al., 2010; Smith et al., 2014)

83 from the amount of carotenoid pigment displayed on his body. Consequently, males are expected to

84 preferentially display their most carotenoid-coloured side during courtship. At the same time,

85 assumptions on the role of carotenoids as sexual signals cannot be uncritically generalized across

86 populations of particular species. Targets of female choice are often variable across populations

87 (Endler & Houde, 1995) due to differences in natural selection (Endler, 1980), signalling

4 88 environment (Seehausen & van Alphen, 1998; Summers et al., 1999; Fuller & Noa, 2010; Maan et

89 al., 2010) or reinforcement (Kirkpatrick & Servedio, 1999; Williams & Mendelson, 2013), resulting

90 in high inter-population variability in preference for visual signals.

91

92 We used Endler's guppy (Poecilia wingei Poeser, Kempkes & Isbrücker) to investigate strategic

93 male signalling in response to FA in their colouration. We were particularly interested in the role of

94 orange carotenoid pigmentation because it has been suggested that female P. wingei prefer males

95 with more orange colour (Lindholm & Breden, 2002; Alexander & Breden, 2004, P. wingei reported

96 as Cumaná guppy). Male P. wingei are generally more colourful than male P. reticulata and

97 additionally possess orange-coloured vestiges of swords in the caudal fin, providing additional

98 circumstantial evidence of the importance of carotenoid colouration in their reproductive success.

99 Indeed, it has been suggested that the female preference for male colouration was the basis of the

100 divergence between P. wingei and P. reticulata (Alexander & Breden, 2004).

101

102 Endler’s guppy, also reported as Cumaná guppy or Campoma guppy (Evans et al., 2011), is a small

103 poecilid fish with ideal characteristics for investigating behavioural laterality and socio-

104 environmental influences on male sexual signalling and courtship behaviour. It is closely related to

105 the common guppy (P. reticulata) (Poeser et al., 2005) and their behaviour is similar (Houde, 1997;

106 Poeser et al., 2005; personal observation), making inferences on P. wingei from studies on P.

107 reticulata possible. While males are generally brightly coloured, their colour may not be distributed

108 evenly on their body sides. The mating system is polygynandrous and non-resource based. After

109 copulation, males provide no parental investment. Alternative mating tactics are present. Male

110 courtship involves sigmoid displays which provide the opportunity to preferentially show their

111 more colourful side. Alternatively, males can mate coercively through gonopodial thrusts (Poeser et

112 al., 2005). Coercive mating has only limited potential to undermine female choice because consent

113 mating following successful courtship enables the transfer of significantly more sperm (Pilastro &

5 114 Bisazza, 1999) which can be stored in the female gonoducts for several months, and used to fertilize

115 more eggs (Houde, 1997; López-Sepulcre et al., 2013). The sexual behaviour of P. wingei is

116 considerably influenced by individual experience and social environment (Řežucha & Reichard,

117 2014). Males devote a significant proportion of their time to courtship and attempts to mate through

118 forced copulations.

119

120 We tested three objectives. First, we examined the association between the overall level of courtship

121 behaviour and the extent of fluctuating asymmetry. Second, we tested laterality of courtship

122 behaviour in relation to fluctuating asymmetry in colouration in young, sexually inexperienced

123 (virgin) males that lacked any feedback from prior female responses to their courtship. We

124 compared this with laterality in courtship behaviour in similarly young but sexually experienced

125 males. Third, we tested laterality in courtship behaviour in experienced males after housing in

126 contrasting social environments – female-biased (FBE, focal males housed with two females) or

127 male-biased (MBE, focal male housed with a rival and a single female). Finally, we explored

128 temporal changes in FA for individual males over a three month period.

129

130 We predicted that overall levels of male courtship rate would be unaffected by fluctuating

131 asymmetry in colouration, including carotenoid colouration. We expected that the level of laterality

132 in courtship displays would be related to the degree of individual FA (positive association). We

133 predicted that experienced males would bias their courtship towards displaying their best side, as a

134 result of positive feedback from females but that virgin males would lack such bias. We expected

135 that males in MBE would preferentially display their best side more often compared to males in

136 FBE due to the presence of mating competition. Finally, we predicted that individual FA scores in

137 colouration might vary slightly during the course of colouration development but would be broadly

138 repeatable across ontogeny.

139

6 140 Methods

141

142 Fish housing

143

144 Subject animals were taken from our breeding stock composed of outbred descendants of P. wingei

145 imported from Laguna de Los Patos (northern Venezuela) in 2007. The stock population was kept in

146 a 120 L aquarium and fed twice a day with commercial flake food and frozen chironomid larvae.

147 Water temperature was maintained at 25°C ± 2°C and the aquarium was subject to a natural daylight

148 regime via a glass rooftop (10-14 hours of light a day). Water was continuously aerated and its

149 quality maintained by air-driven foam filters and regular water exchange. Live plants were provided

150 as refugia. Over a period of one month, all emerging juveniles were collected from the stock

151 aquarium and isolated in a separate 72 L aquarium. Juveniles were frequently sexed, and females

152 were removed and kept separately.

153

154 Sixty-five males (approximately 19 weeks old) were collected from a total of 146 juveniles and

155 placed individually in 2L plastic tanks. The tanks were visually separated from each other,

156 experienced a natural light regime and additional light was provided for 12 hours a day (08:00-

157 20:00) by the 40 W Sun Glo daylight spectrum fluorescent tube. Water temperature fluctuated with

158 ambient temperature between 22 and 25°C. Artificial plants were provided as refugia in each

159 aquarium. Males were fed once a day with commercial flake food and water was exchanged every

160 two weeks.

161

162 Experimental males

163

164 The first behavioural test of male mating behaviour (Trial 1) was performed after 2 months of

165 separation. We tested two groups of males. Virgin males (n=65) had no prior experience with an

7 166 adult receptive female and therefore no feedback on their courtship displays. Experienced males

167 (n=32) were randomly chosen from the social aquarium where they lived from birth. Their age was

168 approximately 6 months (matching the age of virgin males) and they fully interacted with a group of

169 females of various ages. The sex ratio in the social aquarium fluctuated naturally over the their adult

170 period but never departed considerably from parity. All experienced males were separated from

171 females for 3 days prior to testing to standardize their mating effort and to replenish sperm reserves.

172

173 Upon completion of all trials at the first time point, two females were added into 33 randomly

174 assigned tanks with focal males to create a female–biased social environment (1 male, 2 females).

175 In the remaining 32 tanks with focal males, a single male and single female from the stock

176 population were added to create a male–biased social environment (2 males, 1 female). The second

177 test of sexual behaviour of focal males (Trial 2) was completed after 5-weeks housing in the

178 particular social environment for a subset of males who survived until Trial 2 (naïve males in Trial

179 1). Males had unrestricted access to females (and rivals in the MBE) and gained mating experience

180 in their particular social environment. Upon completion of the second trial, males were placed back

181 to their home tank. The social environment treatment was maintained by replacing any dead female

182 or non-experimental male. Focal males were not replaced and focal male mortality caused a minor

183 decrease in sample size during the second trials. Experienced males from Trial 1 (i.e. from the large

184 social aquarium) were not tested in Trial 2.

185

186 Behavioural assays

187

188 Male mating behaviour was scored in a 6 L aquarium with the back and sides covered with black

189 fabric to minimize disturbance. Light was provided by a 25W Sun Glo daylight spectrum bulb

190 positioned above the aquarium, ensuring an even distribution of light. Females used in the

191 experiment were collected from a virgin female aquarium and left with a group of 3 adult non-

8 192 experimental males from the stock aquarium for one day prior to the experimental trials. This

193 standardized female sexual receptivity by ensuring that all females were non-virgin and in the same

194 reproductive state. Each female was tested with 4 experimental males and was used only once in the

195 entire experiment. During Trial 2 (social environment treatments) focal males were isolated from all

196 conspecifics for 1.5 day prior to testing to replenish their sperm reserves and standardize their

197 mating effort.

198 A female was allowed to settle for 5 min in the test aquarium. After this period, a randomly chosen

199 focal male was gently captured, added to the test aquarium and left for 10 min to settle. After

200 settling, male mating behaviour was scored for 15 minutes by a single observer using JWatcher 1.0,

201 software for behavioural scoring (Blumstein & Daniel, 2007). We scored the number and duration

202 of sigmoid displays (courtship) using left and right sides. The sigmoid display is inherently lateral;

203 the male spreads his fins and arches his body into a sigmoid shape and quickly undulates, exposing

204 the bright colouration on one side of his body. Sigmoid displays may be repeated in a succession,

205 with either the same or the opposite side of the body displayed (Houde, 1997). After 15 min, the

206 focal male was gently captured and returned to his home tank.

207

208 Photographing and analysis of colour pattern

209

210 Focal males were photographed at the age of approximately19 weeks (4 weeks prior to Trial 1) and

211 at the age of 32 weeks (1 week after Trial 2). A male was gently captured and immediately released

212 into small photo-aquarium. No anaesthetics were used in order to prevent any distortion of colour

213 pattern (Reynolds et al., 1993; Pélabon et al., 2014) and intensity (Gray et al., 2011). Sedation with

214 cold water was similarly not employed as any stress associated with photographing was considered

215 less than stress associated with cold-water sedation. Each male was fixed in a stable position in the

216 front of the aquarium using a soft sponge and quickly photographed from both sides using a Canon

217 EOS Rebel XTi camera equipped with a Sigma Macro 100 mm lens. The camera was positioned on

9 218 a stationary tripod to ensure a constant distance from the lens. A small ruler was placed next to each

219 individual to provide a scale. Standardized fish colouration and constant light conditions were

220 ensured by placing the photo-aquarium in a dark case and using flash illumination. Orange

221 carotenoid spots do not appear to vary according to the surrounding environment or the emotional

222 state of the fish (Brooks & Caithness, 1995). Black melanin spots can be modified over several

223 minutes in guppies (Baerends et al., 1955). All pictures were taken in the shortest interval possible,

224 typically less than 3-4 min. No obvious colour change or distress was observed during fish

225 handling. Two photographs of each individual were taken from each side and the photograph where

226 the fish was in a better position for each particular side was used for analysis. Where both

227 photographs were of comparable quality the one used for analysis was chosen at random. All

228 pictures were taken in compressed JPG format with 3888 × 2592 pixels resolution and 72 DPI.

229 All image analyses were performed in Fiji, an image processing package (Schindelin et al., 2012)

230 based on ImageJ 1.47n software (Abramoff et al., 2004). We measured the area of each colour

231 category of interest (see below), and the number of separate dots of each defined colour. It was not

232 possible to determine the number of iridescent patches unambiguously (as is the case with P.

233 reticulata, Ruell et al., 2013) and we did not consider the number of iridescent patches in our

234 analysis. The total area of iridescent patches was readily quantified.

235

236 Three colour categories – carotenoid (light orange to dark red), melanin (dark black and fuzzy

237 black) and iridescent (mostly green, blue, purple and silver, formed by guanine crystals in

238 iridiophores) were established. First, we precisely defined the colour scale of each category in RGB

239 colour space. The RGB colour space is defined by the three chromaticities of the red, green and

240 blue primaries (Pascale, 2003). This adjustment allowed us to precisely measure the specific colour

241 areas of each particular side using the Colour Thresholder and Select tools in the Fiji package. The

242 area of specific colour (Ac) was then measured as a proportion of total area of lateral body

243 projection:

10 244

NpxC 245  NpxC  R r1,r2  G g1, g2  B b1,b2 : Ac  100 % NpxS

246

247 where NpxC is number of all pixels of colour within defined intervals of R (red), G (green) and B

248 (blue) in RGB colour space, and NpxS is total side area selected with the Freehand selection tool

249 (under sufficient magnification to minimize measurement error). The dorsal fin, eyes and

250 gonopodium were not in a fixed position during photographing and their areas were excluded from

251 our analysis (i.e. not included in NpxS). Some males did not have evenly spread caudal fins in the

252 photographs and we used mean caudal fin area for their respective body size. We did not exclude

253 caudal fin area as it possessed a non-negligible amount of colouration. The number of clear colour

254 spots was determined visually. Small and discontinuous fuzzy areas were not counted as separate

255 dots but were included in the measure of overall area of the particular colour. Measurement of

256 colour characteristics was made blind with respect to behavioural data.

257

258 Data analysis

259

260 Relative asymmetry (in %) was calculated as absolute asymmetry divided by the larger value for

261 one side (Amcoff et al., 2009). Symmetry of directional fluctuation in male traits was tested by

262 paired t-tests.

263 The effect of colouration on directional bias in male displays was analysed using a Generalized

264 Linear Model (GLM) with a binomial distribution applied to the counts of sigmoid displays of left

265 and right side of the body as the response variable. Males that performed no sexual display within

266 the 15 min observational period were not considered in the analysis. We adopted an IT approach

267 (Burnham & Anderson, 2002). We first selected a set of biologically plausible models (Table 1). We

268 predicted a priori that carotenoid colouration would be targeted by females (Houde, 1997;

269 Alexander & Breden, 2004). However, given the large interpopulation variability in the importance

11 270 of male traits targeted by female choice in the closely related P. reticulata, we considered other

271 traits and their interactions which could potentially affect female choice and directional bias (Fig.

272 1). Collinearity between explanatory variables was evident from an initial data exploration, with

273 clear association between the number of patches of particular colour and its total area. We

274 accounted for this collinearity in the choice of models, but retained models with alternative

275 colouration measures to compare their effect on laterality of sexual displays. We compared 21

276 models of varying complexity (Table 1). Fits of alternative models were compared using the Akaike

277 Information Criterion corrected for small sample size (AICc). Model weights were calculated from

278 relative likelihoods. For the model with the best fit, residuals were examined to ensure model

279 assumptions were met. We considered all models within AIC < 3 to have substantial support.

280 However, given parsimony and interpretation, we further explored only the best-fitting models (see

281 further). Models for Trial 1 and Trial 2 had identical structure but were treated separately (Table 1).

282 Statistical analyses were performed in the R environment (R Development Core Team, 2009). In

283 Trial 1, there were 49 naïve males and 28 experienced males performing sigmoid displays. In Trial

284 2, 17 males in the MBE and 15 males in the FBE performed sigmoid displays during behavioural

285 observation. The overall mean number of sigmoid displays (including males not displaying to

286 females) was 11 (s.d. = 8.8, median = 9) per 15 min observation period. Repeatability was tested as

287 intra-class correlation sensu Lessels & Boag (1987) and visualized as a bivariate plot between

288 individually-based values at the age of 19 and 32 weeks.

289

290 Results

291

292 Description of fluctuating asymmetry

293

294 All colouration variables and sigmoid displays demonstrated fluctuating asymmetry. The relative

295 asymmetry in colouration ranged between 11 and 28 % of the mean value and asymmetry on

12 296 sigmoid displays was even higher (42 %) (Fig. 1). No directional bias was observed for any variable

297 (paired t-test, all P > 0.05), except for Iridescence (t 76 = 3.53, P = 0.001 and t 31 = 3.45, P = 0.002

298 for young and older males, respectively; left side was more iridescent). No directional bias was

299 found in sigmoid displays (paired t-test, both P > 0.50).

300

301 The effects of FA and carotenoid colouration on overall courtship activity

302

303 The intensity of sigmoid displays was not related to the level of overall FA or carotenoid-based FA

304 (Pearson correlation for each male treatment separately, P > 0.15 for all 8 tests) or to the overall

305 score for the number of carotenoid patches or total carotenoid area (Pearson correlation for each

306 male treatment separately, P > 0.14 for 7 tests). The exception was a positive correlation between

307 intensity of sigmoid displays and total carotenoid area in males from the FBE (r 13 = 0.54, P =

308 0.040), though this association was highly non-significant after Bonferroni correction for multiple

309 testing.

310

311 Strategic signalling in response to FA and its modulation by social environment

312

313 Males responded to their FA by displaying their more carotenoid-coloured side preferentially and

314 social environment modulated this response. At both time points, the models with the strongest

315 support contained asymmetry in the percentage of carotenoid colouration, treatment (social

316 environment) and their interaction (Table 1). The response was strongest in naïve males (Trial 1:

317 Fig. 2a) and males in the MBE (Trial 2: Fig. 2c). Experienced males had generally negligible FA

318 (Fig. 2b) that may have resulted in the lack of association with the lateralized displays.

319

320 Other models with substantial empirical support (AIC < 3, relative weight > 0.10) also contained

321 complex associations between Percentage of Carotenoid pigments and Iridescence (Table 1) and

13 322 their interpretation was complex. The area of carotenoid colouration explained laterality in male

323 displays better than the number of carotenoid patches (AICc = 3.87 and 10.49 for Trial 1 in

324 models with and without social environment interaction, respectively, and AICc = 3.40 and 8.87

325 for Trial 2 models).

326

327 Changes in strategic signalling with male age

328

329 There was no temporal change in the magnitude of FA in colouration traits (Wilcoxon paired test

330 calculated on absolute values of differences, N = 33, P > 0.50 for 6 tests and z = 1.71, P = 0.088 for

331 carotenoid patches). The FA in sigmoid displays marginally decreased between Trial 1 and Trial 2

332 (Wilcoxon paired test, z = 2.19, N = 32, P = 0.028).

333

334 Repeatability in sigmoid displays was high (r2 = 0.79) in males that were subjected to the MBE

335 prior to Trial 2 testing, but very modest (r2 = 0.23) in the FBE males (Fig. 3). In colouration traits,

336 repeatability was modest for the number of colour patches but low for the areas of carotenoid,

337 melanin and iridescent colouration and congruent between the FBE and MBE treatments (Fig. 3).

338 The FA in all traits sometimes changed from left bias to right bias and vice versa between the two

339 measurements (i.e. over 13 weeks) (Fig. 3). As for repeatability, the direction of FA was relatively

340 more stable in discrete colouration patches (Fig. 3, top panels) and sigmoid displays (Fig. 3, lower

341 right panel) than in the area of colouration (Fig. 3, mid panels, lower left).

342

343 Discussion

344

345 We experimentally demonstrated the ability of P. wingei males to identify their more colourful side

346 and to use it preferentially during courtship. Carotenoid-based orange colouration was identified as

347 the most important source of side-bias in courtship. These results are in general concordance with a

14 348 previous report for the closely related P. reticulata (Gross et al., 2007) who also found that even

349 sexually naïve males are able to identify their more colourful side. Importantly, we demonstrated

350 that social environment substantially modulates the magnitude of laterality in sexual display. Males

351 from the FBE, who had not competed with any rival for access to mating for a period of 5 weeks,

352 ceased to preferentially display their more colourful side. Young males from mixed social aquarium

353 also did not display their best side - however, the FA in the percentage of carotenoids pigments was

354 negligible in this group of males.

355

356 Another possible explanation for the lack of relationship in experienced males from Trial 1 and

357 males in the FBE in Trial 2 is that such deceptive signalling may not be effective. Males from these

358 two social environments had more interactions with multiple females and may have learned that

359 females have capacity to recognize deceptive signalling. Deceptive signalling is present across taxa,

360 including fish (Plath et al., 2008). It is possible that P. wingei females are able to detect male

361 cheating, as demonstrated in female Lake Malawi cichlids Nyassachromis microcephalus Trewavas

362 (Martin, 2010). However, our study was not designed to test female preference or reproductive

363 success in response to male strategic use of FA and associated courtship behaviour.

364

365 Experienced males in Trial 1 had only negligible colour pattern FA but some still demonstrated a

366 preferential use of one side over the other during sigmoid displays (e.g. left to right ratios of 15:5,

367 8:15, 8:2). This trend could indicate an underlying cerebral lateralizaton of sigmoid displays in P.

368 wingei males. Cerebral lateralization can work in concert, at least to some extent, with lateralization

369 of colouration. This would suggest that in a population there is always a set of behavioural right–

370 siders and left–siders, even if colouration does not significantly differ between the two sides.

371 Cerebral laterality has been shown to influence other aspects of fish behaviour such as turning left

372 or right to detour a barrier in P. reticulata (Bisazza et al., 1997; Irving & Brown, 2013) and the

373 preferential use of a particular side during aggressive displays in Siamese fighting fish (Betta

15 374 splendens Regan) (Takeuchi et al., 2010). It can potentially explain an inherent lateral bias in P.

375 wingei sigmoid displays. At the same time, the number of sigmoid displays was never exclusive to

376 one side in individuals where the total number of displays was greater than 5 (n = 70 trials). The

377 distribution of laterality in displays followed a normal distribution and was therefore congruent with

378 the predictions of FA (Palmer, 1996). Further, we did not record any general directional bias in male

379 sigmoid displays, suggesting there is no species-specific or population-specific pattern in laterality

380 of the behaviour as found in convict cichlid (Amatitlania nigrofasciata Günther) males which

381 showed a right bias in lateral contest displays (Arnott et al., 2011).

382

383 More colourful males did not court at a higher rate than less colourful males, as suggested for P.

384 reticulata (Kiritome et al., 2012), regardless of social environment, age or experience. Neither did

385 we find any association between the level of FA and overall level of courtship. In P. reticulata

386 females do not simply prefer colourful males but they prefer colourful males with the highest

387 courtship intensity (Kodric-Brown, 1993). Colourful males can therefore increase their

388 attractiveness further by increasing the number of sigmoid displays, but this pattern was not

389 confirmed in our study. Courtship is energetically costly (Kotiaho, 2000) and should be used

390 prudently. Any intensification of male courtship may be more evident in longer trials and associated

391 with an initial female interest that we did not measure.

392

393 Importantly, social environment modulated adaptive laterality in sexual displays. Males from the

394 MBE, who experienced constant mating competition, continued to display their best side while

395 males in the female-biased environment did not. This was in agreement with our prediction that

396 rival presence should increase the importance of male courtship quality relative to a non-

397 competitive environment. This was coupled with an overall decline in mating effort in the FBE

398 males, including attempts at forced copulation (Řežucha & Reichard, 2014).

399

16 400 Strategic male signalling can only be possible if a male is able to recognize that the bias in his

401 lateral displays confers higher success. This requires self-awareness in order to identify the best side

402 (Amcoff et al., 2009). A simple mirror test elicited a different response in brains of fish when

403 observing themselves compared to a conspecific (Desjardins & Fernald, 2010). It is therefore

404 possible that P. wingei are, at least to some extent, able to compare themselves with a rival. Thus the

405 strategic lateralization of sigmoid displays in the MBE may be maintained via a process of mutual

406 reinforcement between rivals. In the FBE, intrasexual competition among males was absent

407 (Clutton-Brock & Parker, 1992; Kvarnemo & Ahnesjö, 1996) so males may have ceased to rely on

408 their lateral variation in attractiveness. The strength of sexual selection therefore modulated male

409 signalling effort both in term of courtship rate in general (Řežucha & Reichard, 2014) as well as in

410 the strategic use of FA.

411

412 Iridescent colouration significantly affected lateral bias in sigmoid displays in addition to

413 carotenoid-based colouration. The effect of iridescence was strongest in interaction with carotenoid

414 colouration (Table 1). This suggests that carotenoid pigments are either not the exclusive target of

415 female choice in P. wingei or their clarity may be modified via contrasts with other pigments.

416 Female P. reticulata are sensitive to much subtler variations in signals of male quality (Sathyan &

417 Couldridge, 2012) than generally suggested and the importance of iridescent pigments in female P.

418 reticulata mate choice has been previously highlighted (Endler, 1983).

419

420 Temporal changes in FA have only rarely been investigated (but see e.g. Hallgrímsson, 1999;

421 Kellner & Alford, 2003; Bartoš et al., 2007). We found that the overall magnitude of the FA in

422 colouration remained unchanged between the age of 19 and 32 weeks, but at individual level,

423 measures of FA were relatively unrepeatable and individuals exhibited increases, decreases and

424 even changes in the direction of FA between the two measurements (Fig. 3). This outcome was not

425 predicted and suggests high plasticity in individual FA throughout development rather than either

17 426 simple maintenance of the degree of FA throughout life, or a gradual increase from its initial level

427 (Bartoš et al., 2007) or gradual corrections (Kellner & Alford, 2003). Our data are consistent with

428 major ontogenetic changes in colouration among male P. reticulata (measured only on a single side

429 of the body) (Miller & Brooks, 2005) and suggest that colouration on the left and right side of the

430 body can fluctuate at least partly independently.

431

432 In conclusion, we demonstrated that social environment strongly modulated strategic use of FA in

433 male P. wingei courtship displays. The asymmetry in colouration and behaviour was congruent with

434 the predictions of fluctuating asymmetry (normal distribution) rather than lateralization (bimodal

435 distribution). The degree of FA did not affect overall courtship activity and strategic lateralization in

436 courtship displays was strongest and most repeatable in the social treatment where males competed

437 with rivals. Individual variation in FA between two measurements was high and suggestive of

438 continual changes in colouration, at least within the first 8 months of life, which constitutes much of

439 the typical life expectancy of wild male guppies (Bryant & Reznick, 2004; López-Sepulcre,

440 2013; Arendt et al., 2014). We suggest future work should explore the causal role of social

441 environment on the temporal changes in male signalling traits such as colouration and their

442 asymmetry.

443

444 Acknowledgments

445 We are grateful to Radim Blažek, Matej Polačik, Milan Vrtílek and Rowena Spence for their help

446 with experiments, helpful advices and comments, and to Jiří Farkač for his help with fish

447 maintenance. Financial support came from Czech Science Foundation GACR P505/12/G112.

448

449

18 450 References 451 452 Abramoff, M.D., Magelhaes P.J. &, Ram, S.J. (2004). Image processing with ImageJ. Biophotonics 453 Int. 11: 36–42. 454 Alexander, H.J. & Breden, F. 2004. Sexual isolation and extreme morphological divergence in the 455 Cumaná guppy: a possible case of incipient speciation. J. Evol. Biol. 17: 1238–1254. 456 Amcoff, M., Arnqvist, G. & Kolm, N. 2009. Courtship signalling with a labile bilateral signal: 457 males show their best side. Behav. Ecol. Sociobiol. 63: 1717–1725. 458 Amundsen, T. 2003. Fishes as models in studies of sexual selection and parental care. J. Fish. Biol. 459 63: 17–52. 460 Arendt, J.D., Reznick, D.N. & López-Sepulcre, A. 2014. Replicated origin of female-biased adult 461 sex ratio in introduced populations of the Trinidadian guppy (Poecilia reticulata). Evolution 462 68: 2343–2356. 463 Arnott, G., Ashton. C. & Elwood, R.W. 2011. Lateralization of lateral displays in convict cichlids. 464 Biol. Lett. 7: 683–685. 465 Baerends, G.P., Brouwer, R. & Waterbolk, H.T. 1955. Ethological studies on Lebistes reticulatus 466 (Peters) 1. An analysis of the male courtship pattern. Behaviour 8: 249–334. 467 Barbosa, M., Ojanguren, A.F. & Magurran, A.E. 2013. Courtship display persists despite early 468 social deprivation. Ethology 119: 496–502. 469 Bartoš, L., Bahbouh, R. & Vach, M. 2007. Repeatability of size and fluctuating asymmetry of antler 470 characteristics in red deer (Cervus elaphus) during ontogeny. Biol. J. Linn. Soc. 91: 215–226. 471 Bisazza, A., Pignatti, R. & Vallortigara, G. 1997. Laterality in detour behaviour: interspecific 472 variation in poeciliid fish. Anim. Behav. 54: 1273–1281. 473 Bisazza, A., Rogers, L.J. & Vallortigara, G. 1998. The origins of cerebral asymmetry: a review of 474 evidence of behavioural and brain lateralization in fishes, reptiles and amphibians. Neurosci. 475 Biobehav. Rev. 22: 411–426. 476 Blumstein, D.T., Daniel, J.C. & Evans, C.S. 2007. Quantifying behavior: the JWatcher way. Sinauer 477 Associates, Inc., Sunderland. 478 Boogert, N.J., Fawcett, T.W. & Lefebvre, L. 2011. Mate choice for cognitive traits: a review of the 479 evidence in nonhuman vertebrates. Behav. Ecol. 22: 447–459. 480 Brooks, R. & Caithness, N. 1995. Female choice in a feral guppy population: are there multiple 481 cues? Anim. Behav. 50: 301–307. 482 Bryant, M.J. & Reznick, D. 2004. Comparative studies of senescence in natural populations of 483 guppies. Am. Nat. 163: 55-68.

484 Burnham, K.P. & Anderson, D.R. 2002. Model selection and multimodel inference: a practical 19 485 information-theoretic approach. Springer, New York. 486 Clutton-Brock, T.H. & Parker, G.A. 1992. Potential reproductive rates and the operation of sexual 487 selection. Q. Rev. Biol. 67: 437–456. 488 Desjardins, J.K. & Fernald, R.D. 2010. What do fish make of mirror images? Biol. Lett. 6: 744–747. 489 Endler, J.A. 1980. Natural selection on color patterns in Poecilia reticulata. Evolution 34: 76–91. 490 Endler, J.A. 1983. Natural and sexual selection on color patterns in poeciliid fishes. Environ. Biol. 491 Fishes 9: 173–190. 492 Endler, J.A. & Houde, A.E. 1995. Geographic variation in female preferences for male traits in 493 Poecilia reticulata. Evolution 49: 456–468. 494 Evans, J.P., Pilastro, A. & Schlupp, I. 2011. Ecology and evolution of Poeciliid fishes. The 495 University of Chicago Press, Chicago and London. 496 Fuller, R.C. & Noa, L.A. 2010. Female mating preferences, lighting environment, and a test of the 497 sensory bias hypothesis in the bluefin killifish. Anim. Behav. 80: 23–35. 498 Gray, S.M., Hart, F.L., Tremblay, M.E.M., Lisney, T.J & Hawryshyn, C.W. 2011. The effects of 499 handling time, ambient light, and anaesthetic method, on the standardized measurement of fish 500 colouration. Can. J. Fish. Aquat. Sci. 68: 330–342.

501 Gross, M.R., Suk, H.Y. & Robertson, C.T. 2007. Courtship and genetic quality: asymmetric males 502 show their best side. Proc. R. Soc. B 274: 2115–2122. 503 Guevara-Fiore, P., Svensson, P.A. & Endler, J.A. 2012. Sex as moderator of early life experience: 504 interaction between rearing environment and sexual experience in male guppies. Anim. Behav. 505 84: 1023–1029. 506 Hallgrímsson, B. 1999. Ontogenetic patterning of skeletal fluctuating asymmetry in rhesus 507 macaques and humans: evolutionary and developmental implications. Int. J. Primatol. 20: 508 121–151. 509 Harvey, I.F. & Walsh, K.J. 1993. Fluctuating asymmetry and lifetime mating success are correlated 510 in males of the damselfly Coenagrion puella (Odonata: Coenagrionidae). Ecol. Entomol. 18: 511 198–202. 512 Herdegen, M., Dudka, K. & Radwan, J. 2014. Heterozygosity and orange coloration are associated 513 in the guppy (Poecilia reticulata). J. Evol. Biol. 27: 220–225. 514 Houde, A.E. 1997. Sex, color, and mate choice in guppies. Princeton University Press, New Jersey. 515 Hughes, K.A., Houde, A.E., Price, A.C. & Rodd, F.H. 2013. Mating advantage for rare males in 516 wild guppy populations. Nature 503: 108–110. 517 Irving, E. & Brown, C. 2013. Examining the link between personality and laterality in a feral guppy 518 Poecilia reticulata population. J. Fish. Biol. 83: 311–325. 519 Kahn, A.T., Dolstra, T., Jennions, M.D. & Backwell, P.R.Y. 2013. Strategic male courtship effort

20 520 varies in concert with adaptive shifts in female mating preferences. Behav. Ecol. 24: 906–913. 521 Kiritome, A., Sato, A. & Karino, K. 2012. Influences of orange spot patterns and operational sex 522 ratio on male mating behaviors in the guppy Poecilia reticulata. Ichthyol. Res. 59: 304–313. 523 Kirkpatrick, M. & Servedio, M.R. 1999. The reinforcement of mating preferences on an island. 524 Genetics 151: 865–884. 525 Kodric-Brown, A. 1989. Dietary carotenoids and male mating success in the guppy: an 526 environmental component to female choice. Behav. Ecol. Sociobiol. 25: 393–401. 527 Kodric-Brown, A. 1993. Female choice of multiple male criteria in guppies: interacting effects of 528 dominance, coloration and courtship. Behav. Ecol. Sociobiol. 32: 415–420. 529 Koehler, N., Rhodes, G. & Simmons, L.W. 2002. Are human female preferences for symmetrical 530 male faces enhanced when conception is likely? Anim. Behav. 64: 233–238. 531 Kotiaho, J.S. 2000. Testing the assumptions of conditional handicap theory: costs and condition 532 dependence of a sexually selected trait. Behav. Ecol. Sociobiol. 48: 188–194. 533 Kuijper, B., Pen, I. & Weissing, F.J. 2012. A guide to sexual selection theory. Annu. Rev. Ecol. Evol. 534 Syst. 43: 287–311. 535 Kvarnemo, C. & Ahnesjö, I. 1996. The dynamics of operational sex ratios and competition for 536 mates. Trends. Ecol. Evol. 11: 404–408. 537 Latscha, T. 1990. Carotenoids - their nature and significance in animal feeds / Thierry Latscha. F. 538 Hoffmann-La Roche, Animal Nutrition and Health, Basel.

539 Lessells, C.M. & Boag, P.T. 1987. Unrepeatable repeatabilities: a common mistake. Auk 104: 116– 540 121.

541 Lindholm, A. & Breden, F. 2002. Sex chromosomes and sexual selection in poeciliid fishes. Am. 542 Nat. 160: S214–S224. 543 Locatello, L., Rasotto, M.B., Evans, J.P. & Pilastro, A. 2006. Colourful male guppies produce faster 544 and more viable sperm. J. Evol. Biol. 19: 1595–1602. 545 López-Sepulcre, A., Gordon, S.P., Paterson, I.G., Bentzen, P. & Reznick, D.N. 2013. Beyond 546 lifetime reproductive success: the posthumous reproductive dynamics of male Trinidadian 547 guppies. Proc. R. Soc. B 280: 20131116.

548 Maan, M.E., Seehausen, O. & Van Alphen, J.J.M. 2010. Female mating preferences and male 549 coloration covary with water transparency in a Lake Victoria cichlid fish. Biol. J. Linn. Soc. 550 99: 398–406. 551 Martin, C.H. 2010. Unexploited females and unreliable signals of male quality in a Malawi cichlid 552 bower polymorphism. Behav. Ecol. 21: 1195–1202. 553 Miller, L.K. & Brooks, R. 2005. The effects of genotype, age, and social environment on male

21 554 ornamentation, mating behavior, and attractiveness. Evolution 59: 2414–2425. 555 Møller, A.P. 1992. Female swallow preference for symmetrical male sexual ornaments. Nature 357: 556 238–240. 557 Møller, A.P. 1993. Developmental stability, sexual selection and speciation. J. Evol. Biol. 6: 493– 558 509. 559 Morris, M.R. & Casey, K. 1998. Female swordtail fish prefer symmetrical sexual signal. Anim. 560 Behav. 55: 33–39. 561 Olson, V.A. & Owens, I.P.F. 1998. Costly sexual signals: are carotenoids rare, risky or required? 562 Trends. Ecol. Evol. 13: 510–514. 563 Palmer, A.R. 1996. Waltzing with asymmetry. Bioscience 46: 518–532. 564 Pascale, D. 2003. A review of RGB color spaces... from xyY to R’G’B’. The BabelColor Company, 565 Montreal. 566 Patricelli, G.L., Uy, J.A.C., Walsh, G. & Borgia, G. 2002. Male displays adjusted to female’s 567 response. Nature 415: 279–280. 568 Pélabon, C., Larsen, L.-K., Bolstad, G.H., Viken, Å., Fleming, I.A. & Rosenqvist, G. 2014. The 569 effects of sexual selection on life-history traits: an experimental study on guppies. J. Evol. 570 Biol. 27: 404–416.

571 Pike, T.W., Blount, J.D., Lindström, J. & Metcalfe, N.B. 2010. Dietary carotenoid availability, 572 sexual signalling and functional fertility in sticklebacks. Biol. Lett. 6: 191–193. 573 Pilastro, A. & Bisazza, A. 1999. Insemination efficiency of two alternative male mating tactics in 574 the guppy Poecilia reticulata. Proc. R. Soc. B 266: 1887–1891. 575 Plath, M. & Bierbach, D. 2011. Sex and the public: social eavesdropping, sperm competition risk 576 and male mate choice. Commun. Integr. Biol. 4: 276–280. 577 Plath, M., Richter, S., Tiedemann, R. & Schlupp, I. 2008. Male fish deceive competitors about 578 mating preferences. Curr. Biol. 18: 1138–1141. 579 Poeser, F.N., Kempkes, M. & Isbrücker, I.J.H. 2005. Description of Poecilia (Acanthophacelus) 580 wingei n. sp. from the Paría Peninsula, Venezuela, including notes on Acanthophacelus 581 Eigenmann, 1907 and other subgenera of Poecilia Bloch and Schneider, 1801 (Teleostei, 582 Cyprinodontiformes, Poeciliidae). Contrib. to Zool. 74: 97–115. 583 Pruett-Jones, S. 1992. Independent versus non-independent mate-choice: do females copy each 584 other? Am. Nat. 140: 1000–1009. 585 R Development Core Team 2009. R: a language and environment for statistical computing. R 586 Foundation for Statistical Computing, Vienna. 587 Reynolds, J.D., Gross, M.R. & Coombs, M.J. 1993. Environmental conditions and male 588 morphology determine alternative mating behaviour in Trinidadian guppies. Anim. Behav. 45:

22 589 145–152. 590 Rodriguez, R.L., Haen, C., Cocroft, R.B. & Fowler-Finn, K.D. 2012. Males adjust signaling effort 591 based on female mate-preference cues. Behav. Ecol. 23: 1218–1225. 592 Ruell, E.W., Handelsman, C.A., Hawkins, C.L., Sofaer, H.R., Ghalambor, C.K. & Angeloni, L. 593 2013. Fear, food and sexual ornamentation: plasticity of colour development in Trinidadian 594 guppies. Proc. R. Soc. B 280: 20122019.

595 Řežucha, R. & Reichard, M. 2014. The effect of social environment on alternative mating tactics in 596 male Endler’s guppy, Poecilia wingei. Anim. Behav. 88: 195–202. 597 Sathyan, R. & Couldridge, V.C.K. 2012. Female preference for blue in Japan blue guppies (Poecilia 598 reticulata). Environ. Biol. Fishes. 96: 953–959. 599 Seehausen, O. & van Alphen, J.J.M. 1998. The effect of male coloration on female mate choice in 600 closely related Lake Victoria cichlids (Haplochromis nyererei complex). Behav. Ecol. 601 Sociobiol. 42: 1–8. 602 Sefc, K.M., Brown, A.C. & Clotfelter, E.D. 2014. Carotenoid-based coloration in cichlid fishes. 603 Comp. Biochem. Physiol. Part A 173: 42–51. 604 Sheridan, L. & Pomiankowski, A. 1997. Female choice for spot asymmetry in the Trinidadian 605 guppy. Anim Behav 54: 1523–1530. 606 Schindelin, J. et al. 2012. Fiji: an open-source platform for biological-image analysis. Nat. Methods 607 9: 676–682. 608 Schlüter, A., Parzefall, J. & Schlupp, I. 1998. Female preference for symmetrical vertical bars in 609 male sailfin mollies. Anim. Behav. 56: 147–153. 610 Schuett, W., Tregenza, T. & Dall, S.R.X. 2010. Sexual selection and animal personality. Biol. Rev. 611 85: 217–246. 612 Smith, C., Phillips, A., Polačik, M. & Reichard, M. 2014. Male coloration signals direct benefits in 613 the European bitterling (Rhodeus amarus). Environ. Biol. Fishes. 97: 335–341. 614 Sullivan-Beckers, L. & Hebets, E.A. 2014. Tactical adjustment of signalling leads to increased 615 mating success and survival. Anim. Behav. 93: 111–117. 616 Summers, K., Symula, R., Clough, M. & Cronin, T. 1999. Visual mate choice in poison frogs. Proc. 617 R. Soc. B 266: 2141–2145. 618 Takeuchi, Y., Hori, M., Myint, O. & Kohda, M. 2010. Lateral bias of agonistic responses to mirror 619 images and morphological asymmetry in the Siamese fighting fish (Betta splendens). Behav. 620 Brain. Res. 208: 106–11. 621 Watson, P.J. & Thornhill, R. 1994. Fluctuating asymmetry and sexual selection. Trends. Ecol. Evol. 622 9: 21–25. 623 Williams, T.H. & Mendelson, T.C. 2013. Male and female responses to species-specific coloration

23 624 in darters (Percidae: Etheostoma). Anim. Behav. 85: 1251–1259. 625 626

627 Figure captions

628

629 Fig. 1 Distribution of fluctuating asymmetry in five colouration traits and in sigmoid displays. An

630 arrow indicates the perfect symmetry. Raw difference in values measured for left and right side of

631 the body (and numerical difference between left and right displays, respectively) are displayed. The

632 values of relative asymmetry (in %; calculated as absolute asymmetry divided by the larger value

633 for one side) and its standard error are also included to provide a standardized measure.

634

635 Fig. 2 Relationship between the laterality in sigmoid displays and laterality in carotenoid-based

636 colouration in four treatment groups. Significant correlations are emphasised by the trend line with

637 correlation coefficient and its statistical significance.

638

639 Fig. 3 Ontogenetic changes in fluctuating asymmetry for five colouration traits and for sigmoid

640 displays visualized as bivariate association between raw individually-based values from the first

641 (age of 19 weeks) and second (age of 32 weeks) measurement. Symbols for male-biased

642 environment (closed) and female-biased environment (open) are distinguished.

24 643 Table 1 Set of candidate models for strategic signalling in response to the FA and its modulation by social environment and their relative evaluation.

