Introduction

Over the last century, in many parts of the world wild animals have become increasingly likely to encounter artificial light during their lifetime. Although artificial light has spread at an increasing rate over the last 150 years, reaching all but the most isolated ecosystems on a global scale, it is relatively new on an evolutionary timescale. In its effects artificial light is not limited to nocturnal animals (de Jong et al., 2015) but animals with pronounced night-time activity are more likely to be most severely affected. Strong effects of artificial light have been described for numerous species (Rich and Longcore, 2005), including moths (van Geffen et al., 2014), bats (Stone et al., 2009), marine turtles (Kamrowski et al., 2012), mice (Bird et al., 2004), and amphibians (Perry et al., 2008; Wise, 2007). The knowledge on the impact of artificial light is very limited for many groups of animals and many aspects remain unclear (Hölker et al., 2010) despite a recent increase of interest in this field (Gaston et al., 2015; Spoelstra et al., 2015). Anurans have received little attention, but are likely to be affected as most species are nocturnal, have very sensitive eyesight for hunting and depend on camouflage for safety. Artificial light might thus be an additional pressure on this already endangered group of organisms (Mendelson et al., 2006).

Many amphibians migrate in spring from their terrestrial winter habitats to water bodies to reproduce. This includes the anurans common toad (Bufo bufo) and other toads (Bufo sp.), true frogs (e.g. Rana sp. and Pelophylax sp.) and treefrogs (e.g. Hyla sp.), and urodela with both salamanders (e.g. Ambystoma sp.) and newts (e.g. Triturus sp.) represented (Sinsch, 1990, Russell et al., 2005; Semlitsch, 2008). In landscapes fragmented by roads this is one of the most dangerous periods in their adult life. The number of animals killed by traffic during migration can be substantial, with 23-66% of spawners being killed during spring migration annually (van Gelder, 1973; Elzanowski et al., 2009). This can be a significant threat to the long term survival of amphibian populations especially in densely populated areas (Puky, 2005; Schmidt and Zumbach, 2008; Beebee, 2013). All these studies have however been done on roads without street lighting, while more and more roads are becoming illuminated, especially in peri-urban areas. A number of amphibian species are known to forage or aggregate at light sources. Juvenile common toads aggregate under streetlights (Baker, 1990) and similar behaviour is documented in green toads, Bufo viridis (Freisling, 1948), and juvenile dispersing crested newts, Triturus cristatus, in autumn (Creemers, 1992). They possibly take advantage of the increased abundance of prey underneath street lights or the increased visibility of the prey present. However, if they stay on the road longer adjoining the streetlight, this behaviour may increase the risk of being struck by traffic and detection and capture by predators. The negative effects of habitat fragmentation by roads might thus be increased by artificial illumination and the numbers reported by Beebee (2013) might be an underestimation. Although it is often hypothesized that anthropogenic disturbances can work synergistically (Sala et al., 2000) this has rarely been tested. As artificial illumination (i.e. street lights) typically co-occurs with habitat fragmentation (i.e. roads), they are very likely to interact.

The behaviour of toads in response to street lights during the critical migration period is potentially quite different from the post spawning behaviour (Heusser, 1968; Sinsch, 1988) and from that of juvenile toads (Baker, 1990). Toads migrating to reproduction sites rarely eat and are preoccupied by reaching the water and locating potential mates. Furthermore, they are more strictly nocturnal, preferring to migrate in dark new moon nights (Kusano et al., 2015). It is therefore unclear whether they respond similar to juvenile toads and aggregate under streetlights or behave all together differently. If artificial lights attract toads the placement of streetlights could have a substantial effect on the number off toads killed during migration. Furthermore the efficacy of tunnels for toads to safely cross roads might be influenced by streetlights.

A possibility to mitigate effects of streetlights on migrating toads might be through spectral adjustment. The adjustment of spectral composition can reduce the number of prey attracted to artificial light sources (van Langevelde et al., 2011; van Grunsven et al., 2014). How amphibians respond to light depends on the spectral composition of the light. Anurans are most strongly attracted to blue light in most cases (Donner and Reuter, 1962, 1976; Muntz, 1962). This includes Bufonidae (Hailman and Jaeger, 1974) and therefore it is likely that light sources that emit little short wavelength light affect migrating toads less than light sources rich in short wavelength light.

The amount of illumination is increasing (Hölker et al., 2010) but the type of light is also changing. More and more traditional light sources, such as low pressure sodium lights, are being replaced by LED (Tan et al., 2012) and this is thought to increase the impact of light (Gaston et al., 2012; Davies et al., 2013). However the use of LED allows for spectral adjustment to reduce ecological effects (Longcore et al., 2015; Spoelstra et al., 2015). To realize this potential it is essential to know the role of spectral composition.

