Eutrophication interferes with sand goby mating success, sexual selection, and parental care

(1)

Faculty of Biological and Environmental Sciences University of Helsinki

Finland

Doctoral dissertation

To be presented for public discussion with the permission of the Faculty of Biological and Environmental Sciences of

the University of Helsinki via Zoom, on the 13th of August, 2020 at 12 o’clock noon.

Helsinki 2020

(2)

Supervised by

Professor Kai Lindström | Åbo Akademi University, Finland Dr. Topi Lehtonen | University of Oulu, Finland

Docent Hannu Pietiäinen | University of Helsinki, Finland

Reviewed by

Professor Raine Kortet | University of Eastern Finland, Finland Docent Markus Öst | Åbo Akademi University, Finland

Examined by

Professor Gunilla Rosenqvist | Uppsala University, Sweden

Custos

Professor Emma Vitikainen | University of Helsinki, Finland

•• ••

Photos by Marja järvenpää | layout by lauri järvenpää unigrafia | helsinki | 2020

ISBN 978-951-51-6387-5 (print) ISBN 978-951-51-6388-2 (PDF)

(3)

ABSTRACT

Environmental changes of anthropogenic origin cause serious challenges to animal populations. Animals often first respond to changes in the environment by altering their behaviour. This can have marked consequences for both individuals and populations, in particular when behavioural responses are related to survival or reproduction. In aquatic environments, a major anthropogenic change is eutrophication, with more than 400 eutrophicated coastal systems having been identified globally. Eutrophication, due to enhanced growth of planktonic algae, leads to increased turbidity, decreased visibility and a narrow- ing of the light spectrum in the water column, as well as to increased siltation of decomposing algae and alterations in oxygen levels. All these changes to the environment hold the potential to interfere with the behaviour of aquatic organisms in many ways.

In this thesis, I experimentally studied the effects of increased abundance of planktonic algae and consequent turbidity on fish reproductive behaviour in the severely eutrophicated Baltic Sea. As a model system, I used the sand goby (Pomatoschistus minutus), a small marine fish species with a resource-defence mating system and male parental care. I conducted a series of laboratory studies in which I manipulated water turbidity using cultivated planktonic algae and explored the effects of it on various aspects of sand goby reproduction:

mating system and sexual selection (Chapter I), parental care and offspring survival (Chapter II and III), male mating success (Chapter III) and reproductive decisions under predation risk (Chapter IV).

In Chapter I, I focused on the effects of algal turbidity on the mating system and opportunity for sexual selection in the sand goby. Due to intense male-male competition and female choice, large sand goby males can typically monopolize multiple matings, whereas some males do not get a chance to mate at all. To study how these aspects of sand goby reproduction are influenced by algal turbidity, I allowed four males of different size to establish nests and sequentially introduced them four ripe females either in turbid or clear water.

I found that in turbid conditions matings were more evenly distributed among males, opportunity for sexual selection was lower and mating success was less skewed towards large males than in clear water. These effects of turbidity may, for example, be due to a change in the preferences of the females or their ability

(7)

to express these, or because of changes in male-male competition or male mate attraction effort.

In Chapters II and III, I explored whether algal turbidity affects male parental care behaviour and egg survival. I allowed sand goby males to care for a single female egg clutch (Chapter II) or a large egg clutch spawned by multiple females (Chapter III) and monitored male parental behaviour and egg survival under turbid and clear water conditions. Males caring for a single female egg clutch fanned their eggs less and spent more time away from their nest in turbid than clear water in the beginning of the care period. This difference in parental behaviour disappeared later in the brood cycle. Despite decreased fanning early in the parental phase, egg survival was higher in turbid conditions.

The time spent fanning early in the brood cycle did not affect egg survival in either clear or turbid water, suggesting that fanning early in the parental phase may function more as courtship than care. Decreased fanning in turbid water could be due to decreased efficacy of visual sexual signalling under low visibility. Sand goby females prefer males showing high fanning effort. Under low visibility, this advertisement component of care is likely to be less effective and may not be worth investing in. Instead, searching for females might be more beneficial for mating success, which is in accordance with my observation that males spent more time away from their nest when water was turbid rather than clear. Lesser fanning and higher egg survival under algal turbidity could also be explained by the conditions in turbid water being better for the developing eggs.

Interestingly, when males were caring for a large egg clutch spawned by multiple females (Chapter III), neither parental care nor egg survival was affected by water turbidity. It may be that a large brood required a very high parental effort from the male in both clear and turbid water. Additionally, the high perceived female density may have decreased the need for mate search in turbid conditions. Lastly, males may perceive the value of a large egg mass they already have so highly that they rather invest in care than searching for additional mates.

In Chapter III, I also examined whether turbidity affects male maximal mating success. I introduced five ripe females of random size to one nest-holding

(8)

male in either clear or turbid water and observed how many mates and eggs the male received. When the fish were spawning in clear water, the number of eggs in the male’s nest increased with the total weight of the five females in his aquarium. When spawning took place in turbid water, however, there was no such relationship between the total weight of the females and the number of eggs laid, even though the number of females that spawned was the same as in clear water. It seems that this effect was due to a female decision to adapt spawning to water turbidity rather than a male decision to limit the number of eggs in his nest.

In Chapter IV, I investigated the effects of algal turbidity and predation risk on one particular reproductive decision: the latency of gobies to spawn. I introduced one ripe female to a male either in the presence or absence of a predator in water that was either clear or turbid and recorded the latency to spawning.

I found that gobies were more reluctant to spawn in the presence than in the absence of the predator. Interestingly, latency to spawning was unaffected by turbidity, suggesting that the gobies experienced the predation threat equally high in turbid as in clear water, for instance because they relied on non-visual (e.g. chemical) predator cues at least under turbid conditions.

Taken together, the results of this thesis show that eutrophication can influence important aspects of fish reproductive behaviour: mating success, sexual selection and parental care were all affected by algal turbidity. Most important- ly, I show that sexual selection is relaxed in turbid water. By hampering sexual selection and parental care, eutrophication holds a potential to influence, not only the fitness of individuals, but also population viability and community structure. My results emphasize that human-induced environmental changes can have severe consequences already through subtle alterations in animal behaviour.

(9)

(10)

TIIVISTELMÄ

Rehevöityminen on Itämeren ja lukuisien muiden maailman vesistöjen suu- rimpia ympäristöongelmia. Se johtaa muun muassa planktisten levien runsas- tumiseen, minkä seurauksena vesi samenee, näkyvyys heikkenee, pohjalle las- keutuvan hajoavan eloperäisen aineksen määrä kasvaa ja happipitoisuudet niin vedessä kuin pohjallakin voivat muuttua.

Väitöskirjatyössäni tutkin planktisten levien runsastumisen vaikutuksia ka- lojen lisääntymiskäyttäytymiseen. Mallieläimenä käytin pientä merikalaa hie- tatokkoa (Pomatoschistus minutus). Lisääntymiskauden alussa hietatokkokoiraat valtaavat pesäpaikan ja houkuttelevat naaraita pesäänsä kutemaan. Kudun jäl- keen naaras jättää pesän, ja koiras huolehtii mätimunista yksin, kunnes ne kuo- riutuvat. Koiras suojelee pesää ja hapettaa munia tuulettamalla niitä evillään.