Time 1: naïve / experienced Time 2: MBE / FBE Model ID Model Description d.f. ΔAIC w d.f. ΔAIC w M1 Intercept only (Null) 1 7.06 0.01 1 4.37 0.05 M2 % Carotenoids (%Carot) 2 2.68 0.08 2 2.79 0.11 M3 % Melanin (%Melan) 2 9.04 0.00 2 6.61 0.02 M4 % Iridiscence (%Irid) 2 3.21 0.06 2 4.83 0.04 M5 Number of carotenoid patches (NCarot) 2 6.55 0.01 2 6.19 0.02 M6 Number of melanin patches (NMelan) 2 9.06 0.00 2 6.47 0.02 M7 %Carot + %Melan + %Carot x %Melan 4 4.25 0.03 4 6.96 0.01 M8 %Irid + %Melan + %Irid x %Melan 4 6.74 0.01 4 5.95 0.02 M9 %Carot + %Irid + %Carot x %Irid 4 2.15 0.10 4 1.63 0.19 %Carot + %Melan + %Irid + %Carot x %Melan + %Carot x M10 7 6.41 0.01 7 10.06 0.00 %Irid + %Carot x %Melan M11 NCarot + NMelan + NCarot x NMelan 4 9.96 0.00 4 6.56 0.02 T1 Treatment (Trt) only 2 8.55 0.00 2 4.88 0.04 T2 Trt + %Carot + Trt x %Carot 4 0.00 0.29 4 0.00 0.42 T3 Trt + %Melan + Trt x %Melan 4 9.18 0.00 4 9.54 0.00 T4 Trt + %Irid + Trt x %Irid 4 4.71 0.03 4 8.06 0.01 T5 Trt + NCarot + Trt x NCarot 4 10.49 0.00 4 8.87 0.01 T6 Trt + NMelan+ Trt x NMelan 7 9.52 0.00 7 8.99 0.00 T7 Trt + %Irid + %Carot + %Irid x % Carot + Trt x %Irid + Trt x 7 0.98 0.18 7 6.54 0.02 %Carot T8 Trt + %Irid + %Melan + %Irid x %Melan + Trt x %Irid + Trt x 7 7.04 0.01 7 10.47 0.00 %Melan Trt + %Irid + %Carot + %Irid x %Carot + Trt x %Irid + Trt x T9 13 1.06 0.17 13 6.67 0.02 %Carot T11 Trt + NCarot + NMelan + NCarot x NMelan 8 14.38 0.00 8 17.84 0.00 644

645 * D.f., degrees of freedom, ΔAICc, the difference between the best model and evaluated model; relative model weight (w). Model with evidence ratio <

646 4 (wi/wmax), i.e. having not less than 4 times lower support in data than the best model) are highlighted in bold typeset.

26 647 Table 2 Parameter estimates for the best supported models on the effects of fluctuating asymmetry

648 in colouration on laterality in male sexual displays in relation to social environment treatments.

Model parameter (a) Young males (b) Older males

(Trial 1) (Trial 2)

2 P 2 P

Orange 6.52 0.011 5.74 0.017

Social Environment 0.61 0.434 1.77 0.183

Orange x Social Environment 6.43 0.011 4.21 0.040

649

650 2 denotes test statistics from GLM analysis of deviance for binomial distribution; P is statistical

651 significance of the estimate.

652 653 Fig. 1

654

28 655 Fig. 2

656 657 Fig. 3

658

ZÁVĚR

89

Závěr

Ačkoliv je problematika pohlavního výběru jiţ velmi dobře prostudová- na, je zde stále mnoţství zajímavých otázek čekajících na své zodpovězení. Studium zvířecích personalit a osobnostních znaků a jejich vlivu na fitness jedince je proti tomu teprve na samém začátku. Ve své práci jsem se zaměřil na spojení těchto dvou oblastí a studoval jejich vzájemný vztah. Díky laboratorním experimentům a experimentům v polopřírodních podmínkách s hořavkou duhovou (Rhodeus amarus) jsem zjistil, ţe konzis- tentní behaviorální rozdíly mezi jedinci u tohoto druhu existují. Alternativní reprodukční taktiky samců však u tohoto druhu nejsou ovlivněny ţádným ze zkoumaných osobnostních znaků (odvahou či agresivitou), ale vznikají jako důsledek působení vlivu okolních environmentálních a sociálních faktorů s ohledem na morfologické znaky jedince. Jelikoţ ţádné osobnostní znaky nebyly u pokusných jedinců korelovány, tak se nepotvrdila ani přítomnost behaviorálního syndromu. Stejně tak spolu nebyly korelovány ani znaky morfologické, fyziologické a genetické. Behaviorální syndrom nemusí být pří- tomen u hořavek obecně, nebo můţe chybět pouze u studované populace. Takováto mezipopulační variabilita ve výskytu behaviorálního syndromu je známa u Gasterosteus aculeatus (Bell 2005). V systému reprodukčních tak- tik samců hořavek by navíc takové pevné spojení osobnostních znaků mohlo také působit maladaptivně. Korelace mezi znaky, která je výhodná pro terito- riálního jedince, jiţ nemusí být výhodná pro sneakera a vzhledem k častému střídání obou taktik by tato vazba mohla být spíše kontraproduktivní. Dále jsem zjistil, ţe i teritoriální samci Rhodeus amarus jsou velmi dob- ře adaptováni na kompetici spermií. Tato vlastnost bývá většinou v tomto systému alternativních taktik přisuzována pouze sneakerům, kteří mají také většinou vyšší hodnoty gonadosomatického indexu (GSI, poměr hmotnosti gonád ke zbytku těla) neţ jedinci bránící nějaké zdroje. V případě R. amarus měli ovšem výrazně vysoké GSI také teritoriální samci. To ukazuje na nut- nost samců maximálně investovat do všech moţných aspektů soupeření s rivaly. Energii tak vynakládají nejen na agresivní obranu teritoria, ale i na konkurenceschopnost v kompetici spermií. Zároveň byla zjištěna negativní

ZÁVĚR

90 korelace mezi agresivními výpady a počtem ejakulací. To naznačuje, ţe jed- notliví samci se skutečně lišili v upřednostňování způsobů boje proti riva- lům. V experimentech s ţivorodkou Endlerovou (Poecilia wingei) byla zkou- mána další zanedbávaná oblast pohlavního výběru, a to vliv zkušenosti a so- ciálního prostředí na volbu alternativní reprodukční taktiky. Zkoumal jsem, zda bude mít sociální prostředí vliv na původně zvolenou taktiku a zda bude samec schopen reagovat na změnu v tomto prostředí. Ukázalo se, ţe i nezku- šení jedinci mají preferované taktiky a při prvním setkání se samicí preferují buď sneaking, nebo dvoření ve formě sigmoidního předvádění. Tito jedinci poté, co byli drţeni v určitém prostředí, jím byli silně ovlivněni a své chování, které v tomto prostředí zastávali, si navíc přenášeli dále do budoucích se- tkání s potenciálními sexuálními partnerkami. Jedinci, kteří trávili čas v prostředí bez samčího rivala, navíc vykazovali celkový pokles sexuální akti- vity a relativně zvýšenou míru dvoření. Naopak jedinci drţeni se samčím konkurentem se snaţili získat kopulace za pomoci sneakingu. To vše ukazu- je zásadní vliv konkurenčního prostředí na samčí epigamní chování. Neméně jako významný se ukázal vliv sociálního prostředí na výběr alternativní tak- tiky. Potvrdilo se, ţe časově méně náročný sneaking bývá výhodný v konku- renčních prostředích, kde nezbývá na dvoření dostatek času a kde musí sa- mec navíc soupeřit třeba se zdatnějším konkurentem. Zároveň se za pomoci moderních metod analýzy obrazu potvrdila schopnost samců P. wingei identifikovat v průběhu námluv svou pro samici atraktivnější stranu, a tu jí následně přednostně předvádět v průběhu dvo- ření. Ke kvantifikaci chování i barevných vzorů jsem pouţil moderní metody, které jsem pro vlastní potřeby podle mého názoru ještě vylepšil a účinně za- pracoval do experimentů. S jejich pomocí se mi podařilo získat soubory velmi kvalitních dat. Poznatek, ţe jedinec dokáţe na svém laterálně zploštělém těle identifiko- vat svou atraktivnější stranu, není pro vědu úplně nový (Gross et al. 2007; Amcoff et al. 2009). Nové je však zjištění, ţe tento efekt trvá pouze v konku- renčním prostředí. V konkurenceprostém prostředí tento trend vymizel. To

ZÁVĚR

91 opět ukazuje na silný vliv sociálního prostředí na samčí epigamní chování. To, jak se zdá, je kvalitnější a důslednější, pokud je nablízku konkurent. Svou prací jsem přispěl k zaplnění mezery ve výzkumu zvířecích perso- nalit. Experimentálně jsem potvrdil absenci vlivu osobnostních znaků a behaviorálního syndromu na vyuţívání alternativních reprodukčních taktik u hořavky duhové. Současně se mi podařilo zjistit některá fakta o teritoriál- ních jedincích hořavek neshodující se s obecně uznávanými modely kompe- tice spermií ve studovaném systému alternativních taktik. Dále jsem potvrdil vliv sociálního prostředí a zkušenosti na alternativní taktiky dospělých jedin- ců ţivorodky Endlerovy a ukázal, ţe jedinci si ve snaze maximalizovat svou atraktivitu přenášejí behaviorální návyky z minulosti do budoucích setkání. Tento jev však byl úzce vázaný na demografické sloţení prostředí, ve kterém se předtím vyskytovali.

Identifikace potenciálně zajímavých témat pro další výzkum

Kaţdý výzkum, který nalezne odpověď na nějakou otázku, často zároveň odhalí otázky nové. Také výzkum v rámci mé disertační práce není výjimkou. Ukázalo se, ţe individuální zkušenost jedince a sociální prostředí, které ho obklopuje, má na jeho budoucí chování nezanedbatelný vliv. I kdyţ byly o tomto tématu jiţ nějaké práce napsány, stále lze nalézt oblasti, jejichţ po- tenciální výzkum můţe přinést zajímavé výsledky a přispět k lepšímu pozná- ní mechanismů ovlivňujících zvířecí chování. Jedná se především o spojení zkušenosti a pohlavního výběru. Poutavým tématem pro budoucí studium v oblasti zkušenosti a sociál- ního prostředí je určení relativního vlivu přítomnosti samčích konkurentů a vlivu výměny známé samice za novou na reprodukční chování samce. Je znám tzv. Coolidgeův efekt, kdy snaha samce o páření se stále stejnou sami- cí po čase slábne a je obnovena aţ po jeho setkání s jinou (neznámou) samicí (Wilson et al. 1963). Při své práci se samci P. wingei jsem pozoroval, ţe sna- ha o páření můţe být po předchozím ochabnutí obnovena i v přítomnosti známé samice. Samci se náhle snaţili o páření se samicí obývající s nimi

ZÁVĚR

92 stejnou nádrţ po dobu několika měsíců jiţ při pouhém náhlém výskytu sam- čího konkurenta. Bylo by dozajista přínosné srovnání relativní významnosti náhlého výskytu konkurenta a objevení nové partnerky v souvislosti s obno- vou samcovy motivace pářit se. Ideální modelový druh pro tuto studii by měl splňovat především to, ţe jeho epigamní chování bude do detailů prostudo- váno a bude snadná jeho kvantifikace. Vhodný by byl téţ druh, který není příliš promiskuitní a párovacím systémem je někde na pomezí mezi mono- gamií a promiskuitou. Párovací systém by téţ neměl být ovlivněn rodičov- skou péčí či zdroji. Bylo by ovšem rovněţ moţné vyuţít některý z druhů Poe- cilia wingei či P. reticulata, u kterých překvapivě důslednější studium Coolid- geova efektu chybí. V přírodě a společenských nádrţích jsou sice tyto druhy extrémně promiskuitní, nicméně v laboratorních podmínkách se přizpůsobí experimentálnímu prostředí a jejich chování lze designem pokusu velmi účinně modulovat. U obou druhů je reprodukční chování samců velmi dobře popsáno, a je tedy snadná jeho kvantifikace, co se týče kvality i kvantity (Houde 1997; Poeser et al. 2005). Podle mnou dostupných informací existují pouze dvě práce zabývající se Coolidgeoým efektem u P. reticulata – studie Kelley et al. (1999) a Jordan & Brooks (2010). Vzhledem k tomu, jak dlouho a intenzivně je P. reticulata studována, je to velmi překvapivé. Z vlastního po- zorování v průběhu práce na této disertaci mohu potvrdit, ţe Coolidgeův efekt existuje také u P. wingei. Zůstává tedy otázkou, zda je pro samce důle- ţitější motivací k páření (a obecně k epigamnímu chování) přítomnost nové samice či rivala. Předpokládám, ţe větší sexuální aktivitu budou vykazovat samci, kteří budou muset čelit nové konkurenci i přesto, ţe se budou pářit se stále stejnou partnerkou. Domnívám se, ţe kompetitor bude pro samce ţivorodek v pokusných podmínkách mnohem silnějším stimulem sexuálního chování (i kdyţ půjde o známou samicí) neţ přítomnost nové neznámé sami- ce sama o sobě. Tuto hypotézu je však zapotřebí ověřit. Další relativně neprozkoumanou oblastí se zdá být vliv zkušenosti na výběr partnera. Je známo, ţe jedinci se párují asortativně (nenáhodně) a vy- bírají si partnera podle určitých znaků. Ty jsou specifické pro daný taxon, jedince charakterizují a signalizují partnerovu kvalitu. Do jaké míry se ale nechávají jedinci ovlivnit při výběru partnera svou zkušeností, zůstává stále

ZÁVĚR

93 otázkou. V současné literatuře je toto téma opomíjeno. Vliv zkušenosti a so- ciálního prostředí byl prokázán i v mé práci (Řeţucha & Reichard 2014), kdy sociální prostředí a individuální zkušenosti signifikantně ovlivnily sexuální chování samců P. wingei. Domnívám se, ţe podobné mechanismy, které měly vliv na jejich sexuální chování, by se mohly uplatňovat i při samotné volbě partnera, která tomuto chování většinou předchází. Například nepříliš kva- litní jedinec, ať uţ ve smyslu velikosti, barvy atp., který v průběhu svého on- togenetického vývoje interaguje se samicemi ze svého sociálního prostředí a má od nich zpětnou vazbu o své kvalitě, můţe hypoteticky zaujmout místo v jakési sociosexuální nice (angl. socio – sexual niche). V této pozici pak mů- ţe v budoucnu setrvávat bez ohledu na charakteristiku dostupných partne- rů. Díky svým vlastnostem není vyhledáván a akceptován těmi nejlepšími samicemi, a naopak se nechce pářit s méně kvalitními samicemi neţ je on sám. Nejvíce svého úsilí proto směřuje k samicím podobné úrovně a kvality jako je on. V této sociosexuální nice zaujímá určité místo, protoţe je to pro něj nejvýhodnější z hlediska úspory času a energie. Kvalitnější samice ho neakceptují a jejich námluvami by jen ztrácel čas. Naopak dvořením se ne- kvalitním samicím, které jsou v sociosexuální nice v niţší úrovni neţ je jeho, by zase ztrácel čas a energii (které můţe investovat do hledání kvalitnější samice), popř. gamety, jejichţ mnoţství není neomezené dokonce ani u ryb (Smith et al. 2009). Je pak můţe investovat do hledání kvalitnější samice své úrovně, která bude mít kvalitnější genetickou výbavu a jejíţ potomci budou ţivotaschopnější. To se pak samozřejmě odrazí na samcově fitness. Díky to- mu lze předpokládat, ţe samec ovlivněný zkušeností a svým sociálním pro- středím bude při výběru partnerky nejen vybíravý, jak by se dalo tušit, ale díky zpětné vazbě od samic, kterou měl v průběhu svého ontogenetického vývoje, si bude vybírat konkrétní typy samic (které mají podobnou kvalitu a pozici v sociosexuální nice jako on). Ostatním bude věnovat méně pozor- nosti, nebo je bude přímo ignorovat, i kdyţ budou mnohem kvalitnější, neţ jím preferované samice. Tento koncept sociosexuální niky ideově vychází z konceptu sociální niky (Bergmüller & Taborsky 2010), který se snaţí vy- světlit vznik a udrţování osobnostních znaků jedinců ve skupinách. Podle něj se ve skupině jedinci chovají jednak podle toho, k čemu mají dispozice,

ZÁVĚR

94 ale i podle toho, co jim umoţní ostatní jedinci vyskytující se v jejich sociál- ním prostředí. Sociosexuální nika je ale inovativní svou aplikací na volbu partnera. Pokud by se experimentálně potvrdila její platnost, mohlo by to přispět k pochopení párování jedinců, která byla aţ do této chvíle chápána jako suboptimální. Jako modelový druh se zde opět nabízejí ţivorodky rodu Poecilia, u nichţ lze kvalitu obou pohlaví experimentálně ovlivnit a rozdělit do sobě odpovídajících kategorií, stejně jako jejich sociální prostředí a dobu, po jakou mu budou vystaveni. Oblast pro výzkum je to určitě přitaţlivá a podle mého soudu dosud nedostatečně prozkoumaná s potenciálem vrh- nout nové světlo na fungování pohlavního výběru. Třetí oblastí pro moţný budoucí výzkum, kterou bych rád zmínil, je stu- dium důsledků moţné mezidruhové hybridizace Poecilia reticulata a Poecilia wingei, popř. i s třetím, velmi nedávno objeveným sesterským druhem, P. ob- scura (Schories et al. 2009). V limitovaných experimentálních podmínkách druhy P. reticulata a P. wingei hybridizují a poskytují dále plodné potomstvo (vlastní pozorování). Výzkum by se mohl zaměřit na to, zda tento proces pro- bíhá i v přírodnějších podmínkách, např. na úrovni mezokosmu. Zda i tam se budou jedinci ochotně kříţit, nebo si budou vybírat partnera přednostně ze stejného druhu. V případě absence etologické a dalších prezygotických re- produkčních bariér by bylo vhodné zaměřit se na výsledný fenotyp F1 po- tomstva a sledovat ho potom po několik následujících generací. Je známo, ţe se tyto dva druhy liší v několika morfologických (jiný tvar gonopodia, jíné zbarvení) a behaviorálních (absence několika prvků epigamního chování u P. wingei) znacích (Poeser et al. 2005). Mohlo by být důleţité určit výsledný be- haviorální a morfologický fenotyp potomstva, tedy to, zda je intermediální, nebo zda více připomíná ten či onen druh a s jakým druhem (fenotypem) má pak tento hybrid větší tendence se pářit. Výsledky by mohly napovědět něco o osudu P. wingei v přírodě. Tam je její nevelký areál výskytu ve Venezuele ohraničen areálem P. reticulata. V budoucnu by se mohlo stát, ţe areály těchto druhů splynou, ať uţ vlivem eutrofizace a zvýšené turbidity vody v ob- lasti, záměrnými či náhodnými introdukcemi a nebo přirozeným způsobem bez pomoci člověka. Zároveň by mohla takováto studie přispět k detailnější- mu popisu a pochopení chování P. wingei, které je oproti chování P. reticula-

ZÁVĚR

95 ta neprozkoumané a je nutné k pochopení vzdálenosti nebo naopak blízkosti těchto druhů. To je taky nezbytné vzhledem k moţnému budoucímu uplat- nění P. wingei jakoţto nového modelového druhu v behaviorálním výzkumu díky svým etologickým i ţivotním charakteristikám. Další a podle mě zajímavou oblastí pro výzkum zvířecích personalit je vliv osobnosti souseda na výběr vlastního teritoria. Např. u teritoriálních druhů ptáků hnízdících v koloniích můţe být výhodné mít hnízdo vedle nea- gresivního a třeba tichého jedince, který na sebe nepoutá pozornost predáto- rů, a tím neohroţuje ani své sousedy. Zároveň můţe být výhodné sousedit s někým, kdo je explorativní a dokáţe si vybrat teritorium u kvalitního zdroje potravy. Analogicky můţe být tato skutečnost přenesena i do studia ryb. Podle jakých vlastností a prostorového postavení mlţů si samci hořavek hle- dají teritoria, která pak brání, je jiţ dobře prostudováno (např. Przybylski et al. 2007). Bývá ovšem opomíjena sociální stránka. Zůstává otázkou, zda samci tato teritoria nevolí i na základě charakteristik svého souseda. Můţe být výhodné mít teritorium vedle kvalitativně horšího jedince neţ je sameček sám (menšího, méně zbarveného), který nebude tolik přitahovat pozornost predátorů a zvýší atraktivitu svého souseda, protoţe samičky při své volbě často porovnávají jedince mezi sebou (Gasparini et al. 2013). Na druhou stranu můţe zase kvalitnější jedinec pomoci méně kvalitnějšímu s přiláká- ním více samiček na společné území a zároveň můţe více pomoci s jeho ob- ranou proti sneakerům. Zda se ovšem samci hořavek něčím podobným vů- bec řídí a podle čeho své sousedy snad volí, zůstává nezodpovězenou otáz- kou. Z menších dílčích témat lze zmínit otázky, které spíše neţ ucelené téma by mohly samy přispět k lepšímu pochopení biologie hořavek. O alternativ- ních reprodukčních taktikách samců toho bylo napsáno jiţ hodně, avšak chování samic v době reprodukce bývá spíše opomíjeno. Jednou z moţných otázek by mohlo být sledování moţné fluktuace explorace samic v průběhu třecí sezóny. Není zcela jasné, zda samice s prodlouţeným kladélkem v době tření nacházejí samce s mlţi náhodně, či zda se u nich v té době mění jejich osobnostní znaky, především explorace či odvaha, díky čemuţ dokáţí v době tření prozkoumat větší území, neţ tomu je mimo toto období, kdy nepotřebu-

ZÁVĚR

96 jí vyhledávat samce a jejich teritoria s mlţi. Na to pak vzdáleně navazuje zkoumání osobnostních znaků, jako např. agresivita a odvaha u mladých je- dinců narozených na začátku rozmnoţovací sezóny (květen) a na jejím konci (srpen – září). Čím později se jedinec narodí, tím méně času má na získávání energie pro svůj růst a vytváření zásob před zimním obdobím. V této době by mohli být favorizováni ti jedinci, kteří jsou odváţní a agresivní při obstarává- ní ubývající potravy a nehledí na moţné riziko predace. Naopak u jedinců narozených na začátku by mohla být upřednostňována spíše plachost, která by je ochránila před predátory a zároveň jim nebránila v získávání potravy, které je v tomto období ještě na dlouhou dobu dostatek. Je zřejmé, ţe poutavých témat pro další výzkum oblasti zvířecích perso- nalit a pohlavního výběru, které mohou přispět k lepšímu poznání zvířecího chování a jeho ekologických a evolučních důsledků, lze naléz ještě stále velké mnoţství. Proto si myslím, ţe se zkoumání v tomto oboru bude v příštích le- tech dále rozvíjet a bude doplňovat a zpřesňovat současné poznatky.

