The Effects of Artificial Night Lighting on Tail Regeneration and Prey Consumption in a Nocturnal Salamander (Plethodon cinereus) and on the Behavior of Fruit Fly Prey (Drosophila virilis)

Increased human development has resulted in the introduction of artificial light at night (ALAN) into natural habitats around the globe [1]. This artificial light at night, i.e., light pollution, is a major ecological problem that seems to have a widespread, negative impact on many different species [2,3,4,5,6,7]. Although it is clear that wildlife living in urban areas is exposed to a greater intensity of ALAN from point (direct) sources, such as street lighting (10 lx), ambient sources—particularly sky glow—can also produce artificial light levels of up to 1 lx (comparable to dawn/dusk twilight light intensity) [2] in areas surrounding urban centers far from point sources, such that very few areas in the world are unaffected by ALAN [1]. The global exposure to ALAN may have consequences for biodiversity world-wide [8]; however, there are relatively few studies that examine the effect of ecologically-relevant dim light at night (dALAN) on the ecology, behavior, and physiology of organisms; thus, controlled experiments examining the consequences of dALAN, such as that associated with sky glow are necessary and important [9]. Additionally, studies examining dose responses to ALAN are even scarcer and are essential in determining the threshold levels of ALAN that affect organisms [10]. Most studies of the effects of nocturnal ALAN on amphibians have used dichotomous treatments of nocturnal illuminations that are either very (or even completely) dark or brightly lit, thereby missing the opportunity to assay the effects of intermediate, ecologically relevant amounts of ALAN [11,12]. Thus, studies of the effects of ecologically-relevant dALAN on amphibian biology are essential and necessary for understanding how light pollution affects these animals [10,13].

Dim light intensities, representative of light levels associated with changing lighting at dawn or dusk can be important zeitgebers that synchronize circadian cycles and regulate photoperiodic behavior due to different gene expression in different neurons, e.g., morning (M) and evening (E) oscillators in Drosophila [14,15,16,17] and so it is reasonable to predict that biological systems might respond differently to dALAN than to darker or brighter ALAN. Additionally, ALAN intensity diminishes as the inverse of the squared distance from the light source and so animals in habitats adjacent to sources of lighting would experience different amounts of ALAN (from bright to exceedingly dim) depending on their distance from the source. Thus, in order to understand the effects of ALAN on organisms living in affected habitats, it is important to recognize that those organisms may not respond identically throughout the habitat if light intensities vary.

Nocturnal species may be particularly susceptible to ALAN since they are active under normally dark conditions and have evolved physiological functions and behavior associated with dark ambient illumination [3]. Many species of amphibians are nocturnal, and their numbers have been declining globally over the last century due to environmental perturbations, such as habitat destruction, climate change, environmental pollution, and invasive species [18,19,20,21]. Although multiple drivers are probably responsible for global amphibian decline [20,21,22,23,24], local population-level declines are likely due to a more specific set of stressors unique to that habitat [21,22]. ALAN has the potential to be a nearly ubiquitous stressor for nocturnal amphibians [8], but it is one of the least studied forms of pollution in amphibians, with surprisingly few studies examining its effects [13]. ALAN can affect behavior, reproductive cycles, hormones, and metabolic rates of amphibians through altered photoperiodism and vision [11,12].

Salamanders are important components of many forest and aquatic ecosystems and are sensitive to environmental perturbations; thus, they are excellent indicators of the health of some ecosystems [25,26]. The eastern red-backed salamander, Plethodon cinereus, is a major predator of invertebrates in deciduous forests of eastern North America [27], occurring at extremely high densities in some locations [28], such that the biomass of these salamanders may be higher than that of any other vertebrates in some forests [29,30]. The behavior and ecology of these nocturnal salamanders have been extensively studied [31], making them a model species for studies of the effects of light pollution on an ecologically important nocturnal animal.

