Comparing the predatory impact of captive – bred and free – living yellow spotted mountain newt ( Neurergus microspilotus ) on the larval green toad ( Bufotes variabilis )

espanolLa cria en cautividad de especies amenazadas es un practica importante en la conservacion, pero no esta claro si los programas de cria en cautividad a largo plazo pueden afectar a determinados atributos fundamentales de la eficacia biologica de las especies como la capacidad de depredacion. En el presente estudio se estudian experimentalmente los efectos de la capacidad de depredacion de tritones del Kurdistan adultos nacidos en cautividad (CBN) y de individuos obtenidos directamente de la naturaleza (FLN) en la supervivencia y el crecimiento de renacuajos de sapo verde mantenidos en cuatro tratamientos de densidad de depredadores, que contenian cero, uno, dos y tres tritones. El grupo de tritones FLN mostro una tasa de depredacion rapida dependiente de la densidad de depredadores que conllevo que las tasas de supervivencia de los renacuajos en las densidades de control, baja, media y alta fueran, respectivamente, del 81%, el 74%, el 60% y el 17%. El grupo CBN tuvo una menor tasa media de depredacion sobre los renacuajos de B. variabilis, cuya tasa de supervivencia disminuyo hasta el 83%, el 81%, el 82% y el 77% para los tratamientos con cero, uno, dos y tres tritones, respectivamente. Sin embargo, a diferencia del grupo FLN, esta tasa de depredacion aumento significativamente con el tiempo y paso de 0,37 a 0,60 renacuajos por dia. Por otra parte, la tasa de crecimiento de los renacuajos criados con depredadores del grupo FLN fue significativamente superior a la de los renacuajos criados en contenedores de control sin depredadores. Como conclusion, nuestros resultados sugieren que la exposicion de individuos nacidos en cautividad del triton del Kurdistan a sus presas potenciales enriquece el medio ambiente y puede ser util para elaborar programas mas eficientes de cria en cautividad y reintroduccion de esta especie de anfibio muy amenazada. EnglishCaptive breeding of endangered species is an important conservation tool, but it is not clear how long–term captive breeding can influence fitness attributes such as predatory ability. We experimentally investigated the predatory impact of adult captive–bred newts (CBN) and adult free–living newts (FLN) on the survival and growth of larval green toad (B. variabilis) in four predator density treatments containing none, one, two, or three newts. FLNs performed a rapid density–dependent predation, yielding average survival rates of tadpoles in no, low, medium, and high densities to 81%, 74%, 60% and 17%, respectively. CBNs had an average lower predation rate on B. variabilis tadpoles with a decrease in survival rate of tadpoles to 83%, 81%, 82% and 77% for 0, 1, 2 and 3 predator treatments, respectively. However, contrary to FLNs, they exhibited a significant increase in predation rate with time from 0.37 to 0.60 tadpoles per day. In addition, the growth rate of tadpoles reared with predators for the FLN group was significantly higher than the growth rate of tadpoles reared in control containers without the predator. In conclusion, our findings suggest that exposing captive–born adult yellow spotted mountain newts to their potential prey enriches the environment, and may be a useful approach in the development of more efficient captive breeding and reintroduction programs for this highly endangered amphibian.

Comparing the predatory impact of captive-bred and free-living yellow spotted mountain newt (Neurergus microspilotus) on the larval green toad (Bufotes variabilis) T. Salehi, M. Sharifi Introduction Captive breeding programmes have been used for a number of endangered species to save them from extinction, but the long-term outcome of these programmes has not always been satisfactory for a variety of reasons (Hedrick and Fredrickson, 2008).Some captive breeding programmes in amphibians have been characterized by poor nutritional requirements for growth (Pough, 2007), reduced natural behaviours such as foraging (including finding, identifying, acquiring, and handling food) (Burghardt, 2013), inactivity in the natural environment (Keulen and Janssens, 2017), inability to recognize natural foods (Olfert et al., 1993), a variety of diseases due to nutritional deficiency (Densmore and Green, 2007), chytridiomycosis infections (Parto et al., 2013), unsuccessful reproduction (Browne and Zippel, 2007), loss of social interactions (Rabin, 2003), variations in morphology of natural skin coloration (Ogilvy et al., 2012), loss of anti-predator response to predators (Kraaijeveld-Smit et al., 2006), a reduced immune response (Keulen and Janssens, 2017) and losing genetic diversity due to inbreeding (Zippel et al., 2011).The potential survival of captiveborn amphibians after their reintroduction into the wild is one of the concerns (Michaels et al., 2014).