To test whether migrating toads (Bufo bufo) are affected by street lighting and whether mitigation through spectral adjustment is possible we experimentally illuminated locations where toads are known to cross roads during migration. By using barriers and bucket traps we assessed where toads attempt to cross a road and how different spectra affected this.

View Full Size

A schematic representation of the setup with barriers, buckets (pale green circles) and four light treatments (red light, green light, no light and white light). The arrow indicates the direction of migration. The order of the treatments was randomised over the nights and the order depicted here is merely an example. This figure is published in colour in the online version.

Citation: Amphibia-Reptilia 38, 1 (2017) ; 10.1163/15685381-00003081

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Spectral composition of the used lamps in milliwatts per nm. This figure is published in colour in the online version.

Citation: Amphibia-Reptilia 38, 1 (2017) ; 10.1163/15685381-00003081

• Download Figure
• Download figure as PowerPoint slide

Materials and methods

In the Netherlands volunteers help migrating common toads (Bufo bufo) to cross roads. Toads are caught by digging in a barrier, consisting of a screen parallel to the road in the road verge, and buckets at intervals so that toads will walk along this barrier and fall in the bucket (fig. 1). The volunteers collect the toads in the early night and next morning and transport them to the other side of the road where they can safely continue their migration. The Dutch toad patrols record their daily and annual numbers of migrating amphibians, these numbers are presented in www.padden.nu, a portal with information for toad patrollers. We selected four locations where large numbers of toads are transferred yearly and where in previous years the distribution was rather homogeneous, the vegetation is similar along this part and the waterbody is not so close as to cause convergence (see online Supplement S1). Here we constructed additional 1.5 meter long barriers perpendicular to the road effectively creating 25 meter long sections. Buckets are placed in the middle and the corners of these sections (fig. 1). In the middle of each section, 1.5 meter from the barrier a 1.8 meter wooden pole is placed on which a streetlight can be affixed. We used Philips Fortimo white, Clearsky green and Clearfield red lamps (Philips, Amsterdam, the Netherlands) for the experimental lighting. The spectral composition of the light for the three different treatments is presented in fig. 2. Briefly, all white, green and red lamps emit full spectrum light; green lamps have an increased blue and reduced red light emission, and red lamps have an increased red and strongly reduced blue emission (Spoelstra et al., 2015). All light colours have a negligible UV emission. Illumination directly underneath the lamp was mean ± SE = 52 ± 2.3 lux, at the barrier next to the lamp 10 ± 0.4 lux, this is similar to illumination level on secondary roads. At the barriers separating the treatments light levels were 0.03 ± 0.01 lux. Lamps of different colours were adjusted to have identical luminous flux. Lights were run from a generator placed 40 meter away from the experiment and which had acoustic shielding to avoid noise to interfere with the experiments. The generators were always in the same locations. The position of the four treatments was randomised over the nights to avoid confounding of preferential migration routes of the toads as a result of e.g. local vegetation structure or the effect of the sound of the generator (text S1).

Migration of toads is mainly episodic and concentrated on a number of evenings in early spring when climatic conditions are favourable, temperature above 6°C and an air humidity of at least 75%. This is primarily in February or March but differs between years. Most amphibians migrate in the beginning of the night when it is relatively warm. On evenings when a large number of migrating toads was expected, the lamps were placed on the poles and connected before sunset and switched on at sunset. After two hours the amphibians in all buckets were counted and released on the other side of the road (thus avoiding multiple captures of the same individual), and the lights and generator were removed. Toads in amplexus are counted as 2 individuals as we do not know whether they entered the bucket in amplexus or not. This was performed at Kootwijk (N 52°09′38, E 5°43′05) in 2012, Amerongen (N 51°59′30, E 5°28′48) in 2013 and 2014 and Bloemendaal (N 52°22′33, E 4°35′29) in 2014 (text S1). Observations were made for a total of 24 nights. In 2014 the number of amphibians in the buckets was also recorded in the morning, representing the animals that walked into the buckets after the lights were switched off, this was done for 11 nights.

As all data consists of counts they are analysed using generalized linear mixed models for Poisson distributions, GLMer, lme4 (Bates, 2010) with an observation level random factor to account for over dispersion. Observations were included for individual buckets this allows us to include the position of the buckets. Night and section nested in location are included as random factor. Treatments, the position of the bucket (corner or middle of a section) and the three light colours and dark control, are considered as fixed factor. The interaction between bucket position and light treatment is included in the initial model but removed from the model if insignificant when testing the main effects of position and light treatment. The same analysis is used for the data collected in mornings where the colour of the previous evening is included as factor (see online Supplement S2).