Tutkimuspopulaatiossa pesäpaikkoja on niukasti ja kilpailu niistä on tiukkaa.

Jos pesän koko antaa myöten, yksi suuri kilpailukykyinen koiras voi saada usean naaraan mätimunat hoidettavakseen monen muun koiraan jäädessä kokonaan ilman pariutumisia. Väitöskirjassani selvitin kokeellisesti, miten rehevöitymi- sen aiheuttama planktisten levien runsastuminen vaikuttaa hietatokon pariu- tumisjärjestelmään ja seksuaalivalintaan (Kappale I), koiraiden pariutumis- menestykseen (Kappale III), jälkeläishoivaan ja jälkeläisten selviämiseen, kun koiras huolehti yhden naaraan kudusta (kappale II) tai suuresta kutulaikusta, jossa oli monen naaraan mätimunat (Kappale III) sekä lisääntymispäätöksiin pedon läsnäollessa (Kappale IV).

Osoitin, että planktisten levien runsastuminen vaikuttaa hietatokon lisään- tymiskäyttäytymiseen monin tavoin. Levän samentamassa vedessä hietatokon pariutumisjärjestelmä mureni: pariutumiset jakaantuivat tasaisemmin koiraiden kesken kuin kirkkaassa vedessä ja seksuaalivalinta heikkeni (Kappale I). Toisin kuin kirkkaassa vedessä, levän samentamassa vedessä suuri koko ei taannut koiraille menestystä pariutumismarkkinoilla (Kappale I). Sameassa vedessä myös koiraiden hoivakäyttäytyminen muuttui: yhden naaraan kudusta huolehtivat koiraat tuulettivat mätimuniaan vähemmän ja viettivät enemmän aikaa pois pesiltään hoivajakson alussa kuin koiraat kirkkaassa vedessä (Kap- pale II). Vähäisemmästä hoivasta huolimatta munat selvisivät sameassa vedessä paremmin. Kun koirailla oli huolehdittavanaan usean naaraan mätimunat, hoi- vassa ja munien selviytymisessä ei ollut eroja samean ja kirkkaan veden välillä (Kappale III). Planktisten levien runsastuminen näytti kuitenkin rajoittavan

(11)

munamäärää, jonka koiraat pesäänsä saivat (Kappale III). Kaloilla koko ja he- delmällisyys kulkevat yleensä käsi kädessä, ja isommilla naarailla on enemmän mätimunia. Kun koirailla oli tilaisuus pariutua useiden naaraiden kanssa, mu- namäärä niiden pesässä kasvoi naaraiden yhteispainon myötä kirkkaassa vedes- sä. Levän samentamassa vedessä yhteys munamäärän ja naaraiden koon välillä katosi. Vaikuttaa siltä, että naaraat säätelivät kutemiensa munien määrää veden sameuden mukaan. Hietatokkojen suhtautumiseen pedon läsnäoloon sameus ei näyttänyt vaikuttavan: kalat lykkäsivät riskialttiita lisääntymispuuhia pedon läsnäollessa sekä sameassa että kirkkaassa vedessä (Kappale IV).

Muutokset lisääntymispäätöksissä, seksuaalivalinnassa ja hoivakäyttäyty- misessä voivat vaikuttaa paitsi yksilöihin itseensä ja niiden jälkeläisiin, myös populaatioiden elinkykyyn ja eliöyhteisöihin. Ihmisen aiheuttamilla ympäris- tömuutoksilla voikin olla mittavia seurauksia myös silloin, kun ne näyttävät aiheuttavan vain vähäisiä muutoksia eläinten käyttäytymisessä.

(12)

SUMMARY

1. Introduction

Environmental changes of anthropogenic origin pose various challenges to wildlife. Animals often first respond to changes in the environment by altering their behaviour. Changes in behaviour can have a direct bearing on an individuals’ fitness, in particular when the behavioural responses are related to survival or reproduction. Eutrophication is a major environmental change in aquatic environments. In this thesis, I explore how eutrophication-induced increased abundance of planktonic algae can affect the reproductive behaviour in fish, using the sand goby, Pomatoschistus minutus, as my study species.

For the sake of simplicity, in this thesis I often refer to increased abundance of planktonic algae as algal turbidity.

1.1. Reproductive behaviour is sensitive to environmental and social conditions

An individual’s reproductive success depends both on the number of offspring it produces and on how well these offspring survive to reproduce later. In most species, males, who generate enormous numbers of tiny sperm, could potentially sire eggs of numerous females (Bateman, 1948; Trivers, 1972). In practice, to sire any, males usually must both cope with rivals that are interested in the same mating opportunities and impress females (Alcock, 2005). Females typically produce a limited number of large eggs and often have a lower potential reproductive rate than males (Clutton-Brock and Parker, 1992). This dispari- ty creates a situation where the number of sexually receptive males is higher than the number of sexually receptive females, i.e. the operational sex ratio is male-biased (Clutton-Brock and Parker, 1992; Emlen and Oring, 1977).

Under these circumstances, females are usually choosy regarding who gets to fertilize their eggs, and male reproductive success is limited by the access to receptive females, whereas female reproductive success is limited by access to resources (Trivers, 1972). However, the situation can also be opposite with

(13)

females having a higher potential reproductive rate (Berglund et al., 1989), or with rapid changes over the breeding season (Forsgren et al., 2004).

Sexual selection occurs when individuals of a given sex vary in their ability to attract mates (inter-sexual selection) or in their ability to compete over mating opportunities (intra-sexual selection), i.e. mates or resources necessary for attracting them (Darwin, 1871). Individuals that are successful in mating competition and mate attraction are likely to have a high reproductive success and pass more copies of their genes to the next generation (Andersson, 1994).

Competing over breeding opportunities and being selective about mating partners do not come without costs. Investment in gametes, mating competition, mate attraction and parental care all bear an energetic cost and may, thereby, have a negative impact on the growth and survival of parents and may also keep individuals from attracting further mates (Andersson, 1994).

Reproductive activities also expose individuals to disease and predators and may hence result in lowered survival (Harshman and Zera, 2007; Magnhagen, 1991). Although the costs associated with mate choice and mate competition can be considerable, also the benefits achieved may be sumptuous. For example, in species with a resource-defence mating system, a successful competitor can obtain resources that allow it to monopolize multiple mates, while individuals who fail in resource competition may gain zero reproductive success at that breeding attempt (Emlen and Oring, 1977; Shuster and Wade, 2003).

Choosiness about mates, may, in turn, lead to a loss of time, energy, and mating opportunities, but a suitable mate can provide the partner direct benefits, such as high fertilization success, food, nest site, protection, and parental care, or indirect benefits, i.e. beneficial characteristics the offspring will inherit (Alcock, 2005). The latter can come in the form of good or compatible genes that increase the viability or attractiveness of offspring, thereby increasing their fitness (Hamilton and Zuk, 1982; Petrie, 1994; Zeh and Zeh, 1997).

