Repeated trends in altitudinal gradients of diversity: how habitat filtering and biotic interactions structure ecological communities

This study demonstrates that despite distinct regional evolutionary histories in the Amazonian Andes and Guiana Shield, Ithomiini butterfly communities exhibit consistent elevational diversity patterns driven by local habitat filtering and biotic interactions acting upon regionally distinct species pools.

The Gist

Imagine the tropical rainforest as a giant, multi-story library. Usually, we think of biodiversity (the variety of life) as being richest at the bottom floor (the hot, humid lowlands) and getting thinner as you climb higher up the stairs into the cooler, misty peaks. But this paper asks a fascinating question: Is the way the books are arranged on the shelves the same in two different libraries, even if the libraries themselves have very different histories?

The researchers studied a specific group of butterflies called Ithomiini (think of them as the "colorful librarians" of the forest). They looked at two very different "libraries":

1. The Andes (Peru): An ancient, towering mountain range where these butterflies have lived and evolved for millions of years. It's like a massive, old library with millions of rare, unique books.
2. The Guiana Shield (French Guiana): A younger, flatter region where these butterflies arrived more recently. It's like a newer branch library with fewer books, mostly borrowed from the main Andes branch.

Here is what they discovered, broken down into simple concepts:

1. The "Elevator" Effect: Same Pattern, Different Books

Even though the Andes library had way more total species (more "books") than the French Guiana library, the pattern of how the butterflies changed as you went up the mountain was surprisingly identical in both places.

• The Analogy: Imagine walking up a mountain. In both regions, as you climbed higher, the butterflies didn't just disappear; they changed. The types of butterflies you saw at the bottom were completely different from the ones at the top.
• The Finding: Nature seems to have a "rulebook" for mountains. No matter where you are in the tropics, the environment (temperature, air, plants) acts like a strict filter. It only lets certain types of butterflies survive at specific heights. This "filter" works the same way in both the old Andes library and the newer Guiana library.

2. The "Family Tree" Twist

The researchers looked at the family trees of these butterflies.

• In the Andes: Because the library is so old and diverse, climbing the mountain meant swapping out entire branches of the family tree. You'd see one group of cousins at the bottom and a totally different group of cousins at the top. This is called turnover.
• In French Guiana: Since the library is newer, the butterflies at the top were often just a smaller subset of the ones at the bottom. It was more like taking a few books from the bottom shelf and moving them up, rather than swapping in entirely new genres.

3. The "Weather" vs. The "Neighbors"

The study asked: What drives these changes? Is it the weather (abiotic) or the neighbors (biotic)?

• The Weather (The Big Filter): Temperature and rain act like the building's thermostat. They set the broad rules. If it's too cold or too dry, certain butterflies simply can't survive, no matter how good their neighbors are. This happens on a large scale.
• The Neighbors (The Local Drama): Once the weather allows them to survive, the butterflies have to deal with their neighbors.
  • The Plant Connection: These butterflies eat specific plants. The study found a "Diversity begets Diversity" rule: Where there are more types of host plants, there are more types of butterflies. It's like a restaurant with a bigger menu attracting more diverse chefs.
  • The Predator Game (The Twist): This is the most surprising part. Usually, we think predators (birds) make life hard for prey, forcing them to hide or look the same to confuse the enemy. But here, more predation actually led to MORE diversity.
  • The Analogy: Think of a high-stakes game of "Hide and Seek" with birds. The birds are so good at spotting patterns that the butterflies have to split up into different "teams" (mimicry rings) and hide in different corners of the room (different micro-habitats) to avoid being eaten. The pressure from the birds forces the butterflies to specialize and create more unique groups, rather than just disappearing.

The Big Takeaway

This paper tells us that while history matters (the Andes have more species because they've been around longer), the rules of the game are universal.

Whether you are in an ancient, complex ecosystem or a newer, simpler one, the way nature builds communities on mountains follows a similar script:

1. Climate sets the stage.
2. Plants provide the resources.
3. Predators (birds) act as the ultimate editors, forcing butterflies to diversify and specialize to survive.

It's like two different chefs cooking in different kitchens with different ingredients, but because they are both following the same recipe for "Mountain Life," the final dishes end up tasting surprisingly similar.

Technical Summary

1. Problem Statement and Objectives

The study addresses a central challenge in ecology: disentangling the relative roles of regional evolutionary history (the regional species pool) versus local ecological processes (habitat filtering and biotic interactions) in shaping biodiversity patterns along altitudinal gradients.

While latitudinal diversity gradients are well-documented, elevational gradients remain less generalized across taxa and regions. The authors sought to determine if similar community assembly patterns emerge in distinct biogeographic regions with contrasting evolutionary histories. Specifically, they investigated whether Ithomiini butterflies (a Neotropical tribe known for Müllerian mimicry and host plant specialization) exhibit convergent patterns in taxonomic, mimetic, and phylogenetic structure across two regions:

1. The Amazonian Andes (Peru): An older, highly diverse region where the tribe originated and diversified.
2. The Guiana Shield (French Guiana): A younger region where assemblages are derived from recent colonization events.

Key Hypotheses:

• H1: Despite different regional species pools, taxonomic, phylogenetic, and mimetic structures will follow similar altitudinal patterns, indicating repeatable local assembly processes.
• H2: Abiotic factors (temperature, precipitation) will drive broad-scale community composition via physiological constraints (habitat filtering).
• H3: Biotic interactions (host plant diversity and predation pressure) will locally shape diversity, supporting the "diversity begets diversity" hypothesis and influencing mimicry dynamics.

