In visually-driven seed dispersal mutualisms, natural selection should promote plant strategies that maximize fruit visibility to dispersers. Plants might increase seed dispersal profitability by increasing conspicuousness of fruit display, understood as a plant strategy to maximize fruit detectability by seed dispersers.
The role of different plant traits in fruit choice and consumption by seed dispersers has been broadly studied. However, there is no clear evidence about the importance of the traits that increase conspicuousness of fruit display. Because strategies to maximize conspicuousness of fruit display are diverse, and usually are expected to be costly, we would expect that individual plant species will produce an efficient combination of traits.We explored this prediction with 62 fleshy-fruited plant species of a subtropical Andean forest (Southern Yungas), and using a large dataset of fruit consumption by birds (4,476 records). The conspicuousness of fruit display was characterized by both fruit and plant traits including chromatic contrast, size, exposure, aggregation, and crop size of fruits. We also considered phylogenetic effects on phenotypic variation.Fruit consumption was explained by fruit chromatic contrast depending on fruit crop size. These traits revealed low phylogenetic effects, with the exception of four plant clades at different levels in the phylogenetic tree. Negative correlations between pairs of traits support our assumption that fruit display traits are costly, suggesting natural selection favours parsimonious evolutionary pathways.
Plant species seem rely on conspicuousness of fruit display by a combination of traits that might minimize costs of fruit display. This appears adaptively relevant to improve communication with mutualistic animals, to increase fruit consumption in a community context and, ultimately, to enhance the profitability of seed dispersal.
List of plant species and traits
Table S1. Summary of fruit display traits from plant species in a subtropical Andean montane forest of NW Argentina (Southern Yungas). Fruit consumption is a relative frequency of fruit consumption for each plant species by the community of fruit-eating birds (see Methods); Nc (not consumption) indicates fruit species that are not usually eaten by birds , while a hyphen indicates species eaten by birds but without data in our databases. Fruit size is the mean fresh-fruit mass expressed in grams. Fruit colour type is described as it is perceived by humans (i.e., related to hue). Fruit chromatic contrast (ΔS) is expressed in unities of just noticeable difference (JND) between the fruit and the corresponding leaves of the species (background) (see Methods). Fruit exposure has two categories: Exposed (pendulums, erect and terminals) and Unexposed (sessile fruits and fruits that are not exposed away from the contiguous foliage). Fruit aggregation expresses the mean number of fruit in discrete and differentiable groups. Fruit crop size is the mean number of fruits of each plant species, classified into five categories: 1 (1-10 fruits), 2 (11-100), 3 (101-1000), 4 (1001-10000) and 5 ( >10000). See Methods for details. Plant systematics follows Instituto de Botánica Darwinion (2015).
FE-2016-01072-R2-Appendix-S1.doc
Methods for collection and analysis for data on fruit reflectance
Methods for collection and analysis for data on fruit reflectance.
FE-2016-01072-R2-Appendix-S2.doc
Bases and criteria for fruit consumption data
The first step in the multi-step process we used to make raw data on fruit consumption comparable, consisted in the delimitation of datasets that were homogeneous with respect to their geographic location, forest type, collection date and sampling technique. This entailed, on the one hand, the split of some large raw datasets (e.g. covering various types of forests, or more than one season) and, on the other, the grouping of similar opportunistic records from different sources in a same dataset.
FE-2016-01072-R2-Appendix-S3.doc
Competing models in the GLM procedure
Competing models in the GLM procedure to determine the importance of fruit conspicuousness on fruit consumption by birds.
FE-2016-01072-R2-Appendix-S4.doc
Construction of the phylogenetic tree
We constructed the meta-tree of plants from trees stored in Phylomatic (Webb & Donoghue 2005) and using the R package taxize (Chamberlain & Szöcs 2013). To resolve relationships not defined by Phylomatic for Muehlenbeckia and Phoradendron genera, and for the Myrsinaceae family, we used information from the webpage of the Missouri Botanical Garden (http://tropicos.org/). Phylogenetic relationships among Solanaceae (Solanum, Salpichroa, Vassobia, Dunalia, Eriolarynx and Iochroma) were resolved using trees from Levin, Watson & Bohs (2005), Bohs (2007), Weese & Bohs (2007), Olmstead et al. (2008) and Särkinen et al. (2013). Phylogenetic relationships among Celtidaceae, Myrtaceae and Rosaceae were resolved using tree from Song et al. (2001), Lucas et al (2007), and Potter et al. (2007), respectively.
FE-2016-01072-R2-Appendix-S5.doc
R script for estimating relationships between fruit display traits
R script for estimating relationships between fruit display traits
traitrels.R