The breeding behaviour of individuals is affected by social context, environmental conditions, such as predator abundance or breeding resource avail- ability, and the individual’s own state, for instance age or health. For example, female mate preferences can be influenced both by male-male competition (Lehtonen and Lindström, 2009) and by the number of female competitors (Heubel, 2018), while offspring provisioning can depend both on the parents’

age and time of the season (Arnold et al., 2018). In the sand goby, salinity levels influence male nesting behaviour, and large and small males respond differently to salinity levels (Lehtonen et al., 2016). In humans, women’s judgment of men’s attractiveness varies over the menstrual cycle, so that the women in

(14)

the pre-ovulatory (i.e. fertile) state are more attracted to masculine features of men than women in the non-fertile state (reviewed by Gildersleeve et al., 2014).

Mate preferences are also affected by stress, as shown e.g. in the stalk-eyed fly (Cyrtodiopsis dalmanni), in which stress decreases the preference for more attractive mates (Hingle et al., 2001), and in humans, in which stress reduces female preference for male facial masculinity (Ditzen et al., 2017) and switches male preference from similar mates to dissimilar mates (Lass-Hennemann et al., 2010).

Given that reproductive decisions and behaviours are sensitive to varying environmental, social, and individual-related conditions, it is clear that human-induced changes in the environment have a great potential to influence these too.

1.2. Human-induced changes in environment can interfere with reproductive behaviour

Many human-induced environmental changes, such as noise (Gordon and Uetz, 2012; Huet des Aunay et al., 2017; Schmidt et al., 2014), light pollution (Bird and Parker, 2014), chemicalization (Zala and Penn, 2004), or global warming (Martín and López, 2013) interfere with the natural breeding behaviour of wildlife. Global warming, for example, has resulted in advances in bird breeding dates and in earlier spawning in amphibians (Lane et al., 2011 and the references therein; Walther et al., 2002). Exposure to endocrine-disrupting chemicals, which are present in numerous man-made products ranging from contraceptive pills to non-stick frying pans, has led to changes in reproductive organs, secondary sexual characteristics, and reproductive behaviour in fish, amphibians, aquatic reptiles, and birds, in many cases leading to serious impairment of reproduction (Zala and Penn, 2004 and the references therein). Behaviours affected include nest building (McCarty and Secord, 1999a, 1999b), courtship (Baatrup and Junge, 2001; Tomkins et al., 2017), competition

between rivals (Tomkins et al., 2017) and parental care (McCarty and Secord, 1999a, 1999b). For instance, tree swallows (Tachycineta bicolor) living close to PCB-contamination build smaller and lower-quality nests and are more prone to abandon or bury their eggs than birds from cleaner areas (McCarty and Secord, 1999a, 1999b).

Human-induced environmental changes often affect reproductive behaviour by changing the prerequisites for successful sexual signalling. Anthropogenic

(15)

noise and light pollution, for example, can interfere with both mate attraction and finding a mate, as demonstrated e.g. in tree frogs, field crickets, and glow worms. Noise pollution increases the stress hormone levels in tree frogs (Hyla arborea) resulting in diminished coloration of vocal sacs, which are used to attract females (Troianowski et al., 2017). The ability of female field crickets (Gryllus bimaculatus) to locate singing males decreases when they are exposed to traffic noise (Schmidt et al., 2014). Similarly, traffic noise can mask male advertisement calls in the grey treefrog (Hyla chrysoscelis) (Bee and Swanson, 2007). In common gobies (Pomatoschistus microps), continuous noise decreases mating success, probably due to a reduced ability of females to assess male acoustic signals (Blom et al., 2019). In the glow-worm (Lampyris noctiluca), light pollution, even at very low levels, can block the ability of males to locate the glowing females (Bird and Parker, 2014). Man-made chemicals, in turn, have been reported to disturb pheromone-mediated communication in a number of taxa, including wasps, moths, snails, newts, and fish, resulting in reduced mate responsiveness (Lurling and Scheffer, 2007 and the references therein;

Zala and Penn, 2004 and the references therein). Global warming, too, interferes with sexual signalling, as demonstrated in mountain lizards (Iberolacerta cyreni), in which the detectability and persistence of male scent marks is lower at high temperatures (Martín and López, 2013).

Lowered detectability of and decreased responsiveness to the sexual signals of potential mates can lead to disrupted mate choice and impaired reproduction. This has been demonstrated for example in domestic canaries (Serinus ca- naria domestica), in which sexual receptivity of females is reduced, and they lay fewer eggs when attractive male songs are overlapped by low-frequency noise (Huet des Aunay et al., 2017). In native and invasive Cyprinella fish species, exposure to an endocrine disrupting chemical led to changes in the expression of male secondary sexual characteristics and mate choice, breaking down the sexual isolation between the two (Ward and Blum, 2012). Similarly, it has been suggested that numerous Lake Victoria cichlid sibling species were lost when eutrophication masked the species-specific male colour signals, thereby hindering female mate choice and leading to hybridization between the sub-species (Seehausen et al., 1997).

From the above, it is clear that human-induced changes in environment can interfere with reproductive behaviour of natural populations in many ways, with potentially drastic consequences to individuals and populations.

(16)

1.3. Eutrophication is a major anthropogenic change in aquatic environments

In aquatic environments a major anthropogenic change is eutrophication, which has been identified in more than 400 coastal systems around the globe (Selman et al., 2008). Eutrophication of water bodies is caused by an excess input of nutri- ents, in particular nitrogen and phosphorus, from agriculture, industry, waste waters, transportation, and other human activities (Smith et al., 1999). A surplus of nutrients leads to increased primary production. The growth of planktonic algae is enhanced, which leads to increased turbidity, decreased visibility, and a narrow- ing of the light spectrum in the water column (Sanden and Håkansson, 1996).

Depending on the balance between photosynthesis and respiration of algae, oxygen conditions in eutrophied waters can vary from supersaturation to anoxia (Elmgren, 1989; Muller and Stadelmann, 2004). When algae die, they settle on the bottom, where they can cover breeding grounds, suffocate eggs of demersally spawning species, and cause changes in the benthic ecosystems (Jobling, 1995;

Muller and Stadelmann, 2004). Decomposing algae consume oxygen, which can lead to hypoxic or anoxic conditions on the bottom and in the surrounding water column (Carpenter et al., 1998; Larsson et al., 1985). This can kill animals, com- promise their reproduction, or force them to migrate. Under anoxia, nutrients stored in the bottom sediment are released back into to the water column, which further enhances eutrophication.

The Baltic Sea, in particular, is heavily impacted by eutrophication (Cederwall and Elmgren, 1990; Elmgren, 1989; Perttilä et al., 1995; Sanden and Håkansson, 1996; Sanden and Rahm, 1993). The severity of eutrophication in the Baltic Sea

is, for example, reflected in the largest hypoxic area in the world (Carstensen et al., 2014). The Baltic Sea is especially prone to eutrophication, because it is a shallow and enclosed sea basin with a large and densely populated drainage area. Conse- quently, nutrient load is high but the water changes slowly, and nutrients contrib- uting to eutrophication remain in the Baltic Sea for a long time. Although the anthropogenic nutrient load to the Baltic Sea has recently been decreasing, the symptoms of eutrophication – such as algal blooms, murky water, and anoxic sea floor – remain obvious and severe. Ongoing climate change, which warms up the water and increases precipitation and hence the run-off of nutrients from land, is expected to worsen the state of the Baltic Sea further (Carstensen et al., 2014).

