Competition and Environmental Stress Impacts on Trophic Performance of Three Sympatric Insectivorous Lizard Species in Eastern Spain

Abstract

Trophic ecology is an important aspect to consider when studying interactions between species, especially in ecologically similar species. We studied the trophic ecology of three sympatric insectivorous lizards in a dune system in the Eastern Iberian Peninsula: Acanthodactylus erythrurus, Psammodromus algirus and Psammodromus edwardsianus. We obtained a total of 485 faecal samples and found 18 different prey groups. The trophic niche breath analysis showed that A. erythrurus was the most specialised species of the three. We also considered two different habitat types and, interestingly, both Psammodromus species had wider trophic niches in the more extreme habitat type where arthropod diversity is expected to be lower. Trophic niche overlaps were especially low between P. algirus and A. erythrurus, indicating resource partitioning, and higher between both Psammodromus species in the suboptimal habitat type. Our results lead to the conclusion that environmental stress could favour trophic generalism (increased trophic niche breadth). This is a very interesting result, especially in the context of climate change and habitat alteration.

1. Introduction

Species within an ecosystem interact in many very different ways, but one of the most important connections between species is through trophic interactions. Trophic ecology is not limited to only describing the diet of a species; there are numerous factors affecting it. For instance, there can be variations in diet from one population to another of the same species [1], or even within the same population, given that sex, body size or condition or the phenology can lead to diet variations [2,3,4]. Also, environmental factors can play a major role in the trophic ecology of a species. For example, the presence of other species with a similar trophic niche can result in trophic competition, which affects the general ecological performance of the species, like the competitive interaction between insectivorous birds and lizards on islands in Lake Gatun (Panama), where bird and lizard abundances are negatively correlated [5]. This competition can be avoided by resource partitioning and a trophic specialisation. Also, abiotic factors like temperature, humidity or insulation can affect the trophic performance of a species directly (e.g., affecting prey handling capacity) or indirectly (e.g., affecting prey availability), especially in ectotherm organisms.

Modern studies, in this case in frogs, suggest that niche breadth variations at the individual level are important drivers of general niche dynamics at the population level, crucial for understanding the coexistence of several ecological similar species [6]. In this context, lizards also appear to be excellent study targets, ass they are widely used for ecological studies due to traits like generally high abundances, easy capturing, marking and monitoring [7]. There are studies showing that, usually, lizard communities do not tend to partition resources on the trophic axis [8] but on the spatial [9] or temporal one [10]. However, there are also studies indicating the importance of prey selection and the numerous factors that influence it, thus clearly indicating that prey selection and diet, even in generalist species, are far from being a mere random side effect of habitat use [11]. Thus, it would be very interesting to study the interaction between several lizard species feeding on the same prey pool in two different environmental conditions, where each lizard species has its own distinctive traits, being thermal ecology, body size, hunting strategy, phenology, population dynamics, etc., in order to better understand how these factors and traits can affect prey selection, trophic niche breath or trophic niche overlap.

In the study area, three species of lizards (all belong to the Lacertidae family) cohabited: the spiny-footed lizard Acanthodactylus erythrurus (Schinz, 1834), the Algerian psammodromus Psammodromus algirus (Linnaeus, 1758) and the Edward’s sand racer Psammodromus edwardsianus (Dugès, 1829). These species share the same microhabitats and the same trophic resource. Thus, we would expect a high level of competition between the species, especially in cases when resources are limited. There are also studies that describe a higher specific aggression between lizard species in syntopy [12]. This competition, which would clearly reduce the fitness of the involved species, can be avoided reducing the trophic niche breadth (specialisation) in order to decrease the degree of niche overlap [6,13]. There are numerous previous works on the diet of the three study species: [14] for A. erythrurus, Ref. [15] for P. algirus and [16] for P. edwardsianus. In some cases, the authors even considered the ecological interaction between A. erythrurus and P. algirus, showing that there is indeed a resource partitioning between the two species [10,17]. But most studies are purely descriptive and limited to only one species, failing to capture a larger ecological picture.