POUŽITÉ ZDROJE

97

Použité zdroje

Agbali, M., Reichard, M., Bryjová, A., Bryja, J. & Smith, C. 2010. Mate choice for nonadditive genetic benefits correlate with mhc dissimilarity in the rose bitterling (Rhodeus ocellatus). Evolution, 64, 1683–1696. Amcoff, M., Arnqvist, G. & Kolm, N. 2009. Courtship signalling with a labi- le bilateral signal: males show their best side. Behavioral Ecology and Sociobiology, 63, 1717–1725. Amundsen, T. 2003. Fishes as models in studies of sexual selection and parental care. Journal of Fish Biology, 63, 17–52. Andersson, M. 1994. Sexual selection. London: Princeton University Press. Ariyomo, T. O. & Watt, P. J. 2013. Disassortative mating for boldness decreases reproductive success in the guppy. Behavioral Ecology, 24, 1320–1326. Bell, A. M. 2005. Behavioural differences between individuals and two popu- lations of stickleback (Gasterosteus aculeatus). Journal of Evolutionary Biology, 18, 464–473. Bergmüller, R. & Taborsky, M. 2007. Adaptive behavioural syndromes due to strategic niche specialization. BMC Ecology, 7, Bergmüller, R. & Taborsky, M. 2010. Animal personality due to social ni- che specialisation. Trends in Ecology & Evolution, 25, 504–511. Bisazza, A., Rogers, L. J. & Vallortigara, G. 1998. The origins of cerebral asymmetry: a review of evidence of behavioural and brain lateralization in fishes, reptiles and amphibians. Neuroscience and Biobehavioral Re- views, 22, 411–426. Blumstein, D. T., Daniel, J. C. & Evans, C. S. 2006. JWatcher. Boogert, N. J., Fawcett, T. W. & Lefebvre, L. 2011. Mate choice for cogni- tive traits: a review of the evidence in nonhuman vertebrates. Behavioral Ecology, 22, 447–459. Brown, C., Jones, F. & Braithwaite, V. 2005. In situ examination of bold- ness–shyness traits in the tropical poeciliid, Brachyraphis episcopi. Animal Behaviour, 70, 1003–1009. Budaev, S. & Brown, C. 2011. Personality traits and behaviour. In: Fish co- gnition and behavior, 2nd edn. (Ed. by C. Brown, K. Laland, & J. Krau- se), pp. 135–165. Cambridge: Wiley-Blackwell. Burns, J. G. 2008. The validity of three tests of temperament in guppies (Po- ecilia reticulata). Journal of Comparative Psychology, 122, 344–356. Cain, S. 2012. Quiet: the power of introverts in a world that can’t stop talking. New York: Crown Publishing Group.

POUŽITÉ ZDROJE

98

Candolin, U. 2004. Effects of algae cover on egg acquisition in male three- spined stickleback. Behaviour, 141, 1389–1399. Candolin, U. & Reynolds, J. D. 2001. Sexual signaling in the European bit- terling: females learn the truth by direct inspection of the resource. Be- havioral Ecology, 12, 407–411. Candolin, U. & Reynolds, J. D. 2002. Why do males tolerate sneakers? Tests with the European bitterling, Rhodeus sericeus. Behavioral Ecolo- gy and Sociobiology, 51, 146–152. Carere, C. & Maestripieri, D. Eds. 2013. Animal personalities: behavior, physiology, and evolution. Chicago: University of Chicago Press. Colléter, M. & Brown, C. 2011. Personality traits predict hierarchy rank in male rainbowfish social groups. Animal Behaviour, 81, 1231–1237. Cote, J., Clobert, J., Brodin, T., Fogarty, S. & Sih, A. 2010. Personality- dependent dispersal: characterization, ontogeny and consequences for spatially structured populations. Philosophical Transactions of the Royal Society B, 365, 4065–4076. Darwin, C. 1859. On the origin of species by means of natural selection. London: John Murray. Denoël, M. & Doellen, J. 2010. Displaying in the dark: light-dependent al- ternative mating tactics in the Alpine newt. Behavioral Ecology and So- ciobiology, 64, 1171–1177. Dingemanse, N. J., Both, C., Drent, P. J. & Tinbergen, J. M. 2005. Pairs of extreme avian personalities have highest reproductive success. Jour- nal of Animal Ecology, 74, 667–674. Dzieweczynski, T. L., Russell, A. M., Forrette, L. M. & Mannion, K. L. 2014. Male behavioral type affects female preference in Siamese fighting fish. Behavioral Ecology, 25, 136–141. Endler, J. A. & Houde, A. E. 1995. Geographic variation in female preferen- ces for male traits in Poecilia reticulata. Evolution, 49, 456–468. Evans, J. P., Pilastro, A. & Schlupp, I. Eds. 2011. Ecology and evolution of Poeciliid fishes. Chicago: The University of Chicago Press. Evans, J. P. & Magurran, A. E. 1999. Male mating behaviour and sperm production characteristics under varying sperm competition risk in guppies. Animal Behaviour, 58, 1001–1006. Evans, J. P. & Magurran, A. E. 2001. Patterns of sperm precedence and predictors of paternity in the Trinidadian guppy. Proceedings of the Royal Society B, 268, 719–724. Farr, J. A. 1980. The effects of sexual experience and female receptivity on courtship-rape decisions in male guppies, Poecilia reticulata (Pisces: Po- eciliidae). Animal Behaviour, 28, 1195–1201.

POUŽITÉ ZDROJE

99

Farr, J. A. & Herrnkind, W. F. 1974. A quantitative analysis of social inter- action of the guppy, Poecilia reticulata (pisces: Poeciliidae) as a function of population density. Animal Behaviour, 22, 582–591. Fisher, R. A. 1930. Sexual reproduction and sexual selection. In: The gene- tical theory of natural selection, pp. 121–145. Oxford: Oxford University Press. Freeberg, T. M. 2000. Culture and courtship in vertebrates: a review of so- cial learning and transmission of courtship systems and mating pat- terns. Behavioural Processes, 51, 177–192. Gasparini, C., Serena, S. & Pilastro, A. 2013. Do unattractive friends make you look better? Context-dependent male mating preferences in the guppy. Proceedings of the Royal Society B, 280, 20123072. Godin, J. G. J. 1995. Predation risk and alternative mating tactics in male Trinidadian guppies (Poecilia reticulata). Oecologia, 103, 224–229. Godin, J. G. J. & Dugatkin, L. A. 1996. Female mating preference for bold males in the guppy, Poecilia reticulata. Proceedings of the National Aca- demy of Sciences of the United States of America, 93, 10262–10267. Gross, M. R., Suk, H. Y. & Robertson, C. T. 2007. Courtship and genetic quality: asymmetric males show their best side. Proceedings of the Royal Society B, 274, 2115–2122. Guevara-Fiore, P. 2012. Early social experience significantly affects sexual behaviour in male guppies. Animal Behaviour, 84, 191–195. Guevara-Fiore, P., Svensson, P. A. & Endler, J. A. 2012. Sex as moderator of early life experience: interaction between rearing environment and sexual experience in male guppies. Animal Behaviour, 84, 1023–1029. Harfmann Short, K. & Petren, K. 2008. Boldness underlies foraging suc- cess of invasive Lepidodactylus lugubris geckos in the human landsca- pe. Animal Behaviour, 76, 429–437. Herdman, E. J. E., Kelly, C. D. & Godin, J. J. 2004. Male mate choice in the guppy (Poecilia reticulata): do males prefer larger females as mates? Ethology, 110, 97–111. Houde, A. E. 1997. Sex, color, and mate choice in guppies. New Jersey: Prin- ceton University Press. Huntingford, F. A. 1976. The relationship between anti-predator behaviour and aggression among conspecifics in the three-spined stickleback, Gasterosteus aculeatus. Animal Behaviour, 24, 245–260. Chan, T. Y. & Ribbink, J. 1990. Alternative reproductive behaviour in fishes, with particular reference to Lepomis macrochira and Pseudocreni- labrus philander. Environmental Biology of Fishes, 28, 249–256.

POUŽITÉ ZDROJE

100

Chapman, B. B., Morrell, L. J. & Krause, J. 2009. Plasticity in male court- ship behaviour as a function of light intensity in guppies. Behavioral Ecology and Sociobiology, 63, 1757–1763. Isvaran, K. 2005. Variation in male mating behaviour within ungulate popu- lations: patterns and processes. Current Science, 89, 1192–1199. Jennions, M. D. & Petrie, M. 2000. Why do females mate multiply? A re- view of the genetic benefits. Biological Reviews, 75, 21–64. Jirotkul, M. 2000. Male trait distribution determined alternative mating tactics in guppies. Journal of Fish Biology, 56, 1427–1434. Jordan, L. A. & Brooks, R. C. 2010. The lifetime costs of increased male re- productive effort: courtship, copulation and the Coolidge effect. Journal of Evolutionary Biology, 23, 2403–2409. Jordan, L. A. & Brooks, R. C. 2012. Recent social history alters male court- ship preferences. Evolution, 66, 280–287. Kanoh, Y. 1996. Pre-oviposition ejaculation in externally fertilising fish: How sneaker male rose bitterlings contrive to mate. Ethology, 102, 883–899. Karino, K. & Kamada, N. 2009. Plasticity in courtship and sneaking beha- viors depending on tail length in the male guppy, Poecilia reticulata. Ichthyological Research, 56, 253–259. Karino, K. & Kobayashi, M. 2005. Male alternative mating behaviour de- pending on tail length of the guppy, Poecilia reticulata. Behaviour, 142, 191–202. Kelley, J. L., Graves, J. A. & Magurran, A. E. 1999. Familiarity breeds contempt in guppies. Nature, 401, 661–662. Kelley, J. L., Phillips, S. C. & Evans, J. P. 2013. Individual consistency in exploratory behaviour and mating tactics in male guppies. Naturwis- senschaften, 100, 965–974. Kiritome, A., Sato, A. & Karino, K. 2012. Influences of orange spot pat- terns and operational sex ratio on male mating behaviors in the guppy Poecilia reticulata. Ichthyological Research, 59, 304–313. Kokko, H. & Rankin, D. J. 2006. Lonely hearts or sex in the city? Density- dependent effects in mating systems. Philosophical Transactions of the Royal Society B, 361, 319–334. Kolluru, G. R. & Grether, G. F. 2004. The effects of resource availability on alternative mating tactics in guppies (Poecilia reticulata). Behavioral Eco- logy, 16, 294–300. Kolluru, G. R., Grether, G. F. & Contreras, H. 2007. Environmental and genetic influences on mating strategies along a replicated food availabi- lity gradient in guppies (Poecilia reticulata). Behavioral Ecology and Soci- obiology, 61, 689–701.

POUŽITÉ ZDROJE

101

Kolluru, G. R., Grether, G. F., Dunlop, E. & South, S. H. 2009. Food avai- lability and parasite infection influence mating tactics in guppies (Poeci- lia reticulata). Behavioral Ecology, 20, 131–137. Lank, D. B., Smith, C. M., Hanotte, O., Burke, T. & Cooke, F. 1995. Ge- netic polymorphism for alternative mating behaviour in lekking male ruff Philomachus pugnax. Nature, 378, 59–62. Latscha, T. 1990. Carotenoids - their nature and significance in animal feeds / Thierry Latscha. Basel: F. Hoffmann-La Roche, Animal Nutrition and Health. Liley, N. R. 1966. Ethological isolating mechanisms in four sympatric spe- cies of poeciliid fishes. Behaviour, 13, 1–197. Locatello, L., Rasotto, M. B., Evans, J. P. & Pilastro, A. 2006. Colourful male guppies produce faster and more viable sperm. Journal of Evoluti- onary Biology, 19, 1595–1602. López-Sepulcre, A., Gordon, S. P., Paterson, I. G., Bentzen, P. & Rez- nick, D. N. 2013. Beyond lifetime reproductive success: the posthu- mous reproductive dynamics of male Trinidadian guppies. Proceedings of the Royal Society B, 280, 20131116. Magellan, K. & Magurran, A. E. 2006. Habitat use mediates the conflict of interest between the sexes. Animal Behaviour, 72, 75–81. Magurran, A. E. & Magellan, K. 2007. Behavioural profiles: individual con- sistency in male mating behaviour under varying sex ratios. Animal Be- haviour, 74, 1545–1550. Magurran, A. E. & Seghers, B. H. 1994. A cost of sexual harassment in the guppy, Poecilia reticulata. Proceedings of the Royal Society B, 258, 89– 92. Meredith, R. W., Pires, M. N., Reznick, D. N. & Springer, M. S. 2010. Mo- lecular phylogenetic relationships and the evolution of the placenta in Poecilia (Micropoecilia) (Poeciliidae: Cyprinodontiformes). Molecular Phy- logenetics and Evolution, 55, 631–639. Møller, A. P. & Jennions, M. 2001. How important are direct fitness bene- fits of sexual selection? Naturwissenschaften, 88, 401–415. Neff, B. D. & Pitcher, T. E. 2005. Genetic quality and sexual selection: an integrated framework for good genes and compatible genes. Molecular Ecology, 14, 19–38. Ojanguren, A. F. & Magurran, A. E. 2004. Uncoupling the links between male mating tactics and female attractiveness. Proceedings of the Royal Society B, 271, S427–S429. Oliveira, R. F., Taborsky, M. & Brockmann, H. J. Eds. 2008. Alternative reproductive tactics: an integrative approach. 1. edn. New York: Ca- mbridge University Press.

POUŽITÉ ZDROJE

102

Olson, V. A. & Owens, I. P. F. 1998. Costly sexual signals: are carotenoids rare, risky or required? Trends in Ecology and Evolution, 13, 510–514. Parker, G. A. 1974. Courtship persistence and female-guarding as male time investment strategies. Behaviour, 48, 157–184. Patterson, L. D. & Schulte-Hostedde, A. I. 2011. Behavioural correlates of parasitism and reproductive success in male eastern chipmunks, Ta- mias striatus. Animal Behaviour, 81, 1129–1137. Pilastro, A. & Bisazza, A. 1999. Insemination efficiency of two alternative male mating tactics in the guppy Poecilia reticulata. Proceedings of the Royal Society B, 266, 1887–1891. Poeser, F. N., Kempkes, M. & Isbrücker, I. J. H. 2005. Description of Poe- cilia (Acanthophacelus) wingei n. sp. from the Paría Peninsula, Venezue- la, including notes on Acanthophacelus Eigenmann, 1907 and other subgenera of Poecilia Bloch and Schneider, 1801 (Teleostei, Cyprinodon- tiformes, Poeciliidae). Contributions to Zoology, 74, 97–115. Przybylski, M., Reichard, M., Spence, R. & Smith, C. 2007. Spatial distri- bution of oviposition sites determines variance in the reproductive rate of European bitterling (Rhodeus amarus). Behaviour, 144, 1403–1417. Réale, D., Reader, S. M., McDougall, P. T., Sol, D. & Dingemanse, N. J. 2007. Integrating animal temperament within ecology and evolution. Bi- ological Reviews, 82, 291–318. Réale, D., Martin, J., Coltman, D. W., Poissant, J. & Festa-Bianchet, M. 2009. Male personality, life-history strategies and reproductive success in a promiscuous mammal. Journal of Evolutionary Biology, 22, 1599– 1607. Reaney, L. T. & Backwell, P. R. Y. 2007. Risk-taking behavior predicts aggression and mating success in a fiddler crab. Behavioral Ecology, 18, 521–525. Reichard, M., Smith, C. & Jordan, W. C. 2004. Genetic evidence reveals density-dependent mediated success of alternative mating behaviours in the European bitterling (Rhodeus sericeus). Molecular Ecology, 13, 1569–1578. Reichard, M., Smith, C. & Bryja, J. 2008. Seasonal change in the oppor- tunity for sexual selection. Molecular Ecology, 17, 642–651. Reichard, M., Ondračková, M., Bryjová, A., Smith, C. & Bryja, J. 2009. Breeding resouce distribution affects selection gradients on male pheno- typic traits: experimental study on lifetime reproductive success in the bitterling fish (Rhodeus amarus). Evolution, 63, 377–390.

POUŽITÉ ZDROJE

103

Reichard, M., Spence, R., Bryjová, A., Bryja, J. & Smith, C. 2012. Female rose bitterling prefer MHC-dissimilar males: experimental evidence. PLoS ONE, 7, e40780. Reynolds, J. D., Gross, M. R. & Coombs, M. J. 1993. Environmental con- ditions and male morphology determine alternative mating behaviour in Trinidadian guppies. Animal Behaviour, 45, 145–152. Rodd, F. H. & Sokolowski, M. B. 1995. Complex origins of variation in the sexual behaviour of male Trinidadian guppies, Poecilia reticulata: inter- actions between social environment, heredity, body size and age. Animal Behaviour, 49, 1139–1159. Rousset, F. 2008. Genepop’007: a complete re-implementation of the Gene- pop software for Windows and Linux. Molecular Ecology Resources, 8, 103–106. Ryan, M. J. 1998. Sexual selection, receiver biases, and the evolution of sex differences. Science, 281, 1999–2003. Řěţucha, R. 2010. Vztah mezi osobnostními charakteristikami a repro- dukčními strategiemi samců hořavky duhové (Rhodeus amarus). Diplo- mová práce, Masarykova univerzita, Brno. Řeţucha, R. & Reichard, M. 2014. The effect of social environment on al- ternative mating tactics in male Endler’s guppy, Poecilia wingei. Animal Behaviour, 88, 195–202. Řeţucha, R., Smith, C. & Reichard, M. 2012. Personality traits, reprodu- ctive behaviour and alternative mating tactics in male European bitter- ling, Rhodeus amarus. Behaviour, 149, 531–553. Sathyan, R. & Couldridge, V. C. K. 2013. Female preference for blue in Ja- pan blue guppies (Poecilia reticulata). Environmental Biology of Fishes, 96, 953–959. Sato, A., Shimoichi, A. & Karino, K. 2011. Copulation type affects parturi- tion in the guppy. Zoological Science, 28, 98–104. Shuster, S. M. & Wade, M. J. 1991. Equal mating success among male re- productive strategies in a marine isopod. Nature, 350, 608–610. Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Ti- nevez, J. Y., White, D. J., Hartenstein, V., Eliceiri, K., Tomancak, P. & Cardona, A. 2012. Fiji: an open-source platform for biological- image analysis. Nature Methods, 9, 676–682. Schories, S., Meyer, M. K. & Schartl, M. 2009. Description of Poecilia (Acanthophacelus) obscura n. sp.,(Teleostei: Poeciliidae), a new guppy species from western Trinidad, with remarks on P. wingei and the sta- tus of the “Endler’s guppy.”Zootaxa, 2266, 35–50.

POUŽITÉ ZDROJE

104

Schuett, W., Tregenza, T. & Dall, S. R. X. 2010. Sexual selection and ani- mal personality. Biological Reviews, 85, 217–246. Schuett, W., Godin, J. G. J. & Dall, S. R. X. 2011. Do female zebra finches, Taeniopygia guttata, choose their mates based on their “Personality”? Ethology, 117, 908–917. Sih, A. 2013. Frontiers on the interface between behavioral syndromes and social behavioral ecology. In: Animal personalities: behavior, physiology, and evolution, 1st edn. (Ed. by C. Carere & D. Maestripieri), pp. 221– 251. Chicago: University of Chicago Press. Sih, A., Bell, A. M., Johnson, J. C. & Ziemba, R. E. 2004. Behavioral syn- dromes: an integrative overview. The Quarterly Review of Biology, 79, 241–277. Sih, A., Cote, J., Evans, M., Fogarty, S. & Pruitt, J. 2012. Ecological im- plications of behavioural syndromes. Ecology Letters, 15, 278–289. Smith, C. & Reichard, M. 2005. Females solicit sneakers to improve fertili- zation success in the bitterling fish (Rhodeus sericeus). Proceedings of the Royal Society B, 272, 1683–1688. Smith, C., Rippon, K., Douglas, A. & Jurajda, P. 2001. A proximate cue for oviposition site choice in the bitterling (Rhodeus sericeus). Freshwater Biology, 46, 903–911. Smith, C., Douglas, A. & Jurajda, P. 2002. Sexual conflict, sexual selection and sperm competition in the spawning decisions of bitterling, Rhodeus sericeus. Behavioral Ecology and Sociobiology, 51, 433–439. Smith, C., Reichard, M., Jurajda, P. & Przybylski, M. 2004. The reprodu- ctive ecology of the European bitterling (Rhodeus sericeus). Journal of Zoology, 262, 107–124. Smith, C., Pateman-Jones, C., Zięba, G., Przybylski, M. & Reichard, M. 2009. Sperm depletion as a consequence of increased sperm competiti- on risk in the European bitterling, Rhodeus amarus. Animal Behaviour, 77, 1227–1233. Smith, C., Phillips, A., Polačik, M. & Reichard, M. 2014. Male coloration signals direct benefits in the European bitterling (Rhodeus amarus). En- vironmental Biology of Fishes, 97, 335–341. Taborsky, M. 2001. The evolution of bourgeois, parasitic, and cooperative reproductive behaviors in fishes. Journal of Heredity, 92, 100–110. Thornhill, R. 1981. Panorpa (Mecoptera: Panorpidae) scorpionflies: systems for understanding resource-defense polygyny and alternative male re- productive efforts. Annual Review of Ecology and Systematics, 12, 355– 386.

POUŽITÉ ZDROJE

105

Van Damme, D., Bogutskaya, N., Hoffmann, R. C. & Smith, C. 2007. The introduction of the European bitterling (Rhodeus amarus) to west and central Europe. Fish and Fisheries, 8, 79–106. Wilson, A. D. M. & Godin, J. G. J. 2009. Boldness and behavioral syndro- mes in the bluegill sunfish, Lepomis macrochirus. Behavioral Ecology, 20, 231–237. Wilson, J. R., Kuehn, R. E. & Beach, F. A. 1963. Modification in the sexual behavior of male rats produced by changing the stimulus female. Jour- nal of Comparative and Physiological Psychology, 56, 636–644. Wilson, D. S., Coleman, K., Clark, A. B. & Biederman, L. 1993. Shy-bold continuum in pumpkinseed sunfish (Lepomis gibbosus): An ecological study of a psychological trait. Journal of Comparative Psychology, 107, 250–260. Wilson, D. S., Clark, A. B., Coleman, K. & Dearstyne, T. 1994. Shyness and boldness in humans and other animals. Trends in Ecology & Evolu- tion, 9, 442–446. Wolf, M. & Weissing, F. J. 2012. Animal personalities: consequences for ecology and evolution. Trends in Ecology and Evolution, 27, 452–461. Zahavi, A. 1975. Mate selection - a selection for a handicap. Journal of Theo- retical Biology, 53, 205–214.

NAVŠTÍVENÉ KONFERENCE S VLASTNÍM PŘÍSPĚVKEM

106

Navštívené konference s vlastním příspěvkem

1. ZOOLOGICKÉ DNY 2011. Přednáška: Řeţucha, R., Reichard, M.: Osobnostní charakteristiky a reprodukční strategie u samců ho- řavky duhové (Rhodeus amarus). 17. – 18. 2. 2011, Brno, Czech Republic

2. ISBE 2012 - THE 14TH INTERNATIONAL BEHAVIORAL ECOLOGY CONGRESS. Poster: Řeţucha, R., Reichard, M.: The effect of soci- al environment on mating tactic of male Endler’s guppy (Poecilia wingei). 12. – 17. 8. 2012, Lund, Sweden

NAVŠTÍVENÉ KONFERENCE S VLASTNÍM PŘÍSPĚVKEM

107

3. XIV EUROPEAN CONGRESS OF ICHTHYOLOGY 2012. Přednáška: Řeţucha, R., Reichard, M.: The effect of social environment on mating tactic of male Endler’s guppy (Poecilia wingei). 3. – 8. 7. 2012, Liège, Belgium

NAVŠTÍVENÉ KONFERENCE S VLASTNÍM PŘÍSPĚVKEM

108

4. BIO@SNS SYMPOSIUM SERIES: I NOTHOBRANCHIUS SYMPOSI- UM. FROM BUSH TO BENCH: 10 YEARS OF NOTHOBRANCHIUS FURZERI AS A MODEL SYSTEM IN BIOLOGY. Poster: Řeţucha, R., Reichard, M.: Boldness, aggression and locomotor activity in four Nothobranchius species. 6. – 8. 2. 2014, Pisa, Italy

PŘÍLOHA – PUBLIKACE NESOUVISEJÍCÍ S DISERTACÍ

109

Příloha – publikace nesouvisející s disertací

Článek A

Polačik, M., Blaţek, R., Řeţucha, R., Vrtílek, M., Terzibasi Tozzini, E. & Reichard, M. 2014. Alternative intrapopulation strategies and their trade-offs in an African annual fish. Journal of Evolutionary Biology, 27, 854 – 865.

V této práci jsem navrhl design měření aktivity halančíka Furzerova (No- thobranchius furzeri) a měření také sám realizoval. Získaná data jsem později vyhodnotil za pomoci softwaru na analýzu chování.

doi: 10.1111/jeb.12359

Alternative intrapopulation life-history strategies and their trade-offs in an African annual fish

M. POLACIK*,R.BLAZEK*, R. RE ZUCHA* †,M.VRTILEK*†, E. TERZIBASI TOZZINI‡ & M. REICHARD* *Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Brno, Czech Republic †Department of Botany and Zoology, Masaryk University, Brno, Czech Republic ‡Scuola Normale Superiore, Pisa, Italy

Keywords: Abstract alternative embryonic pathway; In ephemeral habitats, the same genotypes cope with unpredictable environ- lifespan; mental conditions, favouring the evolution of developmental plasticity and periodic habitat; alternative life-history strategies (ALHS). We tested the existence of intra- phenotypic plasticity; population ALHS in an annual killifish, Nothobranchius furzeri, inhabiting polyphenism; temporary pools. The pools are either primary (persisting throughout the senescence rate. whole rainy season) or secondary (refilled after desiccation of the initial pool), representing alternative niches. The unpredictable conditions led to the evolution of reproductive bet-hedging with asynchronous embryonic development. We used a common garden experiment to test whether the duration of embryonic period is associated with post-embryonic life-history traits. Fish with rapid embryonic development (secondary pool strategy, high risk of desiccation) produced phenotypes with more rapid life-history traits than fish with slow embryonic development (primary pool strategy). The fast fish were smaller at hatching but had larger yolk sac reserves. Their post-hatching growth was more rapid, and they matured earlier. Further, fast fish grew to a smaller body size and died earlier than slow fish. No dif- ferences in fecundity, propensity to mate or physiological ageing were found, demonstrating a combination of plastic responses and constraints. Such developmentally related within-population plasticity in life history is exceptional among vertebrates.

developmental plasticity is common in invertebrates Introduction but unusual in vertebrates (Stearns, 1989; West-Eber- Phenotypic plasticity is the capacity of a genotype to hard, 1989; Podrabsky et al., 2010a; Beldade et al., produce different phenotypes in response to environ- 2011) as they more often display highly conserved mental variation (Piersma & Gils, 2011). Particular developmental programmes (Podrabsky et al., 2010a). environments can favour the existence of discrete pat- Ephemeral habitats in which organisms with the terns of variation that can coexist as distinct life-history same genotypes must cope with unpredictable environ- strategies (Stearns, 1989). A typical example of discrete mental conditions promote the evolution of develop- life-history strategies sharing the same gene pool is po- mental plasticity and alternative life-history strategies lyphenism in butterflies (e.g. Brakefield & French, (e.g. Marcus & Weeks, 1997; Pfennig & McGee, 2010). 1999; Van Dyck & Wiklund, 2002) where environmen- Temporary aquatic pools are periodically filled with tal cues trigger distinct developmental pathways water, and their existence or extinction is largely a sto- (Van Dyck & Wiklund, 2002; Oostra et al., 2011) to chastic process (Wanschoenwinkel et al., 2010). These produce seasonally specific phenotypes. Such adaptive unpredictable conditions can give rise to ‘bet-hedging’ strategies that reduce the risk of hatching failure of all offspring by production of offspring with a variable Correspondence: Matej Polacik, Institute of Vertebrate Biology, Academy of Sciences of the Czech Republic, Kvetn a 8, 603 65 Brno, Czech length of embryonic development (Stearns, 1976). Republic. Tel.: +420 543 422 521; fax: +420 543 211 346; e-mail: Asynchronous developmental rates and alternative life [email protected] histories potentially ensure some progeny are able to

ª 2014 THE AUTHORS. J. EVOL. BIOL. JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 1 2 M. POLACIK ET AL.

complete development when the pool becomes filled that reproductive bet-hedging can lead to the evolution with water. of distinct post-embryonic phenotypes associated with Annual killifish are a unique example of this strategy either fast (skipping diapause) or slow (entering dia- among vertebrates (Wourms, 1972; Watters, 2009; pause) embryonic development (hereafter termed ‘fast’ Podrabsky et al., 2010a,b). They inhabit temporary and ‘slow’ fish, respectively). The slow fish show the savannah and pampa pools of Africa and South Amer- phenotype entering a diapause and surviving in dor- ica. Populations survive seasonal drying as embryos in mancy through the entire dry season lasting several the form of eggs buried in sediment (Wildekamp, months. In contrast, fast fish bypass diapause and 2004). Such bet-hedging is manifested through varia- develop rapidly, with an embryonic period of only sev- tion in the length of their embryonic development eral weeks. The latter are hypothesized to be adapted to which spans from 3 weeks to several years (Wourms, exploit secondary pools, emerging after the first genera- 1972; Wildekamp, 2004; Genade, 2005), with wide var- tion of fish in a given rainy season have already repro- iation within a single clutch (Genade, 2005; Podrabsky duced and died and when primary pools undergo et al., 2010a; Blazek et al., 2013). The asynchronous desiccation. If fast fish are adapted to utilize secondary development is achieved through a system of three fac- pools through alterations to their developmental pro- ultative stages of diapause of variable duration (termed gramme, they are expected to express modified life-his- diapause I–III) (Wourms, 1972). Plasticity in embryonic tory traits, with corresponding trade-offs, as predicted timing is adaptive because annual killifish live in by the life-history theory (Charlesworth, 1980; Reznick unpredictable habitats (Wildekamp, 2004; Mazuze, et al., 1990; Stearns, 1992; Roff, 2002). These predic- 2007; Watters, 2009). The precipitation patterns vary tions are summarized in Table 1. among years, producing pools existing from just a few To test these predictions, we performed a common weeks to more than 10 months (Terzibasi-Tozzini et al., garden experiment where fast and slow fish were 2013). In contrast to primary pools, which remain filled simultaneously raised under the same conditions. We with water throughout the rainy season, some pools used Nothobranchius furzeri Jubb, an African annual may desiccate during the rainy season and refill again killifish species from a region with a particularly only following further extensive precipitation (Fig. S1). unpredictable rainfall pattern (Mazuze, 2007; Terzibasi- Such secondary pools may contain a subsequent gener- Tozzini et al., 2013; M. Reichard, unpublished data). All ation of fish (Podrabsky et al., 2010b), likely from the experimental fish originated from a single founding and eggs deposited earlier in the same rainy season, as some recently imported population and were hatched from embryos bypass all diapauses and develop rapidly the eggs with contrasting durations of embryonic devel- (Podrabsky et al., 2010b; Blazek et al., 2013). opment (see below). The species N. furzeri is well A relationship between the duration of embryonic studied (e.g. Genade et al., 2005; Terzibasi et al., 2008; development and post-embryonic phenotypic traits is Reichard et al., 2009; Polacik & Reichard, 2011; expected to be adaptive. The duration of secondary Terzibasi-Tozzini et al., 2013) and expresses remarkable pools is inevitably shorter than that of primary pools, as developmental asynchrony in its embryonic period primary pools exist over an entire season. Hence, (Genade, 2005). It exhibits an exceptionally fast individuals with a short embryonic period should be maturation and an unusually short minimum genera- adapted to a more rapid completion of life cycle given tion time (Blazek et al., 2013) and natural lifespan the expectation of shorter duration of their habitat. For (Valdesalici & Cellerino, 2003). The unpredictable example, two distinct subgroups were identified in character of N. furzeri habitats and interannual variabil- captive populations of Nothobranchius guentheri Pfeffer, ity in habitat duration (Terzibasi-Tozzini et al., 2013) which differed in lifespan (Markofsky & Perlmutter, make it an unusually amenable model for investigating 1972). The short-lived fish initially grew faster, reached phenotypic plasticity and bet-hedging strategies in key maturity earlier, but ceased growth sooner and attained life-history traits. a smaller maximum size than the long-lived fish (Markofsky & Perlmutter, 1973). Regarding embryonic Materials and methods development, differences between fast-developing (bypassing diapause) embryos and those entering dia- Experimental fish pause were described in Neotropical annual killifish Austrofundulus limnaeus Schultz (Podrabsky et al., Experimental fish were noninbred descendants of 2010a). However, none of those studies related the var- wild-caught N. furzeri, imported from the Chefu iability in the embryonic development to post-embry- region in Gaza Province, southern Mozambique (GPS: S onic phenotype or placed their results in the context of 21°52024.84″,E32°4802.34″) under a collection code adaptive phenotypic plasticity and alternative life-his- MZCS 222 in April 2011. The founding stock consisted tory strategies. of 20 males and 40 females. Imported fish were bred Here, we link the length of the embryonic period to according to a laboratory breeding protocol to maximize post-embryonic life-history traits to test the hypothesis offspring outbreeding.

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY Alternative life histories in an annual fish 3

Table 1 Summary of the predictions for the differences in the life-history traits derived from life-history theory and sample size in comparisons between treatment groups. ‘+’ denotes predictions of an increase, and ‘’ denotes predicted decrease in trait value in comparison with the alternative treatment group. The traits which followed initial predictions are in bold.

Predicted difference N slow fish N fast fish

Trait Slow fish Fast fish Juv Males Females Juv Males Females

Size at hatching + 28* 29* Yolk sac area + 13* 19* Growth rate + 23 22 24 21 Sexual maturation + 23 22 24 21 Willingness to spawn + 20 20 20 20 Lipofuscin deposition + 55 64 Locomotor activity + 22 (14) 19 (6) 23 (15) 20 (4) Final body size + 17 15 18 17 Lifespan + 18 NA 18 NA

*Some individuals were not well preserved and therefore not measured, yielding nonmatching N among treatments.

There were two experimental treatment groups of incubation, all developed embryos (detectable by pig- fish – fast and slow. The fish were assigned to a treat- mented or golden eyes) (Genade, 2005) were removed ment based on the natural length of embryonic devel- from the population. This ensured that at the hatching opment they displayed; all embryos were maintained date (day 200 of the incubation), the pool of eggs only under the same conditions, and their developmental consisted of the embryos which had recently completed rate was not altered by environmental manipulation. development and their realized embryonic development The adult lifespan of N. furzeri is much shorter than the was slow. typical duration of embryonic development (here repre- The fast fish in the experiment were the F2 genera- sented by the slow fish treatment) (Valdesalici & Celle- tion of the imported fish. The eggs that gave origin to rino, 2003, this study), and we were not able to the fast fish were spawned 50 days prior to their hatch- produce both slow and fast embryos from the same ing by a group of 14 males and 21 females (more than generation of parental fish. This is because parents of 500 eggs collected) and incubated identically to the the original cohort of embryos used for the selection of slow fish treatment. Notably, both parental generations slow-developing fish were already naturally deceased at (and also later generations) produce fast and slow the time when we needed to produce a new cohort of embryos within the same clutch, which is a common embryos for the selection of fast-developing embryos. feature in annual killifish (Genade, 2005; Podrabsky We prioritized strict common garden conditions for all et al., 2010a; Blazek et al., 2013; M. Polacik, unpub- experimental fish. All fish experienced the same devel- lished data). opmental conditions during their embryonic period (details below), were hatched on the same day and Experimental conditions raised together ensuring identical water quality and feeding regime. To incubate experimental fish, fertilized eggs were The slow fish were the F1 generation of the imported placed in damp peat following standard Nothobranchius founder fish population. Eggs that were spawned culture protocol (Genade, 2005; Genade et al., 2005) within the 6 weeks after the fish were imported and sealed in plastic bags to incubate at a constant tem- were discarded to eliminate potential maternal effects perature (22 °C) in a laboratory incubator (Q-Cell, Pol- linked to the original environment (Green, 2008). lab, www.poll.pl). Embryos of N. furzeri require a dry Nothobranchius species are income breeders, and egg substrate during their development to hatch success- turnover is extremely rapid. Nothobranchius females are fully, and in the laboratory, this is normally achieved capable of regenerating mature oocyte stock every using damp peat moss. To ensure common garden con- 2 days (Polacik & Reichard, 2009), and the period of ditions during embryonic development, individuals 6 weeks included multiple oocyte maturation cycles. from both treatment groups were incubated in peat of The eggs that gave origin to the slow fish were collected the same origin (Kera, Belarus) and under the same from a spawning of all imported fish (20 males, 40 conditions. females) over 2 days. Given a typical daily clutch pro- All experimental embryos (fast fish treatment: duction of 20–40 eggs per female (Polacik & Reichard, 50 days old, slow fish treatment: 200 days old) were 2011; this study) and high number of the collected eggs simultaneously wetted with water, which induces (> 1000 eggs), they certainly originated from a variety hatching at the final stage of development. Four hours of parental combinations. After 150 days of embryo after the wetting (when hatching of all developed fish

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 4 M. POLACIK ET AL.

was attained), a subsample of fish from each group was females. This negligible growth rate (< 0.1 mm) may be anaesthetized with clove oil and preserved in 4% form- largely affected by the measurement error (fish were aldehyde to obtain data on early life-history traits (see measured alive), and we compared growth rates only below). until reaching this asymptotic size. Newly hatched fish were kept in groups housed in Schedule of sexual maturity is predicted to vary 6-litre (L) tanks for 6 days. On the sixth day, 45 fish in response to the rate of extrinsic mortality (Stearns, from either group were transferred into individual 2-L 1992; Marcus & Weeks, 1997). We estimated the onset tanks aligned into two recirculation systems (Fish of sexual maturity by direct methods in females and Boxes, Aqua Medic, www.aqua-medic.de) where they indirect methods for males. Individual females were were housed until their natural death. The position of each placed into a spawning tank (see Polacik & Rei- fast and slow fish on the shelves and across the two chard, 2009 for details) with a randomly chosen male systems was randomized. Each system was supplied from the same treatment group and allowed to spawn with a powerful UV lamp, lighting (light period of for 2 h. After 2 h, the substrate was checked for the 14 h), temperature regulation and water filtration. All presence of eggs. Tests started at the age of 23 days and fish experienced an identical feeding regime. During were repeated every other day for the direct detection the initial 14 days, they were fed Artemia nauplii three of successful production of fertilized eggs. An indirect times per day. The ration was similar to an ad libitum method based on the evaluation of the degree of color- regime as live Artemia nauplii often survived in the ation (e.g. Kotrschal et al., 2012) was used for males. tank until the next feeding. At the age of 15–18 days, Coloration was independently assessed on a categorical fish received a mixture of Artemia nauplii with small basis by four experienced evaluators at the age of bloodworms. Then, after 19 days, fish were fed blood- 30 days. Evaluation was blind with respect to treatment worms and adult brine shrimp twice each day. The fish for three of the four evaluators. As male Nothobranchius received the amount of food they were able to consume develop their coloration very gradually with no sudden within 15 min (Polacik & Reichard, 2009), and any appearance of any particular ornament, each male was uneaten food at that time was removed. Water temper- assigned one of three categories: (i) signs of coloration ature was maintained at 28 °C, with occasional minor – first traces of male colour observable in a close look, decrease during 50% water exchanges twice each week that is, minimum colour but the individual confidently (never below 25 °C). scored as a male, (ii) intermediate colour – male body and fins clearly coloured, but the colours were pale (iii) – Measurements of life-history traits full coloration male fully and brightly coloured. This enabled noninvasive identification of the relative differ- Body size at hatching and the mass of yolk ence in the onset of sexual maturity among treatment reserves represent trade-offs in energy allocation groups. (Table 1) and are closely correlated with the energy Allocation to reproduction was measured as the consumption during embryonic development. Image number of eggs generated per unit time at two time analysis (ImageJ software, Bethesda, MD, USA) was points. Early allocation to reproduction was estimated used to measure the traits using the subsampled indi- as the number of eggs laid during the first definitive viduals preserved immediately after hatching. For yolk mating (i.e. during the period of rapid growth). The test sac measurements, the area of the yolk sac was deter- was repeated at the age of 9 weeks, when female mined as an index of its volume. Sample size is shown growth approached an asymptote (i.e. after cessation of in Table 1. rapid growth; Blazek et al., 2013; result of this study). Growth rate is a crucial component of life history The mean number of eggs was compared between and strongly associated with other life-history traits, treatment groups at each time point. Given that rela- especially in ectotherms (Lee et al., 2013). We regularly tively large numbers of females did not lay any eggs measured each experimental fish from the age of during the second testing, we additionally compared 6 days using analysis of individual digital images. The ‘mean clutch size’, an analysis based on a subset of intervals between measurements were as follows: females, which produced a clutch. 4 days (age 6–28 days), 7 days (age 28–133 days), Specific behavioural schemes may be an integral 15–25 days (age 133–314 days) and 30 days (from age part of a life-history strategy in polyphenic animals 314 days to death). Fish were photographed from (Mellstrom€ et al., 2010; Cullen et al., 2012). We pre- above in a gridded container in shallow water. Growth dicted that secondary pools of shorter duration would rate was calculated as the increase in mean total body predispose fast fish to spawn more readily than the length (TL, from tip of the snout to the end of caudal slow fish (Mellstrom€ et al., 2010). Male Nothobranchius fin) per day. Analysis of growth data showed that the initiate spawning by courting, and females respond to mean growth increment between two subsequent mea- this either positively or flee. We quantified time to ini- surements fell below 3% of actual mean body size of tiate spawning attempts by males and time to a positive the fish at age of 77 days in males and 63 days in female response to male courtship. All fish were tested

