ABSTRACT
Artificial light at night (ALAN) is a growing environmental problem influencing the fitness of individuals through effects on their physiology and behaviour. Research on animals has primarily focused on effects on behaviour during the night, whereas less is known about effects transferred to daytime. Here, we investigated in the lab the impact of ALAN on the mating behaviour of an ecologically important freshwater amphipod, Gammarus pulex, during both daytime and nighttime. We manipulated the presence of ALAN and the intensity of male–male competition for access to females, and found the impact of ALAN on mating activity to be stronger during daytime than during nighttime, independent of male–male competition. At night, ALAN only reduced the probability of precopula pair formation, while during the daytime, it both decreased general activity and increased the probability of pair separation after pair formation. Thus, ALAN reduced mating success in G. pulex not only directly, through effects on mating behaviour at night, but also indirectly through a carry-over effect on daytime activity and the ability to remain in precopula. These results emphasise the importance of considering delayed effects of ALAN on organisms, including daytime activities that can be more important fitness determinants than nighttime activities.
INTRODUCTION
Artificial light at night (ALAN), which is a form of light pollution, is changing the natural light regime (Gaston, 2018). The lit area is currently growing rapidly, accelerated by the increased use of light emitting diodes (LEDs) to reduce energy consumption (Kyba et al., 2023; Sánchez de Miguel et al., 2021; Schulte-Römer et al., 2019). An increasing number of studies is finding artificial light to influence the fitness of individuals through effects on their physiology, behaviour and life history (Gaston et al., 2013; Hölker et al., 2023). These changes in individual fitness can, in turn, alter the spatial and temporal distribution of populations and, hence, have ecological consequences by altering species interactions, biochemical cycles and ecosystem stability (Candolin, 2024; Hirt et al., 2023; Sanders et al., 2021).
Research on the impact of ALAN on the behaviour of animals has concentrated on activities during the night, including crepuscular times (Aulsebrook et al., 2020; Davies et al., 2023; Owens and Lewis, 2018). These investigations show that ALAN often alters the activity of animals, such as their movements, foraging and reproductive behaviour (Elgert et al., 2020; Manfrin et al., 2017; Szaz et al., 2015). However, ALAN could also influence activities during the daytime, especially through effects on physiology (Grunst and Grunst, 2023; Ouyang et al., 2018). In humans, an increasing number of studies has found ALAN to influence daytime activities through negative effects on hormones, especially on melatonin, and thereby on the quantity and quality of rest (Cao et al., 2023; Navara and Nelson, 2007). Whether similar effects occur in other organisms has so far received little attention (Kurvers et al., 2018). Nocturnal species are often found to be negatively affected through reduced activity and increased susceptibility to predators, while diurnal animals are expected to benefit from ALAN in terms of more time for foraging and mating, and improved ability to detect predators (Cremer et al., 2022; Gaston et al., 2013; Minnaar et al., 2015; Nelson et al., 2021; Sanders et al., 2021). However, the lack of rest, increased time active and altered hormonal levels could have negative effects on daytime activities (Grubisic et al., 2019; Ouyang et al., 2018).
Another limitation of previous ALAN research is that it has primarily focused on terrestrial species, whereas less is known about effects on aquatic species (Sanders et al., 2021). Especially little is known about effects on aquatic invertebrates (Davies et al., 2014; Ganguly and Candolin, 2023; Hölker et al., 2023; Marangoni et al., 2022). Yet, these species play key roles in ecosystems by providing food for higher trophic levels and by decomposing organic material (Eisenhauer and Hines, 2021; Murkin and Wrubleski, 1988). Changes in their behaviour and distribution could have far-reaching consequences for ecosystem function (Hölker et al., 2023; Tidau et al., 2022). Invertebrates in urban and suburban streams and ponds are especially often exposed to ALAN, given the shallowness of the water and proximity to lighting systems, and because of sky glow, i.e. when light that radiates upwards from urban areas is diverted back to Earth through scattering and lightens up areas far from the light source (Kyba et al., 2015).