Regarding the distinctive traits of each species, A. erythrurus can be considered the more trophically specialised species, as it is known that myrmecophagy is an important part of a species’ diet [18]. It is also the more thermophile of the three species [19,20]. P. algirus is the species with the longest lifespan [16,21,22], and it stands out by a different foraging and hunting strategy as it usually climbs branches to hunt within the vegetation, while the other two species mainly hunt on the ground [18,23]. P. edwardsianus (former P. hispanicus) is the smallest of the three species and has the shortest lifespan (a near 100% renewal of the population each year) [24].

The aim of this study was to analyse the differences in trophic niche breadths and overlaps between the three sympatric species, and how the different traits and environmental factors affect the trophic performance of each species. The results are expected to give additional insights to better understand the coexistence of ecologically similar species.

2. Material and Methods

2.1. Study Area

The study area is situated about 10 km South from Valencia city and is part of the Albufera de Valencia Natural Park (39°20′20″ N 0°18′43″ W). It represents a coastal line about 10 km long (N–S) and 1 km wide (E–W) (Figure 1). It is a very human-frequented area; in summer, there is much activity related to touristic use of the beach, but it is also an area with a great interest in the ecotouristic and cycling sectors because of its distinct and naturalised landscapes.

The area presents a typical Mediterranean vegetation. It is a dune habitat with an area of dunes with a variable degree of maturation and a vegetation formed mainly by herbaceous and bush species, generally less than 1 m tall, like Ammophila arenaria, Helichrysum stoechas, Euphorbia paralias, Medicago marina or Rhamnus alaternus, among others (henceforth called “Dune” habitat). Behind this habitat, there is an area of fixed dunes with an arboreal vegetation formed mainly by Aleppo pine (Pinus halepensis) and a dense undergrowth with species like Smilax aspera, Asparagus officinalis, Chamaerops humilis or Pistacia lentiscus (henceforth called “forest” habitat) (Figure 1).

Regarding other potentially competing reptile species, the study area is quite isolated from other environments by the Albufera wetland and the only other reptile species detected in the area (except tortoises) were non-insectivorous snakes (Malpolon monspessulanus and Zamenis scalaris), Tarentola mauritanica and Chalcides bedriagai, which are also insectivorous but live in other microhabitats than the studied lizards and are quite rare in the study area. Thus, the collected dietary data from the latter species are not sufficient to be included in this study.

2.2. Fieldwork

Field samplings were carried out between April and October (both included) of 2015 and 2017. In 2015, we obtained one sampling every two weeks in each habitat type, while in 2017, we intensified the sampling effort to four times a week (two in each habitat type).

Every sampling had a duration of about three hours, beginning about three hours after sunrise. We captured lizards by hand or noosing [25]. We put the captured specimens in individual cloth bags until the later processing. At the end of the sampling, we measured the snout–vent length or SVL (to the nearest 0.5 mm), and we identified the species, age class and sex every time possible. Sexes were identified by the development of femoral pores, the presence of hemipenis and a widening of the base of the tail or by palpation of eggs in development inside females (if possible). We obtained faecal samples deposited in the cloth bags or by performing a gentle abdominal massage on each individual. Before releasing the lizards near the corresponding capture point we marked them with a permanent black marker at the ventral surface of the head, in order to recognize recaptures which were not included in this study to avoid self-replication. This marking does not induce any pain nor has any negative effect on the lizards, and it lasts until the next shedding (normally several weeks to months) [26]. We kept the faecal samples frozen at −20 °C to avoid mould until their processing in the laboratory.

2.3. Fecal Sample Processing and Data Analysis

All obtained samples were analysed in the laboratory by the authors. We put every sample on a petri dish with some drops of water, and carefully separated the different items with tweezers. Then, using a binocular magnifier and with the help of an arthropod identification guide [27], we identified every item found in every sample, at least to the order level. Finally, we conserved every sample in an Eppendorf tube with 70% ethanol.

We could distinguish a total of 18 groups (prey groups), which, in general, correspond to arthropod orders, except some lower levels, considered separately because of their peculiar characteristics. For instance, we considered the family Formicidae apart from the rest of Hymenoptera given their small size and very high abundance in the environment; and Scorpiones from the rest of Arachnida given their harder exoskeleton, and generally bigger body size. With the obtained data, we built a presence/absence matrix of each prey group in each sample, assigning 1 to presence and 0 to absence. We noticed that our sample sizes are quite uneven across the different groupings (for example, larger n of A. erythurus, compared to the other species). To avoid this bias, we plotted the saturation curves in each case, and all of them showed that with the current sample size reached the asymptotic part of the curve, confirming that most of the diversity of prey groups was detected (Figure S1).

