The signatures of Anthropocene defaunation: cascading effects of the seed dispersal collapse

Abstract

Anthropogenic activity is driving population declines and extinctions of large-bodied, fruit-eating animals worldwide. Loss of these frugivores is expected to trigger negative cascading effects on plant populations if remnant species fail to replace the seed dispersal services provided by the extinct frugivores. A collapse of seed dispersal may not only affect plant demography (i.e., lack of recruitment), but should also supress gene flow via seed dispersal. Yet little empirical data still exist demonstrating the genetic consequences of defaunation for animal-dispersed plant species. Here, we first document a significant reduction of seed dispersal distances along a gradient of human-driven defaunation, with increasing loss of large- and medium-bodied frugivores. We then show that local plant neighbourhoods have higher genetic similarity, and smaller effective population sizes when large seed dispersers become extinct (i.e., only small frugivores remain) or are even partially downgraded (i.e., medium-sized frugivores providing less efficient seed dispersal). Our results demonstrate that preservation of large frugivores is crucial to maintain functional seed dispersal services and their associated genetic imprints, a central conservation target. Early signals of reduced dispersal distances that accompany the Anthropogenic defaunation forecast multiple, cascading effects on plant populations.

Figure 1. The extinction-driven lizard downsizing gradient in the Canary Islands.

Schematic representation of the human-driven lizard defaunation resulting in present-day variable frugivore body-sizes among islands. Neochamaelea pulverulenta (Rutaceae) relies exclusively on the lizards for seed dispersal and is only found in lowland areas of Gran Canaria, Tenerife, and La Gomera (geographic range shown in green) (Fig. S4). Grey silhouettes illustrate extinct lizard taxa (†); black silhouettes represent the three extant, widely distributed species (photos). The maximum snout-to-vent length (SVL) is indicated. Red dots indicate locations of the main 1-ha study plots, whereas the blue ones indicate the replicated study populations (see Table S1). The potential geographic distribution of N. pulverulenta (green) was redrawn from ATLANTIS 3.1 (available on line, Banco de Datos de Biodiversidad de Canarias; http://www.biodiversidadcanarias.es/atlantis/). For details about natural history of these lizard species see also ref. and .

Figure 2. Seed dispersal patterns of Neochamaelea pulverulenta.

Frequency distributions of seed dispersal distances (2-m bins). Vertical marks along the distance axis represent unique documented dispersal events. We included a non-parametric smoothing spline fit (black line) to the empirical distance distribution together with bootstrapped estimates (grey lines) to allow comparisons across plots. Right inset bars indicate the percentage of seed immigration from plants growing outside the plots (white) and when plants growing in the plot buffer area were also considered (black) (see Methods and Table 1 for sample sizes of dispersed seeds; maps shown in Figs S5 and S6).

Figure 3. Spatial patterns of fine scale genetic structure of Neochamelea pulverulenta.

Autocorrelograms showing the variation of genetic similarity (r_ij) with geographic distance among pairs of adult plants (see Fig. S3 for results with sub-adults). Grey areas represent the 95% confidence intervals for the null hypothesis calculated by permutations of the plant spatial coordinates. Arrows indicate the first distance class at which the spatial autocorrelation becomes non-significant. Insets show the 1-ha plots maps indicating the distribution of adult plants (dots). Colours indicate the assignment of each plant to genetic clusters inferred according to the posterior probability of cluster membership (see Methods). All the study plots are represented at the same spatial scale.