Environmental predictability drives different routes to adaptation
Data files
Dec 30, 2025 version files 4.74 GB
-
Analysis_Code.R
4.36 KB
-
Fly_Data_Dryad.csv
4.74 GB
-
README.md
4.76 KB
Abstract
Climate change is altering thermal environments, yet we know little about how environmental predictability shapes species' adaptive responses. Different species may rely on plasticity or evolution to survive environmental change, but how these strategies depend on environmental predictability remains unclear. Experimental evidence that distinguishes between plastic and evolutionary responses to different patterns of environmental variability has been lacking. Here we present the first experimental demonstration that compares adaptive responses to predictable versus unpredictable thermal variation disentangling plastic from evolutionary changes. Using Drosophila melanogaster populations evolved for 11 generations under constant, predictably fluctuating, and randomly fluctuating thermal regimes, we assessed survival and fecundity: a plasticity assay testing flies directly from their evolutionary environments to capture total phenotypic responses, and a common garden assay after two generations of standardized rearing to isolate genetic changes. Strikingly, environmental predictability shaped divergent life-history strategies that were only revealed by comparing our two assays. Populations from predictably fluctuating environments evolved enhanced survival over generations, but this benefit was only visible in the common garden assay, not when tested directly from their evolutionary environment. Conversely, populations from randomly fluctuating environments showed reduced survival in the plasticity assay and consistently lower fecundity in the common garden assay, though this reproductive cost was completely masked in the plasticity assay. These contrasting responses demonstrate that environmental predictability fundamentally determines life-history evolution: predictable variation favors investment in stress-resistant longevity, while unpredictable variation imposes both immediate survival costs and constitutive reproductive constraints. Our findings challenge the traditional view that environmental variation uniformly selects for increased plasticity, instead revealing that the predictability of environmental change determines both the target and mechanism of adaptation. As climate change increases environmental variability and reduces environmental predictability, these insights provide crucial guidance for predicting species persistence and developing effective conservation strategies.
Dataset DOI: 10.5061/dryad.2jm63xt3v
Description of the data and file structure
README for Data and Code Associated with: "Environmental Predictability Drives Different Routes to Adaptation" Vinton et al. - Evolution Letters
CONTENTS
1. Fly_Data_Dryad.csv - Experimental data
2. Analysis_Code.R - R script to reproduce all statistical analyses
DATA FILE: Fly_Data_Dryad.csv
DESCRIPTION
This dataset contains survival and fecundity measurements from an experimental evolution study using Drosophila melanogaster. Flies were maintained in three thermal regimes (constant, predictably fluctuating, and randomly fluctuating) for 11 generations. Fitness assays were conducted at generations 4, 7, and 10.
Two assay types were performed:
- Plasticity Assay (PA): Flies tested directly from their evolutionary environment
- Common Garden Assay (CG): Flies reared for 3 generations at constant 27°C before testing, to isolate genetic changes from plastic responses
COLUMN DEFINITIONS
- Age : Days since start of assay (integer)
- Assay : Type of fitness assay
"PA" = Plasticity Assay
"CG" = Common Garden Assay
- Parent_Gen : Generation of selection when parental flies were collected
4, 7, or 10
- Incubator_Temp : Evolutionary thermal environment
"27" = Constant 27°C
"Flux" = Predictably fluctuating (daily 21-33°C cycle)
"Random" = Randomly fluctuating (unpredictable daily cycle)
- Pop : Replicate population (1-9)
Populations 1-3: Random treatment
Populations 4-6: Constant (27°C) treatment
Populations 7-9: Predictable (Flux) treatment
- WB_Temp : Test temperature in water bath (°C)
23, 25, 27, 29, or 31
- Vial_Replicate : Unique identifier for each vial
- Male_Num : Number of males alive at observation
- Female_Num : Number of females alive at observation
- Egg_Count : Total eggs counted in vial at observation
- Male_Em – Number of male individuals that successfully emerged (eclosed) from eggs.
- Female_Em – Number of female individuals that successfully emerged.
- Sum – Likely sum of male and female numbers (Male_Num + Female_Num).
- Sum_Em – Sum of male and female emergences (Male_Em + Female_Em).
- Missing values are indicated as blank cells or NA.
EXPERIMENTAL DESIGN
- Species: Drosophila melanogaster
- Selection duration: 11 generations
- Thermal regimes:
* Constant: 27°C continuous
* Predictable: Daily cycle of 21°C (12h) and 33°C (12h)
* Random: Daily cycle with temperatures randomized weekly (mean 27°C)
- Replicate populations: 3 per treatment (9 total)
- Test temperatures: 23, 25, 27, 29, 31°C
- Vial setup: 2 males + 2 females per vial, 4 replicate vials per combination
- Measurements: Survival counts and egg counts taken twice weekly
ANALYSIS CODE: Analysis_Code.R
R script to reproduce statistical analyses reported in the manuscript, including Cox proportional hazards models for survival and negative binomial GLMs for fecundity.
REQUIREMENTS
R version 4.0 or later
Required packages: survival, dplyr, MASS
USAGE
1. Place Fly_Data_Dryad.csv in your working directory
2. Update the file path in the script if needed
3. Run the entire script to reproduce all analyses
ANALYSES INCLUDED
1. Plasticity Assay Survival (Cox proportional hazards model)
- Results correspond to Supplementary Table S3
- Model: surv_obj ~ Parent_Gen * WB_Temp + Incubator_Temp
2. Common Garden Survival (Cox proportional hazards model)
- Results correspond to Supplementary Table S6
- Model: surv_obj ~ Parent_Gen * Incubator_Temp + WB_Temp
3. Plasticity Assay Fecundity (Negative binomial GLM)
- Results correspond to Supplementary Table S9
- Model: Egg_Count_Per_Female ~ Age + Parent_Gen + WB_Temp + Age:WB_Temp
4. Common Garden Fecundity (Negative binomial GLM)
- Results correspond to Supplementary Table S12
- Model: Egg_Count_Per_Female ~ Age + Parent_Gen + WB_Temp + Incubator_Temp + Age:WB_Temp
NOTE ON OUTPUT LABELS
R codes factor levels alphabetically, so in model output:
- Incubator_Temp reference level = "27" (Constant)
- Incubator_Temp1 = "Flux" (Predictably fluctuating)
- Incubator_Temp2 = "Random" (Randomly fluctuating)
CITATION
If you use these data, please cite: [Vinton et. al, 2025, Evolution Letters]
CONTACT
For questions about the data or analysis, please contact:
