The experiment was conducted between March and June 2016 in the laboratory of the Vivarium of the Babes-Bolyai University, Cluj-Napoca, Romania and consisted of two components:

1. The larval experiment: the assessment of the feeding regime on tadpole growth and development.
2. The carry-over experiment: the short-term carry-over effect of the feeding regime on metamorphs.

Three entire Common Toad clutches were collected on 3 April 2016, from a large (ca. 3000 m²) permanent fishpond situated in the Faget forest, Cluj-Napoca, Cluj county, Romania (46°41’48,57’’N 23°32’46,80’’E (DMS), Someș-River Basin, elevation 682 m). They were kept for 11 days, each in a 20 litre container, until the hatching period was completed (14 April 2016). With the start of the feeding-larva stage (Smith-Gill and Berven 1979), we selected 120 healthy-looking tadpoles, 40 of each clutch, at ca. Gosner stage 25 (Gosner 1960) for the start of the experiment. The selected larvae were then randomly assigned to the three diet groups. All the other larvae were released into the original pond.

Three food treatments (hereafter referred as ‘vegetarian diet’, ‘mixed diet’ and ‘carnivore diet’), with five replicates for each treatment were used. The replicates consisted of four-litre opaque containers (9×17 x 26 cm), each holding eight tadpoles in three litres of water (total of 40 larvae per treatment). This larval density corresponds to low densities in natural populations (Glennemeier and Denver 2002; Rot-Nikcevic et al. 2005). The water was kept at constant levels, at 20°C (+/-1) temperature and 12/12 hour light/dark photo period during the entire larval experiment, with light switching on and off at 9 am and 9 pm respectively. Light was generated by four 36 W fluorescent light tubes. Infra-red light was used during observations of night-time activity. Each container was provided with oxygen through electric air pumps. Over 90% of the water was changed every day with aged tap water, with occasional extra sanitation through siphoning, to clean organic materials.

The food for the ‘vegetarian diet’ group consisted of a mix of spirulina 500 mg tablets, protein 63.5%, carbohydrate 16.1%, fat 8.2%. origin: China), rabbit food (Versele Laga Cuni fit pure, protein 15%, carbohydrate 15% lipid 3%. origin: Hungary) and collard greens (protein 3%, carbohydrate 6%, lipid 1%. USDA), allowing them to selectively graze on the preferred food item. Spirulina was used in order to mimic the presence of high-protein microalgae species, e.g. from the genera Chlorella, Scenedesmus, Anabaena, Aphanizomenon, Pediastrum, identified in the original pond (Caraus 2012). The ‘mixed diet’ group was fed with vegetarian (see above) and carnivore diet, consisting of freeze-dried Tubifex worms (Bio-lio, 54% protein, 16% lipid, origin: Hungary) and freeze-dried Chironomidae larvae (Exoti-k, 30% protein, 10% lipid, origin: Romania). The carbohydrate levels for Tubifex worms and Chironomidae larvae were not provided by the manufacturers. However, according to literature, carbohydrate levels, expressed in % glycogen in the aerobic phase and at room temperature are ca. 5% for Tubifex worms and ca. 10% for Chironomidae larvae (Seuß et al. 1983; Hamburger et al. 1998). Diet types were alternated daily in the ‘mixed diet’ group. Finally, the ‘carnivore diet’ group was fed strictly with animal diet (see above). Tadpoles were fed ad libitum, every day, two hours after the daily water change.

The carry-over experiment targeted the growth and development of the toadlets. When the forelimbs emerged (ca. Gosner stage 42), the quantity of water in the holding containers was reduced to 2 litres and the containers were tilted so that a dry area formed at the raised end. Metamorphs were removed one-by-one, when they climbed out of the water at about Gosner stage 43 (starting date: 06 May 2016).

In order to provide a uniform diet, we used a single insect species as food item – Acheta domesticus (house cricket). Appropriate cricket size was judged as roughly the distance between the eyes of toads.

The crickets were raised on a combination of cat food, baby turtle food and Spirulina in equal proportions with added turtle vitamin (Vita-Plus Vit A: 150.000 U.I.; Vit D3: 50.00. U.I.; Vit K: 25 mg). A small number of unconsumed crickets always remained in the enclosure until the following day proving that the toads were satiated.