For the study of diet composition, in each comparison, we calculated the proportion of presence of each prey group with respect to the total of the samples. In order to evidence the statistical significance of the observed differences, we performed a non-parametric Krushkal–Wallis test [28] for each prey group in each comparison.

To study the similarity of the diet composition between species and habitats, we calculated the Bray–Curtis similarity index (in percentage), following Wolda [29] and Somerfield [30]:

$${\mathrm{BC}}_S=\left(1-{\displaystyle \frac{\sum \left|n_{1i}-n_{2i}\right|}{\sum \left|n_{1i}{+\mathrm{n}}_{2\mathrm{i}}\right|}}\right)\ast 100$$

where n_1i and n_2i are the abundances of each prey group (number of individuals where the prey group was present) in each species.

For the study of the trophic ecology, we used the same method as in Sasa and Monrós [31]. We calculated niche breadth in each case as the Levins index [32], using the following formula:

$$\mathrm{B}={\displaystyle \frac{1}{{\sum }_{\mathrm{i}=1}^n{p}_i^2}}$$

where p_i is the proportion of the presence of each prey group with respect to the total. In order to facilitate the interpretation of the results, we standardized the Levins index, following Serafini and Lovari [33]:

$$\mathrm{Bs}={\displaystyle \frac{B-1}{\mathrm{Bmax}-1}}$$

where B is the obtained Levins index, Bmax is the total number of prey groups (18) and Bs is a value between 0 (lowest niche breadth) and 1 (greatest niche breadth).

Once we obtained the values of trophic niche breadth, we proceeded to calculate the degree of niche overlap between species and habitats, using the Pianka index [34], applying the following formula:

$$\mathrm{O}={\displaystyle \frac{\sum p_1\ast p_2}{\sqrt{\sum p_1^2\ast \sum p_2^2}}}$$

where p₁ and p₂ are the proportions of a certain prey group in groupings 1 and 2, respectively. In order to test the randomness of our results, we carried out 30,000 Monte Carlo simulations for each comparison [35], randomizing the utilization matrix with the RA3 algorithm from Lawlor [36], and compared our data to the distribution of the randomized data performing a two-tailed Z-test, considering a significance level of 0.05. All statistical analyses were carried out using R [37] v3.4.1.

3. Results

3.1. Comparison of Diet Between Species

During the samplings, we obtained a total of 485 samples: 246 from A. erythrurus, 101 from P. algirus and 138 from P. edwardsianus. The analysis of diet composition of the three species showed differences in the consumption of the different prey groups (Figure 2). In the case of A. erythrurus, we can remark the proportions of non-Formicidae Hymenoptera, Formicidae, Coleoptera and Hemiptera as the major groups present. In the case of P. algirus, we observed maximum values in Lepidoptera and non-Scorpiones Arachnida, followed by a relatively equivalent composition of Coleoptera, Diptera, Hemiptera and Orthoptera. We can also remark that this species consumes the highest proportion of Scorpiones from the three studied lizards, with the remaining in approximately 15% of the samples. Finally, in the case of P. edwardsianus, the most consumed prey groups were Coleoptera, Hemiptera and non-Scorpiones Arachnida. This species was also the only one consuming Blattodea oothecae (Figure 2).

3.2. Descriptive Diet Analysis by Species

From the detailed diet composition analysis of A. erythrurus, we can remark that there seems to be an increase in Hymenoptera consumption as individuals grow, going from values of about 25% in small individuals (SVL 30–39 mm) to values of about 60% in big individuals (SVL 70–79 mm). The opposite occurs in Formicidae, where the proportions in small individuals (SVL 30–39 mm) lie at about 60%, while in big individuals (SVL 70–79 mm), it is reduced to 30%. In the case of Coleoptera, we can observe the relatively low proportion in small individuals (SVL 30–39 mm) (Figure 3).

