Effects of global warming on sensory ecology of rock lizards: increased temperatures alter the efficacy of sexual chemical signals

Authors


Summary

  1. An almost neglected aspect of climate change is its effects on sensory ecology and sexual signals of animals. Signals used in intraspecific communication are expected to evolve to maximize efficacy under a given climatic condition, but it is unlikely that quick changes in environmental conditions could be compensated by similarly quick evolutionary changes in the design of signals. We predict that global warming will lead to a loss of efficacy of some sexual signals, with important consequences for sexual selection.
  2. We examined experimentally the effects of global warming on the efficacy of chemical signals of a mountain lizard (Iberolacerta cyreni).
  3. We first showed how environmental temperatures in the study area during the mating season of lizards have actually increased in the last years. Then, we tested whether female lizards were able to detect by chemosensory cues the males’ scent marks (i.e. femoral secretions) that were experimentally maintained under current and predicted future temperature conditions.
  4. Results showed that the efficacy (i.e. detectability and persistence) of scent marks is lower at high temperature. Moreover, we showed that scent-marked substrates maintained under high temperatures were not selected by females, in contrast to the selection of areas scent marked by males, when these substrates were maintained under normal temperatures.
  5. Our study suggests that climate warming could lead to negative changes in the efficacy of sexual signals with potential consequences for the sexual selection and conservation of threatened lizard species.

Introduction

One of the most notorious effects of climate change is a global rapid increase in temperature (IPCC 2007). This is known to have already induced ecological responses of many animals and plants, such as changes in phenology or distribution (e.g. Parmesan 2006; Moyes et al. 2011), and even phenotypic plasticity and evolutionary changes in some traits (Charmantier et al. 2008; Hoffman & Sgrò 2011). Most studies on the effects of global warming have focused on how diverse species have shifted, or are predicted to shift, their life cycles or geographical ranges (e.g. Araújo, Thuiller & Pearson 2006; Moyes et al. 2011). Other studies examined the effects of climate change on the biology and physiology of organisms (Huey et al. 2009; Telemeco, Elphick & Shine 2009). Ectotherms seem especially vulnerable to global warming (Araújo, Thuiller & Pearson 2006; Deutsch et al. 2008; Huey et al. 2009; Sinervo et al. 2010). This is mainly because their physiological performance largely depends on their body temperature, which is mainly dependent on environmental temperatures (Huey 1982; Huey et al. 2009; Gunderson & Leal 2012). In reptiles, global warming also affects nest temperatures, which directly determine offspring sex and other phenotypic traits (Aubret & Shine 2010; Dubey & Shine 2011).

An almost neglected aspect of climate change is its effects on sensory ecology and sexual signals of animals (Herbert-Read, Logendran & Ward 2010; Ferrari et al. 2011; Möller 2011). However, this is important because signals used in intraspecific communication are expected to evolve to maximize efficacy of the signal in a given environment (Guilford & Dawkins 1991; Endler & Basolo 1998; Bradbury & Vehrencamp 2011). Factors such as how the signal is transmitted through the environment, durability or persistence of the signal, or how well the signal is detected by the sensory system of receivers, influence the efficacy of the signal and may be shaped by selection to enhance perception by the receiver (Guilford & Dawkins 1991; Alberts 1992; Endler & Basolo 1998; Boughman 2002). All these variables largely depend on the climatic conditions. If sexual signals became less effective due to human-induced environmental changes, this may result in the alteration of some crucial aspects of sexual selection such as sex and species recognition, territorial defence, mate attraction or mate choice (Lurling & Scheffer 2007; Sih, Ferrari & Harris 2011; Tuomainen & Candolin 2011).

Chemoreception is one of the main sensory systems used by many animals, including many vertebrates, and chemical signals (pheromones) play an important role in communication and sexual selection (Wyatt 2003; Müller-Schwarze 2006; Mason & Parker 2010). In terrestrial vertebrates, pheromones are very often incorporated into faeces, urine or other scent marks left on the substrate with the purpose of marking territories boundaries or attracting mates (López, Aragón & Martín 1998; Brennan & Kendrick 2006; Martins et al. 2006). For example, many lizards produce chemical secretions with pheromonal activity (Mason 1992). These chemical signals may give information on sex, size, age or familiarity recognition (reviewed in Mason 1992; Mason & Parker 2010; Martín & López 2011) or even provide more detailed information on morphological traits and health condition of the signaller (López, Amo & Martín 2006; Martín et al. 2007). This information is important in intrasexual relationships between males (López & Martín 2002, 2011; Carazo, Font & Desfilis 2007; Martín & López 2007) and in female mate choice (Martín & López 2000, 2006a,b; Olsson et al. 2003; Martín et al. 2007).

