A giant’s appetite: How body size drives the diet and trophic position of the Japanese giant salamander
Data files
Sep 17, 2025 version files 40.91 KB
-
data_diet_trophic_position_giant_salamanders.zip
33.13 KB
-
README.md
7.78 KB
Abstract
In predators, growth often drives ontogenetic dietary shifts (ODSs), leading to increasing trophic position (TP) with body size as growing individuals gradually incorporate larger prey in their diet. In species exhibiting extreme size variation, particularly those with gigantism, TP may increase markedly with body size, as large individuals might gain access to prey considerably higher in the food chain and inaccessible to smaller conspecifics. This can ultimately lead to apex predator status in the largest individuals. In this study, we investigated for the first time ODSs in one of the world’s largest amphibians, the Japanese giant salamander (Andrias japonicus). We combined stomach content and stable isotope analysis (δ¹⁵N, δ¹³C) from 160 individuals across a broad size range to quantify dietary patterns and TP changes. We found a non-linear increase in TP with body size, from approximately 3.0 to 5.1, with a marked inflection point at a snout–vent length of 39 cm. This threshold corresponded to a clear dietary transition: from primarily consuming aquatic insects, to feeding predominantly on fish, anurans, and freshwater crabs. This transition likely reflects morphological and physiological adaptations associated with gigantism, enabling the exploitation of large prey. Our findings suggest that gigantism may be adaptive in predators such as giant salamanders by promoting ecological opportunities, allowing individuals to access high trophic levels through extensive growth and ultimately function as apex predators. These results contribute to a broader understanding of the ecological consequences of body size evolution in predatory vertebrates, highlighting how extreme growth can reshape species' ecological roles.
Dataset DOI: https://doi.org/10.5061/dryad.w9ghx3g1x
This work was originally performed using R 4.3.3. The list of packages used to perform the analyses is provided in each script document.
Main folder: data_diet_trophic_position_giant_salamanders.zip
Folder 1: Food web structure
In this folder, data and scripts are provided to present the plot of the structure of the food web using stable isotope data (Figure 3 in the manuscript).
File: data_isotopic_space.csv
Columns:
id: species groups. For giant salamanders, includes individual values (aj) and the mean (andrias_all)dC: δ13C values (‰)dN: δ15N values (‰)dNd+,dCd+: standard deviation values ofdCanddNper species group (not for individuals) (‰)
Folder 2: Trophic position
In this folder, data and scripts are provided to calculate the trophic position of each Japanese giant salamander using stable isotope data of individuals and baselines.
File: data_tRophic_Position.csv
Columns:
LocationFG: functional group (ajfor individuals,aquatic_BL,terrestrial_BL)d13C,d15N: isotope values (‰)
Folder 3: Stable isotope mixing models
In this folder, data and scripts are provided to run stable isotope mixing models for each of the 160 giant salamanders.
Files:
salamanders_consumers.csv: stable isotope data of salamandersSite: the river section where the salamander was foundid: identification number of each salamanderSVL: body size (snout-vent length) of each individual (cm)d13C,d15N: isotope values (‰)
sources_salamanders.csv: stable isotope data of food sourcesSite: the river section where the salamander was foundSource: food source: 4 possibilities: crab (freshwater crabs), fish, frog, invert (aquatic insects)Meand13C: δ13C values of food sources (‰)SDd13C: standard deviation of δ13C values of food sources (‰)Meand15N: δ15N values of food sources (‰)SDd15N: δ15N values of food sources (‰)
TEFs_salamanders.csv: Trophic Enrichment FactorsSource: food source: 4 possibilities: crab (freshwater crabs), fish, frog, invert (aquatic insects)Meand13C: TEF values of food sources for carbon (‰)SDd13C: standard deviation of TEF values of food sources for carbon (‰)Meand15N: TEF values of food sources for nitrogen (‰)SDd15N: standard deviation of TEF values of food sources for nitrogen (‰)
Folder 4: GAMs for δ15N/TP and Body Size
In this folder, data and scripts are provided to develop the GAMs used to highlight the relation between body size and δ15N and trophic position. It also allows for estimating a breakpoint in the δ15N - body size and trophic position - body size relationships using piecewise linear models.
