Effects of habitat disturbance on the reproductive ecology of Arum pictum ssp. sagittifolium: from pollination to seedling recruitment

Abstract

Background and aims: The loss of natural habitats may strongly affect the fitness of plants that depend on animals for reproduction. However, very little is known regarding the differential effects of habitat disturbance on the distinct phases of the reproductive cycle of plants, especially in non-rewarding species.

Methods: We assessed the effects of habitat disturbance on the entire reproductive cycle of Arum pictum ssp. sagittifolium, a species with deceptive pollination that is endemic to the western Mediterranean Basin. For this, we performed hand-pollination and bagging experiments, evaluated the role of pollinators and dispersers on reproduction, and estimated seedling recruitment in three natural and three disturbed populations (according to their surrounding natural habitat) in Mallorca Island.

Key results: Pollinators were sphaerocerid flies (mainly Coproica, with ~50 % of visits) and staphylinid beetles, and were required for sexual reproduction. Habitat disturbance differently affected the reproductive phases of A. pictum ssp. sagittifolium. Habitat disturbance had a positive effect on Shannon pollinator diversity (but not on pollinator richness), and total pollinator and Coproica abundance were three times higher in disturbed habitats, where overall seed production was also ~30 % higher in natural habitats. Seed production increased with Coproica abundance, but only in natural habitats. Seed dispersers of A. pictum ssp. sagittifolium were birds, mainly Sylvia atricapilla. Although habitat disturbance did not influence disperser diversity or abundance, the majority of seedlings appeared under adult plants and in natural habitats.

Conclusions: Overall recruitment was higher in natural habitats, but this effect could have been masked by only assessing pollinator and disperser numbers, as processes related to the quality of these interactions might be influencing fitness. Our study highlights the need to study different reproductive phases and their multiple components and processes to properly understand the effects of habitat disturbance on the regeneration of plant populations.

Keywords: Arum pictum ssp. sagittifolium; Mediterranean ecosystems; bird dispersal; deceptive pollination; habitat loss; land-use changes; reproductive output; seed germination; seedling establishment.

Fig. 1.

Pollen limitation in Arum pictum ssp. sagittifolium. Mean and standard error (s.e.) of the pollination treatments [open (squares) and hand-pollination (circles)] in disturbed and natural habitats for infructescence set (A), berry (B) and seeds per individual (C). There were no significant differences between natural and disturbed habitats for any of the studied variables. Asterisks show significant differences between pollination treatments (**, 0.01 < P ≤ 0.001; ***, 0.001 < P ≤ 0.0001).

Fig. 2.

Reproductive output of Arum pictum ssp. sagittifolium individuals. Partial residual plots showing the relationships between Coproica abundance and the number of berries (A) and seeds (B) per individual. Significant interactions are shown separately for disturbed (triangle) and natural (circle) habitats. Lines represent the estimates of the best model and the shadow area represents confidence intervals: dashed line and dark grey for natural habitats and solid line and light grey area for disturbed ones.

Fig. 3.

Seedling recruitment in disturbed and natural habitats. Mean density values (+s.e.) of (A) seedlings and (B) juveniles per square metre in each microhabitat. Different letters show significant differences among microhabitats, and asterisks show significant differences between natural (circle) and disturbed (triangle) habitats (***, 0.001 < P ≤ 0.0001), after Tukey’s post hoc test (α < 0.05).