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY Alternative life histories in an annual fish 5

at the age of 8 weeks. Female reproductive condition groups whenever possible. Gamma distribution was was standardized by allowing them to completely used to analyse data on latency of spawning (glm func- release their present egg stocks (by spawning with a tion). Nonsignificant interactions between sex and randomly chosen male) and to recover their eggs before treatment and the nonsignificant term ‘sex’ were the test (isolated for a standard period of 2 days; for removed from final models. Activity in the open-field details, see Polacik & Reichard, 2009). For the test, each test was first analysed on pairwise differences to female was placed in a 20-L tank for 15 min with a account for paired design. This included only a subset randomly chosen male from the same treatment. Time of fish surviving until the age of 129 days (n = 19 for to the onset of sex-specific mating activities was fast and n = 20 for slow fish). The response variable recorded by direct observation. was the difference between squares of the grids crossed Ageing rate is predicted to be faster with an at young and old age, with treatment and sex as fixed increased rate of extrinsic mortality (e.g. Williams, factors. To enable the use of entire data set (n = 43 for 1957; Marcus & Weeks, 1997; Dudycha & Tessier, fast and n = 41 for slow fish) at young age, an addi- 1999). To compare the rate of ageing between the fast tional analysis was completed for young and old age and slow fish, we used two previously identified ageing separately. The response variable was the number of markers for this species: locomotor activity and the grids crossed rather than their pairwise differences; accumulation of lipofuscin (an autofluorescent product otherwise, the analysis was identical. A nonparametric of the oxidation of unsaturated fatty acids) in the liver Mann–Whitney test was used to compare the onset of (Terzibasi et al., 2008; Terzibasi-Tozzini et al., 2013). male sexual maturity based on coloration (ordinal scale Locomotor activity of experimental fish (Table 1) 1–3) and repeatability of ranking by four independent was measured at both the age of 63 and 129 days. Each evaluators measured as intraclass correlation sensu Les- fish was placed in the middle of a 300 9 300 mm sells & Boag (1987). Repeated-measures ANOVA was used aquarium with a 20 9 20 mm square grid drawn on to test differences in body size of individual males and the bottom, and the fish’s movements were recorded females. Given the relatively small sample sizes, life- from above with a video camera. The recording started span was compared using a nonparametric log-rank 15 s after the fish had been placed into the aquarium test. All mean values are given with 1 standard error in (no startle response was observed) and continued for parenthesis. 4 min. Locomotor activity was quantified by counting all squares within the grid that were crossed by the Results head of the fish. For lipofuscin quantification, liver tissue fixed in Baker’s solution was embedded in Parap- Body size at hatching was significantly higher in slow last, sectioned at 5 lm, dehydrated and mounted with fish than in fast fish (F1,55 = 4.58, P = 0.037, Fig. 1a), DEPX without counterstaining. Images were acquired and the yolk sac of slow fish was smaller (F1,30 = 5.13, using a Leica confocal microscope at an excitation P = 0.031, Fig. 1b). Both treatment groups followed wavelength of 488 nm with fixed confocal parameters similar growth trajectories. Growth rate was high fol- (pinhole, photo-multiplier, laser intensity, etc.). For lowing hatching, gradually decreased but peaked again details, see Terzibasi et al., 2009. Ten fish from each after fish were weaned onto the adult diet. Growth rate group were killed (Table 1) at the age of 20 weeks gradually decreased after sexual maturity (Fig. 2). (Terzibasi et al., 2008). Within the period of intensive growth (until the age Survival and final body size. We predicted that a of 63 and 77 days for males and females, respectively), trade-off with other life-history traits would impair life- fast males and females remained significantly larger span and final adult size of the fast fish (e.g. Stearns, than slow fish (RM ANOVA, F14,30 = 3.90 for males and 1992; Roff, 2002; Lee et al., 2013), which were pre- F12,26 = 5.33 for females, both P < 0.001, Fig. 3). Nota- dicted to show more rapid growth and earlier matura- bly, despite being smaller at hatching, fast fish outsized tion. We recorded the date of natural death for each slow fish at the age of 6 days (time of the first body size experimental individual. Unexpectedly, most experi- measurement) and were consistently larger until sexual mental females likely died due to a failure to lay eggs maturity (in males) or even beyond (in females) in the absence of males (compare Graf et al., 2010; see (Fig. 3). Results for more details). Consequently, differences in All females reached sexual maturity between the ages survival and lifespan were only tested in males. Final of 25 and 37 days. Fast females matured significantly (maximum) body size was the size at the last measure- earlier than slow females (F1,40 = 6.06, P = 0.018). For ment before death. fast females, the mean (SE) age at maturation was 29.5 (0.60) days, whereas for slow females, it was 31.6 (0.85) Data analysis days. Females from both groups matured at the same size (F1,40 = 0.56, P = 0.46), measuring 30.26 (0.47) mm and General linear models (lm function in R 3.0.0) were 30.76 (0.67) mm for the fast and slow females, respec- used to test differences between the two treatment tively. The number of eggs laid in the first spawning did

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 6 M. POLACIK ET AL.

(a) ability of the degree of male coloration by four evaluators 3.78 was high (N = 4 evaluators, R = 0.720). At the age of 30 days, fast males had developed significantly more col- 3.76 oration (median colour category = 3) than slow males 3.74 (median colour category = 2) (Mann–Whitney test, 3.72 W = 398.5, P = 0.006). The latency to spawn did not differ between treat- 3.70 ment groups. Fast males initiated spawning attempts at 3.68 a similar time as slow males (Gamma distribution, 3.66 F1,30 = 0.99, P = 0.328, median time was 138 and 160 s for fast and slow males, respectively). Similarly, there 3.64 was no difference in female response (Gamma distribu-

Standard length (mm) 3.62 tion, F1,28 = 1.46, P = 0.237; median time was 216 s for 3.60 fast and 310 s for slow females). There was no differ- ence in allocation to reproduction between fast and 3.58 slow females (quasi-Poisson distribution, F1,38 = 1.85, 3.56 P = 0.182, median of 5 and 18 eggs for slow and fast females, respectively). A total of six females from the (b) slow fish treatment (N = 20) did not lay any eggs com- 0.44 pared to only two females (N = 20) from the fast fish treatment. However, clutch size was not different 0.42 between treatments (F1,30 = 0.38, P = 0.545). Fast

) females produced a mean of 21.72 (3.37) and slow 2 0.40 females 18.57 (3.91) eggs. Both groups showed overall comparable activity in 0.38 the open-field test. There was no decline in locomotor activity in fast or slow fish (F1,36 = 0.58, P = 0.453, 0.36 Fig. 4a). Young males were more active than young = = Yolk sac area (mm females (F1,81 5.51, P 0.021), but this difference = = 0. 34 was not observed in older fish (F1,36 1.41, P 0.243). Lipofuscin accumulated in the livers at the same rate 0.32 in fast and slow fish at an age of 20 weeks (F1,17 = 0.001, P = 0.983, Fig. 4b). Females tended to accumulate lipofuscin at a higher rate than males, Slow fish Fast fish though this difference was not statistically significant (F1,18 = 3.94, P = 0.063, Fig. 4b). Fig. 1 Mean body size (a) and yolk sac area (b) at hatching. Boxes Fast males had significantly shorter lifespans (median represent one SE, and whiskers represent 95% confidence survival of fast males = 140.5 days) than slow males intervals. (median survival of slow males = 290.5 days, log-rank test, P = 0.019, Fig. 5). Female lifespan was consider- not differ between treatments (log-transformed data, ably shorter than male lifespan (median survival F1,40 = 0.32, P = 0.574). Fast females produced a mean 100 days), but 82 % of the females died unnaturally. of 9.95 (1.73) and slow females 9.10 (1.86) eggs. Repeat- After the final experimental spawning at the age of

1.8 Fast males 1.6 Weaning on the Fast females 1.4 adult diet Slow males 1.2 Female maturity 1 Slow females

(mm) 0.8 0.6 0.4 0.2 Onset of male colouration Fig. 2 Daily growth rates of fast and Mean daily growth increment 0 slow males and females with indications 10 14 17 21 24 28 35 42 49 56 63 70 77 84 of weaning onto an adult diet and the Age (days) onset of sexual maturation.

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY Alternative life histories in an annual fish 7

35 Fast males * 30 Slow males ** 0.06 * Fast females 0.07 25 Slow females ns

20 ** * ** ns 15 ** *** 10 *** *** 5

0 6 101417212428

60 Body size Body size (mm) 50 ns ns ns ns ns ns ns 40 ns 0.07 * * * Fig. 3 Mean body size of fast (white 30 bars) and slow (grey bars) males and females at the age of 6–77 days. Error 20 bars denote 1 SE. Statistical significance in pairwise comparisons is denoted by 10 asterisks (***: P < 0.001, **: P < 0.01, Not available Not available *: P < 0.05); significance at 0 0.05 > P < 0.10 is shown by the exact 35 42 49 56 63 70 77 P-value, and ‘ns’ denotes P > 0.10. Age (days)

63 days, females had no access to males and did not or short (1.5 months) embryonic developmental trajec- spawn. Given previous reports of spontaneous egg tories was predicted, allowing fish to maximize their deposition in the absence of males under comparable reproductive success under specific habitat conditions. experimental conditions (Graf et al., 2010), the failure Secondary pools (pools that dried and subsequently of spontaneous egg deposition was not anticipated refilled within a single rainy season) are populated by under our experimental conditions. However, most fish which undergo rapid embryonic development and females were not capable of releasing all their eggs in which typically experience an extremely brief period of the absence of males. Eggs were retained in ovaries habitat persistence. These fish were assumed to produce resulting in later tissue rupture and premature mortal- ‘faster’ phenotypes, characteristic of rapid growth and an ity. In contrast, male mortality followed a predicted pat- earlier attainment of sexual maturity, at a cost of smaller tern and was associated with a senescent phenotype. final body size, lower fecundity, rapid phenotypic deteri- Males were consistently larger than females even prior oration and shorter lifespan. to sexual maturity (Fig. 3) and grew to ultimately a lar- Overall, the data supported our main hypothesis that ger size than females (Fig. 4c). Slow males and females N. furzeri with rapid embryonic development and reached a significantly larger maximum size than fast exposed to a relatively higher risk of habitat desiccation males and females (F1,33 = 8.18, P = 0.007 for males, produced phenotypes with a more rapid life history. F1,30 = 4.95, P = 0.034 for females). Slow males were on We found a combination of adaptive plastic responses average 13.6 % larger than fast males, and slow females and constraints. Fast fish were smaller at hatching but were 8.3 % larger than fast females (Fig. 4c) despite the had larger yolk sac reserves, contributing to a rapid fact that the majority of them died prematurely. post-hatching growth. The fast fish matured signifi- cantly earlier but grew to a smaller final body size and Discussion died sooner than the slow fish. In contrast, there were no differences in female fecundity, the propensity to We hypothesized that unpredictable and erratic environ- mate and in the markers of ageing (decrease in locomo- mental conditions encountered by N. furzeri would lead tor activity and lipofuscin accumulation). to the evolution of adaptive phenotypic plasticity with Seasonal environments facilitate selection for adapta- alternative life-history strategies. Plasticity in life-history tions enabling organisms to cope with cyclically traits in response to either long (typically, 6–7 months) changing conditions (Wanschoenwinkel et al., 2010).

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 8 M. POLACIK ET AL.

160 (a) 1.0 140 Slow males 0.8 Slow females 120 Fast males Fast females 100 0.6 80 60 0.4 40 0.2 20

0.0

0 Cumulative proportion of surviving fish

ocomotor activity (N of grid crossing) Young Old Young Old Young Old Young Old L 0 50 100 150 200 250 300 350 400 450 500 550 Fast fish Slow fish Age (days)

18 (b) Fig. 5 Survival curves of slow and fast male and female fish 16 visualized as the cumulative proportion of surviving fish using a Kaplan–Meier estimate. Circles represent deaths; crosses illustrate ) 5 14 census events (e.g. removal of fish for histology). 12 10 more optimal thermoregulation at a given season (Hazel, 2002; Van Dyck & Wiklund, 2002). Conversely, unpre- 8 dictable conditions select either for highly plastic 6 responses or for developmental bet-hedging. For exam- 4 ple, some species of anurans spawning in ephemeral Integrated density (10 pools can reversibly accelerate their premetamorphic 2 development (e.g. Denver et al., 1998) or commence 0 Fast Slow Fast Slow either carnivorous or omnivorous feeding on the basis of environmental cues (Pfenning, 1992). Many inverte- Males Females brate and some vertebrate species adopt a bet-hedging 70 (c) strategy and produce offspring with varying rates of ** development within a single clutch (e.g. frogs: Lane & 60 Mahony, 2002; crustaceans: Ripley et al., 2004). In N. furzeri, a combination of both developmental 50 * bet-hedging (Wourms, 1972; this study) and high plas- ticity (Blazek et al., 2013; this study) is apparent. Alter- 40 native developmental pathways give rise to alternative post-hatching life-history strategies. The difference is 30 based on bimodality in life-history traits, similarly to 20 that seen in spadefoot toads (Storz et al., 2011), rather than on strict morphological differences, as is typical for Maximum body size (mm) 10 some other taxa (e.g. insects: reviewed in Simpson et al., 2011; amphibians: reviewed in Denoel et al., 0 2005). In annual killifish, the mechanism triggering Fast Slow Fast Slow Males Females rapid embryonic development, which results in a subse- quent generation of fish within a single rainy season, is Fig. 4 Locomotor activity (a), density of lipofuscin deposition in facilitated by the response of the embryos to several liver (b) and maximum body size (c) in fast (white bars) and slow environmental cues (e.g. Wourms, 1972). (grey bars) males and females. Locomotor activity in young Rapid maturation is a crucial life-history trait of (9 weeks) and old (18 weeks) fish is shown separately. Statistical N. furzeri. At the intraspecific level, an accelerated ** < * < significance is denoted by asterisks ( : P 0.01, : P 0.05). developmental schedule is typical for relatively more time-constrained individuals across a range of Patterns of change that are predictable and stable across other organisms. In the common frog tadpoles, Rana years enable the evolution of alternative phenotypes temporaria, decreased habitat water levels lead to earlier with qualitative, nonplastic traits. For example, polyphe- metamorphosis at a smaller body size (Johansson et al., nic butterflies with regularly shifting spring and summer 2010). A drought-escape strategy in terms of early generations possess dark or light colour that results in flowering was documented in an annual grass (Sherrard

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY Alternative life histories in an annual fish 9

& Maherali, 2006). In N. furzeri, the difference in female Lee et al., 2013), we have confirmed that a growth maturation of 2 days represents a 7 % difference which rate–lifespan trade-off is also possible under strict com- we believe is quite substantial given the constraints mon garden experimental conditions, presumably aris- imposed by a common genome of fast and slow fish. It is ing as an intrinsic consequence of metabolic profiles. also biologically relevant in the savannah environment This outcome may have consequences for the use of where daily evaporation is high (Sacramento et al., N. furzeri in the field of ageing research (Valenzano 2012). The robustness of an alternatively rapid matura- et al., 2006) and prompt developmental history of tion rate in fast fish is further evidenced by maturation experimental individuals to be considered. rate displayed in males. In the present study, females suffered from an unex- In agreement with the correlation between matura- pected physiological defect caused by their inability to tion schedule and growth rate (Engen & Sæther, 1994), oviposit, which ultimately led to their premature death. fast fish group of our study grew more rapidly than the This outcome was not anticipated as Graf et al. (2010) slow fish early after hatching and remained signifi- reported no physiological defect in females kept in the cantly rapid until sexual maturity and, in the case of absence of males in N. furzeri. It is also notable that due females, even continued beyond sexual maturity to the higher mortality of males and incidences of (Fig. 3). This finding is in contrast to moor frogs extremely female-biased sex ratios (> 90% of females) (R. arvalis) in which individuals undergoing prolonged in the wild (Reichard et al., 2009), the lack of males embryonic development show accelerated premetamor- may be relevant at the end of the rainy season. In nat- phic (‘catch up’) growth to compensate for the time ural habitats of N. furzeri, we have observed females spent in the egg (Orizaola et al., 2010). However, the with abnormally swollen abdomens in pools with low discrepancy likely stems from a difference in adult life fish densities during the latter part of the rainy season history; frogs leave the aquatic environment after meta- (M. Reichard, unpublished data). Therefore, our results morphosis and are not further constrained by the risk demonstrate that, as well as costs of reproduction (e.g. of habitat desiccation. Furthermore, to our knowledge, Stearns, 1992), there may also be costs of nonreproduc- nothing is known about schedules of sexual maturation tion under specific conditions. in adult anurans expressing contrasting premetamor- Our initial prediction was that the fast fish group in phic development. Overall, the results show how finely our study would be larger at hatching given their invest- key life-history traits may respond to embryonic devel- ment in rapid growth and early maturation (Lindholm opment trajectory, despite matching environmental et al., 2006); however, our data showed the reverse. This conditions and a common gene pool. outcome may be because having a longer embryonic per- Rapid development and growth is associated with iod simply enables embryos to grow to an ultimately lar- other biological costs (e.g. Arendt & Wilson, 2000). In ger size, beyond the context of adaptation. Despite Pieris napi, a polyphenic butterfly, individuals of the fas- hatching at a smaller body size, the fast fish demon- ter-developing cohort eclose less sexually mature than strated a rapid growth and were larger than the slow fish those with longer development. Males of this species are at the initial measurement at an age of 6 days (Fig. 3). not fully capable of synthesizing sex pheromones syn- Their accelerated early growth may have been partly thesis, whereas females suffer from fecundity losses supported energetically by the larger yolk reserves (Moo- (Mellstrom€ et al., 2010). In N. furzeri, we detected no die et al., 1989). Other mechanisms such as higher meta- cost of rapid growth and development on fecundity. bolic rates played also a role as increased growth rate Nothobranchius are ‘income breeders’ (sensu Bonnet et al., extended over the entire juvenile period and juvenile 1998) and readily convert available resources into con- N. furzeri start feeding immediately after hatching. Devel- tinuous reproduction (Polacik & Reichard, 2011). They opmental history may result in differences in metabolic are capable of reproduction as soon as they achieve a rates with consequences for growth (Burton et al., 2011), minimum physiologically suitable size (Blazek et al., and individuals with relatively higher metabolic rates 2013). We found that costs of accelerated growth in the can benefit from superior conditions and grow more rap- fast fish were expressed as relatively smaller ultimate idly (e.g. McCarthy, 2000; Alvarez & Nicieza, 2005). body size and shorter lifespan rather than as decreased The clear differences in durations of lifespan between fecundity. treatment groups (with males from the slow treatment The theory of oxidative ageing predicts reduced lon- living twice as long as those from the fast treatment) gevity as a cost of fast growth due to a side effect of were not reflected in behavioural or histological mark- increased metabolic damage (e.g. Almroth et al., 2012). ers of ageing. The behaviour of N. furzeri in open-field A negative relationship between individual growth rate tests appears to have high interpopulation variation and lifespan has been recognized across many taxa (Terzibasi et al., 2008), and it is possible that the (Rollo, 2002; Metcalfe & Monaghan, 2003). While N. furzeri population used in our study is robust to age- altered growth rates (along with the consequences on ing-related locomotor decay. Nonetheless, we observed lifespan) are typically due to manipulation of resources clear impairment of locomotor activity in senescent fish or ambient temperature in the laboratory setting (e.g. at a much later age during the tests (performed when

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 10 M. POLACIK ET AL.