Darwinian selection will rapidly act to maximize the efficacy of sexual signals in different environments (Boughman 2002; Bradbury & Vehrencamp 2011). For example, chemical secretions used for scent marking are expected to change, in the evolutionary time, to ensure that signals are perfectly tuned to local humidity and temperature (i.e. affecting their volatility and therefore, their persistence and transmission through the environment) (Alberts 1992). Thus, interspecific differences in chemical signals of lizards might reflect selection for the efficacy of signals in different climatic conditions (Alberts 1992; Escobar, Labra & Niemeyer 2001; Boughman 2002; Escobar et al. 2003; Martín & López 2006c). However, it is unlikely that quick changes in climate, such as the current global warming, could be compensated by similarly quick evolutionary changes in the design of the sexual signals (Hoffman & Sgrò 2011; Sih, Ferrari & Harris 2011). Although predicting the behavioural responses of organisms facing climatic changes is very difficult (Sih, Ferrari & Harris 2011; Tuomainen & Candolin 2011), terrestrial ectotherms might show short-term, although limited, flexibility in thermoregulatory or nesting behaviour to respond to ambient changes (Kearney, Porter & Shine 2009; Telemeco, Elphick & Shine 2009; Aubret & Shine 2010). However, it is unlikely that animals could use behaviour to entirely compensate for changes in the efficacy of sexual signals. Thus, we predict that increased temperatures will result in a higher volatilization and degradation of the chemical compounds found in the scent marks, which will lead to a loss of efficacy and of the signalling function of sexual chemical signals. This can have important consequences for communication and sexual selection.

The Carpetan rock lizard, Iberolacerta cyreni (formerly Lacerta monticola cyreni), is a lacertid lizard restricted to rocky habitats of high mountains of Central Spain (Martín & Salvador 1997a). This is a threatened species, mainly because of its restricted geographical range and the vulnerability of high-mountains ecosystems to global change (Nogués-Bravo et al. 2008; Monasterio et al. 2009). It is a polygynandrous species, where older males defend territories that overlap partly with those of other males (Aragón, López & Martín 2004), and where few males obtain most of the successful matings, siring the offspring of several females (Salvador et al. 2008). This indicates that sexual selection is a strong selective force affecting the evolution of these lizards. Males scent-mark substrates with femoral gland secretions, which contain a mixture of proteic and lipophilic compounds, mainly fatty acids and steroids (López & Martín 2005). Female I. cyreni show strong chemosensory responses and prefer areas scent marked by males that allocate more cholesta-5,7-dien-3-ol and ergosterol to femoral secretions (Martín & López 2006a,b, 2008, 2012), which are of presumably high quality (i.e. those more symmetric and with a higher immune response). These results suggest that scent marks of males affect space use of females, increasing the probability of mating with the male that has scent marked the selected area (Martín & López 2012).

In this paper, we examined experimentally the effects of global warming on the persistence and efficacy of chemical sexual signals of rock lizards (I. cyreni). We first showed how air temperatures in the study area have actually increased in the last 42 years. Then, we tested the chemosensory ability of female lizards to detect males’ scent marks (i.e. femoral secretions) that were experimentally maintained under different temperature regimes. We mimicked current temperature conditions and, because variance and intensity of temperature are expected to increase under the global change scenario, an accidental brief increase of temperature after scent marks had been deposited, rather than the effect of a permanent warming, which would require that scent production and deposition would also be under warm conditions. We examined whether the temporal attenuation of the chemical stimuli depended on temperature. We predicted that efficacy (i.e. detectability and persistence) of scent marks would decrease with increasing temperatures. Finally, we maintained scent marks of males at different temperatures, and tested the effects on female choice of areas scent marked by males. We discuss the potential negative effects of changes in the efficacy of sexual signals for the sexual selection and conservation of this lizard.

