Common seed dispersers contribute most to the persistence of a fleshy-fruited tree

Abstract

Mutualistic interactions are by definition beneficial for each contributing partner. However, it is insufficiently understood how mutualistic interactions influence partners throughout their lives. Here, we used animal species-explicit, microhabitat-structured integral projection models to quantify the effect of seed dispersal by 20 animal species on the full life cycle of the tree Frangula alnus in Białowieża Forest, Eastern Poland. Our analysis showed that animal seed dispersal increased population growth by 2.5%. The effectiveness of animals as seed dispersers was strongly related to the interaction frequency but not the quality of seed dispersal. Consequently, the projected population decline due to simulated species extinction was driven by the loss of common rather than rare mutualist species. Our results support the notion that frequently interacting mutualists contribute most to the persistence of the populations of their partners, underscoring the role of common species for ecosystem functioning and nature conservation.

Fig. 1. The seed dispersal loop of animal-dispersed plants.

a Animal seed dispersal as a demographic bridge between reproductive adults and new plant recruits. b Fruit removal: the number of visits of animal species to the fleshy-fruited tree Frangula alnus (‘Removal’) and the number of scats with seeds of F. alnus (‘Deposition’) in the Białowieża Forest, Eastern Poland. Two more disperser species were only observed using camera traps (Supplementary Table 3). Means ± 95% prediction interval based on 500 bootstraps of dispersal data. c Seed deposition: the effect of different animal species on relative scat deposition density along the canopy cover gradient. The relative availability of microhabitats is shown for comparison. A vertical line indicates the median of a distribution. d Plant performance: relationship between the population growth rate of F. alnus and canopy cover when seeds are solely dispersed by gravity. The 95% prediction interval is based on 500 bootstraps of plant data. The bird silhouette (public domain) was obtained from phylopic.org.

Fig. 2. Benefits of animal seed dispersal.

The effect of seed dispersal by 20 animal species on the population growth rate of F. alnus in Białowieża Forest, Eastern Poland. Two scenarios are presented: (1) Plant individuals occur along the entire canopy gradient (‘fully established’, short-dashed line) or (2) occur only in closed forest (the 50%-darkest microhabitats, i.e., 92.7% of available area) and depend therefore on animal seed dispersal for establishment in forest gaps (‘gap colonization’, long-dashed line). In a ‘fully established’ population occurring along the full canopy gradient, gravity dispersal and fruits dropped by animals can result in successful plant regeneration where the canopy is not closed. In comparison, the population growth rate is reduced in a ‘gap colonization’ population if gravity dispersal is predominant, because plant establishment and growth is much reduced under a closed canopy. When 100% of seeds are dispersed by animals, the long-term stable stage distribution of plants and their distribution along the canopy cover is identical, and so is the population growth rate λ. Prediction intervals (95%) based on 500 non-parametric bootstraps of dispersal data were calculated for the effect of the proportion of seeds dispersed on λ. Please note that uncertainty in the effect of seed dispersal resulted in only extremely small differences in the population growth rate of F. alnus.

Fig. 3. Total effectiveness of seed dispersal by animals.

a Projected consequences of the extinction of each of 20 frugivorous animals on the population growth rate of F. alnus assuming complete interaction deficit (i.e., the interactions with an animal are lost completely). Animal species are in order of declining importance as dispersers for F. alnus. Animal species depicted by the same color (‘other’ frugivores, see Supplementary Table 3) were assigned the same values for seed removal and for seed deposition, resulting in the same estimate for seed dispersal effectiveness. b The ‘landscape’ of seed dispersal effectiveness of animal species and gravity dispersal for F. alnus. The y-axis represents the quantity of seed dispersal (i.e., the relative interaction frequency of each animal species with F. alnus) and the x-axis represents the quality of seed dispersal (i.e., the probability of a seed to produce a mature plant when dispersed by a certain animal species). The two vertical lines depict the consequences of exclusive gravity dispersal in the two scenarios shown in Fig. 2 for seed dispersal quality. c The relationship between seed dispersal effectiveness (i.e., change in population growth after interaction loss) and the relative interaction frequency of different animal species with F. alnus. Color coding in (b) and (c) is the same as in (a) and Fig. 1c. Mean ± 95% prediction intervals based on 500 bootstraps of dispersal data.

Fig. 4. Interaction compensation after frugivore loss.

Projected consequences of the extinction of each of the four main frugivorous animals on population growth of F. alnus assuming complete interaction compensation after their loss (i.e., a species is lost, but the remaining animal community compensates interactions). Mean ± 95% prediction intervals based on 500 bootstraps of dispersal data.

Fig. 5. Study area and sites.

Map showing the location of the 17 study sites in the Białowieża Forest, Eastern Poland, and the studies conducted at these locations. The map was based on OpenStreetMap and created using QGIS.