Co-extinction

The co-extinction is the extinction of a species triggered not by direct human impact but by the extinction of another species on which it depended. In an ecosystem, species are nodes in a network of interactions — mutualisms, predation, parasitism, facilitation — and the loss of one node can cause the collapse of connected nodes. The pollinator that goes extinct takes with it the plants it pollinated; the host-specific parasite that loses its host has no alternative resource; the predator that loses its prey starves.

The phenomenon is a network property. Its magnitude depends on the network's topology: how specialized the interactions are, how many alternative partners each species has, and how central the lost species was to the network's structure. Specialized mutualisms are particularly vulnerable: a plant pollinated by a single insect species will go extinct with that insect. Generalist species, which have many alternative partners, are more resilient. The structure of the interaction network is thus a predictor of the system's vulnerability to co-extinction cascades.

Empirical Evidence

Direct empirical evidence for co-extinction is difficult to obtain because it requires documenting the extinction of a species that was not directly targeted by human activity. Most extinctions are attributed to habitat loss, overexploitation, or climate change — the direct drivers. But indirect extinctions are increasingly inferred from historical data. The extinction of the dodo on Mauritius was followed by the decline of the tambalacoque tree, which may have depended on the dodo for seed dispersal. The loss of passenger pigeons in North America may have affected acorn dispersal and forest composition. These cases are suggestive but not conclusive; the interactions are inferred from historical records rather than observed directly.

Theoretical Projections

Theoretical models project that co-extinctions will be the dominant form of biodiversity loss in the coming decades. A 2004 study by Koh et al. estimated that the number of co-extinctions could exceed the number of direct extinctions by a factor of two to ten, depending on the interaction network's structure. The projection is sensitive to assumptions about interaction specificity: if most species are generalists, co-extinctions are rare; if most are specialists, co-extinctions are the primary mode of loss. The empirical data on interaction specificity are still incomplete, and the projections are therefore uncertain — but the direction is clear: the network structure of ecosystems amplifies the impact of direct extinctions.

Co-extinction reveals that biodiversity loss is not a sum of independent tragedies but a network collapse. The species that disappears is not the only one that matters. The web that unravels with it matters more.