Materials and methods

Study Animals

We captured by noosing adult lizards (18 males and 18 females) during May–June 2011 at ‘Alto del Telégrafo’ (40° 47′ N, 04° 00′ W; elevation 1900 m; Guadarrama Mountains, central Spain). Granite rock boulders interspersed with shrubs (Cytisus oromediterraneus and Juniperus communis) and meadows predominated at the study site (Martín & Salvador 1997a). Lizards are active from late April to early October, mating in May–June and producing a single clutch in July (Salvador et al. 2008). Lizards were individually housed at ‘El Ventorrillo’ Field Station, 5 km from the capture site in outdoor 80 × 50 cm PVC terraria containing rocks for cover, and food (mealworms) and water ad libitum. All animals were healthy and were returned to their capture sites at the end of trials.

Weather Data

We summarized the temporal changes in available environmental temperatures in the study area by using long-term (from 1970 to 2011) daily data from a nearby meteorological station (‘Puerto de Navacerrada’; 40° 46′ 50″ N, 04° 00′ 37″ W; elevation 1894 m; Madrid province) (data available from the web of the Spanish Meteorological Agency, ‘Agencia Española de Metereología, AEMET’; http://www.aemet.es). Monthly temperatures were measured as means of daily mean temperatures. We also used means of daily maximum air temperatures, as advised for reptiles (Huey 1982). We used data from May and June, which coincides with the mating season of lizards, when males have the highest rates of femoral secretions, used for scent marking, and all matings occur (Salvador et al. 2008).

Chemosensory Trials

We used tongue-flick (TF) bioassays as a test of detection of chemical cues based on the differential rates of tongue extensions of lizards to the different chemical stimuli (Cooper & Burghardt 1990; Cooper 1994, 1998). We compared TF rate by female lizards in response to stimuli arising from cotton applicators bearing femoral secretions of males. We examined the effect of temperature on the temporal fading of the chemical stimuli after they had been deposited on the cotton swabs. We prepared stimuli by dipping the cotton tip (1 cm) of a wooden applicator in deionized water. Then, we took out femoral secretions of males by pressing around the femoral pores with a pair of forceps and collected the secretion directly on the cotton tips. We used approximately the same amount of femoral secretion in each stimulus (2 × 1 mm of solid secretion from each of three pores) to minimize the likelihood that differences in TF rates were due to differences in the amount of secretion presented. To avoid differences in responses to different individual males (Martín & López 2006a), every female was always tested with secretions from the same individual male.

We placed swabs impregnated with chemical stimuli and blank control swabs in two incubator chambers at 16 and 22 °C (‘cold’ vs. ‘warm’, respectively), and left them there for 1 min (initial time), 1 or 3 h before being used in chemosensory tests. The cold temperature corresponded approximately to the mean of daily maximum temperatures in the study area during the mating season of the last twelve years (2000–2011). Whereas the warm temperature was the expected mean of daily maximum temperatures after a period of about 50 years under the current rate of warming increase (see results). The cotton swabs were not in contact with anything inside the chambers, and after being used in a single test were discarded.

Each individual female was tested in twelve treatments: three periods of time since the femoral secretion was deposited × two temperatures × two stimuli (femoral secretion vs. control), but participated in only one test every day in a random order. Trials were conducted in outdoor conditions at the beginning of June, coinciding with the mating season, and between 8:00–12:00 h (GMT) when lizards were fully active. Before the tests, females were allowed to bask and attain an optimal body temperature (around 29·4 °C; Martín & Salvador 1993). In each trial, one experimenter slowly approached a lizard's home cage and slowly moved the cotton swab applicator attached to a long stick (50 cm) to a position 2 cm anterior to the lizard's snout. Lizards allowed approach and testing without fleeing. All female lizards responded to swabs by tongue flicking. We recorded latency to the first TF and numbers of TFs directed to the swab during 1 min, beginning with the first TF.

To examine differences in latencies or number of directed TFs of the same individual females among treatments, we used Generalized Linear Models (GLMs) with type of stimulus (femoral secretions vs. control), time (initial vs. 1 h vs. 3 h) and temperature treatment (cool vs. warm) as within-subject factors, including the interactions in the models. Data were log-transformed to ensure normality (Shapiro–Wilk's test). Tests of homogeneity of variances (Levene's test) showed that in all cases, variances were not significantly heterogeneous after transformation. Pairwise comparisons used Tukey's honestly significant difference (HSD) tests (Sokal & Rohlf 1995). All the statistical analyses were performed with statistica v8.0 (Statsoft Inc., Tulsa, OK, USA).