Results

The number of common toads migrating differed substantially between nights, ranging from nights with hardly any migration resulting in no caught animals (6 nights), to nights with a large number of migrating toads (18 nights), with up to 91 animals caught in one location in one night. The number of toads caught per evening was higher in Amerongen (on average 24) then in Kootwijk (6.2) and Bloemendaal (2.0). In Amerongen the number was higher in 2013 (on average 32.2 toads) then in 2014 (13.2). In the morning counts in 2014 more toads were found in Amerongen (32.2 on average) then in Bloemendaal (7.7).

There was no interaction between position of the bucket (corner or middle) and light treatment (Wald ${\mathit{\chi }}_{\mathrm{df}=3}^2=5.17$, $\mathit{p}=0.16$) and this interaction was subsequently removed from the model. The number of toads caught was affected by the light treatment (Wald ${\mathit{\chi }}_{\mathrm{df}=3}^2=12.62$, $\mathit{p}\mathrm{<}0.01$). Fewer toads were caught under white and green light than in the dark control ($\mathit{p}\mathrm{<}0.01$, fig. 3). The number of toads caught near red light did not differ significantly from the dark control but did differ from the white ($\mathit{p}=0.02$). The difference between the green and red treatment was no longer significant after Benjamini-Hochberg correction ($\mathit{p}=0.04$, corrected $\mathit{\alpha }=0.033$) (Supplement S2). Position of the bucket was near significant (Wald ${\mathit{\chi }}_{\mathrm{df}=1}^2=3.13$, $\mathit{p}=0.077$) with on average more toads in the middle than in corner buckets (1.47 vs. 0.87 per night).

View Full Size

The mean number of toads caught per night, per bucket is given for every light treatment. Data are presented both for animals caught in the evening with the illumination turned on and the animals caught in the morning, i.e. after the light was turned off. Letters indicate significant contrasts in GLMMs and whiskers are standard deviations based on 10 000 sample bootstrapping.

Citation: Amphibia-Reptilia 38, 1 (2017) ; 10.1163/15685381-00003081

• Download Figure
• Download figure as PowerPoint slide

For two locations in 2014 we also collected data on the number of toads caught in the morning after removing the light. There was no interaction between light treatment and position of the buckets (Wald ${\mathit{\chi }}_{\mathrm{df}=3}^2=1.57$, $\mathit{p}=0.66$). Nor was there a main effect of either the light treatment the evening before (Wald ${\mathit{\chi }}_{\mathrm{df}=3}^2=2.43$, $\mathit{p}=0.49$) or position of the buckets (Wald ${\mathit{\chi }}_{\mathrm{df}=1}^2=0.29$, $\mathit{p}=0.59$).

Discussion

The response of toads to artificial light during migration contrasts with the behaviour in summer and autumn. Where toads are attracted to streetlights and aggregate near light sources in summer (Baker, 1990) they avoid light during migration. Most toads tried to cross the road at the dark control and least at white and green light, with red being intermediate. There are several possible non-exclusive explanations for this difference. The toads might be truly phototactic in summer, attracted as a direct response to light. Additionally the main attractant in the summer may not be the light itself but toads might be merely attracted by prey that itself has been attracted, and made more visible, by the light (Frank, 1988; Eisenbeis, 2006). During spring migration toads are in a different physiological state than in summer, driven by reproduction and not feeding and therefore have a different response to light.

There was a trend to more toads being caught in middle than corner buckets. The corner buckets can only be reached from one side while the middle buckets can be entered from both sides, therefore this can be understood from random movement of the toads after reaching the barrier. There was no interaction between the light treatment and the position, the toads did thus not strongly avoid moving to the more strongly illuminated middle part of the section once near the barrier.

The different light colours were not separated by dark controls, and therefore we can not quantitatively asses the effect of the different light sources. However, the number of toads in the dark section was not affected by the position (unpublished results) and the difference between the three light colours is as expected, with red light having the least impact. The red light has very little output in the blue and green part of the spectrum, where the eyes of toads are most sensitive. Therefore, the brightness of this light, as experienced by the toads, is much less than that of the other two light sources. The eyes of toads do contain a relatively large numbers of so-called “red” rods (rods in the retina that are sensitive to red light), however, these have their main spectral sensitivity around 500 nm while the red lamp in our experiment emits light from around 600 nm where the sensitivity is much lower (differing approximately by a factor of 300) (Fain, 1976). Red light also had the least overlap with the photosensitivity of the pineal gland (Eldred and Nolte, 1978). The importance of spectral sensitivity is also clear in attraction of amphibians to light where the strongest effects are also caused by light with a high level of short wavelength light even at light levels where colour vision is no longer possible (Donner and Reuter, 1962, 1976).

In this study most toads tended to avoid green and white light heading towards the dark and red areas, although some of them were caught at the white and green treatment. We did not find significant differences in the distribution of toads caught after the lights were switched off. Therefore we do not have support for the hypothesis that the toads halt their migration when they come to the light.