Although the above-mentioned studies cast doubts on the value of captive breeding and subsequent reintroduction as a conservation tool for threatened amphibian species (Griffiths and Pavajeau, 2008), there are always situations in which the captive breeding is the only conservation option available (Stuart et al., 2004).Harding et al. (2016) reviewed captive breeding programmes involving 213 amphibian species and found captive breeding affected various characteristics that could be taken into account in efforts to improve captivity management.Additionally, efforts to expand current understanding of ecology and behaviour of re-introduced species are growing (Pough, 2007).Moreover, research focusing on issues related to the release of captive-bred animals to the wild, such as acclimation to the new environments, pre-release health condition, genetic management, and long-term post-release monitoring, is greatly needed (Armstrong and Seddon, 2008).Questions regarding the importance of captive breeding and subsequent reintroduction of threatened species will continue to appear as more studies are carried out on these programmes (Armstrong and Seddon, 2008), and, especially in view of the practical irreversibility of many current threats to amphibians in their natural environments, which makes captive breeding and reintroduction indispensable conservation methods (Griffiths and Pavajeau, 2008;Harding et al., 2016).
The yellow spotted mountain newt (Neurergus microspilotus) is included on the IUCN Red List as a critically endangered species for many reasons: its very small area of occupancy in its breeding streams (˂ 10 km), fragmented habitats, continuing decline in the quality and extent of aquatic habitats, habitat degradation, drought, and the pet trade (Sharifi et al., 2009).A conservation management plan funded by the Mohamed bin Zayed Species Conservation Fund was initiated for this species in 2010.Part of this plan included the development of a captive breeding facility at Razi University, Kermanshah, Iran (Vaissi and Sharifi, 2015).The project also launched several field and laboratory studies aiming to provide critical information for the management plan.These studies include delimiting the species range (Afroosheh et al., 2016), feeding habits (Farasat and Sharifi, 2014), activity pattern and home range (Sharifi and Afroosheh, 2014), reports on emergent diseases such as chytridiomycosis (Parto et al., 2013;Sharifi et al., 2014) and red-leg syndrome (Parto et al., 2014), and, finally, a first trial re-introduction of captive-born newts (Sharifi and Vaissi, 2014).In addition, several laboratory studies have provided information on ontogenetic changes in spot configuration (Vaissi et al., 2017), cannibalism (Vaissi and Sharifi, 2016a), and growth and development of N. microspilotus in the captive breeding facility (Vaissi andSharifi, 2016a, 2016b).
N. microspilotus in its natural environment acts as a top predator and feeds on a variety of prey, including benthic macroinvertebrates, amphibian eggs and tadpoles (Farasat and Sharifi, 2014), while the adult captive-born newts consume readily available food such as blood worm (Glycera dibranchiate), earthworm (Lumbricus terrestris) and chopped mealworm (Tenebrio molitor) ad libitum.The green toad, Bufotes variabilis, is a common toad in Iran with a wide distribution in most temperate areas, including parts of the distribution of the N. microspilotus in western Iran (Dastansara et al., 2017).We hypothesized that captive born N. microspilotus may lose their predatory ability to search, seize and engulf their natural prey (Burghardt, 2013).But we also postulated that captive born N. microspilotus could restore their predatory capability if they encountered natural prey.We therefore aimed to measure the predatory impact of captive-born adults at different densities and compare their impact with the free living newts under similar laboratory conditions in order: i) to determine whether there is a difference in the predatory impact of captive-born and free-living newts by examining the survival and growth rate of tadpoles of green toad (Bufotes variabilis) as prey, and ii) to measure whether captive newts can increase their predatory ability on tadpoles over time.