Plethodon cinereus, similar to many other species of salamanders, can autotomize and completely regenerate its tail [32]. Tail autotomy in this and similar species may serve as a predatory defense mechanism allowing the animal to escape capture by sacrificing its tail [33,34]; and tail elongation following autotomy is an energetic priority for salamanders [35]. However, tail loss may put a salamander at a disadvantage because the tail is important for fat storage during winter brumation [36]; tail loss reduces reproductive output, with fewer ova produced following regeneration [37]; the tail is involved in territorial communication and defense, including scent marking [38], threat displays [39,40], and fighting [41]; and subsequent escape from predation is reduced, while the tail is regenerating. Thus, factors influencing regeneration rates may, in turn, affect territoriality, communication, spatial dynamics, reproduction, and even survival of individuals within populations.

Although the effect of ALAN on tail regeneration has not previously been studied in salamanders, constant and relatively high levels of light at night enhanced tail regeneration in a gekkonid lizard (Hemidactylus flaviviridis) and limb regeneration in eastern red-spotted newts (Notophthalmus viridescens), whereas constant darkness or pinealectomy slowed regeneration [42,43,44]. This indicates an inhibition of melatonin production [45] that is responsible for increased rates of regeneration, perhaps because (1) DNA synthesis and mitotic activity fluctuate on a daily serotonin–melatonin rhythm regulated by light [46,47], with melatonin slowing mitotic activity [48,49,50] and (2) production of prolactin is increased, stimulating regeneration [51,52]. Conversely, ALAN may decrease the rate of tail regeneration by increasing corticosterone levels as a stress response, as has been demonstrated in tadpoles of Bufo valliceps and Rana berlandieri [53]. Elevated levels of corticosterone reduced rates of regeneration in the plethodontid salamander, Desmognathus ochrophaeus [54].

ALAN may also affect prey capture and consumption, which are necessary to provide energy and nutrients for tail regeneration [55], by altering vision or activity. The eastern red-backed salamander, similar to many other species of salamanders, is nocturnally active and negatively phototactic, such that these animals emerge from the leaf litter and from under cover objects (rocks and logs) to forage at night [12,56] and avoid brightly lit areas [57,58], such as those that might be associated with ALAN. Therefore, if ALAN is sufficient to trigger photonegative behaviors, salamanders may be more likely to express escape behaviors and may be less apt to forage in brightly lit conditions. Additionally, visual cues are important to foraging P. cinereus and under some conditions, e.g., when olfactory cues were not available, individuals of P. cinereus oriented toward prey sooner at higher nocturnal illuminations (unpublished data cited in [12,59]). Because these salamanders orient toward and use vision to detect and capture moving prey, if ALAN alters the activity of insect prey, it may indirectly impact tail regeneration rates in salamanders by influencing prey detection and capture. In our study, we fed salamanders fruit flies, Drosophila virilis. Although literature examining the effect of dALAN on the activity of D. virilis is not available, the activity of wild-type Drosophila (Sophophora) melanogaster was greater when exposed to dALAN (0.03 lx) than when exposed to complete darkness [60]. Wild-type D. (S.) melanogaster have distinctly bimodal morning and evening activity under laboratory conditions, whereas D. virilis, displayed 96% of their locomotor activity during daylight, and although seemingly crepuscular, may show more activity closer to dusk [61,62,63]. It is possible that the addition of night lighting may increase the nocturnal activity level of D. virilis in our study thereby directly or indirectly influencing the foraging behavior and tail regeneration rates of the salamanders. Therefore, we examined the nocturnal behavior of prey as part of this study.

We hypothesized that an increase in the nocturnal ambient illumination equivalent to ecologically-relevant levels of light pollution would alter the rate of tail regeneration in P. cinereus, although we did not predict a direction because there are multiple physiological and behavioral factors that may influence tail regeneration as previously discussed. Furthermore, because the rate of regeneration (tail growth) may be influenced by food consumption and prey behavior, we hypothesized that consumption of prey items (D. virilis) would be positively correlated with the size of the salamander but would not be influenced by different nocturnal lighting treatments or behavior of the prey, as there may be opposing effects of light at night on successful prey capture (i.e., increased illumination may increase prey capture efficiency but avoidance of lighted areas by the salamanders may reduce time spent foraging). We also hypothesized that ALAN would increase the nocturnal behavior (activity) of D. virilis over what would be exhibited under natural dark illuminations.