Species in the Baltic Sea are living in a challenging brackish water environment where species of both marine and fresh water origin are subjected to salinity

(17)

stress. This can make them particularly vulnerable to any further changes in the environment, including those caused by eutrophication (Breitholz et al., 2001).

1.4. Eutrophication interferes with fish reproductive behaviour

Reproduction is a critical life-history stage that is likely to be highly affected by eutrophication (Winfield, 2004). An increased abundance of planktonic algae can decrease the visibility and change the prevailing light conditions in water bodies. This can interfere with the use of visual information and, thereby, affect various aspects of animal reproduction, such as communication, mate detection, competition for mates and mate choice (Guthrie and Muntz, 1993), as well as responses to predation threat (e.g. Lehtiniemi et al., 2005; Sohel and Lindström 2015). Increased siltation of decomposing algae and consequent alterations in oxygen levels can influence parental care and offspring survival (Jobling, 1995; Jones and Reynolds, 1999a; Potts et al., 1988). Both the above can have remarkable consequences on individuals and populations. Altered parental care, for example, can directly affect the fitness of both the parents and their offspring, while changes in mate detection, competition for mates and mate choice can weaken sexual selection on traits that improve mating success and offspring quality.

Fish are vulnerable to disturbances during their juvenile and reproductive stages (Winfield, 2004; Wootton et al., 1995). Fish species with complicated reproductive systems seem to be especially prone to disturbances (Bruton, 1995).

Parental care occurs in around 22% of teleost fish families (Sargent and Gross, 1986), and it is particularly common in species reproducing in fresh waters, estuaries and marine littoral zones (Jones and Reynolds, 1997). These environments are also most susceptible to human impact, including eutrophication.

Recent work has shown that eutrophication has the potential to seriously interfere with fish reproduction. One of the first and most dramatic examples comes from Lake Victoria, where numerous sympatric rocky-shore cichlid species have been lost, apparently due to hybridization, when turbidity masked the species-specific colour signals hindering female mate choice for the males of their own sub-species (Seehausen et al., 1997). Around the same time, other studies indicated that eutrophication may affect fish reproductive behaviour via increased siltation and altered oxygen conditions. Fifteen-spine stickleback (Spinachia spinachia) males were found to increase parental activities following

(18)

silt addition (Potts et al., 1988). Similarly, common goby males were shown to increase parental activities when subjected to low oxygen conditions (Jones and Reynolds, 1999a). Consequently, males breeding under low oxygen lost more weight than control males and were more likely to abandon brood care during consecutive brood cycles (Jones and Reynolds, 1999b). Moreover, common goby females who normally prefer already mated males, reversed their preference to favour males that were not guarding eggs (Reynolds and Jones, 1999).

Findings such as the above were the starting point for my thesis. Eutroph- ication was, and still is, a major environmental issue in the Baltic Sea, as it is in numerous other water bodies around the world (Selman et al., 2008). As mentioned above, there already were indications that eutrophication interferes with reproductive behaviour. However, at the time I started the work presented in this thesis, there were only few studies of the direct effects of algal turbidity on fish behaviour (e.g. Cobcroft et al., 2001; Utne-Palm, 2002) and, to my knowledge, none of these concerned reproduction. Hence, a completely new area of research waited for to be explored. Could enhanced algal abundance influence mating decisions, sexual selection, mating systems, parental care, and other aspects of reproductive behaviour in fish? Could it affect the reproductive output of individuals or offspring survival?

Since I started to work on this topic, a growing body of evidence has been gathered that now shows that eutrophication-induced increased algal abundance can interfere with various aspects of reproduction ranging from nest building to mate choice, parental care, and the offspring survival. Below, I will briefly go through some of the demonstrated effects.

Nest building

The use of some kind of a nest site is common in many taxa of teleost fish (Barber, 2013; Hansell, 2005). A nest provides the offspring with a protected environment to develop and may also act as an extended phenotype that reveals information about the quality of the nest owner (Barber, 2013; Schaedelin and Taborsky, 2009).

In aquatic environments, oxygen levels can vary greatly for a number of reasons, including eutrophication. Insufficient oxygenation can hamper both egg development and egg survival (Hale et al., 2003). Many fish species, which lay their eggs in enclosed nests or cavities, where water movement is limited, fan their eggs using pectoral and tail fins (Barber, 2013). Fanning oxygenates

(19)

the developing eggs but can also be energetically demanding and costly to the parent (Jones and Reynolds, 1999b; Lissåker et al., 2003).

Nest site and/or nest characteristics are often adjusted according to prev- alent environmental conditions (Eggers et al., 2006; Lehtonen et al., 2016;

Lissåker and Kvarnemo, 2006; Lissåker et al., 2003; Rushbrook et al., 2010).

Such adjustments can affect the costs of nest building and maintenance as well as the conditions inside the nest and their suitability for the offspring (Jones and Reynolds, 1999a; Lissåker and Kvarnemo, 2006).

Increased abundance of algae has been shown to influence nest building in the three-spine stickleback (Gasterosteus aculatus) and in the sand goby. In the sand goby, algal turbidity breaks down the association between male size and nest elaboration, i.e. the amount of sand piled on top of the nest (Lehto- nen et al., 2015). In the three-spine stickleback, nest building takes longer and completed nests are smaller under algal turbidity than clear water conditions (Wong et al., 2012). In turbid conditions, nests are built with wider entrances, possibly as means to increase water flow through the nest (Wong et al., 2012).

Enhanced water flow could maintain higher oxygen levels in the nest and also help preventing algae from smothering the eggs. However, it could also be in conflict with the protectiveness of the nest (Barber, 2013).

Mate search, mate attraction, and mate choice

Mate choice requires individuals to encounter potential mates, attract them, detect and evaluate signals delivering information about their quality, and to make a choice (Bateson, 1987; Candolin, 2019).

Increased abundance of planktonic algae, resulting in decreased visibility, has been shown to interfere with perception and reliability of signals used in mate choice (Sundin et al., 2016; Wong et al., 2007), to reduce courtship activity (Candolin et al., 2016; Michelangeli et al., 2015), and to prolong mate evaluation time (Candolin et al., 2016). Visual conditions mimicking algal turbidity have, in turn, been reported to decrease the time associated with potential mates and to hinder mate preference for large size (Sundin et al., 2010).

Moreover, in the three-spine stickleback, algal turbidity affects the cues used in mate choice, such that in clear water, visual cues are more important than olfactory cues, whereas in turbid water this situation is reversed (Heuschele et al., 2009). Interestingly, stickleback females choose males with higher hatching success less often in turbid than in clear water (Candolin et al., 2016).