Material and methods
To experimentally assess the predatory ability of captiveborn newts (CBN), we used eighteen adult individuals (6-year-old) with SVL (mean ± SD) of 64.37 ± 4.99 mm and body mass (mean ± SD) of 8.03 ± 1.40 g.The captive newt stock originated from two gravid females captured from Kavat stream (35º 21′ N, 46º 24′ E) on 5 February 2010.Experimental individuals were raised to maturity in aquaria boxes with electrical ventilators, waste pumps, and water temperature ranging from 15 to 21 ºC in summer and from 3 to 12 ºC in winter.Embryonic development (from oviposition to hatching) usually takes 28 d.Once hatched, larval newts were placed in separate aerated plastic containers (length x width x height: 22 × 16 × 12 cm) to prevent cannibalism, and they were fed by Artemia egg and shredded blood worms (Glycera dibranchiate).The larval period lasted six to eight months, until metamorphosis (loss of gills) was reached.Metamorphosed individuals were transferred to an aquarium (length x width x height: 49 × 21 × 16 cm) that contained small pebbles taken from their natural habitat to create approximately one third of their aquarium (300-350 cm 2 ) as a terrestrial habitat.Recently metamorphosed individuals left the water, mostly occupying the terrestrial portion of their aquaria.They were fed ad libitum with gradually larger food items (blood worms, earthworm, and live mealworms).Adults were kept together, numbering six to nine newts, in a single aquarium (length x width x height: 75 × 25 × 40 cm) filled with mosses and some aquatic plants, with a sex ratio of two males to one female.Eighteen wild-caught adult individuals (FLN): SVL (60.79 ± 5.04 mm) and body mass (7.59 ± 1.30 g), were collected from the same population (Kavat stream) on 14 May 2016 and brought into the laboratory.They were kept, for one week before testing, in an aerated aquarium containing dechlorinated tap water and allowed to acclimatize in the laboratory, in similar conditions to CBN individuals.
We collected three clutches of B. variabilis from Sarable wetland (34º 32' N, 47º 01' E), Kermanshah Province, Iran, in April 2016.Embryos were transferred to plastic containers until the experiments started, and they were allowed to develop and hatch in the laboratory in 12:12 light-dark cycles at approximately 20 ºC.Tadpoles were fed boiled spinach daily, and the water in all containers was changed weekly.The experiment was started when tadpoles were at the Gosner's developmental stages 25-26 (Gosner, 1960) and within the consumable size range for the predators.SVL and body mass at the start of the experiments were 1.33 ± 0.58 cm and 0.035 ± 0.004 g, n = 50.The surviving tadpoles and all eighteen FLN newts were released back to their original habitats after the experiments were finished.The study was conducted with the approval of the Razi University ethics committee under permit number: 3962022.
Our experimental design tests for two sources of variation: predator origin, involving two levels: CBN and FLN, and predator density, with three levels: low (L), medium (M), high (H) and a control (1, 2, 3 and 0 newts, respectively).This design determines eight different predator origin and density combinations (including controls) and was replicated three times, for a total of twenty-four containers.Each container (49 × 21 × 16 cm) was filled with 550 ml of dechlorinated tap water.Fifty larvae of green toad, B. variabilis, were added in each experimental container.Tadpoles and predators were visually matched for size so that the initial differences in size would be minimal.We started the experiment on 21 May 2016 and randomly selected tadpoles and the newts to be used in the twenty-four experimental containers of predator origin and predator density treatment combinations: CBF-H, CBF-M, CBF-L, FLN-H, FLN-M, FLN-L and controls.For each container, we added four similar size stones to provide spatial heterogeneity.All containers were covered with mosquito cloth mesh to prevent newts from crawling out of the containers.
Predators were weighed and measured prior to the start of the experiments.During the twenty-five days of the experiment, containers were checked daily, and the number of surviving tadpoles in each container was recorded.At each observation, dead and partly eaten tadpoles were removed from the containers and absent tadpoles were considered as preyed.Every four days, tadpole body length was measured in each container.Photographs were taken with a digital camera (Sony, DSCHX9V) on a tripod at a fixed height (30 cm).The tadpoles were put in a bowl which was located over a graph paper.Immediately after photography, the tadpoles were released into their containers.All pictures were loaded into Digimizer 4 software and the tadpole total length (TL) was measured by drawing a line from the tip of the snout to the tip of the tail.The containers were cleaned and water was renewed with dechlorinated tap water every eight days.During the experiment, tadpoles were fed boiled spinach every day ad libitum.
All variables were checked for normality and homogeneity using the Kolmogorov-Smirnov normality test and Levene's test, respectively.We used two-way analysis of variance (ANOVA) to examine the effect of both predator origin and predator density on predation rate (the mean number of preyed tadpoles in each container during twenty-five days) and growth rate of tadpoles total body length (mm/day).Tukey's HSD post hoc pairwise comparisons were used to determine differences in predation rate between different newt densities.We compared the average number of preyed tadpoles in five-day periods for CBNs using repeated measure analysis of variance (repeated measure ANOVA) to determine whether predatory ability can improve with time.All statistical analyses were carried out using SPSS statistical software, version 16.0 (SPSS Inc, 2007).The statistical significance level was 0.05.