Female Choice of Areas Scent Marked By Males

At the beginning of the experiments, we placed in the males’ cages several absorbent paper strips (35 × 10 cm) fixed to the floor, and left them there for 1 week to allow males to scent mark them. The ventral location of the femoral pores allowed that secretions were passively deposited on the paper substrate as male lizards moved through their terraria, but active rubbing of the legs on the substrate has also been observed in this and other lizards (Martín & López 2006a; Martins et al. 2006). We also placed similar paper strips in empty terraria that were located close to the males’ terraria, to be used as blank controls.

Female choice of scent experiments were performed at the end of May in outdoor conditions. At the beginning of each test (07:00 h GMT), when females were still inactive and hidden in refuges, we took with fresh gloves the papers strip scent marked by males and blank papers from empty terraria. We placed papers in two incubator chambers under two different temperatures as above (16 and 22 °C, ‘cold’ vs. ‘warm’), and left them there for 30 min before being used in tests. Females’ cages had two basking platforms (two identical flat tiles) placed symmetrically at each end of the cage, and rocks for cover and basking in the centre. We took the papers from the incubators and fixed on one tile one paper strip scent marked by a male, and a clean blank paper on the other tile, both of the same temperature treatment. Each female was tested in the two treatments (cold vs. warm) once a day in a random order, with papers from the same individual male to avoid differences in responses to different individual males (Martín & López 2006a). The males tested the positions of papers, and order of treatments was randomly determined.

Each trial lasted 6 h (from 09:00 h GMT, shortly after females appeared from night refuges, and until 15:00 h GMT, when females hid again). Females were monitored each 30 min from a hidden point recording their location in the cages. If a female was located on a tile with the paper strip, she was designated as having chosen temporarily that particular paper, whereas if she was located on the central area outside of the tiles she was designated as having made no choice (for a similar procedure see Martín & López 2000, 2006a; Olsson et al. 2003). We excluded scan observations when females were hidden in the refuge. Female I. cyreni have limited movement rates, spending long time periods stationary, and use more a sit-and-wait foraging strategy (Martín & Salvador 1997b). Thus, the locations observed on each of the 17 scans were considered to be representative of females’ space use of the cages. To ensure that females were exposed to both paper stimuli, at least two recordings in each terrarium's section were considered necessary for a trial to be valid. This presumption was fulfilled in all tests. At the end of the trials, the papers were removed and the cage was thoroughly rinsed with alcohol and clean water and let to dry outdoor.

To compare the number of observations (counts) of each female on the paper with femoral secretions of a male between temperature treatments (cold vs. warm), we used a Generalized Linear Model (GLZ) with a logit link and a Poisson's distribution (Green & Silverman 1994). Also, to test for differences between temperature treatments in the number of observations of females on the paper with femoral secretions of a male divided by the total number of observations on the three sections of the terraria, we used a logistic regression with a logit link and a binomial distribution (Hosmer & Lemeshow 1989).

Results

Changes in Climatic Conditions

Over a 42-year period, 1970–2011, the mean air temperatures rose significantly in May (Ta = −179 + 0·094*year; = 0·49, F1,40 = 12·77, = 0·0009) and in June (Ta = −190 + 0·10*year; = 0·66, F1,40 = 18·51, = 0·0001) (Fig. 1a). Similarly, the means of daily maximum air temperatures rose significantly in May (Ta = −257 + 0·13*year; = 0·56, F1,40 = 17·89, = 0·00013) and in June (Ta = −267 + 0·14*year; = 0·62, F1,40 = 25·48, < 0·0001) (Fig. 1b). Therefore, mean and maximum air temperature during the mating season of lizards showed a warming trend of 4·7–5·1 °C and 5·7–6·1 °C, respectively, over the period we examined (i.e. an increase of about 0·1 °C/year−1).

Figure 1.

Interannual variation from 1970 to 2011 in means of daily (a) mean and (b) maximum air temperatures in the study area (Guadarrama Mts., Central Spain) in the months of May (solid line, black dots) and June (dashed line, open dots), coinciding with the mating season of Iberolacerta cyreni lizards.

