Life stage shapes condition-dependent foraging under predation risk in an electric fish
Data files
Sep 09, 2026 version files 1.16 MB
-
1_wild_length_mass_condition.csv
20.08 KB
-
2_wild_stomach_fullness.csv
28.42 KB
-
3_captive_monthly_condition_measurements.csv
17.82 KB
-
4_captive_worm_consumption_over_lunar_cycles.csv
26.12 KB
-
5_captive_EOD_rate_over_one_lunar_cycle.csv
986.90 KB
-
6_captive_active_minute.csv
22.34 KB
-
7_captive_light_ramping_tolerance.csv
13.80 KB
-
README.md
8.76 KB
-
script_model_1_2.R
4.75 KB
-
script_model_3_4.R
6.96 KB
-
script_model_5_6.R
9.52 KB
-
script_model_7.R
3.72 KB
-
script_model_8.R
4.98 KB
-
script_wild_LEA_distribution.R
2.33 KB
Abstract
Low body condition can promote risky foraging, but whether condition dependence differs across life stages remains unclear. Using field and laboratory studies, we tested whether residual body condition (RBC) predicted foraging under moonlight-mediated predation risk in the Amazonian sand knifefish, Gymnorhamphichthys rondoni. This species has a two-year lifespan and one late second-year breeding season. Because adults have less time than juveniles to recover condition before breeding, we predicted stronger condition dependence in adults. Among wild fish pooled across life stages, stomach fullness increased overnight on dark (lower-risk) but not bright moonlit (higher-risk) nights, when most stomachs were empty and non-empty stomachs occurred mainly in low-RBC fish. In experiments under laboratory-recreated moonlight, lower-RBC adults were active longer than higher-RBC adults on full-moon nights, as quantified by electric organ discharge rate. RBC did not predict adult activity on new-moon nights or juvenile activity on full- or new-moon nights. A full-moon interaction model showed a significantly stronger RBC–activity relationship in adults than in juveniles. In light-ramping trials, lower-RBC adults remained active at higher illuminance before hiding, whereas most juveniles did not hide. These results are consistent with the asset-protection prediction of stronger condition-dependent risk-taking in adults facing late-life constraints than in juveniles.
Dataset DOI: 10.5061/dryad.3tx95x6x4
Description of the data and file structure
Wild sand knifefish datasets:
1. 1_wild_length_mass_condition.csv (data for figure 2)
2. 2_wild_stomach_fullness.csv (data for figure 4)
Captive sand knifefish datasets:
3. 3_captive_monthly_condition_measurements.csv (data for figure 3c,d)
4. 4_captive_worm_consumption_over_lunar_cycles.csv (data for figure 3e)
5. 5_captive_EOD_rate_over_one_lunar_cycle.csv (data for figure 3f)
6. 6_captive_active_minute.csv (data for figure 5a-e)
7. 7_captive_light_ramping_tolerance.csv (data for figure 6a,b)
R scripts for statistical modeling and plotting:
- script_model_1_2.R: Script for establishing average body condition in the wild population, calculating individual residual body condition, and generating figures 2b and 2c for the manuscript.
- script_model_3_4.R: Script for analyzing stomach data from wild populations and generating figures 4c and 4d for the manuscript.
- script_model_5_6.R: Script for analyzing the active-minutes dataset and generating figures 5a–e for the manuscript.
- script_model_7.R: Script for analyzing the active-minutes dataset with body condition x life stage interaction.
- script_model_8.R: Script for analyzing the light-tolerance dataset and generating figure 6a for the manuscript.
- script_wild_LEA_distribution.R: Script for fitting the sand knifefish LEA-frequency histogram to inform size-group classification and generating figure 2a for the manuscript.
Files and variables
1_wild_length_mass_condition.csv: Data for plotting figure 2. Measurements of length, mass, and condition from a wild population of sand knifefish. Also includes measurements of captive fish at the end of the experiments to show the level of condition achieved relative to the condition range of the wild population.
- fish_group: Whether the measurement is from the wild population, captive adults, or captive juveniles.
- raw_mass_(g): Raw mass measurement in g.
- raw_LEA_(mm): Raw length measurement (snout-to-anal-fin terminus [LEA]) in mm.
- log10_mass: log10 of mass.
- log10_LEA: log10 of LEA.
- expected_log10_mass: The expected log10(mass) given the individual’s log10(LEA) under the SMA log10(LEA)~log10(mass) regression based on the wild population. Residual_body_condition (RBC) calculated as measured log10(mass) minus expected log10(mass).
2_wild_stomach_fullness.csv: Data for plotting figure 4. Measurements of quantitative stomach fullness index and EOD rate for a wild population of sand knifefish, sampled across a lunar cycle.
- wild_fish_id: A unique identification code for each collected individual.
- date: Night of sampling (M/D/YY)
- recording_time: Local time at which EOD rate was recorded, 1–2 minutes before the fish was captured (hh:mm).
- minute_since_nightfall: Capture time of the fish, expressed as the number of minutes since the end of nautical dusk twilight.