Results
The results of two-way ANOVA indicated a significant effect of newt origin and density and their interaction on the predation rate (table 1).In all densities of newts, the predation rates on tadpoles were higher in FLNs than in CBNs (table 2).High predator density greatly decreased tadpole survival, but this decrease was more drastic in FLN newts than in CBN newts, as revealed by the significant interaction term.Tukey's post hoc test showed that the highest predation rate was found for the FLN high density treatment.A repeated measure ANOVA at 5-day intervals showed that the predatory ability in CBNs, although lower than that of FLNs, increased with time (F 4, 32 = 4.38, P = 0.006).FLN predation rate was maximal (18.11 ± 13.91 eaten tadpoles) in the first interval but did not increase later (fig.1B), whereas the predatory ability of CBNs was low at the beginning but increased over time (3.66 ± 2.50 eaten tadpoles in the first interval to 7.33 ± 4.30 eaten tadpoles in the third interval) (fig.1A).
In addition, two-way ANOVA showed that the predator origin and predator density did not have a significant effect on the growth rate of the green toad, B. variabilis (table 1).However, there was a significant difference between the predator origin and the control on the growth rate of tadpoles (ANOVA, P-value = 0.037).Tukey's post hoc test indicated that free-living predators caused a significant increase in the growth rate compared with the controls (P-value = 0.03).

Discussion
Our results are in agreement with our expectations.First, we found that newts raised in captivity had lower predation rates than the free-living newts.In addition, we observed that the predation rate in CBNs improved when they were exposed to the green toad B. variabilis tadpoles during the short experimental period (25 days).This finding suggests the predatory abilities of the captive stock newts increased.As in accordance with out study design we did not replace the consumed prey to keep the original prey densities constant, we could expect lower predation rates simply by reducing the chance of prey encounter by newts.However, the observation that CBNs newts increased their tadpole consumption rate is probably explained by increased capacities in prey recognition and hunting abilities over time.These results indicate that husbandry conditions of captive born newts can improve predatory ability by initiating prey-predator interaction in the breeding facilities.Under natural conditions, N. microspilotus is an active predator moving through the water column and ground feeding on macroinvertebrates, eggs and tadpoles (Farasat and Sharifi, 2014).As a result, active movements in aquatic environments increase the chances of capturing prey (Hossie and Murray, 2010).In our study, CBNs were fed several food items, such as earthworms, bloodworms and mealworms ad libitum, possibly meeting optimal nutritional need.However, such an approach may cause inactivity and obesity (McWilliams, 2008).Therefore, the initial assumption of predation rates being higher among FLN newts was demonstrated, probably because the prey capture efficiency and foraging was lower in the newts of captive origin.By comparing both captive-born and free-living N. microspilotus, we found both groups were similar in performing stereotype behavior of approaching, seizing and sudden engulfing tadpoles.However, the newts reared in the captive environment differed in their boldness and speed of their attack and showed less success in capturing the tadpoles than their free-living conspecifics (T.Salehi and M. Sharifi, pers. observ.).This may explain their constrasting predatory rates.
Secondly, this study shows that exposing captive-born adult newts to a potential prey improve their predatory ability, suggesting that an environmental enrichment such as that used in this study, can be effective in developing more efficient captive breeding and reintroduction programs in N. microspilotus.Burghardt (2013) and Michaels et al. (2014) reviewed studies on how changes in different aspects of the environmental enrichment may improve the individual welfare of captive-born amphibian species.For instance, feeding enrichment in Dendrobates tinctorius, D. azureus, D. auratus, D. leucomelas, and Oophaga pumilio increases their activity levels and foraging behaviour (Campbell-Palmer et al., 2006).A study on Golden Mantella frogs (Mantella aurantiaca) showed that a diverse diet and a variety of live invertebrates, together with vitamins and mineral supplementations, can affect body condition and normal behaviour (Pas-Table 1. Results of univariate ANOVA for assessing effects of predatory origin (captiveborn vs free-living) and density of newts (3, 2, and 1) on predation rate and prey growth rate.