Chemosensory Responses of Females to Femoral Secretions of Males

The latency to the first TF differed significantly between types of stimulus and among time periods, and the interactions between stimulus and time and between stimulus and temperature were significant (Table 1; Fig. 2a). Latencies to blank control swabs did not differ significantly between time periods nor between temperature treatments (Tukey's tests, = 0·99 in all cases). In contrast, latencies to femoral secretions were significantly longer when temperature was warm and when the time since the secretion was deposited increased (Tukey's tests, < 0·05 in all cases). Latencies to the first TF were significantly longer to blank control swabs than to swabs with femoral secretions in the initial tests and 1 h after (Tukey's tests, < 0·0001 in all cases), but did not significantly differ 3 h after (> 0·40 in all cases) (Table 1; Fig. 2a).

Table 1. Results of full factorial Generalized Linear Models (GLMs) examining variation in latencies or number of directed tongue-flicks of the same individual females among treatments, with type of stimulus (control vs. femoral secretions), time (initial vs. 1 h vs. 3 h) and temperature treatment (cold vs. warm) as within-subject factors
Effectd.f.LatencyTongue-flicks
F P F P
Stimulus1,16199·53<0·00012520·35<0·0001
Time2,3248·52<0·000131·79<0·0001
Temperature1,161·790·208·96<0·001
Stimulus × Time2,3252·32<0·000185·69<0·0001
Stimulus × Temperature1,168·61<0·0110·65<0·005
Time × Temperature2,321·640·212·560·09
Stimulus × Time × Temperature2,321·230·302·200·13
Figure 2.

Mean (± SE) (a) latency and (b) number of tongue-flicks (TF) directed to swabs by female Iberolacerta cyreni lizards in response to control blank cotton-tipped applicators (triangles) or swabs bearing femoral secretions of males (circles) immediately after they were secreted (‘initial’), or 1 h or 3 h since deposition and maintained under two temperature regimes (cold: white; warm: black).

The number of TF directed by females differed significantly between types of stimulus, among time periods, and between temperatures, and the interactions between stimulus and time and between stimulus and temperature were significant (Table 1; Fig. 2b). Females made significantly less TFs directed to blank control swabs than to swabs bearing femoral secretions. Responses to blank control swabs did not differ significantly between time periods nor between temperature treatments (Tukey's tests, = 0·99 in all cases). However, females made significantly less TFs directed to femoral secretions as the time since secretions were deposited increased. Initial TF rates to femoral secretions did not vary significantly between temperatures (Tukey's test, = 0·99), but TF rates were significantly higher in the cold than in the warm treatment 1 h (< 0·05) and 3 h (= 0·008) since secretions were deposited (Table 1; Fig. 2b).

Female Choice of Areas Scent Marked By Males

The number of observations of each female on the paper with femoral secretions of a male depended significantly on the temperature treatment (GLZ with a Poisson's distribution, Wald's χ2 = 16·59, d.f. = 1, < 0·0001) (Fig. 3). There were also significant differences between temperature treatments in the probability of observation of a female on the paper scent marked by a male in relation to the total number of observations on the terrarium (Logistic regression with a binomial distribution, χ2 = 4·83, d.f. = 1, = 0·028).

Figure 3.

Mean (± SE) number of observations of female Iberolacerta cyreni lizards on sections of a terraria containing a paper scent marked by a male (black bars), or a blank control paper (open bars), both maintained under two temperature regimes (cold vs. warm) 30 min before the tests, or on a nonchoice area (hatched bars).

Discussion

Climatic data from these high mountains of Central Spain showed that there is a continuous increase in environmental temperatures during the last 40 years, at least during the spring months when the mating season of rock lizards occurs. This confirms the general trend noted and expected for Mediterranean mountains (IPCC 2007; Nogués-Bravo et al. 2008). Global warming could have serious consequences for I. cyreni lizards, which are adapted to the relatively cold temperatures prevailing in montane environments (Monasterio et al. 2009). In addition to the previously reported negative effects on lizard physiology and the higher risk of extinctions of lizard populations (Huey et al. 2009; Sinervo et al. 2010), our study shows that scent marks of male lizards will have a lower stability and persistence, resulting in a lower detectability by females. This loss of efficacy of chemical sexual signals may negatively affect social behaviour and sexual selection of rock lizards.

