Data associated with: A phylogenetic perspective on ecological specialisation reveals hummingbird and insect pollinators have generalist diets
Data files
Jan 22, 2024 version files 79.50 KB
-
Oikos_Data___analyses_Maglianesi.zip
-
README.md
Abstract
Specialisation in food resource use is a crucial process that fosters species coexistence in plant-animal networks, contributing to the maintenance of biodiversity, ecological complexity, and community stability. Notably, although there is a vast literature on ecological specialisation in pollination systems, the evolutionary similarity among the plant species visited by particular pollinators has been largely ignored. Here, we apply a robust phylogenetic approach to analyse whether the evolutionary relatedness of plant species is a significant factor in mediating pollinator visits and how it relates to the morphology of interacting species. We quantified ecological and clade specialisation of hummingbird and insect species in three mutualistic networks from the Costa Rican highlands and associated these metrics with species traits. We found that hummingbirds were overall ecologically more specialised than insects (i.e. visited a less diverse set of plant species). However, when evaluating the phylogenetic relatedness among the visited plant species, all hummingbird species and most insects had overdispersed diets, which indicates they visited phylogenetically distant plant species in the community. Moreover, a great proportion of these clade generalists visited plant species with a great variation in corolla length, showing a lack of preference for this morphological trait. Altogether, our results demonstrate that by incorporating plant phylogeny to network analysis, pollinator species were generalists and that corolla length weakly influences plant-pollinator interactions in the three studied networks. A phylogenetic perspective should occupy a central role in the study of specialisation since it contributes to understanding the interplay between ecological and evolutionary processes in mutualistic networks. Future research should focus on evaluating whether the phylogenetic structure of animal diets mediates patterns of interactions in different types of mutualisms and environmental contexts, linking these patterns to other floral traits. This knowledge may be valuable for deepening our comprehension of the underlying mechanisms shaping ecological networks.
README
Data associated with: A phylogenetic perspective on ecological specialisation reveals hummingbird and insect pollinators have generalist diets
This repository contains three folders, each of them includes R codes and the respective data files in csv format; some of these data files include cells with "NA", which means data not available. The folders are named: DSI & phylo signal, Index d' and Models.
DSI & phylo signal. It includes the R scripts “1_DSI pollination.R” where the DSI* index is calculated for each pollinator species at each of the three study sites and “2_Phylo Signal.R” where the phylogenetic signal of pollinator richness is calculated for the three studied communities.
- The source files used in “1_DSI pollination.R” are:
(a) inter.matrix.reserva.2016.csv, inter.matrix.sakira.2016.csv, and interact.matrix.all.BuenaVista.csv: quantitative bipartite matrices for each of the three study sites, with rows corresponding to plant species and columns to pollinator species, where cell entries are integers representing the observed frequency of pairwise interactions.
(b) Abund_flowers.csv is a data frame with the following variables:
Year: year of the data collection
Perio: 15-days periods
Date: date of sampling
Site: sampling site (Reserva, Sakira or BuenaVista)
Trans: transect (plot) for flower counts
PlCode: number assigned at each plant species
PlOrder: taxonomic order of the plant species
PlFamily: taxonomic family of the plant species
PlantSp: scientific name of the plant species
Abund_1: number of open flowers counted in the transects where in some plant species (for example from the family Asteraceae) inflorescences were considered as unit count
Unit_Abund: whether the unit count was an individual flower or an inflorescence
Abund_2: Abund_1 is multiplied by the mean number of flowers in a set of inflorescences previously sampled, so the unit count in this case is always the individual flower
- The source files used in “2_Phylo Signal.R” are:
(a) NBV.nex, Nreserva.nex and Nsakira.nex: plant phylogenies for each community, obtained using the phylocom software based on the species list of each site.
(b) Plant_traits.csv including the variables
PlantSp: scientific name of the plant species
Type: whether the flower corolla is open or tubular
corola_length: corolla length in mm
curvature: corolla curvature in degrees (calculated from trigonometry)
diameter: diameter of the corolla opening in mm.
(c) inter.matrix.bv.csv, inter.matrix.res.csv, inter.matrix.sk.csv: the three quantitative bipartite matrices outlined before with a slight modification in their format to be used in phylogenetic analysis.