Amphipods (Gammarus spp.) are common detritivores in streams and ponds in the Holarctic region. They are active during both daytime and nighttime, but avoid artificial light by moving away from it (Kohler et al., 2018; van den Berg et al., 2023). A few recent studies have shown that ALAN increases their vulnerability to fish predators by reducing their use of shelters (Czarnecka et al., 2019), and suppresses their body growth and activity (Czarnecka et al., 2022, 2021). However, the impact on their mating behaviour is unknown, despite mating success being a key determinant of fitness and regulated by physiological state and hormones, which are known to be sensitive to light conditions in a range of invertebrates (Ganguly and Candolin, 2023; Kivelä et al., 2023; Owens and Lewis, 2022; Zapata et al., 2019). During the reproductive season, gammarid males search for receptive females using visual and chemical cues that reveal receptive status (Breithaupt and Thiel, 2011). They clasp encountered females to form a precopulatory pair, which prevents other males from mating with the female (Bollache and Cezilly, 2004; Iltis et al., 2017; Plaistow et al., 2003; Sutcliffe, 1993). A pair may stay together for several days until the female moults and mating can take place (Cornet et al., 2012; Sutcliffe, 1993). Longer mate guarding is costly in terms of higher predation risk and increased risk of interruption from competing males (Plaistow et al., 2003). Thus, changes in any of these steps in the mating process because of light pollution could influence the reproductive success of individuals and, hence, the dynamics of the population and the ecological function of the species (Candolin, 2019).
We investigated in the lab whether ALAN influences the mating behaviour of the common European freshwater amphipod G. pulex (Chaumot et al., 2015), and whether exposure to light at night could have delayed effects on daytime mating activity. We exposed individuals to ALAN for 1 week and then observed their behaviour during both daytime and nighttime. We considered the impact of male competition by manipulating the number of males, given that male competition could influence mating activity and the impact of ALAN.
MATERIALS AND METHODS
Collection and maintenance
We collected Gammarus pulex (Linnaeus 1758) through kick-sampling from an urban stream, Longinoja, in the Helsinki area (60°14′44.82″N, 25°0′39.40″E) in June and July 2022. The sampled site has 100% canopy cover in the summer and the light level at the water surface during cloudy nights without moonlight is approximately 5.05×1011 photons cm–2 s−1 (0.4 lx; measured using a FLAME-S spectrometer, Ocean Optics, Dunedin, FL, USA). We selected individuals with no signs of infection with acanthocephalans (i.e. no orange spots under the cuticle), the main parasite influencing fitness in the species and which can alter its behaviour (Durieux et al., 2012). We transported the gammarids to the University of Helsinki, males and females in separate tanks. The individuals were first kept in stream water for 24 h, aerated with airstones, and then moved to aerated tanks with 5 l of mixed stream and aerated tap water. The bottom of the tanks was covered by gravel and dried maple leaves (Acer platanoides) for food and shelter. The temperature of the water was 14°C, and the light:dark cycle followed the natural rhythm for the latitude, 20 h:4 h, using programmable LEDs above the tanks (Aquarius 120 plus, Aqua Medic, Germany) (Fig. S1). Daylight was set to gradually increase from 03:00 to 07:00 h and gradually decrease from 19:00 to 23:00 h, to imitate natural conditions. We renewed one-third of the water every 2 days to reduce waste accumulation and produce optimal conditions for the species.
Light treatment
Prior to recording mating behaviour, we exposed the individuals to one of two nightlight treatments for 1 week: control and ALAN, using ∼640 individuals in total. We selected individuals of similar body length (males:15.6±1.5 mm, females: 12.2±1.1 mm) to prevent size-assortative mating (Bollache et al., 2000; Franceschi et al., 2010), and placed 20 individuals into each of eight tanks with 5 l of water (35×40 cm, water depth 4 cm), with males and females in different tanks. This resulted in four control and four ALAN tanks. Each tank was exposed to the same daylight conditions as during maintenance, but individuals in the ALAN treatment were exposed to light at night with an intensity of 1.9×1012 photons cm–2 s−1 (2.2 lx, FLAME-S) at 449 nm using LEDs above the tanks. This corresponds to more strongly lit areas than the individuals have experienced in the field during nights. Individuals in the control experienced no measurable artificial light during nights (<0.1 lx, FLAME-S). Owing to limited space, we repeated this procedure four times during June–July 2022, each time having four control and four ALAN treatments and using new individuals. To avoid the spillage of light between treatments, which were performed in the same room, we used blackout fabrics.