When comparing the diet of A. erythrurus in the two different habitat types, we found differences in the proportions of Formicidae, Coleoptera and Diplopoda, with Formicidae being more abundant in the dune habitat and Coleoptera more abundant in the forest habitat. Diplopoda was only found in the forest habitat. Finally, the phenological analysis of diet composition showed differences in the consumption of Scorpiones, non-Formicidae Hymenoptera, Formicidae, Coleoptera and Hemiptera. In this case, Scorpiones presented maximum values in April; non-Formicidae Hymenoptera were consumed more in spring and the beginning of summer; Coleoptera also had maximum values in spring (reaching values up to 90% in April) and Hemiptera, in contrast, were consumed more in summer (with values up to 80% in July). We also detected individual cases of the ingestion of vegetal items (small sticks, flowers, seeds), which were not included in the analysis, as we think they were accidental ingestions while hunting prey sitting on the vegetation.

In the case of P. algirus, we can remark on the consumption of Scorpiones, which were more consumed by large individuals (SVL > 50 mm), and Diptera, more consumed by small individuals (SVL < 50 mm) (Figure 4). In the comparison of diet composition between habitats, we found differences in the consumption of Diptera, presenting higher values (about 35%) in the dune habitat, and Hemiptera, presenting higher values in the forest habitat (about 40%). Finally, the phenological analysis showed differences in Arachnida (being especially abundant in April, with about 65%), Scorpiones (being quite abundant in June, with more than 50%) and Coleoptera (being less abundant in June, August and October) (Figure 4). We also recorded individual ingestions of plant matter and some snail shells, what we interpret as accidental ingestions, and thus did not include it in the analysis. In addition, we recorded one case of interspecific predation: an adult male of P. algirus consumed a juvenile of A. erythrurus.

Finally, in the detailed analysis of diet composition in P. edwardsianus, considering body size, we can observe differences in the consumption of Coleoptera, which were more consumed by large individuals (SVL > 44 mm) (Figure 5). The comparison of diet composition in the different habitat types showed differences in the consumption of Formicidae, Coleoptera and Diptera. While Coleoptera were especially abundant in the forest habitat (about 65%), Formicidae and Diptera were more abundant in the diet of individuals in the dune habitat (Figure 5).

3.3. Niche Breadth and Overlap

The analysis of trophic niche breadth shows that all three species have similar niche breadths considering the overall data, with the values for P. algirus being slightly higher. We did not observe great differences between sexes in neither species. We did detect noticeable differences comparing habitat types, with both Psammodromus species having greater niche breadths in the dunes habitat (almost doubling the values in the case of P. edwardsianus), while in the case of A. erythrurus, the values for both habitats were practically identical (

Table 1. Sample size (n) and the values of the standardized Levins index (Bs) for the overall data of each species, and considering the data of each sex and habitat type separately.

			n	Bs
A. erythrurus			246	0.34
P. algirus			101	0.41
P. edwardsianus			138	0.34
Sex	A. erythrurus	F	37	0.31
		M	139	0.35
	P. algirus	F	20	0.44
		M	22	0.4
	P. edwardsianus	F	22	0.33
		M	40	0.35
Habitat	A. erythrurus	forest	142	0.33
		dunes	104	0.34
	P. algirus	forest	32	0.33
		dunes	69	0.39
	P. edwardsianus	forest	82	0.25
		dunes	56	0.42

).

The comparison of niche overlap values to the mean of the randomized distribution showed that our results were significantly higher than random values in all cases (Z-test, p < 0.001, for all comparisons) (

Table 2. Values of the Pianka trophic niche overlap and the Bray–Curtis similarity index (respectively above the diagonal), and the mean and standard deviation of the Monte Carlo simulations (below the diagonal) for the overall data of each species and the data for each habitat type separately, and also the comparison between habitats for each species.