Source of variation
In addition, to hide or scatter food items throughout captive rearing enclosures can encourage natural foraging behaviour and predatory ability (Poole and Grow, 2012).Summarizing determining best feeding regimes for captive-born species can provide opportunities to mimic the situation where these species may be introduced into the wild (Keulen and Janssens, 2017).Interaction between predator origin and predator density influences tadpole predation rates.Predator density decreases prey survival in FLN newts but not in the captive reared newts.This will imply a much lower individual ingestion rate at higher densities in captive origin newts, which may have implications on net growth and breeding performance.A range of spatial organizations with increasing densities of amphibians and reptiles in the wild has been reported, but captive environment changes these organizations due to spatial constraints (Hayes et al., 1998).Thus, a lower ingestion rate in high density CBNs may help to choose lower raising densities in order to enhance growth rates and ultimately increase the viability of the captive stock.
The effects of the density of predatory newts on anuran larvae as prey are not always predictable.Morin (1986) has shown that relative abundance of larval anurans may be positively or negatively altered by densities of predatory salamanders in pond communities.Morin (1983) showed that relative abundances of prey tadpoles of some species (Scaphiopus holbrooki, Rana sphenocephala, Bufo terrestris, and Hyla chrysoscelis) at high predator salamander densities (Notophthalmus viridescens dorsalis) were reduced, whereas the relative abundances of H. crucifer and H. gratiosa tadpoles in the presence of newts were not affected.Various explanations have been proposed for the differences found in the intensities of salamander predation on amphibian tadpoles (Morin, 1986).Individual predators may influence one another's predation rate when foraging on prey (Ramos and Van Buskirk, 2012), or changes in prey behaviour (Charnov et al., 1976) which can reduce predator success at higher predator densities.Our results suggest that predatory ability on tadpoles increased with density in the FLN group, thus suggesting that individual predators did not affect each other.One possibility of larval salamander interference is cannibalism reported for larvae of N. microspilotus (Vaissi et al., 2017).However, neither cannibalism nor  Overall, predator origin and density did not influence the growth rate of surviving tadpoles.However, by comparing the growth rate of tadpoles between control and predator treatments, we observed that predators induced higher and significant growth rates in FLN newts.Current evidence suggests that the presence and origin of predators have a great influence on prey growth rates (e.g.McPeek et al., 2001;Relyea, 2004;Van Buskirk, 2009).Prey responses to predators are widespread among larval amphibians.Rapid growth, for example, may reduce or even avoid larval predation by becoming too large for predators (Brodie and Formanowicz, 1983).However, other evidence has indicated that the presence of the non-lethal predator may induce reductions in growth rate (Van Buskirk, 2002).The increase in tadpole size in the presence of predators may be an adaptive strategy of prey known as the 'gap limited predator', a mechanism that reduces predation intensity when a predator cannot consume larger tadpoles.Another possible explanation for the higher growth rate in prey species is 'thinning', that is, an indirect effect of predators by removing competitors, leading to an increase in the per capita food resources for surviving prey (Abrams and Rowe, 1996).Although tadpole survival decreased at high FLN newt density, we did not find an expected increase in growth rates at high density, suggesting that no thinning mechanism is occurring.A possible explanation for the absence of competition in tadpoles under high densities of newt predators could be a strong reduction in activity by surviving tadpoles that will reduce their ingestion rates and, therefore, exhibiting a decline in their growth rates (e.g.McPeek et al., 2001).In conclusion, the significant increase in growth rate of FLN tadpoles suggests that N. microspilotus may be a 'gap limited predator' able to reduce predation intensity when they cannot consume large tadpoles.
In conclusion, our results emphasize that exposing captive-born adult N. microspilotus newts to their potential prey can enrich the environment and may be a promising approach in the development of efficient captive breeding and reintroduction programs for this highly endangered amphibian.

Fig. 1 .
Fig. 1.Average and standard deviation of the predation rate (total number of tadpoles consumed in 5-day intervals: by captive-bred (A) and free-living (B) Neurergus microspilotus exposed to different treatments (H, 3 newts; M, 2 newts; L, 1 newt) over the twenty-five days of the experiment.

Table 2 .
Mean ± SD for predation rate (%), and growth rate (mm/day) on fifty larvae of green toad, Bufotes variabilis under different predator origin (CBN and FLN) and predator density (High, Medium and Low) during the twenty-five days of the experiment.