Male competition and behavioural recordings
After 1 week of ALAN or control treatment, we placed (i) one male and one female (no male competition), or (ii) two males and one female (male competition) into a tank (35×40 cm, water depth 4 cm) containing 5 l of water under the same light conditions as during the ALAN or control treatment. Prior experimentation had shown that such conditions allow normal mating behaviour in the species. The bottom was covered by gravel, and a dried leaf was added as a shelter (5×5 cm) to reflect the natural environment with hiding places. Individuals were randomly selected and the trials were performed either during daytime (between 11:00 and 14:00 h) or during nighttime (between 23:00 and 02:00 h). Each individual was first enclosed in a 3.5 cm metal mesh sphere (a tea ball) (Arnal et al., 2015) for 10 min to acclimatise to the conditions (Durieux et al., 2012), with the male(s) and female placed at opposite corners of the tank. After releasing the individuals, we recorded their behaviour for 10 min, as this was found to be an appropriate time period in pilot observations, and as 10 min has been used in earlier research on the behaviour of the species (Truhlar and Aldridge, 2015). During nights, we used an infrared-sensitive video camera (SIONYX Aurora PRO), combined with an infrared light source (940 nm) installed above the tanks. Gammarids express normal behaviour under infrared light (Czarnecka et al., 2021). We performed eight treatment combinations (light×male competition×daytime/nighttime observation), each repeated at least 18 times using different individuals (Fig. 1).
Video analysis
We analysed the behaviour of G. pulex from the video using the software BORIS (Behavioral Observation Research Interactive Software, v. 7.12.2, http://www.boris.unito.it/) (Friard and Gamba, 2016). We recorded for each individual the total time they were actively moving (given as percent of total experimental time), their use of the shelter, contacts between males and females, and precopula formation and separation, as detailed in Table 1. We considered the sex of the individuals in the analyses, as research on other gammarids found sex to influence behaviour (Sornom et al., 2010). Details on the video analysis and the use of BORIS are given in the Supplementary Materials and Methods. Owing to poor visibility, behaviour could not be reliably analysed for some nighttime recordings, and these replicates were discarded (see final sample sizes in Fig. 1).
Statistics analysis
We used generalised linear mixed models (GLMMs) to analyse the impact of ALAN and male competition on behaviour, separately for daytime and nighttime recordings. The presence or absence of ALAN and male competition were added as fixed factors. The person that recorded the behaviours (using the software BORIS) was added as random factor. To consider that the behaviour of individuals in a tank are dependent, ‘tank’ was added as a random factor when analysing the behaviour of multiple individuals. The sex of an individual was included as a covariate in models, but removed when no significant effect was detected and the removal did not affect the significance of the main factors. To select the appropriate error distribution, we inspected the structure of the data using the function ‘str()’, histograms and Q-Q plots. We used GLMM with a beta distribution to analyse the effects of the fixed factors on total time actively moving, and GLMM with a gamma distribution to analyse the effects on time spent in the shelter and the duration of contacts between males and females. To analyse the effects of the fixed factors on number of contacts between males and females, between two males and number of visits to the shelter, we used GLMM with zero-inflated Poisson distribution (given the large number of zero values). We used GLMM with binomial distribution to analyse the probability of pair formation and pair separation. We included all factors and interactions between fixed factors in the initial models, and iteratively removed nonsignificant interaction terms (P>0.1) when this did not influence the significance of the main factors. The analyses were performed using R 4.2.2 (https://www.r-project.org/) and the package ‘glmmTMB’ (Brooks et al., 2017).
RESULTS
Behaviour during nighttime
ALAN had no significant impact on time actively moving (z=−1.432, P=0.152), but male competition reduced activity (z=−3.820, P<0.001) (Fig. 2A). Females were less active than males, independent of ALAN and male competition (z=2.374, P=0.018). Neither ALAN nor male competition influenced number of visits to the shelter (ALAN: z=0.514, P=0.607; male competition: z=1.850, P=0.064) (Fig. 3A), or the duration of the visits (ALAN: z=−0.301, P=0.763; male competition: z=1.376, P=0.169) (Fig. 3C). Females spent more time than males in the shelter, independent of ALAN or male competition (z=2.166, P=0.030). ALAN did not significantly influence male–male interactions (GLMM, z=−1.746, P=0.081) and their duration (z=−1.750, P=0.080) (Fig. S2).
