Yong-Deng He and Zhong-Ming Ye, Wuhan Botanical Garden of the Chinese Academy of Sciences, discuss their article: Disentangling the mechanisms behind indirect interactions between plants via shared pollinators: Effects of neutral and niche-based processes

In biodiversity hotspots like the alpine meadows of north-western Yunnan, more than 100 flowering species can bloom in a single season. Walking into such a meadow is dazzling: colours everywhere, insects zigzagging from flower to flower. Faced with this richness, one wonders—how do so many plants share pollinators without driving each other extinct? Our team set out to uncover the hidden web of indirect interactions that links plants through their shared visitors.

Hidden connections dominate the network

When a bee visits two different plant species, it creates an invisible link between them. These “indirect interactions” can be competitive (one plant diverts pollinators from another) or facilitative (one plant attracts pollinators that then visit neighbours).

Recording nearly 7,000 flower visits, we discovered something remarkable: these indirect interactions were everywhere, occurring six times more frequently than direct plant-pollinator interactions. It’s like discovering that in a social network, the connections between people through mutual friends vastly outnumber direct friendships. These hidden pathways contributed to network “nestedness”—a pattern where specialist plants that rely on just a few pollinators tend to share those pollinators with generalist plants that attract many species. This structure reduces competition and helps maintain biodiversity.

Plant traits shape who connects with whom

We then asked: what determines whether two plants will share pollinators? Examining plant characteristics alone, we found distinct patterns. For bee-pollinated networks, flower morphology was key. Plants with similar corolla-tube lengths were far more likely to share pollinators. This pattern emerged because short-tongued bees physically cannot access deep flowers, while long-tongued bees avoid shallow flowers to reduce competition. This morphological barrier acted as a natural filter, determining which plants could potentially compete for the same visitors. For fly-pollinated networks, flowering duration mattered most. Flies have notoriously short lifespans and must maximise foraging quickly. Plants blooming longer had more opportunities to interact with diverse fly species, creating more indirect connections.

Once plants were connected through shared pollinators, flower abundance determined the strength of these indirect effects. Abundant species acted as network hubs—like busy restaurants attracting customers who then discover neighbouring establishments.

The surprise: Abundance matters more than matching

But here’s where our findings took an unexpected turn. When we incorporated pollinator traits alongside plant traits—examining both sides of the interaction—we tested whether “neutral processes” (interactions driven by species abundance) or “niche-based processes” (interactions shaped by morphological matching and phenological overlap) better predicted these indirect connections.

Surprisingly, neutral processes dominated. The combined abundance of flowers and pollinators predicted indirect interactions better than morphological matching (similarity in size and shape between corolla tubes and pollinator mouthparts) or phenological overlap (temporal co-occurrence of flowers and pollinators). In our alpine meadow, being common mattered more than having perfectly matched traits.

This doesn’t mean morphological matching is unimportant—it still shapes which connections are possible. But once those barriers are overcome, sheer numbers drive the system. Think of it as a city’s transport network: road design determines which routes are possible, but traffic volume determines which routes actually matter.

Why this matters?

These findings reshape how we think about pollination networks. They’re not just collections of pairwise interactions, but complex webs where indirect pathways outnumber direct ones. Every species loss doesn’t just remove direct interactions—it eliminates all the indirect connections that species facilitated between others.

Our study reveals that in nature’s networks, the hidden connections matter as much as—or even more than—the visible ones. Understanding these indirect pathways helps us predict how communities respond to environmental changes. Our findings suggest that maintaining diverse abundance patterns, alongside the full spectrum of flowering schedules and morphological diversity in both plants and pollinators, is crucial for ecosystem stability. These elements together form the invisible scaffolding that maintains biodiversity in species-rich communities like our alpine meadow.