The chemosensory experiments indicated that females detected later (i.e. longer latency times) and had lower chemosensory TF responses to the femoral secretions of males when the time since deposition increased. This indicated that the chemical stimuli in secretions faded with time, very likely because chemical compounds that elicit responses evaporated and degraded with time since they were secreted (Epple et al. 1980). In addition, the loss of detectability and efficacy was faster under warm temperature. This is because high temperatures increase evaporation and diffusion rates of chemicals, affecting to their persistence (Regnier & Goodwin 1977; McDonough, Brown & Aller 1989). Therefore, under future warmer climatic conditions, the efficacy in communication of scent marks of rock lizards will decrease because their durability and persistence will be lower.

Consequences of The Loss of Efficacy of Sexual Signals

The experiment of selection of areas scent marked by males showed which can be the result of the lower detectability of chemical signals. When scent marks of males had been maintained under normal cold temperatures, females selected and spent more time on areas scent marked by males. This indicated that scent marks of males attract females to a given area, which may increase mating opportunities of males that had marked that area (Martín & López 2006a, 2012). In contrast, when scent marks were maintained under warm temperatures, the ‘attractive’ of scent marks decreased, very likely because decreased their detectability and persistence, and females did not select areas marked by males.

Moreover, also the information content of the signal may probably be lost quicker under high temperatures, with scent marks providing limited information about the characteristics of the signaller. Thus, it is likely that females may still detect that a degraded scent mark belongs to a conspecific male, but females might not discriminate more subtle details, such as body size, health state, etc., which can be readily discriminated from fresh secretions (López, Amo & Martín 2006; Martín & López 2006a, 2011). Similarly, the scent marks of saddle back tamarins (Saguinus fuscicollis) remain attractive to conspecifics for 1–3 days, but the monkeys discriminate between gender on the basis of fresh, 1- and 2-day-old marks but not on the basis of older marks (Epple et al. 1980).

What are the potential negative effects of this loss of efficacy of chemical sexual signals for I. cyreni lizards? Scent marks have a significant role in sexual selection in this and many other lizard species (Mason & Parker 2010; Martín & López 2011). Chemical signals are used in both intra- and intersexual relationships. If the efficacy of the chemical signal decreased, on the one hand, neighbour males entering a territory might not use degraded scent marks to identify the owner of the territory (i.e. ‘scent matching’ mechanism), which would increase costs of escalated agonistic fights (Carazo, Font & Desfilis 2007; López & Martín 2011). On the other hand, females would not obtain enough information on the quality of males and might not be attracted to the territories of the best potential partners (Martín & López 2006a, 2012). Therefore, if males of high quality invested in costly chemical signals for attracting females but these signals result unreliable, any male could access to females independently of his quality. Then, sexual selection would be disrupted because if mating is random, the strength of selection on male's quality signalled by chemical signals would be relaxed. Similar disruptions of sexual selection induced by anthropogenic environmental changes, such as chemical pollution and eutrophication, which render sexual signals ineffective, occur in other animals (Candolin, Salesto & Evers 2007; Lurling & Scheffer 2007). Therefore, global warming, through its effects on sexual selection pressures, could have long-term effects on the viability and evolution of the populations.

Compensation for the Loss of Efficacy of Sexual Signals

In contrast to the effects of global warming on thermal physiology of lizards, which could be partially compensated with flexible thermoregulatory behaviour (Kearney, Porter & Shine 2009; Aubret & Shine 2010), it is unlikely that changes in behaviour could compensate for the loss of efficacy of chemical signals. Male lizards might select to scent mark in shady or cold microhabitats to compensate for the higher temperatures, as some lizards do when selecting nesting places (Telemeco, Elphick & Shine 2009). However, this strategy would greatly restrict the surface of the areas that could be scent marked, decreasing the advertisement function of the signal. Another possible compensation might be to increase the frequency of scent marking to replace old scent marks (Martins et al. 2006). However, this would require increasing production of femoral secretions, which may be very costly, as secretions are made of lipids that, in many cases, can only be obtained from the diet and be diverted from other important metabolic functions (Martín & López 2006b; López, Gabirot & Martín 2009). Therefore, behavioural flexibility of signalling males is unlikely to be useful to compensate this negative effect of global warming.

There is, however, a potential behavioural change in the conspecific receivers of the signal that could allow compensating, at least partly, for the loss of efficiency of signals. Assuming that lizards could detect even very small concentrations of chemicals, but respond only when stimuli are above a given threshold, then some change in threshold could allow lizards to respond to smaller concentrations (e.g. partly evaporated or degraded scent marks). This might be possible if the threshold was plastic (e.g. influenced by previous experience or learning), although changes in the threshold might be precluded by physiological limitations in the vomeronasal system functioning.