Each surviving toadlet (n=106) was included in the experiment. The toadlets were housed in specially designed containers at a density of ca. 100 cm²/individual, keeping the same replications as for the larval experiment.

Enclosures were specially designed opaque plastic containers with textured sidewalls, to ensure better accuracy of the visual functions (Ewert et al. 2004). The textures consisted in random patterns of yellow, black, green and opaque colours. As substrate, we used a washcloth-type material with good water retention capabilities (Craioveanu et al. 2017). A shelter was provided on the right side of the enclosure and consisted of two green artificial forest plant branches with seven leaves each (Exoterra forest plant) mimicking natural conditions (Craioveanu et al. 2017). Each container was cleaned once a week and dechlorinated tap water was used to maintain humidity. The initial environmental temperature was set to 20°C and was progressively raised over the period of seven weeks to 25°C to mimic natural conditions. The photoperiod was maintained the same as in the larval experiment (see above). The holding tanks’ position was rotated/changed once a week.

In order to keep the individual density in the containers, dead toadlets were replaced with marked individuals (toe clipping) that were not used in the measurements.

All containers were held in a 7 x 5 m laboratory room with artificial ventilation. Lighting was provided by four 36 W fluorescent light tubes. Infra-red light was used during observations of night-time activity.

The carry-over experiment lasted for seven weeks and, at the end of the experiment, all the toadlets were released.

All length measurements and developmental stage determinations were conducted on digital photographs using the free image analysis software IMAGEJ (http://imagej.nih.gov/ij). Photographs were taken using a Nikon D 3200 camera mounted 30 cm above the specimens.

In the larval experiment, we measured: total body length (snout to tail end in mm) and assessed developmental stages according to Gosner (1960) (multiple images from different angles). We also recorded the duration of the larval period (days) and mortality until and during metamorphosis. Due to the small size and fragility of the larvae at day 1 of the experiment (Gosner 25), the first measurements were made on day 7 (20 April 2016) with a total of three measurements performed weekly (days 7, 15 and 21 of the experiment).

The tadpoles were gently removed from their containers using a shallow net and placed in a scaled Petri dish containing tap water at a depth that approximately equalled the maximum dorso-ventral diameter of the tadpole (Davis et al. 2008). For the length measurements, care was taken to ensure that the animal was not moving during the photography session and the whole dorsal area was fully visible. After the photographs were taken, the tadpole was immediately returned to the tank.

We considered Gosner stage 43 as the time of metamorphosis. This was ecologically meaningful, as it marked the moment the metamorphosing animals first emerged from the water. We measured the body length (BL in mm) (the distance from the snout to the base of the remnant tail) (Altig 2007) daily, throughout the entire period of metamorphosis (06–14 May 2016; days 23–31 of the experiment).

In the carry-over experiment, we measured: snout to urostyle length (SUL in mm) and body mass (g). We used these measurements to calculate body condition (residual condition index) of individuals. The residual condition index was calculated as the value of the residuals resulting from the linear regression of body mass against the SUL (Băncilă et al. 2010) (R² varied between 0.78–0.84, computed in RSTUDIO version 1.1.463 (2016)). We also recorded mortality (% deaths) during metamorphosis and during the whole extent of the carry-over experiment.

To avoid excessive handling, for the metamorphosis and carry-over measurements we used the digital photographic technique combined with a scaled glass Petri dish containing 2-3 mm of water. We ensured that the animal was in a flat position, by replacing the lid (Antwis and Browne in http://www.amphibianark.org/). Immediately after the photograph was taken, the specimens were dried, by placing them briefly on a filter paper and weighed on an electronic scale to the nearest of 0.01g. The first measurement for the carry-over experiment took place 18 days after the last specimen left the water (2 June 2016; day 50 of the experiment) with a total of five measurements performed (days 50, 56, 64, 71 and 77 of the experiment).

For the larval experiment, we analysed the effect of ‘Diet’ (Fixed effect) on the development (Gosner stage) and growth (total body length, mm) of the Common Toad tadpoles with linear mixed models followed by Type I ANOVA, for each of the three measurements performed in the tadpole phase (we used car, MASS, nlme, lme4 packages in RSTUDIO Team (2016)). Subsequently, we tested for significant differences between the developmental progress of larvae fed with different diets, expressed in values of Gosner stages and total body length of larvae, at each measurement. To test for differences, we used analysis of variance (aov function in R) followed by Tukey HSD tests.