ALAN did not influence number of contacts between the male(s) and female (z=−1.180, P=0.238), but male competition increased number of contacts (z=3.115, P=0.002) (Fig. 4A). Neither ALAN nor male competition influenced the duration of contacts (ALAN: z=−0.234, P=0.815; male competition: z=−0.832, P=0.405) (Fig. 4C). ALAN reduced the probability of pair formation (z=2.302, P=0.021), while male competition increased the probability (z=2.008, P=0.044) (Fig. 5A). ALAN also reduced the probability of pair formation after contact (z=2.152, P=0.031) (Fig. S2), but male competition showed no significant effect (z=1.118, P=0.263). Neither factor influenced the probability of pair separation (ALAN: z=0.796, P=0.426; male competition: z=0.930, P=0.353) (Fig. 5A).
Behaviour during daytime
ALAN reduced the time actively moving (z=3.963, P<0.001), and the effect was weaker in the presence of male competition (z=2.589, P=0.001) (Fig. 2B). ALAN also reduced visits to the shelter in the absence of male competition (z=2.112, P=0.035), but increased visits in the presence of male competition (z=−2.240, P=0.025) (Fig. 3B). ALAN increased time spent in the shelter (z=−2.726, P=0.006), whereas male competition had no significant effect (z=1.376, P=0.169) (Fig. 3D). ALAN did not influence the number of male–male interactions (z=−1.167, P=0.243) or their duration (z=−0.394, P=0.694) (Fig. S2).
Both ALAN and male competition reduced the probability of contacts between the male(s) and female (ALAN: z=3.320, P<0.001; male competition: z=3.371, P<0.001) (Fig. 4B). Neither factor affected the duration of contacts (ALAN: z=−1.117, P=0.264; male competition: z=0.202, P=0.840) (Fig. 4D) or the probability of pair formation (ALAN: z=−1.003, P=0.316; male competition: z=1.003, P=0.316), but ALAN increased the probability of pair separation (z=–2.297, P=0.022), and so did male competition (z=2.297, P=0.022) (Fig. 5B). Neither ALAN (z=−1.339, P=0.181) nor male competition (z=−0.897, P=0.370) affected the probability of pair formation after female–male contacts (Fig. S2).
See Tables S1–S4 for full details of the statistical results.
DISCUSSION
The results show that G. pulex is active during both daytime and nighttime, and that exposure to ALAN reduces activity and shelter use during daytime, for both males and females, but not during nighttime. Thus, ALAN has a carry-over effect from the nighttime to daytime activity. This differs from most earlier studies that found ALAN to have a direct effect on animal activity during nighttime (e.g. Luarte et al., 2016; Navarro-Barranco and Hughes, 2015; Underwood et al., 2017). However, studies on ALAN have generally not considered effects on daytime activity, which suggest that full effects of ALAN might often not be recorded and that the impact of ALAN could be larger than generally reported.
Our results further show that male competition influences the impact of ALAN on daytime activity; an increase in the abundance of males from one to two counteracted the reducing impact of ALAN on daytime activity. The counteracting effect was probably due to ALAN increasing male–male interactions during daytime. Interestingly, ALAN did not increase male–male interactions during nighttime, further supporting a delayed effect of exposure to light at night on daytime activity.
Exposure to ALAN also reduced contacts between males and females during daytime, independent of the number of males present. It further reduced the duration of the contacts when two males were present, i.e. under male–male competition, while prolonging the duration when only one male was present. The reduction in the duration under male competition was probably due to ALAN increasing the frequency of male–male interactions and, hence, mating interruptions. ALAN did not influence the probability of pair formation during daytime, but increased the probability of pair separation. During nighttime, ALAN reduced contacts between males and females when only one male was present, and increased contacts when two males were present. The increase in contacts was probably due to males investing more in securing a mate when competitors for mates were visible under the artificial light. Yet, ALAN reduced the probability of precopula pair formation during nighttime, independent of the number of males present. This could be because the light enhances the ability of females to detect approaching males and avoid their mating attempts (Sutcliffe, 1993), or because the light induces stress in males that reduces their ability to attach and hold on to females.