Alternatively, lizards may shift their distribution to match changes in ambient temperature, or start their mating season earlier, as it occurs in other animals (e.g. Parmesan 2006; Moyes et al. 2011). However, in montane environments the only option for lizards is to disperse to higher elevations, and this would restrict greatly their geographical range. Similarly, an earlier initiation of the mating season may be limited by other factors such as photoperiod and insolation, which restrict activity of lizards in the high mountain. In particular, the amount of insolation, which affects the time available for effective basking, needed to achieve optimal body temperatures (Huey 1982), is lower in April (42·5 ± 1·8% isolation), when lizards emerge from hibernation, than during May (47·6 ± 1·9% isolation) and June (62·4 + 1·7% isolation), when the mating season occurs (climatic data from AEMET; see methods). This is due to a high number of cloudy days in April, a parameter that has not changed in this area in the last 42 years (= 0·12, F1,40 = 0·66, = 0·50). Therefore, shifting ranges or the phenology may be unsuitable or too costly in this species.

The loss of efficacy of chemical signals could probably be compensated only with phenotypic plasticity or evolutionary changes. In fact, there are variations in chemical signals between species or populations inhabiting areas with different environmental conditions (Escobar et al. 2003; Martín & López 2006c). For example, different populations of Podarcis hispanica lizards differ in the proportion of compounds in secretions, which seems related to different climatic regimes (Martín & López 2006c). Lizards from warmer areas secrete less volatile compounds and with a higher chemical stability, which may counteract a higher volatilization and decomposition (Alberts 1992; Martín & López 2006c). Lizards could also increase the number or size of femoral pores to enhance the overall production of secretions. Thus, species of Liolaemus lizards inhabiting warm habitats have more femoral glands and produce greater amounts of secretions than species of Liolaemus inhabiting colder environments (Escobar, Labra & Niemeyer 2001). Similarly, Psammodromus algirus lizards from warmer low-elevation populations have larger pores, which could increase the effectiveness to spread the femoral secretions, than lizards from colder high-elevation populations (Iraeta et al. 2011). Finally, when unfavourable environmental conditions render chemical signals very costly or not useful, an alternative communication system (i.e. visual) may be favoured (Fox & Shipman 2003). Therefore, evolutionary changes in the chemical signals or in secretory glands can occur as an adaptation to different environments. The question that arises is whether lizards could be able to respond fast enough to the quick environmental changes imposed by global warming (Visser 2008; Hoffman & Sgrò 2011; Sih, Ferrari & Harris 2011). In addition, it is expected that global warming would induce changes in mean temperature but also in variance and intensity of accidental brief increases in temperature. Therefore, lizards could not be able to predict when they have to change their behavioural strategies and neither selective pressures could induce consistent evolutionary changes allowing to respond to any environmental situation.

In summary, in addition to the previously described effects of global warming on thermal physiology of lizards, our study shows that global warming could also decrease the efficacy of their sexual signals. This might disrupt sexual selection, affecting quality of the offspring and survivorship of populations. Moreover, it is unlikely that the loss of efficacy of signals could be compensated with behavioural changes, and, at least in these mountain lizards, shifts in geographical ranges or phenology will be unsuitable or too costly. Only evolutionary changes or phenotypic plasticity could change the chemical signals, or the glands that produce them, as it may occur in the evolutionary time in response to different environments. However, it is also unlikely that these changes will occur fast enough to keep up with the quick human-induced changing environment. More studies are needed to analyse the actual consequences of global warming on the biology and conservation of mountain lizards.

Acknowledgements

We thank two anonymous reviewers for helpful comments and ‘El Ventorrillo’ MNCN Field Station for use of their facilities. Financial support was provided by the project MICIIN-CGL2011-24150/BOS. The Spanish Meteorological Agency ‘Agencia Estatal de Metereología (AEMET)’ kindly provided climatic data. Captures and observations of lizards were performed under licence (permit numbers: 10/111708.9/10 and 10/142790.9/11) from the Environmental Agency of Madrid Government (‘Consejería del Medio Ambiente de la Comunidad de Madrid’, Spain).

Ancillary