Data collected during metamorphosis were used to test the effect of ‘Diet’ (Fixed effect) on the time until metamorphosis with linear mixed models followed by Type I ANOVA (we used car, MASS, nlme, lme4 packages in RSTUDIO Team (2016)). We checked for differences in time until metamorphosis (number of days between the start of the experiment and metamorphosis), mortality during metamorphosis and BL at metamorphosis, in the three treatments. We tested for differences in mortality and BL with the Mann-Whitney U-test and analysis of variance followed by Tukey HSD tests, respectively.

For the carry-over experiment, we tested the effect of larval diet (‘diet’ = fixed effect) and period (‘period’ = number of days between metamorphosis and end of the experiment or death of the toadlets, fixed effect) on the body condition of the Common Toad metamorphs with linear mixed models followed by Type I ANOVA. Subsequently, we tested whether there were differences in residual BCI of the toads fed on different larval diets with the help of a one-way ANOVA. We also analysed differences in the percentage of deaths occurring during metamorphosis and over seven weeks after metamorphosis with a one-way ANOVA followed by a Tukey HSD test. All analyses were performed in RSTUDIO version 1.1.463 (RSTUDIO Team 2016). Plots for Fig. 1 were drawn in RSTUDIO using ggplot2 package (RSTUDIO Team 2016).

The linear mixed models, followed by type I ANOVA, showed that the fixed effect “diet” had a strong effect on the development of larvae at all measurements (Table 2).

At the first measurement, we found significant differences between the Gosner stages of larvae fed with different diets (F (2,117) = 79.88, P < 0.0001). There was no significant difference between the Gosner stage of the larvae from the ‘mixed’ and ‘carnivore’ diet categories (Fig. 1, Tukey HSD test: P = 0.81). The larvae developing under ‘vegetarian’ diet had significantly lower Gosner stages than those from the ‘mixed’ (Tukey HSD test: P <0.0001) and the ‘carnivore’ (Tukey HSD test: P <0.0001) groups (Fig. 1).

Figure 1. 

Mean developmental stage (Gosner) and mean total length of Bufo bufo larvae. Mean developmental stage (Gosner) and mean total length in the three diet treatments at the three larval measurement days. Bars represent 95% confidence intervals.

At the second measurement, we found significant differences between the Gosner stages of larvae fed with different diets (F (2,117) = 28.48, P < 0.0001). We detected significant differences in the average Gosner stage for all three diet treatments (Fig. 1). The larvae, developing under the ‘mixed’ diet, had a higher average Gosner stage than those from the ‘carnivore’ (Tukey HSD test: P&nbsp;= 0.04) and ‘vegetarian’ diets (Tukey HSD test: P < 0.0001). The larvae from the ‘vegetarian’ diet had a significantly lower average Gosner stage than those from the ‘carnivore’ and ‘mixed’ diet categories (Tukey HSD test: &nbsp;P&nbsp;< 0.001 in both comparisons) (Fig. 1).

At the third measurement, we found significant differences between the Gosner stages of larvae fed with different diets (F (2,117) = 5.78, P = 0.004). The average Gosner stage of the larvae from the ‘carnivore’ diet was significantly lower than that of the larvae from the other two diet categories (Fig. 1, Tukey HSD test: ‘mixed’ P = 0.02, ‘vegetarian’ P = 0.008). There was no significant difference between the average Gosner stage of the ‘mixed’ and ‘vegetarian’ diets (Tukey HSD test: P = 0.97).

The 95% confidence interval (CI) of the average in the ‘carnivore’ category shows a constant increase in time, while in the other two diet categories, no trend in CIs was observable (Fig. 1).

The linear mixed models followed by type I ANOVA also showed that the fixed effect “diet” had a strong effect on the growth of larvae in all measurements (Table 2).

At the first measurement, on day 7, we found significant differences between the length of larvae fed with different diets (F (2,117) = 5.33, P = 0.006). The average length of the larvae from the ‘carnivore’ diet was the smallest and the ‘vegetarian’ diet was the largest (Fig. 1) (Tukey HSD test: P = 0.005), while the ‘mixed’ diet category was intermediate (not significantly different from the other two categories, P > 0.05 in both cases).