the fast fish suffered 50% mortality). At that late age, with compressed lifespan - a model for addressing aging. Free however, the number of surviving individuals was too Radic. Res. 46: 1183–1189. low to permit rigorous testing. Alvarez, D. & Nicieza, A. 2005. Is metabolic rate a reliable pre- Lipofuscin is autofluorescent marker that is used for dictor of growth and survival of brown trout (Salmo trutta) – identifying age-related accumulation of oxidative dam- in the wild? Can. J. Fish. Aquat. Sci. 62: 643 649. Arendt, J.D. & Wilson, D.S. 2000. Population differences in the age, and it typically accurately reflects histological age- onset of cranial ossification in pumpkinseed (Lepomis gibbo- ing across various organs (Ding et al., 2010; Bosley & sus), a potential cost of rapid growth. Can. J. Fish. Aquat. Sci. Dumbauld, 2011; Yu & Li, 2012; Terzibasi-Tozzini et al., 57: 351–356. 2013). Despite differing durations in lifespan, lipofuscin Beldade, P., Mateus, A.R.A. & Keller, R.A. 2011. Evolution was deposited in the liver at a comparable level and molecular mechanisms of adaptive developmental plas- between treatment groups, implicating a similar rate of ticity. Mol. Ecol. 20: 1347–1363. intrinsic ageing in fast and slow fish, measured at the Blazek, R., Polacik, M. & Reichard, M. 2013. Rapid growth, histological level. A large difference in lipofuscin early maturation and short generation time in African accumulation was been reported in several species of annual fishes. Evol. Dev. 4: 24. Nothobranchius at the interpopulation level (Terzibasi Bonnet, X., Bradshaw, D. & Shine, R. 1998. Capital versus income breeding: an ectothermic perspective. Oikos 83: 333– et al., 2008; Terzibasi-Tozzini et al., 2013). In our study, 342. all fish came from the same population and a common Bosley, K.M. & Dumbauld, B.R. 2011. Use of extractable lipo- garden approach ruled out potential sources of bias in fuscin to estimate age structure of ghost shrimp populations lipofuscin accumulation, such as different temperature in west coast estuaries of the USA. Mar. Ecol. Prog. Ser. 428: (Valenzano et al., 2006) or diet (Castro et al., 2002). 161–176. The observed decoupling of lifespan and lipofuscin Brakefield, P.M. & French, V. 1999. Butterfly wings: the evolu- accumulation and a tendency for dissimilar deposition tion of development of colour patterns. BioEssays 21: 391–401. between sexes (Fig. 5) warrant further research. Burton, T., Killen, S.S., Armstrong, J.D. & Metcalfe, N.B. 2011. We documented developmentally related plasticity in What causes intraspecific variation in resting metabolic rate the life history of N. furzeri, which is an unusual feature and what are its ecological consequences? Proc. Biol. Sci. 278: 3465–3473. among vertebrates. The existence of alternative life-his- Castro, M., Encarnacßao,~ P. & Tully, O. 2002. The effect of die- tory strategies in annual killifish is a response to unpre- tary antioxidants on lipofuscin accumulation in the crusta- dictable environmental conditions and effectively cean brain. J. Exp. Mar. Biol. Ecol. 269:53–64. spreads the risk of failure of offspring survival. A subse- Charlesworth, B. 1980. Evolution in Age-Structured Populations. quent generation of fish may be generated within a sin- Cambridge University Press, Cambridge. gle rainy season by the production of embryos with Cullen, D.A., Sword, G.A. & Simpson, S.J. 2012. Optimizing direct and rapid embryonic development. These fish multivariate behavioural syndrome models in locusts using have the prospect of an unusually short temporal dura- automated video tracking. Anim. Behav. 84: 771–784. tion to their habitat, and we demonstrated that their Denoel, M., Joly, P. & Whiteman, H.H. 2005. Evolutionary embryonic developmental trajectory results in the pro- ecology of facultative paedomorphosis in newts and sala- manders. Biol. Rev. Camb. Philos. Soc. 80: 663–671. duction of phenotypes with a rapid life history. The Denver, R.J., Mirhadi, N. & Phillips, M. 1998. Adaptive pattern of life-history trade-offs at the intrapopulation plasticity in amphibian metamorphosis: response of Scaphi- level was comparable to the pattern documented in opus hammondii tadpoles to habitat desiccation. Ecology 79: Nothobranchius at the interpopulation (Terzibasi et al., 1859–1872. 2008) and interspecific level (Terzibasi-Tozzini et al., Ding, L.L., Kuhne, W.W., Hinton, D.E., Song, J. & Dynan, 2013), indicating that genetic differences are not neces- W.S. 2010. Quantifiable biomarkers of normal aging in the sarily the main cause of life-history divergence. Ongo- Japanese medaka fish (Oryzias latipes). PLoS One 5: e13287. ing studies are focused on the mechanisms linking Dudycha, J.L. & Tessier, A.J. 1999. Natural genetic variation of embryologic development with post-hatching ecological life span, reproduction and juvenile growth in Daphnia. Evo- – and physiological performance. lution 53: 1744 1756. Engen, S. & Sæther, B.E. 1994. Optimal allocation of resources to growth and reproduction. Theor. Popul. Biol. 46: 232–248. Acknowledgments Genade, T. 2005. Laboratory manual for culturing N. furzeri. See http://www.nothobranchius.info/pdfs/lab_protocols_1.pdf. This work was supported by CSF grant P505/11/P646. Genade, T., Benedetti, M., Terzibasi, E., Roncaglia, P., Valenz- The authors would like to thank C. Smith and A. Rom- ano, D.R., Cattaneo, A. et al. 2005. Annual fishes of the ney for English corrections of the MS. genus Nothobranchius as a model system for aging research. Aging Cell 4: 223–233. Graf, M., Cellerino, A. & Englert, C. 2010. Gender separation References increases somatic growth in females but does not affect life- span in Nothobranchius furzeri. PLoS One 5: e11958. Almroth, B.C., Johnsson, J.I., Devlin, R. & Sturve, J. 2012. Green, B.S. 2008. Maternal effects in fish populations. Adv. Oxidative stress in growth hormone transgenic coho salmon Mar. Biol. 54:1–105.

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY Alternative life histories in an annual fish 11

Hazel, W.N. 2002. The environmental and genetic control of Pfenning, D.W. 1992. Polyphenism in spadefoot toad tadpoles seasonal polyphenism in larval color and its adaptive signifi- as a locally adjusted evolutionarily stable strategy. Evolution cance in a swallowtail butterfly. Evolution 56: 342–348. 46: 1408–1420. Johansson, F., Lederer, B. & Lind, M.I. 2010. Trait perfor- Piersma, T. & Gils, J.A. 2011. The Flexible Phenotype. Oxford mance correlations across life stages under environmental University Press Inc., New York. stress conditions in the common frog, Rana temporaria. PLoS Podrabsky, J.E., Garrett, I.D.F. & Kohl, Z.F. 2010a. Alternative One 5: e11680. developmental pathways associated with diapause regulated Kotrschal, A., Rogell, B., Maklakov, A.A. & Kolm, N. 2012. by temperature and maternal influences in embryos of the Sex-specific plasticity in brain morphology depends on social annual killifish Austrofundulus limnaeus. J. Exp. Zool. 213: environment of the guppy, Poecilia reticulata. Behav. Ecol. 3280–3288. Sociobiol. 66: 1485–1492. Podrabsky, J.E., Tingaud-Sequeira, A. & Cerda, J. 2010b. Met- Lane, S.J. & Mahony, M.J. 2002. Larval anurans with synchro- abolic dormancy and responses to environmental dessication nous and asynchronous development periods: contrasting in fish embryos. In: Dormancy and Resistance in Harsh Environ- responses to water reduction and predator presence. J. Anim. ments (E. Lubzens, J. Cerda & M.S. Clark, eds), pp. 283. Ecol. 71: 780–792. Springer-Verlag, Berlin, Heidelberg. Lee, W.S., Monaghan, P. & Metcalfe, N.B. 2013. Experimental Polacik, M. & Reichard, M. 2009. Indirect fitness benefits are demonstration of the growth rate - lifespan trade-off. Proc. not related to male dominance in a killifish. Behav. Ecol. So- Biol. Sci. 280: 20122370. ciobiol. 63: 1427–1435. Lessells, C.M. & Boag, P.T. 1987. Unrepeatable repeatabilities: Polacik, M. & Reichard, M. 2011. Asymmetric reproductive a common mistake. Auk 104: 116–121. isolation between two sympatric annual killifish with extre- Lindholm, A.K., Hunt, J. & Brooks, R. 2006. Where do all the mely short lifespans. PLoS One 6: e2268. maternal effects go? Variation in offspring body size through Reichard, M., Polacik, M. & Sedlacek, O. 2009. Distribution, ontogeny in the live-bearing fish Poecilia parae. Biol. Lett. 2: colour polymorphism and habitat use of the African killifish, 86–589. Nothobranchius furzeri, the vertebrate with the shortest life- Marcus, V. & Weeks, S.C. 1997. The effects of pond duration span. J. Fish Biol. 74: 198–212. on the life history traits of an ephemeral pond crustacean, Reznick, D.A., Bryga, H. & Endler, J.A. 1990. Experimentally Eulimnadia texana. Hydrobiologia 359: 213–221. induced life-history evolution in a natural population. Nature Markofsky, J. & Perlmutter, A. 1972. Age at sexual maturity 346: 357–359. and relationship to longevity in the annual cyprinodontid Ripley, B.J., Holtz, J. & Simovich, M.A. 2004. Cyst bank fish, Nothobranchius guentheri. Exp. Gerontol. 7: 131–135. life-history model for a fairy shrimp from ephemeral ponds. Markofsky, J. & Perlmutter, A. 1973. Growth differences in Freshw. Biol. 49: 221–231. subgroups of varying longevities in a laboratory population Roff, D.A. 2002. Life History Evolution. Sinauer Associates, Sun- of the male annual cyprinodont fish, Nothobranchius guentheri derland. (Pfeffer). Exp. Gerontol. 8:65–73. Rollo, C.D. 2002. Growth negatively impacts the life span of Mazuze, F.M. 2007. Analysis of adoption of orange-fleshed mammals. Evol. Dev. 4:55–61. sweetpotatoes: the case study of Gaza province in Mozam- Sacramento, A., Matavel, A., Basılio, M. & Bila, S. 2012. Cli- bique. See http://fsg.afre.msu.edu/mozambique/MAZUZE_ mate change impacts and copind strategies in Chicualacuala MS.pdf. district, Gaza province, Mozambique. See http://www.unep. McCarthy, I.D. 2000. Temporal repeatability of relative stan- org/climatechange/adaptation/Portals/133/documents/Chicua dard metabolic rate in juvenile Atlantic salmon and its rela- lacuala_Report-Climate_Change_Impacts_n_Coping_Strateg tion to life history variation. J. Fish Biol. 57: 224–238. ies.pdf. Mellstrom,€ H.L., Friberg, M., Borg-Karlson, A.K., Murtazina, Sherrard, M.E. & Maherali, H. 2006. The adaptive significance R., Palm, M. & Wiklund, C. 2010. Seasonal polyphenism in of drought escape in Avena barbata, an annual grass. Evolu- life history traits: time costs of direct development in a but- tion 60: 2478–2489. terfly. Behav. Ecol. Sociobiol. 64: 1377–1383. Simpson, S.J., Sword, G.A. & Lo, N. 2011. Polyphenism in Metcalfe, N.B. & Monaghan, P. 2003. Growth versus lifespan: insects. Curr. Biol. 21: R738–R749. perspectives from evolutionary ecology. Exp. Gerontol. 38: Stearns, S.C. 1976. Life history tactics: a review of the ideas. Q. 935–940. Rev. Biol. 51:3–47. Moodie, G.E.E., Loadman, N.L., Wiegand, M.D. & Mathais, Stearns, S.C. 1989. The evolutionary significance of phenotypic J.A. 1989. Influence of egg characteristics on survival, plasticity: phenotypic sources of variation among organisms growth and feeding in larval walleye (Stizostedion vitreum). can be described by developmental switches and reaction Can. J. Fish. Aquat. Sci. 46: 516–521. norms. Bioscience 39: 436–445. Oostra, V., de Jong, M.A., Invergo, B.M., Kesbeke, F., Wende, Stearns, S.C. 1992. The Evolution of Life Histories. Oxford Uni- F., Brakefield, P.M. et al. 2011. Translating environmental versity Press, Oxford. gradients into discontinuous reaction norms via hormone sig- Storz, B.L., Heinrichs, J., Yazdani, A., Phillips, R.D., Mulvey, nalling in a polyphenic butterfly. Proc. Biol. Sci. 278: 789–797. B.B., Arendt, J.D. et al. 2011. Reassessment of the environ- Orizaola, G., Dahl, E. & Laurila, A. 2010. Compensating for mental model of developmental polyphenism in spadefoot delayed hatching across consecutive life-history stages in an toad tadpoles. Oecologia 156:55–66. amphibian. Oikos 119: 980–987. Terzibasi, E., Valenzano, D.R., Benedetti, M., Roncaglia, P., Pfennig, D.W. & McGee, M. 2010. Resource polyphenism Cattaneo, A., Domenici, L. et al. 2008. Large differences in increases species richness: a test of the hypothesis. Philos. aging phenotype between strains of the short-lived annual Trans. R. Soc. Lond. B Biol. Sci. 365: 577–591. fish Nothobranchius furzeri. PLoS One 3: e3866.

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY 12 M. POLACIK ET AL.

Terzibasi, E., Lefrancois, C., Domenici, P., Hratmann, N., Graf, West-Eberhard, M.J. 1989. Phenotypic plasticity and the ori- M. & Cellerino, A. 2009. Effects of dietary restriction on gins of diversity. Annu. Rev. Ecol. Syst. 20: 249–278. mortality and age-related phenotypes in the short-lived fish Wildekamp, R.H. 2004. A World of Killies: Atlas of the Oviparous Nothobranchius furzeri. Aging Cell 8:88–99. Cyprinodontiform Fishes of the World, vol. 4. American Killifish Terzibasi-Tozzini, E., Dorn, A., Ng’oma, E., Polacik, M., Association, Elyria, OH. Blazek, R., Reichwald, K. et al. 2013. Parallel evolution of Williams, G.C. 1957. Pleiotropy, natural selection and the evo- senescence in annual fishes in response to extrinsic mortal- lution of senescence. Evolution 11: 398–411. ity. BMC Evol. Biol. 13: 77. Wourms, J.P. 1972. Developmental biology of annual fishes. III. Valdesalici, S. & Cellerino, A. 2003. Extremely short lifespan Pre-embryonic and embryonic diapause on variable duration in the annual fish Nothobranchius furzeri. Proc. Biol. Sci. 270: in the eggs of annual fishes. J. Exp. Zool. 182: 389–414. 189–191. Yu, X. & Li, G.R. 2012. Effects of resveratrol on longevity, cog- Valenzano, D.R., Terzibasi, E., Genade, T., Cattaneo, A., Dome- nitive ability and aging-related histological markers in the nici, L. & Cellerino, A. 2006. Resveratrol prolongs lifespan annual fish Nothobranchius guentheri. Exp. Gerontol. 47: 940– and retards the onset of age-related markers in a short-lived 949. vertebrate. Curr. Biol. 16: 296–300. Van Dyck, H. & Wiklund, C. 2002. Seasonal butterfly design: morphological plasticity among three developmental path- Supporting information ways relative to sex, flight and thermoregulation. J. Evol. Additional Supporting Information may be found in the – Biol. 15: 216 225. online version of this article: Wanschoenwinkel, B., Waterkeyn, A., Jocque, M., Boven, L., Figure S1 Seaman, M. & Brendonck, L. 2010. Species sorting in space Temperature fluctuations (logging every and time-the impact of disturbance regime on community 3 h) at a pool where a primary pool has desiccated and assembly in a temporary pool metacommunity. J. North Am. was refilled again within a single rainy season. Benthol. Soc. 29: 1267–1278. Watters, B.R. 2009. The ecology and distribution of Notho- Received 14 November 2013; revised 13 February 2014; accepted 20 branchius fishes. J. Am. Killifish Assoc. 42:37–76. February 2014

ª 2014 THE AUTHORS. J. EVOL. BIOL. doi: 10.1111/jeb.12359 JOURNAL OF EVOLUTIONARY BIOLOGY ª 2014 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

PŘÍLOHA – PUBLIKACE NESOUVISEJÍCÍ S DISERTACÍ

122

Článek B

Michalko, R., Řeţucha, R. & Košulič, O. 2014. Link between aggressive- ness and shyness in the spider Philodromus albidus (Araneae, Philod- romidae): state dependency over stability. Journal of Insect Behavior, submitováno po větší revizi.

V této práci jsem se podílel na tvorbě prvních verzí manuskriptu a poz- ději i finální verze článku.

Manuscript Click here to download Manuscript: Link between aggressiveness and shyness in the spider Philodromus albidus (Araneae, Philodromidae) state dependency over stability - revised manuscript.docx

1 Link between aggressiveness and shyness in the spider Philodromus albidus (Araneae, Philodromidae): state

2 dependency over stability

3

4 Radek Michalko, Radomil Řežucha, Ondřej Košulič

5

6 Radek Michalko (corresponding author), Radomil Řežucha

7 Department of Botany and Zoology, Masaryk University, Kotlářská 2, 611 37, Brno, Czech Republic; E-mail:

8 [email protected]

9

10 Ondřej Košulič

11 Department of Forest Protection and Wildlife Management, Mendel University in Brno, Zemědělská 3, 613 00,

12 Brno, Czech Republic

13

14 Abstract Behavioral syndromes (BSs), seen as correlations between behaviors through time and across contexts, are

15 well documented in many different taxa. BSs may not be present automatically, however, and their consistency

16 within populations and individuals also varies among studies. Here, we studied a BS comprising aggressiveness

17 toward prey and boldness/shyness and its time consistency in the cursorial spider Philodromus albidus. We measured

18 aggressiveness toward prey as the number of prey killed per 2-hour period. Boldness/shyness was assessed as a

19 latency of initiating exploration in a novel environment. We found the presence of a BS in P. albidus, as bold

20 individuals were also more aggressive toward prey than were shy individuals. The BS was consistent through time

21 within the population but not consistent for individuals. The behavior of individuals was influenced by their feeding

22 history rather than their behavioral type. We argue that as individuals tried to improve their nutritional and/or

23 energetic states, the BS in P. albidus was state-dependent and adaptive. However, the exact proximate mechanism is

24 not known.

25

26 Keywords State dependent behavior, Flexibility, Interindividual differences, Behavioral syndrome, Boldness,

27 Personality type

28

1

29 Introduction

30 Interindividual differences in animal behavior play an important role in ecological and evolutionary

31 processes (Réale et al. 2007; Pruitt and Riechert 2012; Wolf and Weissing 2012). Behavioral types (BTs) can be seen

32 as differences in a behavioral trait among individuals (e.g., aggressive-docile). Boldness and shyness are among the

33 widely studied personality traits in animals and form one of the main axes of animal personality, known as the

34 boldness–shyness continuum (Gosling 2001). Behavioral syndromes (BSs), meanwhile, are significant correlations

35 in behavioral traits across situations and/or contexts, such as when individuals bolder toward predators are also more

36 aggressive toward conspecifics (Huntingford 1976; Sih et al. 2010). Heterogeneity in behavior among individuals is

37 caused by interplay among various intrinsic and extrinsic factors. Intrinsic factors commonly reported included

38 differences in metabolic rate (Cerau et al. 2008), levels of hormones (Jones et al. 2011), size (Dowling and Godin

39 2002), and heritability (Ariyomo et al. 2013). Extrinsic factors, then, comprise externally induced differences in

40 states (e.g., in hunger level), negative frequency selection (e.g., rare phenotype advantage), spatio-temporal variation

41 in the environment (e.g., level of predation), nonequilibrium dynamics (Wolf and Weissing 2010), or social niche

42 specialization (Bergmüller and Taborsky 2010).

43 Individuals that are more aggressive toward prey are also usually bolder generally (e.g., Johnson and Sih

44 2007; Pruitt and Riechert 2012; Sweeney et al. 2013). As boldness is often defined as willingness to take risk

45 (Wilson et al. 1994), an aggressive/bold BS is advantageous for resource extraction. Its presence is also risky,

46 however, as such individuals more frequently expose themselves to predators. The less aggressive and shy behavioral

47 type is safer, but it is less effective in resource acquisition (Sih et al. 2010). Different personality traits are therefore

48 favored in different environments (Carere and Maestripieri 2013). The aggressive/bold BT will be advantageous in a

49 low-risk environment but disadvantageous in a high-risk environment, while the opposite will be true for the less

50 aggressive/shy BT (Riechert and Hedrick 1993). Although BSs are widely reported from many animal species, they

51 are not present automatically. Bell (2005), for example, found a BS in only one of two studied populations of

52 sticklebacks Gasterosteus aculeatus (L. 1758) (Gasterosteidae). Řežucha et al. (2012) found no BS in the European

53 bitterling Rhodeus amarus (Bloch 1782) (Cyprinidae). It also seems that the presence and magnitude of a BS depend

54 upon the environmental conditions which individuals encounter and thus are affected by biotic and abiotic factors

55 (Dingemanse et al. 2007; Pruitt et al. 2011a; Kralj-Fišer and Schneider 2012; Sweeney et al. 2013). Duration of

56 between- and within-individual consistency of a BS may also differ by studied species as well as by studied traits

2

57 (Johnson and Sih 2005; Pruitt et al. 2011b; Kralj-Fišer and Schneider 2012; Sweeney et al. 2013). Spiders constitute

58 an invertebrate taxon frequently used in behavioral syndrome research (reviewed in Pruitt and Riechert 2012).

59 Possible correlation between boldness/shyness and aggressiveness toward prey is a commonly studied BS in spiders,

60 as these traits are important in predator–prey interactions (Pruitt and Riechert 2012).

61 This study investigates the presence of a BS between boldness/shyness and aggressiveness toward prey and

62 its consistency within a period of 3 weeks in the cursorial spider Philodromus albidus (Kulczyński 1911)

63 (Philodromidae). This species shows high potential to be a new model species for studying the role that animal

64 personalities play in intraguild predation. In a previous study, Michalko and Pekár (2014) found a significant

65 presence of individual specialization sensu Bolnick et al. (2003) in a population of P. albidus that was caused by

66 intraguild predation of two other Philodromus species. Individuals that were more aggressive toward prey were less

67 individually specialized than less aggressive individuals. Nonetheless, the presence of a BS between boldness and

68 aggressiveness was not addressed in that study.

69

70 Material and Methods

71 Studied species

72 Spiders of the genus Philodromus belong to the guild “other hunters” (Cardoso et al. 2011). Philodromids

73 catch their prey with their forelegs while waiting in ambush, actively searching for the prey, or actively pursuing it

74 once detected (Haynes and Sisojević 1966). P. albidus is distributed across Europe (Platnick 2014). It occurs on

75 foliage and bark of scrub and trees in various types of habitats such as orchards, forest steppes, and oak forests

76 (Buchar and Růžička 2002; Michalko and Pekár in press). Adults of P. albidus can be found from May to July in the

77 Czech Republic (Michalko 2012).

78 We used 29 adult females of P. albidus with approximate carapace width of 1.4 mm to investigate the

79 presence of a BS between boldness/shyness and aggressiveness toward prey. The females were collected in an

80 abandoned fruit orchard near Brno, Czech Republic (N 49°9'18.74", E 16°33'37.85") at the beginning of June 2013.