(20)

In Lake Victoria cichlids, algal turbidity affects both male coloration and female perception of species-specific male colour signals (Maan et al., 2010;

Seehausen and van Alphen, 1998). As already referred to earlier, this hinders female mate choice for the males of their own species resulting in hybridization between species (Seehausen and van Alphen, 1998; Seehausen et al., 1997).

Sexual selection

Sexual selection occurs when individuals differ in their ability to compete over mating opportunities or in their ability to attract mates (Darwin, 1871).

Interestingly, increased abundance of algae has been shown both to intensify and to relax sexual selection, the latter being the more common outcome.

Decreased sexual selection on male size has been demonstrated in the sand goby exposed to algal turbidity (Järvenpää and Lindström, 2004). Similarly, increased growth of filamentous algae has been shown to relax selection on male nuptial colouration and courtship activity in the three-spine stickleback (Candolin et al., 2007). Weakened sexual selection in response to low visibility might also explain why heavily parasitized three-spine stickleback males acquire more eggs than less parasitized males in conditions mimicking dense growth of filamentous algae (Heuschele and Candolin, 2010) and why Lake Victoria cichlids from eutrophicated parts of the lake are duller in coloration than the other males (Maan et al., 2010; Selz et al., 2014; Selz et al., 2016).

In some cases, turbidity has been shown to intensify sexual selection. Sun- din et al. (2017) have shown in the broad-nose pipefish (Syngnathus typhle) that algal turbidity enhances sexual selection on mate size, as well as mating success and reproductive success.

Parental behaviour

Parental care is of critical importance for the survival of the offspring. In fish, parental behaviours include protecting the offspring against predators and rivals who might consume or steal the eggs, keeping the eggs safe from patho- gens, and fanning the eggs to provide them with oxygen (Clutton-Brock, 1991).

Care does not only influence offspring survival but may also affect mating decisions, as care behaviour can be used as a cue in mate choice (Lindström et al., 2006).

Algal turbidity has been shown to reduce male fanning behaviour but to improve hatching success in the sand goby (Järvenpää and Lindström, 2011)

(21)

and in the three-spine stickleback (Candolin et al., 2008). In the three-spine stickleback, enhanced macro algal growth reduces aggressive interactions between the males during the parental phase, possibly due to decreased visibility, enabling them to maintain a better body condition and to complete more breeding cycles (Candolin et al., 2008).

Increased algal growth affects fish reproductive behaviour also indirectly

Evidence shows that also the indirect effects of increased algal growth, such as oxygen depletion, increased siltation of organic material, and higher pH can affect fish reproductive behaviour. For example, when oxygen levels are low, male sand gobies build nests with larger openings and they are less likely to rebuild destructed nests (Lissåker and Kvarnemo, 2006; Lissåker et al., 2003).

Larger openings probably help to improve the oxygen conditions inside the nest but may also make nests easier to detect for egg predators and harder to defend for the nest owners. In broad-nose pipefish, latency to courtship and latency to mating increase in mildly hypoxic conditions (Sundin et al., 2015).

Both lowered oxygen concentrations and increased siltation have been shown to increase egg fanning (Jones and Reynolds, 1999a; Lissåker and Kvarnemo, 2006; Takegaki and Nakazono, 1999; Torricelli et al., 1985), which likely increases the costs of care (Jones and Reynolds, 1999a; Lissåker et al., 2003).

Increased water pH resulting from increased algal growth, in turn, boosts olfactory communication in the three-spine stickleback, such that ripe females are more attracted to male olfactory cues when pH is higher (Heuschele and Candolin, 2007). Improved olfaction could to some extent compensate for the deteriorated visibility in eutrophied waters.

Taken together, the examples above show that eutrophication-induced increased algal growth can interfere with fish reproductive behaviour in many ways.

(22)

1.5. Aims of the thesis

In this thesis, I set out to experimentally study the effects of increased abundance of planktonic algae on fish reproductive behaviour, mating success, and sexual selection in the severely eutrophicated Baltic Sea. To gain understanding of the topic, I conducted a series of laboratory studies in which I manipulated water turbidity using cultivated planktonic algae. I explored the effects of turbidity on various aspects of reproduction: mating system and sexual selection (Chapter I), parental care and offspring survival (Chapters II and III), male mating success (Chapter III), and reproductive decisions under predation risk (Chapter IV). In all these studies, I used the sand goby, a small marine, littoral fish species with a resource-defence mating system and male parental care, as a model system. In study IV, in which I explored the effects of predation risk and turbidity on mating decisions, I used the European perch, Perca fluviatilis, to impose a perceived risk of predation.

More specifically, my aims were the following:

• to explore how increased abundance of planktonic algae and consequent turbidity affect the mating system and opportunity for sexual selection in the sand goby by measuring the distribution of mating success among males in clear water and under algal turbidity (Chapter I),

• to test whether increased abundance of planktonic algae affects parental care and egg survival in the sand goby, by comparing male fanning behaviour and egg survival in clear and turbid water, when males were caring for an egg clutch spawned by a single female or a large egg clutch spawned by multiple females (Chapters II and III),

• to examine the effects of increased abundance of planktonic algae on female spawning and male mating success by allowing individual males to spawn with several females in either clear or turbid water (Chapter III),

• to compare the latency to spawning in both the presence and absence of a visually oriented piscivorous predator under both algal turbidity and clear water conditions (Chapter IV).

(23)

MATERIAL & METHODS 2.

2.1. The model system: the sand goby

Sand gobies are small (3.5 – 6.5 cm), sand-coloured marine fish that spend the only breeding season of their short life on the shallow water bottoms where they are easily hidden from sight. They are common and have a wide distribution area ranging from the Baltic Sea to the North Sea, the North Atlantic, and the Mediterranean Sea (Kovacic and Patzner, 2011). At first sight, sand gobies may seem dull and inactive, but a closer look reveals bulging eyes that gleam in metallic green, red, and blue, male ornaments that are moderate in size but shiny blue and black in colour, the beautiful rounded, yellowish bellies of ripe females, and an utterly interesting behaviour.

The sand goby is an excellent model system for studying the effects of human-induced environmental changes on reproduction for several reasons. The

Figure 1. The experiments were conducted using the sand goby (Pomatoschistus minutus) as a model system.

(24)

sand goby is common throughout its wide distribution area, it is practically small and readily breeds and displays other natural behaviours in captivity (Le- htonen, 2006). It has a complicated reproductive system, allowing for studies on many aspects of reproductive behaviour, including mating system, mate choice and parental care. Of the more than 20 000 species of fish, the sand goby has been the fifth most used in behavioural studies of sexual selection (Amundsen, 2003), and there now is extensive background information on the sand goby reproductive behaviour and mating system (see e.g. Forsgren, 1999;

Lehtonen, 2012).