We found that even very small amounts of ALAN significantly affected tail regeneration in the eastern red-backed salamander, P. cinereus. However, the effect of ALAN on tail regeneration was not monotonic but much more complex; regeneration was highest in the dark control (0.0001 lx) and 1 lx treatments and significantly lower in the 0.01 and 100 lx light treatments (Figure 2). These data are inconsistent with our hypothesis that the effect of nocturnal illumination on tail regeneration would be a dose-dependent response with either increasing or decreasing rates of tail regeneration with increasing nocturnal illumination, suggesting a more complex physiological or behavioral response to ALAN. The tail has multiple functions in P. cinereus as well as other salamanders, including energy storage [34,36,37,76,77], signaling and territorial defense [38,39,40,41], and antipredator behavior [33,34,78,79], and the ability to regenerate the tail quickly may impact survival and reproductive success [34,35]. An important function of the tail is its use as an escape mechanism when the salamander is attacked by a predator and the tail is autotomized and produces noxious and adhesive secretions that deter and increase handling time for predators [33,34,78,79]. The tail is also important for fat storage during winter when salamanders do not feed; for example, salamanders at higher elevations sequester more fat than salamanders at lower elevations [36]. Fat storage also seems important for reproductive success and oocyte maturation [77] and brooding females that have intact tails produce more ova the following year than females that have lost their tails [37]. Tail loss also impacts territorial behavior including scent marking (rate of deposition of pheromones via post-cloacal press [38] and aggressive displays [39,40], such that salamanders with shortened tails are at a disadvantage during territorial contests and likely subjected to increased aggression from opponents [41]. Presumably, because of the importance of having a complete tail, regeneration and elongation of the tail occurs even in salamanders that are not fed for extended periods of time (>9 mo) [35]. Thus, ALAN-induced differences in tail regeneration are expected to have consequences for the population ecology of these territorial salamanders as well as other species of salamanders with similar life histories.

Although ALAN altered tail regeneration rates in salamanders with slower rates of regeneration at nocturnal illuminations of 0.01 and 100 lx, this effect was unlikely the result of differences in activity of the fruit fly prey. We did find that the activity of fruit flies (D. virilis) was profoundly affected by ALAN, with overall nocturnal movement increasing as nocturnal illuminations increased. This result was most obvious at lower levels of ALAN (dALAN) with flies significantly more active in the 0.01 lx treatment than in the dark control and with flies even more active in the 1 and 100 lx treatments than in the 0.0001 and 0.01 lx treatments (Figure 4). It should be noted that our brightest night lighting treatment was 100 lx, which is equivalent to dim room lighting, typical of what might be experienced under the forest canopy on an overcast day, and not nearly as bright as unobscured day lighting (sunlight). Studies in another species of fruit fly, D. (S.) melanogaster, found an increase in activity when flies were exposed to dALAN (0.03 lx) [60] as opposed to virtual darkness as well as avoidance of activity in brightly lit areas (100 lx or greater) [63]. We saw similar patterns of activity in D. virilis, with significantly more activity in dALAN (0.01 lx) than in the dark control treatment (0.0001 lx). We also found increased activity in the 1 lx treatment under our lighting regimen, similar to D. (S.) melanogaster (with most activity at 7.5 lx regardless of time of day) [63]. However, our data differed from [63] in that we found that activity was similar in both the 1 lx and in the brighter 100 lx treatment, which may be due to species–specific variation in rhythmic behavior or differences in experimental design.