2. Methodology

The study employed a comparative, multi-scale approach across 12 sites in Peru and French Guiana, stratified into three altitudinal bands: Low (<200m), Mid (250–500m), and High (>700m).

• Data Collection:

  • Butterflies: Active sampling (4 days/site) yielded 1,967 individuals (96 taxa, 29 genera, 18 mimicry rings).
  • Predation: 200 artificial cryptic butterfly models were placed at each site to measure avian predation rates.
  • Environment: Temperature and precipitation were recorded via HOBO sensors and WorldClim/AgERA5 data. Forest cover was estimated via photography.
  • Host Plants: Species Distribution Models (SDMs) were built for potential host plants (Solanaceae and Gesneriaceae) using GBIF and herbarium data to estimate host richness.

• Analytical Framework:

  • Diversity Metrics: Calculated taxonomic richness ($S$), Shannon diversity, and Pielou's evenness. Mimicry-ring richness was also quantified.
  • Phylogenetic Structure: Used the most recent Ithomiini phylogeny to calculate the Net Relatedness Index (NRI) and Nearest Taxon Index (NTI) to assess phylogenetic clustering.
  • Beta-Diversity: Partitioned phylogenetic $\beta$-diversity into turnover (lineage replacement) and nestedness (subset loss) using UniFrac distances and the betapart package.
  • Statistical Modeling:
    • GLMMs/GLMs: To test effects of altitude, region, and seasonality on richness and composition.
    • PERMANOVA: To analyze community dissimilarity (Bray-Curtis) and phylogenetic $\beta$-diversity.
    • Mixed MDMR (Distance Matrix Regression): To assess the simultaneous effect of multiple ecological predictors (temperature, precipitation, predation, forest cover) on community dissimilarity while controlling for region as a random effect.
    • GAMs: To test non-linear relationships between host plant richness and butterfly richness.

3. Key Results

A. Regional vs. Altitudinal Patterns

• Regional Differences: Peru (Andes) exhibited significantly higher taxa richness ($\beta = 15.8$) and mimicry richness ($\beta = 4.5$) than French Guiana, reflecting the longer evolutionary history and diversification in the Andes.
• Convergent Altitudinal Trends: Despite regional differences, both regions showed a consistent pattern: taxonomic and mimicry richness increased with elevation, peaking at high altitudes.
• Phylogenetic Structure: High-altitude communities in both regions showed significant phylogenetic clustering (higher NRI), indicating that habitat filtering selects for closely related lineages adapted to cooler conditions.
• Beta-Diversity Drivers: Phylogenetic $\beta$-diversity was primarily driven by turnover (lineage replacement) rather than nestedness, particularly in the Andes. This suggests that different elevations filter distinct subsets of the regional pool rather than simply losing species as elevation increases.

B. Abiotic vs. Biotic Drivers

• Abiotic Factors: Temperature and precipitation declined with altitude. While temperature correlated with richness, its direct effect on community composition in multivariate models was marginal. This suggests temperature acts as a broad-scale filter (defining the regional pool) rather than a fine-scale driver of local assembly.
• Biotic Factors (Predation):
  • Predation rates did not vary significantly with altitude.
  • Crucially, predation pressure was the strongest predictor of community dissimilarity and was positively correlated with taxa richness, mimicry-ring richness, and the number of taxa per mimicry ring.
  • This contradicts the expectation that strong predation limits diversity; instead, it suggests predation drives diversification and coexistence.
• Biotic Factors (Host Plants): In French Guiana, potential host plant richness was strongly correlated with Ithomiini richness ($R^2 = 0.87$), supporting the "diversity begets diversity" hypothesis.

4. Key Contributions

1. Decoupling History from Ecology: The study demonstrates that while regional evolutionary history dictates the total species pool size (Andes > Guiana), local ecological filtering (altitude) drives convergent community assembly patterns (richness peaks and phylogenetic clustering at high elevations) regardless of the source pool.
2. Role of Predation in Mimicry: It challenges the classical view that predation constrains diversity via frequency-dependent selection. Instead, the authors provide evidence that predation pressure promotes mimicry-mediated mutualism and niche partitioning, allowing for higher diversity and more taxa per mimicry ring.
3. Mechanism of Beta-Diversity: The study highlights that elevational gradients in the tropics are driven primarily by phylogenetic turnover (replacement of lineages) rather than simple species loss (nestedness), emphasizing the role of environmental filtering in selecting specific evolutionary lineages.
4. Integration of Scales: By combining phylogenetic metrics, mimicry rings, and field-based predation experiments, the study bridges the gap between macro-evolutionary history and micro-ecological interactions.

5. Significance

This research provides a robust framework for understanding tropical biodiversity conservation. It suggests that:

• Conservation Priorities: Protecting elevational gradients is critical not just for species counts, but for maintaining the evolutionary processes (turnover) that generate diversity.
• Resilience: The repeatability of altitudinal patterns across distinct biogeographic regions implies that tropical montane ecosystems possess robust, generalizable assembly rules.
• Biotic Interactions: The findings underscore that biotic interactions (predation and host availability) are not secondary to abiotic factors but are primary drivers of local community structure, potentially acting as engines of diversification in tropical systems.

In summary, the paper concludes that habitat filtering acts locally on regionally distinct species pools to produce repeatable elevational patterns, while biotic interactions (specifically predation and host diversity) further refine community composition, driving both the maintenance and generation of tropical biodiversity.