Sand gobies have a resource-defence mating system, in which males occupy and defend nest sites, such as mussel shells or stones, attract females to spawn in their nests and then take care of the eggs alone until they hatch (Fonds, 1973;

Hesthagen, 1977; Lehtonen and Lindström, 2004; Lindström, 1988). In my study area, the northern Baltic Sea, nest sites are scarce, and males vigorously compete over them. Large males are usually successful in male-male competition (Lindström, 1988; Lindström and Pampoulie, 2005), and one large male can guard eggs from multiple females simultaneously if his nest is large enough (Lindström, 1992; Lindström and Seppä, 1996). Because sand goby females lay their eggs in a single layer in the ceiling of the nest cavity, the surface area of the nest substrate largely determines the number of females that could place their eggs in the same nest simultaneously, and hence the maximum reproductive success of the nest owner at any breeding attempt (Lindström, 1992).

Once the male has occupied a nest site, he builds a nest by excavating under- neath the nest substrate and piling sand on top of it with his tail. The amount of sand piled on the top of the nest can be considerable and may exceed the body mass of the builder by 100 times (Lehtonen et al., 2013; Lehtonen et al., 2015). The male leaves a single narrow entrance to the nest and can often be seen posing either at the nest opening or on the top of the nest. The nest-building male prepares the interior of his nest by covering the ceiling with mucus that contains sperm; females will later lay their eggs onto the ceiling (Svensson and Kvarnemo, 2005). In other gobies, the mucus has been shown to contain pheromones and to have antimicrobial function, both of which can influence female mate choice (Blom et al., 2016 and the references therein).

A male in spawning condition develops a darkened tail, a dark, bluish-black coloration on the pectoral and anal fins, and a bright blue spot lined by white on the first dorsal fin (Forsgren, 1999; Pedroso et al., 2013). The male attracts females to spawn in his nest with an active courtship display. He poses showing his erected fins which are decorated with spawning colours, swims rapidly in

(25)

short bouts close to the female, and tries to lead her into his nest (“lead swim”) (Forsgren, 1999). He also advertises his parental abilities with courtship fanning at the entrance of the nest (Pampoulie et al., 2004). Inside the nest, he utters sounds which reveal information on his size and body condition (Blom et al., 2016; Lindström and Lugli, 2000; Pedroso et al., 2013).

Females that are ready to spawn have a rounded belly (García-Berro et al., 2019), which they present to a male by making small up and down hops close to him (Blom et al., 2016). During courtship, females develop dark eyes and eye surroundings, the meaning of which remains unclear (Olsson et al., 2017).

Sand goby females base their mate choice on multiple cues (Lehtonen et al., 2007), the importance of which varies depending on the situation (Leht- onen and Lindström, 2009; Lehtonen et al., 2010). Females favour good fa- thers; males that fan their eggs eagerly thereby showing their parental abilities (Lindström et al., 2006) and males with high egg-hatching success (Forsgren, 1997; Lehtonen and Lindström, 2007). Also, the male nest, where females lay their eggs, is of importance to females, who fancy males with well-built nests (Svensson and Kvarnemo, 2005) and large males in large nests (Lehtonen et al., 2007). In the study population, large male size alone does not seem to be attractive to females (Lehtonen et al., 2010). However, females prefer large males when competition between males is intense (Lehtonen et al., 2015). Sand goby females have also been suggested to prefer males that are colourful (Forsgren, 1992) and have a large blue spot on the dorsal fin (Kangas, 2000). Further- more, they have been reported to prefer intensively courting males (Forsgren, 1997) but the evidence here is mixed, suggesting that male courtship may be important for turning the heads of females, after which they can evaluate the other qualities of the courting male (Lehtonen, 2012). Recent research has re- vealed that also odour cues are important in female mate choice, and sand goby females strongly avoid laying eggs into nests that smell like an egg infection (Saprolegnia sp water mould), even when the nest belongs to a male they find visually attractive (Lehtonen and Kvarnemo, 2015).

If the female accepts the male and his nest, she lays her eggs in a single layer in the ceiling of the nest. The female leaves the nest after spawning, and the male then cares for the eggs alone until they hatch. He protects the nest and offspring against intruders, fans and cleans the eggs and removes dead and infected eggs from the nest. Fanning aerates the eggs and may also prevent sediment from accumulating on them. For the sand goby females, resources necessary for breeding and the male’s ability to successfully care for the offspring are the key factors defining her reproductive success. Male reproductive

(26)

success, in turn, is largely determined by four factors: success in competition over a nest site, the size of his nest, which can determine his mating success, access to receptive females, and his ability to keep the eggs alive and healthy.

(Forsgren, 1999; Lindström, 1998; Lindström and St.Mary, 2008)

2.2. Study site

The experiments presented in this thesis were conducted at Tvärminne Zoo- logical Station, southern Finland, during the sand goby breeding seasons in the years 2001-2006. The station is located by the Baltic Sea, at the entrance to the Gulf of Finland, which suffers from severe eutrophication.

2.3. Catching and maintaining fish

Sand gobies for the experiments were caught close by Tvärminne Zoological Station shortly before the experiments started. I caught females by trawling with a small hand trawl, while males were either trawled or caught directly from the nests using a hand net. In the lab, gobies were housed in separate-sex storage tanks with a continuous flow-through of fresh sea-water and fed ad libitum with live crustacea Neomysis integer and frozen chironomid larvae. Be- cause the storage tanks were placed in an outdoor lab with transparent walls and roof, they had natural light conditions. After the experiments all fish were either released back into their natural breeding area or used in other, unrelated experiments.

Perch for the predation risk experiment (Chapter IV) were collected by either fish traps or nets close by the station. Perch were housed in large, plant- filled tubs in an outdoor facility with natural light and temperature conditions and fed ad libitum with live brown shrimp. After the experiment, the perch were released back to their natural habitat.

(27)

2.4. Creating and measuring turbidity

Turbidity can be caused by many different factors, varying from clay or silt particles brought in by rivers or mixed in the water column due to dredging to turbidity created by increased abundance of planktonic algae. Consequently, research on the effects of turbidity on fish behaviour has been conducted either by using planktonic algae or inorganic substances, such as clay, depending on what is relevant regarding the aims of each study. Clay has been used in many feeding studies (e.g. Abrahams and Kattenfeld, 1997; Chamberlain and Ioannou, 2019; Sekhar et al., 2019; Sweka and Hartman, 2003; but see Shaw et al., 2006), while much of the research focusing on the effects of eutrophica- tion-induced turbidity on reproductive behaviour has used planktonic algae to create turbidity (e.g. Candolin et al., 2016; Heuschele et al., 2009; Lehtonen et al., 2015; Sundin et al., 2017). Here, I focused on the effects of eutrophi- cation-induced increased abundance of planktonic algae on reproductive behaviour. Therefore, I used cultivated unicellular planktonic algae, Brachiomonas submarina (10–15 µm in size), which is commonly found in the Baltic Sea to produce turbidity.

The initial culture originated from the pure monoculture maintained by Tvärminne zoological station. I grew the algae in aerated seawater in uncovered 50 l buckets under natural light conditions. I filtered seawater through a 20 µm sieve to eliminate most of the grazers, and added nitrogen and phosphorus, 0.046 and 0.036 g l-1, respectively, to the cultures to ensure sufficient nutrients for algal growth.