Overall Data	A. erythrurus	P. algirus	P. edwardsianus
A. erythrurus	-	0.61 | 36.83	0.82 | 54.94
P. algirus	0.41 ± 0.14	-	0.81 | 69.07
P. edwardsianus	0.37 ± 0.15	0.40 ± 0.14	-
forest
A. erythrurus	-	0.78 | 32.39	0.83 | 65.22
P. algirus	0.35 ± 0.16	-	0.85 | 50.43
P. edwardsianus	0.32 ± 0.16	0.31 ± 0.17	-
dunes
A. erythrurus	-	0.50 | 36.40	0.78 | 38.77
P. algirus	0.39 ± 0.15	-	0.84 | 67.73
P. edwardsianus	0.40 ± 0.15	0.43 ± 0.14	-
forest vs. dunes
Pianka | B-C	0.96 | 79.89	0.82 | 62.00	0.84 | 66.67
Monte Carlo	0.36 ± 0.16	0.37 ± 0.15	0.36 ± 0.15

). The analysis of trophic niche overlap shows that the most important differences can be observed between P. algirus and A. erythrurus, independent of the habitat type. In the case of P. edwardsianus, the difference with A. erythrurus depends on the habitat type, being greater in the dune habitat. Considering both Psammodromus species, they generally present similar diets, especially in the dune habitat. Comparing both habitat types for each species, we observed that A. erythrurus presented very similar diets in both habitats, while P. algirus was the species with the most divergent diet between habitats (Table 2).

4. Discussion

This study used faecal samples to analyse the diet of three lizard species, which has the advantage of being able to gather and process a large number of samples without harming the lizards. However, this method also comes at the cost of not being able to identify prey parts at lower levels than the order level (with some exceptions). Other methodologies, with a higher resolution, may render different values of niche breadth and overlap, allowing for alternative interpretations. The overall diet analysis accords to data obtained in previous studies in the case of A. erythrurus [38]. The major prey groups in this species were non-Formicidae Hymenoptera, Formicidae, Coleoptera and Hemiptera, the same as in a similar study carried out in Alicante [17], although our results show a higher diversity of minor prey groups. In the case of P. algirus, we observed some discrepancies with the results obtained in previous works. The most remarkable is the relatively high proportion of Lepidoptera, a group which has not been described as so important in the diet of this species, yet, in fact, in some studies it is not even mentioned [39,40]. In addition, we found other minor prey groups which were not described previously in the diet of P. algirus, like Scorpiones and Neuroptera. The single case of predation between P. algirus and A. erythrurus observed in this study is also coherent with previous observations by other authors [41]. However, the very low rate of such events detected lead to the conclusion that, although they happen eventually, there does not seem to be a noticeable predation pressure and a consequent implication for the community structure. Finally, in the case of P. edwardsianus, our results, in general, adjust to the ones obtained in previous studies, although we can outline some differences. On the one hand, non-Formicidae Hymenoptera had been described as a very abundant prey group in this species [24], but in our case, it was more a minor prey group. On the other hand, in none of the previous studies was the consumption of Blattodea oothecae (species unknown) by P. edwardsianus described, while in our case, they appeared in about 15% of the samples.

In our case, all trophic overlaps were higher than expected from a random population. Although a worldwide study of lizard communities and their trophic resource partitioning in relation to interspecific competition showed that resource partitioning is rare and appears mostly in tropical ecosystems [8], high values of niche overlap have the potential of strong competition when resource availability is limited [42]. Our results regarding trophic niche overlap show that there was a resource partitioning between the species, at least between A. erythrururs and P. algirus. While the first tended to consume a greater proportion of hard preys like Coleoptera and Hemiptera, the second one, although it also consumes these prey groups, tended to consume more soft preys like Lepidoptera, Arachnida or Diptera. These differences were especially important in the dune habitat. These results are interesting given that in another area, where both species coexist, a high degree of trophic niche overlap and diet similarity was described [10]. But in older studies [15,17], the authors found niche overlaps of 63% and 66%, respectively, very similar values to our case (61%). In his work, Seva [17] explains these differences in diet composition by differences in feeding strategies. A. erythrurus is more specialized (mirmecophagy) and P. algirus is more opportunistic. Additionally, there is a behavioural factor, as P. algirus tends to climb through the vegetation, while A. erythrurus remains mostly on the ground. Finally, A. erythrurus has a relatively bigger head than the other two species which grants it a higher bite force [43] and, subsequently, a domination in interspecific encounters.