These results show that exposure to nocturnal lighting for 1 week reduces mating success during both nighttime and daytime. During nighttime, it reduces precopula pair formation, and during daytime, it increases the probability of pair separation. These effects, combined with the negative effect of ALAN on daytime activity, suggest that nocturnal light either influences the ability of individuals to engage in mating activities or their motivation to do so. Although we cannot determine the mechanism in the present study, earlier research suggests that it is likely related to physiological changes. Research on other invertebrates shows that ALAN influences physiological processes such as metabolic rate, stress, hormonal levels and gene expression, which in turn can influence sexual behaviour (Ganguly and Candolin, 2023; Levy et al., 2022; McLay et al., 2018; Owens and Lewis, 2022; van Geffen et al., 2015). For example, exposure of the fruit fly Drosophila melanogaster to 10 lx of ALAN reduces their concentration of reactive oxygen species (ROS) in ovaries and prolongs their courtship times (McLay et al., 2018), whereas exposure of the cricket Gryllus bimaculatus to ALAN upregulates several of their circadian clock genes that regulate physiological and behavioural rhythms (Levy et al., 2022).
The result of a negative effect of ALAN on gammarid mating behaviour indicates that ALAN can reduce the reproductive success of individuals. This could, in turn, have negative impacts on the population dynamics of the species and, hence, on its ecological functions. Gammarus pulex is a key species in many aquatic ecosystems, and contributes to the recycling of nutrients and energy and serves as an important food source for higher ecological levels (Macneil et al., 1997). Thus, changes in its mating behaviour could have ecological consequences. However, to assess the ecological consequences of altered mating behaviour, the longer-term effects need to be assessed, i.e. effects on lifetime reproductive success, population dynamics and ecological functions. A few studies on other taxa have found ALAN to have longer-term effects on fitness (Botha et al., 2017; McLay et al., 2018; Schligler et al., 2021; Vardi-Naim et al., 2022), but little research has so far been devoted to the topic.
The effect of ALAN on population dynamics and ecological function of gammarids could further depend on how the light influences species interactions, such as predation risk, food availability and parasite abundance, as well as interactions with other environmental disturbances. Reproductive behaviours are generally sensitive to environmental conditions, such as predation risk, food availability and temperature, and changes in these could modify the ultimate impact of ALAN on reproductive success (Candolin and Wong, 2019). Thus, to determine the ultimate impact of ALAN on the population and the consequences for the ecosystem, future research needs to consider the complexity of population dynamics, species interactions and ecosystem processes.
To conclude, our study fills a gap in ALAN research in that it considers carry-over effects of ALAN on daytime activities. The results show that carry-over effects are profound in a gammarid and even larger than nighttime effects. Thus, measuring only nighttime effects would have underestimated the effect of ALAN on mating behaviour. These results highlight the importance of considering both direct effects of ALAN on nighttime activity, and indirect carry-over effects on daytime activity. To gain a holistic view of the impacts of light pollution on organisms, we need to consider all effects, also delayed effects that in the end could have more profound effects on populations and their ecological functions than the short-term direct effects.
Acknowledgements
We are grateful to Yaren Gungor for her help in sampling animal and video recording during experiments, and to Alexandra Seger, Joep Jansen and Kaisa Oksanen for their assistance in analysing behavioural traits with the software BORIS.
Footnotes
Funding
This research was funded by the Swedish Cultural Foundation in Finland (to U.C.), an EDUFI fellowship (to A.G.) and the China Scholarship Council (to Y.H.).
Author contributions
Conceptualization: Y.H., A.G.; Methodology: Y.H., L.Q.; Software: Y.H., A.G., S.L., L.Q., C.S.; Validation: Y.H.; Formal analysis: Y.H., S.L., K.Z.; Investigation: Y.H., A.G., S.L., L.Q., C.S., K.Z.; Resources: Y.H.; Data curation: Y.H.; Writing - original draft: Y.H.; Writing - review & editing: Y.H., U.C.; Visualization: Y.H.; Supervision: U.C.; Project administration: U.C.; Funding acquisition: U.C.
References
Competing interests
The authors declare no competing or financial interests.