Index d'. It includes the R script named “1_Ecol specializ index d.R” where the specialization index d' of pollinator species in the three communities is calculated accounting for the flower abundance of interacting plant species. The source files used in this code are:
(a) Abund_flowers_site is a data frame with the following variables:
Site: study site
PlantSp: scientific name of the plant species
OverallNumFl: total number of flowers counted across transects
(b) d_pollinators: an output from the R code including the variables:
d: specialization index d’
PollinatorSp: scientific name of the pollinator species
Site: study site
marg.total: total number of interactions between a given pollinator species and all the plant species it visited.
(c) inter.matrix.buenavista.2016.csv, inter.matrix.reserva.2016.csv, and inter.matrix.sakira.2016.csv: the quantitative bipartite matrices for each of the three study sites outlined before, with rows corresponding to plant species and columns to pollinator species, where cell entries are integers representing the observed frequency of pairwise interactions.
Models. It includes three R codes: “1_Models d-DSI” that contains comparison of specialization indices (index d' and DSI*) between birds and insects, and across pollinator orders; “2_Ecol Specializ-originality” where the relationship between index d' and morphological originality is analysed; and “3_Flower morphology” where the plant phylogeny is built.
- The source file used in “1_Models d-DSI” is: d_DSI with the following variables:
Site: sampling site (Reserva, Sakira, BuenaVista)
Group: whether the pollinator species belong to bird or insect group
Order: taxonomic order of the pollinator species
PollinatorSp: scientific name of the pollinator species
CorollaType: whether the pollinator visited plant species with flower corolla open, tubular or both
CorollaShape: whether the pollinator visited plant species with flower shape straight, curved or both
Family: taxonomic family of the pollinator species
d: specialization index d’
marg.total: total number of interactions between a given pollinator species and all the plant species it visited
DSI.st: distance-based specialisation index
Index: a variable used to aggregate the data in the R code
- The source file used in “2_Ecol Specializ-originality” is d_DSI_originality including the following variables:
Site, Group, Order, Family, PollinatorSp, d, marg.total and DSI.st as outlined before
st.originality: standardized originality (distance of a species to the community-weighted mean centroid normalised to zero mean and unit variance)
log.st.originality: log-transformed standardized originality
- The source files used in “3_Flower morphology” are:
(a) Plant_all_visited_2016: including the variables as outlined before for the csv Abund_flowers in the DSI & phylo signal: Year, Date, Site, PlOrder, PlFamily, PlantSp. Additionally, in this csv file are included the following variables:
PoOrder, PoFamily, PollinatorSp: taxonomic order, family and scientific name of the pollinator species, respectively
NFlVisit: number of flowers visited by each pollinator
Index: a variable used to aggregate the data in the R code
(b) Plant_traits: including the three variables PlantSp (plant species), Type (whether the flower corolla is open or tubular) and corola_length (corolla length in mm)
(c) d_DSI: including the variables
Site: study site
Group: whether the pollinator belongs to bird or insect group
Order: taxonomic order of the pollinator
PollinatorSp: scientific name of the pollinator species
CorollaType: whether the pollinator visited plant species with flower corolla open, tubular or both)
CorollaShape: whether the pollinator visited plant species with flower shape straight, curved or both
Family: taxonomic family of the pollinator species
d: specialization index d´
marg.total: total number of interactions between a given pollinator species and all the plant species that it visited
DSI.st: distance-based specialisation index
Index: a variable used to aggregate the data in the R code.
Methods
The study was conducted in central–southern Costa Rica on the Caribbean and Pacific slopes of the Talamanca mountains, a volcanic mountain range that forms the spine of the Central American Isthmus from central Costa Rica through western Panama. This area includes the Cerro de la Muerte, which has the largest extent of páramo in the country (09° 60' N, 83° 76' W). The páramo is a neotropical grass- and shrub-dominated ecosystem that occupies the cool and wet upper slopes of neotropical mountains. Field data collection was conducted at three sites of about 12 ha each: Cerro las Vueltas Biological Reserve, Cerro Sákira, and Cerro Buena Vista (hereafter Reserve, Sákira, and Buena Vista, respectively) at elevations between 3000 and 3400 m a.s.l. We collected data on plant-pollinator interactions as well as on abundance and morphological traits of interacting species. Because the two types of studied organisms (plants and mobile pollinators) differ in space use, different approaches for sampling them are required. We used different plots distributed across each study site to sample plants and pollinators, as in other studies (López-Segoviano et al. 2023). In order to accurately record interactions and pollinator species abundance, given the great differences in size and behaviour between hummingbirds and insects, we used different survey techniques for these groups following similar protocols used in previous research (Lehmann et al. 2019). Field data were collected during sampling periods ca. every 12 days in every month at each study site over an entire year, from January to December 2016.