- predicted_moonlight_ground_illumination_(lx): Nocturnal ground-level illuminance in lx at the time of capture, predicted by the R package MoonShineR for the time and location of collection.
- night_brightness_category: Whether the sampling night was categorized as bright or dark. See main text for further details.
- EOD_rate_(Hz): Electric organ discharge rate of the fish in Hz, recorded ~1–2 minutes before capture, indicating activity level. NA = data not available.
- combined_stomach_plus_food_mass_(mg): Mass in mg of the stomach plus its food content.
- food_mass_(mg): Mass in mg of the food content within the stomach only.
- QSF_stomach_fullness: Quantitative stomach fullness index, calculated as food mass divided by combined stomach-plus-food mass, multiplied by 100.
- log10_LEA: log10 of LEA.
- log10_mass: Log10(mass) of the individual.
- expected_log10_mass: The expected log10(mass) given the individual’s log10(LEA) under the SMA log10(LEA)~log10(mass) regression based on the wild population.
- residual_body_condition: RBC calculated as measured log10(mass) minus expected log10(mass).
3_captive_monthly_condition_measurements.csv: Data for plotting figure 3c,d. Measurements of residual body condition (RBC) in captive fish, recorded monthly during the experimental period.Note: NA values = not applicable (for linear interpolations of RBC)
- captive_fish_life_stage: Whether the individual belongs to the adult or juvenile group.
- captive_fish_id: A unique identification code for each captive individual, with adult or juvenile life-stage designation.
- RBC_measurement_month_number: The measurement number, in monthly increments. The 1st represents the measurement at the start of the experiment.
- feeding_group: Whether the individual belongs to the well-fed or food-restricted group.
- raw_mass_(g): Raw mass measurement in g.
- raw_LEA_(mm): Raw length measurement (snout-to-anal-fin terminus) in mm.
- log10_mass: Log10 of mass.
- log10_LEA: Log10 of LEA.
- expected_log10_mass: The expected log10(mass) given the individual’s log10(LEA) under the SMA log10(LEA)~log10(mass) regression based on the wild population.
- residual_body_condition: RBC calculated as measured log10(mass) minus expected log10(mass).
4_captive_worm_consumption_over_lunar_cycles.csv: Data for plotting figure 3e. Measurements of nightly worm consumption over approximately 3.5 lunar cycles in captive adults. Measurements include all 23 adults before condition manipulation began, and only the 11 well-fed adults thereafter.
- feeding_date: Date of feeding (MM/DD/YY). Worm consumption was counted the following morning.
- captive_fish_id: A unique identification code for each captive individual, with adult or juvenile life-stage designation.
- worm_consumed_percentage: Percentage of worms consumed overnight. Each fish was provided daily with a number of worms equal to its known maximum consumable amount, placed in a feeding dish; worms remaining the following morning were counted to calculate the proportion consumed.
5_captive_EOD_rate_over_one_lunar_cycle.csv: Data for plotting figure 3f. Measurements of captive adult EOD rate over one lunar cycle before the start of body condition manipulation.
- datetime: Date and time of EOD recording (MM/DD/YY hh:mm).
- average_mean_EOD_rate_(Hz): Mean EOD rate in Hz per minute, averaged across all 23 adults.
- ground_illuminance_(lx): Measured ground-level illuminance in lx at the floor of the experimental room. The room was maintained under a 12:12 h solar light-dark photoperiod with 30-min transitions at lights-on and lights-off. Lunar illumination was recreated using MoonShineP. See main text for further details.
6_captive_active_minute.csv: Data for plotting figure 5a-e. Number of active minutes per night for each captive individual.
- lunar_cycle_timeline: Measurement timing, expressed as the number of months since the start of the experiment and whether the measurement was taken around the new moon or full moon.
- measurement_date: Date of measurement in YYMMDD format.
- captive_fish_id: A unique identification code for each captive individual, with adult or juvenile life-stage designation.
- feeding_group: Whether the individual belongs to the well-fed or food-restricted group.
- active_minute: Number of minutes active, determined by the number of minutes during which the EOD rate exceeded a threshold. See main text for further details.
- residual_body_condition: RBC of the individual, calculated as measured log10(mass) minus expected log10(mass).
- sex: Sex of the individual.
7_captive_light_ramping_tolerance.csv: Data for plotting figure 6a. Illuminance level at which captive sand knifefish transitioned from active swimming to burrowing into the sand to hide during 30-min light-ramping trials.
- measurement_date: Date of measurement in YYMMDD format.
- captive_fish_id: A unique identification code for each captive individual, with adult or juvenile life-stage designation.
- residual_body_condition: RBC of the individual, calculated as measured log10(mass) minus expected log10(mass).
- illuminance_at_hiding_(lx): Experimental-room ground-level illuminance in lx during the 30-min light ramp at the moment the fish initiated hiding. Trials in which the individual did not hide were assigned the maximum illuminance reached during the ramp, 0.062 lx.