I added algae to the treatment tanks by first decreasing the water level in treatment and control tanks and then adding a mixture of water from the individual cultures to the treatment tanks until an original water level was reached.

Control tanks similarly received clear seawater. I made turbidity in the treatment tanks to visually match intense algal blooms that occur naturally in the Baltic Sea during the sand goby breeding season.

I monitored turbidity in the experimental tanks to verify that turbid pools remained turbid and turbidity corresponded to values found in nature. In Chap- ter I, I observed turbidity by sampling the treatment and control aquaria every 12 h and analysing the absorption spectrum on a Shimadzu UV-2101 PC–

spectrophotometer. Total absorption was obtained by summing the absorbance values from 380 to 800 nm (0.5 nm intervals). These values were used to obtain a mean absorbance from the five measurements taken from each aquarium. In the other three experiments, I measured turbidity as nephelometric turbidity

(28)

units (NTUs) using a Hach 2100P portable turbidimeter. Turbidity levels used in the experiments fell within the values observed in the Baltic Sea during the sand goby breeding season (Salonen et al., 2009; Sohel et al., 2017; Veneranta, 2007).

2.5. Behavioural experiments

In the experiment described in Chapter I, I placed the experimental aquaria outside; in the other experiments, described in Chapters II, III and IV, I placed the experimental aquaria in an outdoor lab with transparent walls and roof.

Therefore, all aquaria experienced natural light conditions. Experimental tanks always had a 4 cm layer of fine sand on the bottom offering a hiding place for the bottom-dwelling gobies and facilitating nest building. Halved clay flowerpots served as nesting substrates. Because sand goby females lay their eggs in a single layer in the ceiling of the nest cavity, I lined the ceiling of each nest with a piece of acetate sheet. Any eggs spawned in the nest would attach to the sheet.

Sand gobies are tolerant to handling, and after disturbance they will readily resume nest guarding. Therefore, the sheet could be gently removed from the nest for photographing the eggs and returned in place afterwards. The number of eggs were then counted from the digital images.

2.5.1. Mating system and sexual selection (I)

The sand goby exhibits a resource-defence mating system where males occupy and defend nest sites. Due to scarcity of nest sites in my study population, competition between males for breeding opportunities is intense. Large males are dominant and can typically monopolize multiple matings, especially when the females arrive sequentially, while some males do not get a chance to mate at all (Lindström and Seppä, 1996). In Chapter I, I sequentially introduced four ripe females to four different-sized males either in turbid or in clear water conditions to investigate the effects of algal turbidity on the mating system and opportunity for sexual selection in the sand goby.

I ran the experiment in large glass aquaria, each of which had four halved clay flowerpots as nest sites. The nest sites were large enough to accommodate eggs from four or more females. At the start of a replicate, I added four males of different size to each aquarium and allowed them to occupy and establish nests.

Once the males had constructed nests, I randomly assigned half of the tanks to

(29)

turbid and the other half to clear water treatments. I then sequentially added in total four ripe females with twelve-hour intervals to each of the aquaria containing the males. Before adding the next female, I checked all four nests and photographed the eggs if they were present. At the end of each replicate, I caught the males from their nests and measured them.

I used four different measures of sexual selection: 1) the number of mated males, that is, the number of males that received eggs, 2) the size of the male receiving the first mating, 3) the selection differential estimated as the size difference between mated and unmated males, and 4) the opportunity for sexual selection (Arnold and Wade, 1984) on males, which was estimated by the coefficient of variation (CV) in the number of eggs received by individual males. I expected that if turbidity interferes with sexual selection, it should lead to a more even distribution of mating success among males and hence, a lower opportunity for sexual selection in turbid than in clear water.

2.5.2. Paternal care and egg survival (II)

In the sand goby, males alone care for the eggs until they hatch. Increased abundance of planktonic algae may affect both the need and the costs and benefits of parental care. Increased siltation of organic material can increase the need for egg cleaning (Potts et al., 1988). Fluctuating oxygen levels due to photosynthesis and respiration of algae may either increase or decrease the need for egg fanning. Both the above can affect the costs and benefits of parental care and the amount of care provided.

In the sand goby, care is not only important for egg survival but also plays a role in male courtship (Pampoulie et al., 2004), and females have been shown to prefer males showing more care (Lindström et al., 2006). Reduced visibility due to increased algal abundance can decrease the efficacy of care as a sexual signal, which may lead to decreased care. Algal turbidity may also reduce the apparent mate encounter rate, thereby creating a situation that resembles low population density and hence low future mating opportunities. If a male per- ceives his possibilities to mate in the future as low, then he is expected to invest more in current reproductive success, i.e. to take better care of his current offspring (Roff, 1992; Williams, 1966).

In Chapter II, I experimentally explored whether increased abundance of planktonic algae affects male parental care and egg survival in the sand goby.

I allowed sand goby males to mate with one female in clear water and then care for their eggs either in clear or in turbid water conditions for six days.

(30)

I observed the time males spent fanning and attending the nest in clear, in comparison to turbid water, and monitored egg survival in both conditions. All replicates contained clear seawater in the beginning to assure similar spawning conditions. I placed one male and one ripe female into an aquarium containing clear water and allowed them to construct the nest and spawn. When spawning was first recorded, I removed the female, and photographed and counted the eggs. Then I randomly assigned each tank to either an algal turbidity treatment or a control treatment and replaced part of the clear water with either turbid or clear water correspondingly. The reason I imposed algal turbidity only after spawning is that in this study, I wanted to exclude any effects of algal turbidity on spawning. I video-recorded male parental behaviour (i.e. the time spent fanning, the time spent in the nest excluding fanning time, and the time spent away from the nest) both early and late in the brood cycle. After recording these behaviours, I photographed the eggs and counted the number of them. I compared the egg survival over the brood cycle in turbid and clear water and explored whether the time spent fanning early in the brood cycle influenced the proportion of eggs that survived until day six.

2.5.3. Male maximal mating success and care for a large brood (III)

In Chapter I of this thesis, I show that mating success (the number of mates and eggs) is more evenly distributed among males in turbid than in clear water (Järvenpää and Lindström, 2004). This could be at least partially explained if males in turbid water accepted or attracted fewer females than males in clear water, or if females avoided spawning with males guarding a large number of eggs in turbid conditions. Males could stop courting additional females for example if the cost of care for a large egg clutch increased in turbid water.

Females, in turn, could avoid mated males for instance if they found algal conditions adverse. To determine whether turbidity indeed imposes a mating cost on males and females, I examined in Chapter III whether male maximal mating success is lower in turbid than in clear water. I also investigated whether turbidity influences male parental care behaviour and the survival of eggs in a large brood. In order to do so, I introduced five randomly sized ripe females to one male with a large nest in either clear or turbid water and allowed the fish to spawn, after which I observed how many mates and eggs the male received.

This enabled me to study how algal turbidity affects male mating success.

(31)

To additionally and independently determine effects of algal turbidity on male parental behaviour and egg survival under turbid and clear water conditions, I subsequently manipulated the water clarity for part of the clear water tanks, so that approximately half of the aquaria that had clear water before spawning held turbid water after spawning (see Chapter II for details). Then I followed male parental behaviour and egg survival under turbid and clear water conditions.