File: data_N_TP_results.csv
Columns:
- 'svl': body size (snout-vent length) of each salamander (cm)
- 'd15N': δ15N values for each salamander (‰)
- 'tp': trophic position of each salamander calculated in 'Folder 2'`
Folder 5: GAMs for stomach content and mixing model proportions
In this folder, data and scripts are provided to develop the GAMs used to emphasize (1) the relation between the body size of giant salamanders and the proportion of prey found in their stomach content and (2) the relation between the body size of giant salamanders and the contribution of prey to the assimilated diet of giant salamanders calculated through stable isotope mixing models.
File: results_SCA_mixing_models.csv
Columns:
- 'SVL': body size (snout-vent length) of each salamander (cm)
- 'SCA_crab': proportion of freshwater crabs found in the stomach content of each salamander
- 'SCA_fish': proportion of fish found in the stomach content of each salamander
- 'SCA_frog': proportion of frogs found in the stomach content of each salamander
- 'SCA_aq.ins': proportion of aquatic insects found in the stomach content of each salamander
- 'mix_crab': contribution of freshwater crabs to the assimilated diet of giant salamanders calculated through stable isotope mixing models
- 'mix_fish': contribution of fish to the assimilated diet of giant salamanders calculated through stable isotope mixing models
- 'mix_frog': contribution of frogs to the assimilated diet of giant salamanders calculated through stable isotope mixing models
- 'mix_aq.ins': contribution of aquatic insects to the assimilated diet of giant salamanders calculated through stable isotope mixing models
Folder 6: Steps for Performing dbRDA in PRIMER with PERMANOVA+
To run dbRDA in PRIMER:
Required files:
- Response Data: the file "diet_proportions.csv" in our case, containing a matrix containing the proportions of each prey type (e.g., crabs, fish, frogs, aq.ins (aquatic insects)) consumed by each salamander. Each row represents a prey category, and each column represents an individual.
- Explanatory Variables: the file "env.variables_primer.xlsx" in our case, containing a matrix of environmental variables (de: water depth; rw: river width; cu: water velocity; pr: relative position in the river; agri: proportion of agricultural areas in the surrounding environment). Each column corresponds to the same individual in the response dataset.
Step-by-step:
- Import files into PRIMER
- Import the environmental variable matrix (env.variables_primer.csv):
- Go to File > Open,
- In the Data type window, tick Sample data, then click Next.
- In the Shape section, tick Rectangular and Samples as columns.
- For Data type, select Environmental.
- Set Blank = Missing value, then click Finish.
- Import the prey proportion matrix (diet_proportions.csv):
- Go to File > Open,
- In the Data type window, tick Sample data, then click Next.
- In the Shape section, tick Rectangular and Samples as columns.
- For Data type, select Unknown/Other.
- Set Blank = Missing value, then click Next.
- In the Text delimiters window, tick Comma, then click Finish.
- Square Root Transformation
- Since prey proportions are compositional data, they may contain highly abundant and rare prey types. To reduce the influence of dominant prey categories, apply a square root transformation:
- Select your prey proportion matrix in PRIMER.
- Go to Pre-treatment > Transform(overall).
- Choose the Square root transformation.
- Click OK to apply the transformation.
- Distance Matrix
- Since dbRDA is based on distances, you must first calculate a resemblance matrix for your prey proportion data:
- Select your prey proportion matrix (response data).
- Go to Analyse > Resemblance.
- Choose an appropriate resemblance measure: Bray-Curtis similarity here
- Click OK to generate the resemblance matrix.
- Run dbRDA
- Click on the previously created resemblance matrix (prey proportions).
- Go to PERMANOVA+ > DistLM (Distance-based Linear Model).
- Choose the environmental variables file as the predictor dataset.
4.1. Model selection
- Under the "Selection Procedure" section, choose: Forward
- Set the model selection criterion to adjusted R².
- Set the number of permutations to 9,999
- Check the box 'Do dbRDA plot' to visualise the results
4.2. Running the model
- Click OK to perform the dbRDA.
- The output will show which environmental variables explain the most variation in prey consumption patterns.
References:
- Clarke KR, Gorley RN. PRIMER v7: User manual/tutorial. PRIMER-E, 2015.
- Anderson MJ, Gorley RN, Clarke KR. PERMANOVA+ for PRIMER. PRIMER-E, 2008.