Timing of fly activity is controlled, in part, by the circadian system [80]. Research has demonstrated that a nocturnal illumination of 0.03 lx was sufficient to shift the ‘temporal niche’ of D. (S.) melanogaster to becoming active nocturnally rather than being crepuscular [60]. The flies in that study appeared to have a dim light preference and did not simply become active for more hours a day but shifted their initiation of activity from the morning period (M) to the evening period (E). In another species of fruit fly, Drosophila (Sophophora) biarmipes, dim light (0.03 lx) at night sped up re-entraintment (recovery of normal circadian pattern) after rhythm disruption more effectively than did total darkness [81]. Similar effects of dim lighting have been observed in Syrian hamsters (Mesocricetus auratus), where nocturnal illumination of 0.004 lx (equivalent to dim moonlight) was sufficient to accelerate re-entrainment relative to total darkness after disruption of circadian cycles [82], demonstrating that even small amounts of light during scotophase can alter periodic locomotor rhythms, when compared to what occurs in studies that employ total darkness as a control treatment. It has been suggested that such an effect might be caused by dim lighting disrupting the coupling between the M and E oscillators [81], though data for mice (Mus musculus) at least indicate that this effect is mediated by clock genes outside the SCN [83]. Regardless of the mechanism, dALAN is able to alter periodic locomotor behavior in multiple species of fruit flies and in mammals. We did not include a total darkness control treatment opting instead for what we perceived to be a more natural 0.0001 lx dark control equivalent to starlight. Unless animals are fossorial, they are unlikely to encounter total darkness above ground at night [84] and, therefore, total darkness may not be the most effective control lighting treatment in studies of organisms that do not regularly experience total darkness at night. Without the total darkness control we are unable to assess how well or poorly fly activity in our dark control might match that of total darkness in other studies. What is clear from our study, however, is that even small increases in dim lighting can alter fly nocturnal behavior.

Salamanders provided prey (D. virilis) with different activity levels (amount of movement) at different nocturnal illuminations may experience different rates of prey encounter and/or detection and these differences could potentially affect prey consumption rates and, ultimately, tail regeneration rates; however, that is not what we found. Tail regeneration was not detectably changed by prey consumption because we saw no significant differences in prey capture and consumption by salamanders in different nocturnal lighting treatments, when examined for a 10-day-period. Additionally, food consumption does not seem to alter the regeneration rate in P. cinereus monotonically. One study found that length of tail regeneration was not statistically different in individuals of P. cinereus given low or high quantities of prey items (D. melanogaster), rather quantity of food affected the volume of fat in the regenerated tails [35], which we did not assess in our study. Thus, in spite of substantial differences in fly activity in the different lighting treatments, consumption of fruit fly prey was most likely not a factor influencing tail regeneration because (1) salamanders ate similar quantities of prey in our experiment (Figure 3), (2) regeneration lengths are not impacted by food consumption in this species of salamander [35], and (3) fruit fly activity did not vary in the same way as tail regeneration among light treatments (Figure 2 and Figure 4). We speculate that the complex, non-linear response of regeneration to different amounts of dALAN seen in our experiment is not due to prey behavior but rather is due to (1) physiological factors that are under hormonal (e.g., melatonin, serotonin, prolactin, and corticosterone) control or (2) genetic control where different sets of genes may be active at different scotopic illuminations and that their products may cause differences in endocrine responses under different levels of dALAN.

ALAN can, in some species, result in increased levels of glucocorticoid hormones, such as corticosterone, through the hypothalamic–pituitary–adrenal (HPA) axis [85] (the HPA/intrarenal (I) axis in amphibians [86]). For example, corticosterone levels were higher in tadpoles reared in ALAN treatments compared to dark controls [53]. Additionally, P. cinereus is nocturnally active on the forest floor (Figure 1 in [59]) and negatively phototactic, avoiding bright light (e.g., 1400 lx), presumably as a mechanism to avoid high temperatures, desiccation, or predation, although light avoidance occurs even under controlled conditions of temperature and humidity [58]. Thus, under higher illuminations (i.e., constant 100 lx), salamanders may exhibit stress-related responses, including increases in corticosterone, such as that demonstrated in frog tadpoles after a 14-day-exposure to ALAN (constant light at 250 lx) [53]. This is of particular significance because corticosterone can impede regeneration. In another plethodontid salamander, D. ochrophaeus, exogenous corticosterone delivered via skin patches slowed wound healing [87] and tail regeneration [53]. If ALAN results in a chronic increase in corticosterone, we would predict slower regeneration rates at higher light levels. Such a stress response could explain our lower rates of regeneration occurring at 0.01 and 100 lx, but would not be consistent with the higher rate of regeneration in the 1 lx treatment that had similar regeneration rates as the dark control. However, evidence summarized in [86] suggests that in at least some species of plethodontid salamanders, elevated corticosterone levels may not be associated with chronic stress; although the stressors they investigated did not include ALAN. Species–specific modulation of a corticosterone response to ALAN (e.g., downregulation) can occur, as was found for tadpoles exposed to pulsed ALAN [53]. Additional research examining corticosterone levels in plethodontid salamanders under various ALAN conditions is important for understanding the role of ALAN in influencing corticosterone levels and tail regeneration in this species. This is an area with little research and is an important topic in understanding factors influencing tissue regeneration [88].