In all the treatments, the male was first allowed to construct a nest in clear water. Once the nest was established, the clear water was partly replaced by either turbid water or clear water, after which the females were introduced to the male. Females were allowed 24 hours to spawn, after which they were removed and their spawning status (spawned/not spawned) was assessed based on their roundness. Eggs in the nest were photographed and counted. I then applied the second turbidity treatment: part of the tanks, in which fish had spawned in clear water were now made turbid. Males were allowed to care for their eggs for six days in turbid and clear water conditions, and their parental behaviour (the time spent fanning and the time males spent out of their nest) and egg survival was monitored both early and late in the brood cycle. Behavioural observations were made through blinds with observation holes thus ensuring minimal disturbance to the fish. I predicted that if increased abundance of planktonic algae has a negative effect on the number of eggs males are willing or able to attract, then males should acquire smaller egg clutches when spawning takes place in turbid water. If turbidity interfered with parental care, males should lose more eggs under turbid than clear water conditions.

2.5.4. Spawning under predation risk (IV)

Conspicuous reproductive activities may attract predators, and local environmental conditions can be important for reproductive decisions under predation risk. For a prey-species, decreased visibility due to algal turbidity may decrease the risk of detection by predators; on the other hand, it may also make it more difficult for the prey to spot the predator. In Chapter IV, I investigated the effects of algal turbidity and predation risk on the reproductive decisions in the sand goby. I introduced one ripe female to a male either in clear or in turbid water in the presence or in the absence of a predator and recorded the time it took for the sand gobies to spawn (latency to spawning). I ran the experiment in aquaria divided into two compartments using a clear plastic di- vider with small holes to allow transfer of water and chemical cues. The back

(32)

compartment contained a clump of artificial vegetation and, in the treatments with a predator, was used for housing a perch. The front compartment had a nest site and came to house the gobies.

In the beginning of each replicate, I introduced a male sand goby into the front compartment and allowed him 24 hours to construct a nest. Then, I randomly assigned each aquarium to a treatment. Correspondingly, I replaced part of the water with either turbid or clear water and I either introduced a perch into the back compartment or not. Next, I introduced a ripe female in the male compartment and allowed 24 h for spawning. Within these 24 hours, I then observed when spawning took place by regularly inspecting the nest for eggs with a LED light. I expected that, in the presence of the predator, latency to spawning would be greater in clear water where sand gobies could rely on more cues when assessing the risk of predation in comparison to turbid water, where use of visual information was limited.

(33)

3. MAIN RESULTS &

DISCUSSION

3.1. The sand goby mating system breaks down under algal turbidity

In Chapter I (Järvenpää and Lindström, 2004), I explored the potential effects of increased algal abundance on the sand goby mating system, and sequentially introduced four females to four nest-holding males either in clear water or in water made turbid by planktonic algae. I got to witness a break-down of the sand goby mating system under algal turbidity. Matings were more evenly distributed among males and the opportunity for sexual selection was lower in turbid than in clear water (Chapter I, Fig. 2). Also, in turbid conditions, large size did not bring males the same reproductive benefit as it did in clear water;

mating success was less skewed towards large males and the first mated male was smaller in turbid than clear water (Chapter I, Fig. 3).

Figure 2. Matings were more evenly distributed among males in turbid than clear water, as shown by the number of nests containing eggs (left image) and the coefficient of variation in male mating success based on number of eggs (right image). Error bars indicate + 1 S.E. Modified from Järvenpää and Lind- ström (2004) with the kind permission from the Royal Society.

(34)

Such break-down of the sand goby mating system in turbid conditions could, for example, be due to changes in male-male competition, turbidity in- terfering with sexual signalling, or changes in the preferences of females or their ability to express them.

Algal turbidity could have affected competition among males either by decreasing encounter rates between males and/or by restricting the possibilities for one male to interfere with the courtship of another male (Wong et al., 2007).

In the sand goby, large males are not necessarily preferred by females, but under clear water conditions large males are more successful in mating competition, probably due to their higher competitive ability (Lehtonen and Lindström, 2009). If decreased visibility in the turbid water conditions constrained the possibilities of large males to interfere with the mate attraction attempts of other males, this could have led to the decreased mating benefit for large males, as well as to the more even distribution of matings among males (Kangas and Lindström, 2001). Indeed, algal turbidity has previously been shown to lessen aggressive male-male interactions in three-spine sticklebacks (Candolin et al., 2008), thereby allowing for dishonest signalling by less fit males (Candolin, 2000; Wong et al., 2007).

Female preferences, or their ability to express these, could change, for instance, if turbidity hampered mate detection or interfered with signals used in mate attraction and mate choice. For example, the above-mentioned dishonest signalling of male sticklebacks may interfere with the ability of females to evaluate the condition of males reliably. Restricted visibility is known to constrain the ability of stickleback females to compare males using visual cues and to prolong the time females spend evaluating potential mates (Candolin et al.,

Figure 3. Mated males (grey bars) were on average larger than unmated males (white bars) in clear but not in turbid water. Error bars indicate + 1 S.E. Rep- licated from Järvenpää and Lindström (2004) with the kind permission from the Royal Society.

(35)

2007; Candolin et al., 2016; Engström-Öst and Candolin, 2007). In broad- nose pipefish, visual conditions mimicking algal turbidity decrease the time spent associating with potential mates and hinder the preference for large mate size (Sundin et al., 2010). In the same species, algal turbidity has also been shown to hamper ornament perception leading to randomized mate choice, while in Lake Victoria cichlids turbidity masks species-specific colour signals used in assortative mating, resulting in hybridization between the species (See- hausen et al., 1997; Sundin et al., 2016). Given the above, it is possible that turbidity also interferes with the ability of sand goby females to evaluate males, making the females less selective.

Mating decisions can also be sensitive to the number and the sex of con- specifics (Heubel, 2018; Heubel et al., 2008; Kokko and Rankin, 2006; Leht- onen and Lindström, 2009). Sand goby females have been shown to adjust their mating preferences according to the competitive situation, such that they prefer large males under intense male-male competition (Lehtonen and Lind- ström, 2009). Under low visibility in turbid water, females may perceive intra- sexual competition differently than in clear water, resulting in different mate preferences and dissimilar distribution of matings among males in turbid and clear water conditions.

A more even distribution of matings among males could also be explained if males in turbid water accepted or attracted fewer females to spawn in their nest than males in clear water do, or if females avoided spawning with males guarding a large number of eggs under turbid conditions. Males could stop attracting additional females, for example, if the costs of courting further females and/or taking care of a larger egg clutch exceeded the benefits thereof, or if the trade-off between courtship and care was stronger in turbid water, as suggested in Chapter II (Järvenpää and Lindström, 2011). Females, in turn, could avoid mated males in turbid water, for instance, if increased algal abundance required excess parental effort from the males due to increased siltation, altered oxygen levels or for some other reason, possibly diminishing the ability of males to successfully care for a large clutch (Lissåker et al., 2003; Potts et al., 1988;

Reynolds and Jones, 1999).