At the second measurement, on day 15, we found significant differences between the length of larvae fed with different diets (F (2,117) = 17.71, P < 0.0001). The body length of the larvae from the ‘carnivore’ diet category was significantly smaller than that from the ‘mixed’ (Tukey HSD test: P = 0.0006) and ‘vegetarian’ diet (Tukey HSD test: P < 0.0001). Furthermore, there was no significant difference between the ‘mixed’ and ‘vegetarian’ categories (Tukey HSD test: P = 0.10, Fig. 1).

At the third measurement, on day 21, we found significant differences between the length of larvae fed with different diets (F (2,117) = 9.27, P = 0.0002). We found the same pattern in average size distribution as in the second measurement, the larvae from the ‘carnivore’ diet category being significantly smaller than those from the ‘mixed’ (Tukey HSD test: P = 0.005) and ‘vegetarian’ (z = 3.23, P = 0.0002) categories, while no significant differences were found between the ‘vegetarian’ and ‘mixed’ groups (Tukey HSD test: P = 0.62) (Fig. 1).

Different diets did not have a significant effect on the time until metamorphosis (Gosner stage 43, Table 3). Mortality during metamorphosis was 0 for the individuals fed with a vegetarian diet, 0.05% (2 individuals) for the ones fed with a mixed diet and 0.1% (4 individuals) for those fed on a carnivore diet. These mortality rates were not significantly different from each other (Mann-Whitney U-test: P > 0.05 in all comparisons).

The body length of Bufo bufo at metamorphosis varied significantly amongst the three treatments (F (2,111) = 7.14, P = 0.001). Body length of individuals fed with a carnivore diet was significantly lower than that of individuals fed with mixed (Tukey HSD test: P = 0.005) and vegetarian (Tukey HSD test: P = 0.003) diets, whereas those fed with mixed and vegetarian diets did not differ significantly from each other (Tukey HSD test: P = 0.08) (Fig. 2).

Figure 2. 

Body length of Bufo bufo at metamorphosis. Body length (mm) of Bufo bufo at metamorphosis (Gosner stage 43) for the three larval diets. Letters “a” and “b” indicate significant differences (one-way analysis of variance followed by Tukey HSD tests, P < 0.05).

We found a significant effect of ‘period’ on the metamorphs BCI but the effect of ‘diet’ was only marginally (P < 0.1) significant (Table 3).

At the first post-metamorphic measurement (day 50), we found significant differences between BCI of toads from different treatments (F (2,103) = 5.48, P = 0.006). The BCI of the toads from the ‘vegetarian’ group differed from those from the ‘mixed’ group (Tukey HSD test: P = 0.004).

At the second post-metamorphic measurement (day 56), we found no significant differences between BCI of toads from different treatments (Kruskal-Wallis rank sum test: Chi² = 81.49, DF = 80, P = 0.43).

At the third post-metamorphic measurement (day 64), we found significant differences between BCI of toads from different treatments (F (2,59) = 6.31, P = 0.003). Toads from the ‘mixed’ group had different BCI from those from the ‘vegetarian’ (Tukey HSD test: P = 0.03) and ‘carnivore’ groups (Tukey HSD test: P = 0.004).

At the fourth post-metamorphic measurement (day 71), we found no significant differences between BCI of toads from different treatments (Kruskal-Wallis rank sum test: Chi² = 61.00, DF = 61, P = 0.48).

At the fifth post-metamorphic measurement (day 77), we found no significant differences between BCI of toads from different treatments (F (2,59) = 2.47, P = 0.09).

Sixty-four days after the experiment started, mortality of the metamorphs in the ‘vegetarian’ diet reached 15%, in the ‘mixed’ diet 52.5% and in the ‘carnivore’ diet 77.5% (Fig. 3). Mortality between diets differed significantly after metamorphosis (ANOVA: F (2,12) = 4.04, P = 0.046). Mortality in the carnivore diet was significantly higher than that in the vegetarian diet (Tukey HSD test: P = 0.04). The other two diets did not differ significantly regarding metamorph mortality (Tukey HSD test: Pveg-mix = 0.23, Pmix-car = 0.55).

Figure 3. 

Mortality of Bufo bufo. Mortality (%) during larval development (day 1–23), metamorphosis (day 24–31) and post-metamorphic stage (day 32–77) in Bufo bufo for the three larval diets.