Salamanders, as all vertebrates, show peak levels of melatonin during naturally dark periods and peak serotonin levels during photophase [45,89]. The pineal gland converts serotonin to melatonin during dark periods and production and secretion of melatonin is suppressed during daylight hours, resulting in higher levels of melatonin at night under natural lighting conditions. Although in non-mammalian vertebrates melatonin can be synthesized by the eyes and other organs [13,85], melatonin produced in the eyes does not seem to be as important as pineal melatonin in tissue regeneration in at least one species of nocturnal lizard [42]. The natural light:dark cycles have resulted in a conserved and crucial role for the production of melatonin in regulating many physiological functions, especially those associated with daily and seasonal rhythms [13,85,90,91,92]. ALAN, even at low levels, can disrupt the melatonin cycle by suppressing melatonin production at night [13,45,87]. While no studies have directly investigated the effect of melatonin on tail regeneration in salamanders, studies on limb regeneration in newts provides some evidence for the role of melatonin in salamander tissue regeneration. Eastern red-spotted newts (N. viridescens) kept under constant light at night regenerated lost limbs faster than those maintained in complete darkness [43]. Additionally, newts (N. viridescens) with blocked pineal glands regenerated limbs more slowly than animals with normal, exposed pineal glands, providing evidence of the importance of the melatonin cycle in tissue regeneration [44]. Research on tail regeneration in the nocturnal gekkonid lizard, H. flaviviridis provides similar results. In a controlled experiment, tail regeneration was greatest in lizards kept under bright continuous daylight of either 2500 or 638 lx, compared to a 12L:12D photoperiod, and was slowest in continuous dark [42]. Pinealectomized lizards exhibited reduced growth no matter the lighting treatment compared to a sham control [42], and in another experiment [93], exogenous melatonin enhanced regeneration rates in lizards on a natural photoperiod when given at 1700 h and suppressed regeneration when administered at 0700 h, indicating that the natural, nocturnal production of melatonin is important in tail regeneration.

Rhythmicity of melatonin and serotonin secretion may influence regeneration by modulating the secretion of prolactin [88]. Studies have shown that prolactin significantly increases regeneration in newt forelimbs and newt and lizard tails [51,52,94,95]. Thus, an increase in regeneration occurring under conditions of constant light may be due to high levels of serotonin, which stimulates the release of prolactin [44,88]. Interestingly, pinealetomized lizards given exogenous prolactin regenerated tails faster than those not given prolactin but did not regenerate tails as much as intact lizards receiving prolactin, suggesting that prolactin stimulates tail regeneration but that the cyclical production of melatonin is also important in the process [52]. In our experiment, we did not find support for constant light increasing tail regeneration rates. At constant light levels of 100 lx (substantially lower than day-lighting illuminations used in other experiments we discussed), we found reduced tail regeneration lengths compared to tail lengths of salamanders exposed to a normal dark night (0.0001 lx). However, at intermediate levels of ALAN, we did find more regeneration at 1 lx than 100 lx. Because of the complex relationship between internal and external factors that may influence regeneration in different ways within and among species [88], constant dALAN may result in a more complex response to light levels in our salamanders than can be predicted by individual factors.

Dim ALAN, at particular illuminations, may simulate conditions of dawn or dusk and may differentially regulate the expression of genes associated with the morning (M) or evening (E) oscillators. Under natural conditions, illuminations increase gradually at the end of scotophase triggering regulation of the genes associated with the morning (M) oscillator. If expression of different genes is regulated by different environmental illuminations, then different genes will be expressed as light levels increase at dawn or decrease at dusk. ALAN illuminations of 1 lx and below have been recorded in natural habitats [11,59]. Salamanders exposed to 1 lx in our experiment were effectively temporally trapped in perpetual twilight, greatly increasing the duration of subjective morning (M) or evening (E) conditions. Under normal, clear-sky, natural conditions, twilight illuminations of 1 lx at dawn and dusk only happen for a few minutes each day (though they may happen under dense forest canopy during the day under heavy cloud cover). If nocturnal illuminations are held at 1 lx for the entirety of scotophase, then the normal pattern of gene expression at twilight would be maintained constantly through the night and salamanders under those conditions would never experience gene expression typical of normal scotopic conditions. This same potential problem also exists at 0.01 lx (full moonlight) but organisms are likely to experience full moon lighting for longer periods during months when the full moon is present above the horizon during scotophase.