81 They were housed in vials (5.5 × 1.5 cm) with a piece of rolled filter paper inside. The paper was sprayed twice a

82 week with water to assure adequate moisture. Individuals were kept at room temperature (22 ± 2 °C) and under

83 natural lightning. They were fed ad libitum with Drosophila flies 1 week before the start of experiments in order to

3

84 standardize their hunger levels. All trials were carried out between 19 June and 3 July 2013. All experiments were

85 performed between 10:00 AM and 4:00 PM.

86

87 Measurement of boldness/shyness

88 We use the term shyness here as it is more intuitive in this context than is boldness. That is because shyness

89 increases with increasing time. We used latency to initiate forward movement in a novel environment as a measure

90 of shyness. To do so, we introduced a female into a Petri dish (perimeter 4.5 cm) and measured the time until her

91 first initiation of forward movement. Upon introduction into the Petri dishes, all females began running around the

92 dish for a few seconds, but then became motionless. We began to measure time only after placing the lid onto the

93 dish. We therefore included the time during which the females ran around into the measurement of motionless time.

94 This introduces only a small amount of noise into the data, as it lasted no more than a few seconds (up to 5 s) while

95 the range of the data was 602 s. The time was measured with a digital stopwatch with accuracy of 0.01 s. We

96 supposed that bolder individuals would initiate movement earlier and shyer ones later. Shyness was measured three

97 times in total at 1-week intervals.

98

99 Measurement of aggressiveness toward prey

100 The measurement of aggressiveness toward prey (hereinafter just aggressiveness) followed approximately 1

101 h after the shyness test. We scored aggressiveness as the number of prey killed. We placed 10 fruit flies into a dish

102 with a female and then counted the number of killed flies after 2 h. We assumed that more aggressive individuals

103 would kill more fruit flies than would less aggressive ones. We then allowed the females to eat the flies until the

104 following day. The next day, the spiders were returned to their housing vials and the fruit flies were removed from

105 the dishes. This enabled us to standardize hunger levels every time. The Petri dishes were then cleaned with 75%

106 ethanol in order to eliminate any chemical olfactory cues left by the females. The aggressiveness toward prey was

107 measured three times overall at 1-week intervals.

108

109 Statistical analyses

110 Statistical analyses were performed within the R 3.0.2 environment (R Development Core Team 2013). We

111 used generalized estimating equations (GEE) as an extension of generalized linear models (GLM) for the repeated

4

112 measurements to test the overall relationship between aggressiveness and shyness (Pekár and Brabec 2012). The

113 shyness and voracity acted as the explanatory and response variable, respectively. We used the Poisson error

114 structure (GEE-p) and log link, as the response variable was of counts data type (Pekár and Brabec 2009). We used

115 the autoregressive correlation structure, as the measurements in an individual took place sequentially during 3 weeks

116 at regular 1-week intervals (Pekár and Brabec 2012). The relationship between voracity and shyness in a particular

117 week was tested by the GLM with Poisson error structure (GLM-p) and log link function. The linear predictors of the

118 GEE and GLM were of linear regression type. We compared the levels of aggressiveness among the 3 weeks with

119 GEE-p. We compared levels of shyness among the 3 weeks by GEE with gamma error structure (GEE-g) and log

120 link, as time can be considered to have gamma distribution of errors (Pekár and Brabec 2009). The linear predictors

121 were of one-way ANOVA type. We estimated two types of behavior repeatability: consistency repeatability and

122 agreement repeatability (Nakagawa and Schielzeth 2010). We tested the consistency repeatability of shyness by

123 means of Pearson correlation coefficient (r). The time was log-transformed, as the Pearson coefficient requires a

124 normal distribution of errors. The consistency of aggressiveness was tested by means of Spearman correlation

125 coefficient (rs), as aggressiveness was count data. The agreement repeatability (R) was estimated as interclass

126 correlation (R) using linear mixed effects models (LMEs). LMEs were fitted with restricted maximum likelihood.

127 Significance was tested by parametric bootstrap and we performed 1000 permutations (Nakagawa and Schielzeth

128 2010). The time until movement initiation and number of killed flies was log- and square root-transformed to

129 approach normal distribution. The numbers of individuals differed among the 3 weeks. This was because we

130 excluded females that laid egg sacs or died. Therefore, measurements on 29, 15, and 7 individuals were obtained in

131 the first, second, and third week of study, respectively.

132

133 Results

2 134 We found a significant overall negative relationship between shyness and voracity (GEE-p, Х 1=5.5,

135 P=0.019, Fig. 1a). There was a marginally significant negative relationship between shyness and voracity in the first

2 136 week (GLM-p, Х 1=3.7, P=0.054, Fig. 1b) and also a significantly negative relationship between shyness and

2 137 voracity in the second week (GLM-p, Х 1=7, P=0.008, Fig. 1c). On the other hand, we observed no significant

2 138 relationship between voracity and shyness in the third week (GLM-p, Х 1=0.7, P=0.41). There was also no significant

139 consistency repeatability in shyness (r, P>0.24, Table 1) or individual aggressiveness (rs, P>0.08, Table 2). Nor was

5

140 there significant agreement repeatability in shyness (LME, 1000 permutations, P=0.62, Table 3) or aggressiveness

141 (LME, 1000 permutations P=0.3, Table 3). Furthermore, there were no significant differences in aggressiveness

2 2 142 (GEE-p, Х 2=3.9, P=0.143, Fig. 2) or shyness (GEE-g, Х 2=4, P=0.137, Fig. 3) among the 3 weeks.

143

144 Discussion

145 In the present study, we explored the presence of a behavioral syndrome (BS) between boldness/shyness

146 and aggressiveness toward prey in the spider Philodromus albidus. There was an overall negative relationship

147 between the number of killed fruit flies and the latency to initiate forward movement in a novel environment. In

148 other words, bold individuals were also more aggressive toward prey than were shy individuals. The relationship was

149 marginally significant and significant in the first and second weeks, respectively. The absence of a significant

150 relationship in the third week was not unexpected, as only seven individuals were used in the analysis. Therefore, a

151 BS was absent in the third week because of the small number of individuals rather than due to its absence in general.

152 Contrary to the significant presence of a BS, there was no significant repeatability in the two behavioral

153 traits through time within individuals. This indicates that different individuals (or at least some of them) were

154 responsible for the relationship between the behavioral traits in the two consecutive weeks. As a result, a BS between

155 boldness/shyness and aggressiveness toward prey was consistent during the 3 weeks within the population of P.

156 albidus as a whole, but it was not consistent for individuals. This pattern suggests that changes in personality types

157 were induced by a flexible switch in the behavior of individuals within the studied population of P. albidus. In

158 contrast to our results, Kralj-Fišer and Schneider (2012) had shown in the araneid spider Larinioides sclopetarius

159 (Clerck 1757) (Araneidae) that repeatability of activity within a novel environment and aggressiveness toward prey

160 were not correlated. The non-consistency of boldness/shyness and aggressiveness toward prey for individual females

161 of P. albidus is also at odds with other studies that documented consistency in these two behavioral traits in adult

162 females of cursorial and web-building spider species (e.g., Johnson and Sih 2005, 2007; Pruitt et al. 2011b).

163 The consistency of a BS within the population accompanied by non-consistency for individuals signifies

164 that the individual behavior of P. albidus in our study was influenced by feeding history and that this overcame

165 individual consistency in behavioral traits. The exact proximate reason is not clear, because energy state as well as

166 nutritional state influence spider behavior independently (Mayntz and Toft 2006; Wilder 2011; Toft 2013). The

167 negative trends in consistency repeatability, although nonsignificant, may indicate that individuals that had eaten less

6

168 in the first week could consequently have been hungrier in the second week than were spiders that had eaten more in

169 the first week. Hungrier spiders could have been motivated to search more for prey than were less hungry

170 individuals. Once the prey was found, the more hungry individuals ate more than did the less hungry individuals.

171 Haynes and Sisojević (1966) mention that individuals of the closely related species Philodromus rufus (Walckenaer

172 1826) (Philodromidae) employed a more active hunting strategy when they became hungrier.

173 The repeatability in behavior can be influenced by habituation or sensitization to the experimental assay

174 (Bell et al. 2009). In such case, however, there would be a change in the level of shyness and aggressiveness and/or a

175 breaking off of the BS.

176 The changing pattern in behavior of individuals of P. albidus was also not caused by a worsening of their

177 condition, as aggressiveness and shyness levels did not differ among the 3 weeks. Therefore, the differences

178 originated from the active decisions by spiders to initiate movement and to kill their prey. Consequently, the pattern

179 wherein the BS was consistent among individuals but the behavioral traits were not consistent for a single individual

180 was a product of behavioral flexibility in P. albidus.

181 There is growing evidence of parasite influence on host personality traits (e.g., Barber and Dingemanse

182 2010; Koprivnikar et al. 2012), and a question remains as to whether that could be the case also in our study.

183 Although there is a lack of data concerning parasite load in philodromids, P. albidus and philodromids in general are

184 probably very rarely parasitized in the studied area. In our observation, we have noted only one individual out of

185 about 500 spiders of the genus Philodromus (including three species collected over several years) to be parasitized

186 by the parasitoid Ogcodes fumatus (Erichson, 1846) (Diptera: Acroceridae) (Kehlmaier et al. 2012) and not by any

187 other parasitoids. We therefore believe that our results were not affected by parasitoid manipulation. This topic

188 definitely needs to be investigated further, however, as to our knowledge this question has not yet been studied in

189 spiders.

190 In conclusion, we found evidence for the presence of a behavioral link between aggressiveness toward prey

191 and boldness/shyness in the spider P. albidus. The behavioral syndrome was consistent over time at the population

192 level but individual personality traits forming this syndrome were not consistent for individuals. The behavior of

193 individuals was probably influenced by their feeding history, thus indicating significant adaptive behavioral

194 plasticity. The BS in P. albidus was therefore state-dependent and it seems to be adaptive, as it probably serves to

195 optimize the current state and to minimize potential predation risk.

7

196

197 Acknowledgements

198 We are very grateful to Petr Dolejš and two anonymous reviewers for their comments that greatly improved our

199 manuscript. RM was supported by grant no. QJ1210209 provided by the National Agency for Agricultural Research.

200 RM and RŘ were supported by grant no. MUNI/A/0937/2012 from Masaryk University. OK was supported by the

201 European Social Fund and state budget of the Czech Republic, the project “Indicators of Trees Vitality” (Reg. No.

202 CZ.1.07/2.3.00/20.0265), and by the Internal Grant Agency of Mendel University (TP7/2014).

203

204 Conflicts of Interest

205 The authors declare that they have no conflicts of interest.

206

207 References

208 Ariyomo TO, Carter M, Watt PJ (2013) Heritability of boldness and aggressiveness in the zebrafish. Behav Genet

209 43: 161–167

210 Barber I, Dingemanse NJ (2010) Parasitism and the evolutionary ecology of animal personality. Phil Trans R Soc B

211 365: 4077–4088.

212 Bell AM (2005) Behavioural differences between individuals and two populations of stickleback (Gasterosteus

213 aculeatus). J Evolution Biol 18: 464–473

214 Bell AM, Hankison SJ, Laskowski KL (2009) The repeatability of behaviour: a meta-analysis. Anim Behav 77: 771-

215 783

216 Bergmüller R, Taborsky M (2010) Animal personality due to social niche specialisation. Trends Ecol Evol 25: 504–

217 511

218 Bolnick DI, Svanbäck R., Fordyce JA, Yang LH, Davis JM, Hulsey CD, Forister ML (2003) The ecology of

219 individuals: incidence and implications of individual specialization. Am Nat 161: 1–28

220 Buchar J, Růžička V (2002) Catalogue of Spiders of the Czech Republic. Peres publishers, Praha, Czech Republic.

221 Cardoso P, Pekár S, Jocqué R, Coddington JA (2011) Global patterns of guild composition and functional diversity

222 of spiders. Plos One 6: e21710

8

223 Carere C, Maestripieri D (2013) Animal personalities: behavior, physiology, and evolution. University of Chicago

224 Press.

225 Cerau V, Thomas D, Humphries MM, Réale D (2008) Energy metabolism and animal personality. Oikos 117: 641–

226 653

227 Dingemanse NJ, Wright J, Kazem AJN, Thomas DK, Hickling R, Dawney N (2007) Behavioural syndromes differ

228 predictably between 12 populations of three-spined stickleback. J Anim Ecol 76: 1128–1138

229 Dowling LM, Godin JJ (2002) Refuge use in a killifish: influence of body size and nutritional state. Can J Zool 80:

230 782–788

231 Gosling SD (2001) From mice to men: what can we learn about personality from animal research? Psychol Bull 127:

232 45–86

233 Haynes DL, Sisojević P (1966) Predatory behavior of Philodromus rufus Walckenaer (Araneae: Thomisidae). Can

234 Entomol 98: 113–133

235 Huntingford FA (1976) The relationship between anti-predator behaviour and aggression among conspecifics in the

236 three-spined stickleback, Gasterosteus aculeatus. Anim Behav 24: 245–260.

237 Johnson JC, Sih A (2005) Precopulatory sexual cannibalism in fishing spiders (Dolomedes triton): a role for

238 behavioral syndromes. Behav Ecol Sociobiol 58: 390–396

239 Johnson JC, Sih A (2007) Fear, food, sex and parental care: a syndrome of boldness in the fishing spider, Dolomedes

240 triton. Anim Behav 74: 1131–1138

241 Jones TC, Akoury TS, Hauser CK, Neblett MF, Linville BJ, Edge AA, Weber NO (2011) Octopamine and serotonin

242 have opposite effect on antipredator behavior in the orb-weaving spider, Larinioides cornutus. J Comp

243 Physiol A 197: 819–825

244 Kehlmaier C, Michalko R, Korenko S (2012) Ogcodes fumatus (Diptera: Acroceridae) reared from Philodromus

245 cespitum (Araneae: Philodromidae), and first evidence of Wolbachia Alphaproteobacteria in Acroceridae.

246 Ann Zool 62: 281–286

247 Koprivnikar J, Gibson CH, Redfern JC (2012) Infectious personalities: behavioural syndromes and disease risk in

248 larval amphibians. Phil Trans R Soc B 279: 1544–1550

249 Kralj-Fišer S, Schneider JM (2012) Individual behavioural consistency and plasticity in an urban spider. Anim Behav

250 84: 197–204

9

251 Mayntz D, Toft S (2006) Nutritional value of cannibalism and the role of starvation and nutrient imbalance for

252 cannibalistic tendencies in a generalist predator. J Anim Ecol 75: 288–297

253 Michalko R (2012) Coexistence and niche partitioning in spiders of the genus Philodromus. MSc. thesis, Masaryk

254 University

255 Michalko R, Pekár S (2014) Is different degree of individual specialisation in three closely related spider species

256 caused by different selection pressures? Bass Appl Ecol. http://dx.doi.org/10.1016/j.baae.2014.08.003

257 Michalko R, Pekár S (in press) Niche partitioning and niche filtering jointly mediate coexistence among three closely

258 related spider species (Araneae, Philodromidae). Ecol Entomol

259 Nakagawa S, Schielzeth H (2010) Repeatability for Gaussian and non-Gaussian data: a practical guide for biologists.

260 Biol Rev 85: 935–956

261 Pekár S, Brabec M (2009) Moderní analýza biologických dat 1. Zobecněné lineární modely v prostředí R [Modern

262 analysis of biological data 1. Generalized linear models in R]. Scientia, Praha (in Czech)

263 Pekár S, Brabec M (2012) Moderní analýza biologických dat 2. Lineární modely s korelacemi v prostředí R (Modern

264 analysis of biological data 2. Linear models with correlation in R). Masaryk University Press, Brno (in

265 Czech)

266 Platnick NI (2014) The world spider catalog, 14.5. http://research.amnh.org/iz/spiders/catalog. Accessed 20 January

267 2014

268 Pruitt JN, Demes KW, Dittrich-Reed DR (2011a) Temperature mediates shifts in individual aggressiveness, activity

269 level, and social behavior in a spider. Ethology 117: 318–325

270 Pruitt JN, Riechert JE, Harris DJ (2011b) Reproductive consequences of male body mass and aggressiveness depend

271 on female’ behavioral types. Behav Ecol Sociobiol 65: 1957–1966

272 Pruitt JN, Riechert JE (2012) The ecological consequences of temperament in spiders. Curr Zool, 58: 589–596

273 R Core Team (2013) R: A language and environment for statistical computing. R Foundation for Statistical

274 Computing, Vienna, Austria. http://www.R-project.org. Accessed 20 January 2014

275 Réale D, Reader SM, Sol D, McDougall PT, Dingemanse NJ (2007) Integrating animal temperament within ecology

276 and evolution. Biol Rev 82: 291–318

277 Řežucha R, Smith C, Reichard M (2012) Personality traits, reproductive behaviour and alternative mating tactics in

278 male European bitterling, Rhodeus amarus. Behaviour 149: 531–553

10

279 Riechert SE, Hedrick AV (1993) A test for correlation among fitness-linked behavioural traits in the spider

280 Agelenopsis aperta (Araneae, Agelenidae). Anim Behav 46: 669–675

281 Sih A, Bell A, Johnson JC (2010): Behavioral Syndromes. In: Westneat DE, Fox CW (eds.) Evolutionary Behavioral

282 Ecology, Oxford University Press

283 Sweeney K, Gadd RDH, Hess ZL, McDermott DR, MacDonald L, Cotter P, Armagost F, Chen JZ, Berning AW,

284 DiRienzo N, Pruitt JN (2013) Assessing the effects of rearing environment, natural selection, and

285 developmental stage on the emergence of a behavioral syndrome. Ethology 119: 436–447

286 Toft S (2013) Nutritional Aspects of Spider Feeding. In: Nentwig W (ed.) Spider Ecophysiology, Springer-Verlag,

287 Berlin

288 Wilder SM (2011) Spider nutrition: an integrative perspective. Adv Insect Physiol 40: 87–136

289 Wilson DS, Clark AB, Coleman K., Dearstyne T. (1994) Shyness and boldness in humans and other animals. Trends

290 Ecol Evol 9: 442–446

291 Wolf M, Weissing FJ (2010) An explanatory framework for animal personalities differences. Phil Trans R Soc B

292 365: 3959–3968

293 Wolf M, Weissing FJ (2012) Animal personalities: consequences for ecology and evolution. Trends Ecol Evol 27:

294 452–460

295

11

296 Figure captions

297

298 Fig. 1 Relationship between aggressiveness toward prey and shyness across 3 weeks (a), in the first week (b), and in

299 the second week (c) of the study.

300

301 Fig. 2 Comparison as to level of aggressiveness toward prey of P. albidus among 3 weeks. Thick lines show

302 medians, boxes represent quartiles, and whiskers indicate 1.5 interquartile range.

303

304 Fig. 3 Comparison of shyness level of P. albidus among 3 weeks. Thick lines show medians, boxes represent

305 quartiles, whiskers indicate 1.5 interquartile range, and points identify outliers.

306

12

307 Table captions

308

309 Table 1 Pearson correlations expressing consistency repeatability (r) of individual shyness between specific weeks.

310

311 Table 2 Spearman correlations expressing consistency repeatability (rs) of individual aggressiveness toward prey

312 between specific weeks.

313

314 Table 3 Interclass correlation expressing agreement repeatability (R) of individual (n=15) behavioral traits.

13

Table 1

Weeks r P n

1 and 2 -0.187 0.542 15 1 and 3 -0.399 0.375 7

2 and 3 0.505 0.245 7

Table 2

Weeks rs P n

1 and 2 -0.128 0.677 15

1 and 3 -0.154 0.742 7

2 and 3 0.686 0.089 7

Table 3

Trait R 95% CI P

Shyness 0 0–0.38 0.62

Aggressiveness toward prey 0.1 0–0.51 0.3

14

a) 8 b)

8 7

]

h

2 exp (1.62 - 0.001x)

/

exp (1.715 - 0.001x)

d 6

e

l

l

i

k

6

y

e

r 5

p

[

s

s

e

n 4

e

v

i 4

s

s

e

r 3

g

g

A

2

2 1

100 200 300 400 500 600 100 200 300 400 500 600

Shyness - time until movement initiation [s ] Shyness - time until movement initiation [s]

8 c)

] 7

h

2

/ exp (2.1 - 0.0025x)

d

e

l 6

l

i

k

y

e

r

p

[ 5

s

s

e

n

e 4

v

i

s

s

e

r

g 3

g

A

2 1

100 200 300 400 500

Shyness - time until movement initiation [s]

Fig. 1

15

8

]

h

2

/

d

e

l

l

i

6

k

y

e

r

p

[

s

s

e

n

e

4

v

i

s

s

e

r

g

g

A 2

1 2 3

Week

Fig. 2

16

0

0

6

]

s

[

n

o

i

0

t

0

a

i

5

t

i

n

i

t

n

e 0

0

m

4

e

v

o

m

l

i 0

t

0

n

3

u

e

m

i

t

0

-

0

s

2

s

e

n

y

h

0

S

0 1

1 2 3

Week

Fig. 3

17