The values of trophic niche breadth seem to corroborate that A. erythrururs is the species with the most specialized diet. It presents the lowest values of niche breadth of all three species and it also has the highest intraspecific niche overlap. This is coherent with previous studies [17]. The niche overlap of both Psammodromus species with A. erythrurus was especially low in the dune habitat. Comparing both habitat types, our results show that A. erythrururs presented very similar values in both habitats, and in the other two species, the most diverse diet was found in the dune habitat. This is interesting, as this habitat presents extremer environmental characteristics, with less vegetation and a high degree of insulation which leads to higher temperatures, especially in summer. Sadly, we do not have data on arthropod diversity, but it is known that the abundance and diversity of arthropods tend to increase in dune systems with increasing complexity of the vegetation [44,45]. Our data, in contrast, show the complete opposite. One possible explanation could be the temperature, which plays a very important role, especially in ectothermal organisms, and it was demonstrated that it also affects foraging and feeding behaviour, and thus diet composition [20,46]. It is also known that A. erythrurus is a much more thermophile species compared to P. algirus and P. edwardsianus [19,47,48,49]. Thus, a dune environment can be considered an optimal environment for this species, while for the other two species, this type of habitat presents suboptimal conditions [48].

These results seem to indicate that an extremer environment, and the resulting environmental stress, induces a lower selectivity in the diet of both Psammodromus species, favouring opportunism, and thus increasing the trophic niche breath. This has been described in other organisms, like the snake Natrix tessellata [50], where an increase in trophic niche breadth facilitates the survival in suboptimal habitats. Furthermore, recently, it was described in insectivorous Sceloporus lizards [51]. In this case, the authors found that prey diversity was higher during the more adverse (dry) season where prey availability is expected to be lower. Other possible explanations could be different foraging strategies or adaptation to predation risk. It is known that P. algirus and P. edwardsianus can be considered more as active foragers, while A. erythrurus has a more sit-and-wait character [52]. However, if this factor prevails, we expect the same pattern, independent of the habitat type, and thus similar high overlap values between habitats for all species. It could also be argued that because the dune habitat is more open, predation risk is higher. This could induce a dietary shift towards smaller, easier-to-handle prey [53]. If this was true, we would expect a higher effect on large species (as larger individuals are more prone to predation) and also a reduction in trophic niche breadth in the dune habitat, as the prey spectrum would be size limited. In our case, we observed neither of these patterns.

There is another (physiological) factor that can be discussed as it affects diet composition: bite force. It is already known that A. erythrurus has a generally higher bite force than P. algirus [43], and this grants it better access to harder prey types. In P. algirus, the proportions of hard preys (Orthoptera, Hemiptera, Coleoptera) increase with body size, while soft preys (Diptera, Arachnida, Lepidoptera) decrease with body size. The same can be observed in P. edwardsianus, where the proportion of Coleoptera increases with body size and the proportion of Lepidoptera decreases. In this context, we can also remark that although it had been described that A. erythrurus has a marked tendency to mirmecophagy [15,17], the consumption of Formicidae seems to be especially important for juvenile individuals, appearing in over 60% of the samples. In adult individuals (>60 mm SVL), the proportion of Formicidae in the diet decreases about 20–30%. This dietary shift has been observed before and, although the exact reason remains unknown, a possible explanation could be that A. erythrurus originally evolved in arid environments, where Formicidae tend to be more abundant than other insects. Thus, juveniles hunt instinctively for “ant-like” prey, but with more experience, they start to hunt for other, energetically more efficient prey, like Coleoptera or other Hymenoptera [11,14].

This leads to the conclusion that in unfavourable periods, environmental factors like the temperature seem to be mainly responsible for shaping the diet, while in favourable periods, when species can choose their prey, they do it based on internal factors (like bite force). This is in accordance with previous works; for example, it was described for Zootoca vivipara, a species preferring colder and wetter (atlantic) environments, that diet diversity increases during summer [11]. Supporting this hypothesis, we can conclude that the two Psammodromus species presented in the forest habitat, which is preferred and optimal for both species, more different diets than in the dune habitat where both species may suffer environmental stress, and present a more generalist character, very similar results to the previously mentioned case of Sceloporus lizards [51]. If this hypothesis is further confirmed, it could have crucial importance in the context of climate change and species conservation.