Additional studies addressing the physiological mechanisms responsible for variation in rates of regeneration that we have observed at ecologically-relevant levels of light pollution would greatly benefit our understanding of how regeneration rates are influenced by ALAN. The most studied models for tissue regeneration in vertebrates are lizards and newts (N. viridescens). Adult eastern red-spotted newts are aquatic and lizards have epidermal scales for reducing desiccation, whereas P. cinereus is terrestrial and thin-skinned. Thus, P. cinereus is much more likely than lizards and newts to be impacted by stress associated with water loss [86] that may influence its response, especially a stress response, to lighted conditions above ground. Additionally, no other studies of terrestrial salamanders have been published examining the effect of ALAN at ecologically relevant levels; therefore, it is unknown if the complex, nonlinear pattern seen in our tail regeneration data can be generalized to other species or whether it might be unique to this system.

How likely is it that eastern red-backed salamanders, or fruit flies for that matter, might encounter these levels of ALAN in their natural habitats? Direct glare from streetlamps along roadways, edificarian lighting, such as security lighting, or other sources does not stop when it hits the forest edge. Light from such sources can travel great distances, though it diminishes with distance as the inverse of the squared distance. Skyglow from cities, industrial sites, and sports or other large illuminated complexes occurs when light from those sources is reflected off moisture in the atmosphere back to the ground far from its sources. We have measured combined contributions of glare and sky glow in a protected wetland (Utica Marsh adjacent to Utica, NY) of 0.1–1.0 lx across large areas of that wetland, which contains many species of amphibians and insects. It seems that 23% of the land surface of the Earth (between 75 degrees N and 60 degrees S) is seriously light-polluted and that 46.9% of the United States and 88.4% of the European Union land area experience significant levels of sky glow [1]. Overlaying the Artificial Night Sky Brightness Atlas [1] for the Northeastern United States onto the range map for P. cinereus [96] demonstrates that almost the entire range of the species experiences sky glow on a nightly basis. Such pervasive dALAN, which may widely affect ecologically important nocturnal species, such as P. cinereus, raises significant conservation concerns.

Very small amounts of dALAN that is pervasive across much of the range of P. cinereus caused a complex, non-monotonic effect on tail regeneration rates in the eastern red-backed salamander, P. cinereus, that most likely involves the interaction of several hormones and metabolic processes. ALAN also altered the activity of fruit fly prey (monotonic response) but not in a way that obviously impacted food consumption and tail regeneration in salamanders. Thus, dALAN has the ability to affect tail regrowth, which can in turn affect the survival and reproduction of this important species of salamander.

We would like to thank Meagan Sudik, Praise Austin, Elizabeth Lopez, Jasmina Samardzic, Jessica Nyirinkwaya, and Jaedyn LeLonde for help in the lab and maintenance of fly cultures. The equipment used for this study was funded partially by the Herkimer and Oneida Community Foundation and the Alden Trust. Salamanders were collected under a New York State Department of Environmental Conservation License to Collect, permit number 973.

Conceptualization, S.E.W. and B.W.B.; methodology and investigation, S.E.W., B.W.B., A.R., T.C. and A.E.S.; formal analysis, S.E.W., A.R., T.C. and A.E.S.; data curation, S.E.W. and B.W.B.; writing—original draft preparation, S.E.W., B.W.B. and A.R.; writing—review and editing, S.E.W., B.W.B. and A.R.; supervision, S.E.W. and B.W.B. All authors have read and agreed to the published version of the manuscript.

The animal study protocol was approved by the Institutional Animal Care and Use Committee of Utica University (protocol number 2007-002, approved March 2007).

The data presented in this study are available upon request.

The authors declare no conflict of interest.

Equipment and supplies for this research was partially funded by the Herkimer and Oneida Community Foundation, the Alden Trust, and Utica University. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.