Project Topics
Get Complete Project Material File(s) Now! »
Taxonomic complementarity
We first visually represented the complementarity with Venn diagrams using the VennDiagram package in R, showcasing the number of shared species in each comparison, for each study site. We then compared the species richness between dispersal modes for each ungulate, and between ungulates for each dispersal mode. Then, to understand for which ungulate, the complementarity is stronger, we compared the similarity index, calculated as the number of species dispersed by both dispersal modes divided to the total number of species, between red deer and wild boar. The same comparison cannot be done between dispersal modes because of a lack of data. The comparisons were done by the Wilcoxon rank sum test using the wilcox.test function in R.
Functional complementarity
In order to analyse the functional complementarity, 17 plant attributes relevant to seed dispersal were assigned for each species recorded in the dataset (Annex 1). These plant traits contain plant traits such as habitat preference, Ellenberg indices of light and nitrogen, seed production and diaspore traits such as diaspore releasing height, diaspore morphology (shape, length, the presence of appendage, the characteristics of the appendage present, the presence of pulp, mucilage, elaiosome) and seed bank longevity index (SLI).
The plant traits were first compared between dispersal modes for each ungulate and between ungulates for each dispersal mode, with study sites as sampling units. The comparisons were done by the Kruskall-Wallis rank sum test using the kruskal.test function in R.
The plant attributes were also used to build PCA models to visualize the tendency of the traits in regards to ungulate species or to dispersal modes. The PCA analysis was done using the vegan package and the cleanplot.pca function (built by Borcard et al., 2018). The correlation between plant traits and dispersal vectors or dispersal modes are later calculated by Pearson’s correlation test, using the function rcorr from the package « Hmisc ». All statistical analyses were performed using R Studio software (RStudio Team, 2015).
Taxonomic complementarity
In order to analyse the complementarity between dispersal vectors (red deer – wild boar), we selected study sites having data of plant species that are dispersed by either red deer, wild boar, or by both vectors. The same thing was done with the analysis of the complementarity between dispersal modes (endozoochory – epizoochory). In total, we extracted data from 16 sites for these analyses, the name of the sites are detailed in Fig.1 and Fig.2.
Complementarity endozoochory-epizoochory
Regarding the complementarity between dispersal modes, the number of species shared between endozoochory and epizoochory is represented in Fig.1.
As presented on the figure, the number of species that are dispersed via both endozoochory and epizoochory varies from 0 (Lorris, France) to 9 (Lille Vildmose, Danmark) plant species for red deer, and from 1 (Lorris, France) to 19 (two sites, Germany) for wild boar.
Despite the tendency of red deer dispersing more species than wild boar via endozoochory, the species richness between endozoochory and epizoochory are not significantly different (red deer: p=0.95; wild boar: p=0.69) (Fig.3a,b). Also, the level of complementarity between dispersal modes does not significantly differ between ungulate species (p=0.53) (Fig.3c).
Complementarity red deer – wild boar
For each study site, the number of species shared by red deer and wild boar is represented in Fig.2. Concerning plant species dispersed via endozoochory, the number of plant species dispersed by both red deer and wild boar is highest for the Golestan National Park’s Scrub and Juniper woodland (24 species) and lowest for the New Zealand’s Alpine ecosystems (1 species). Only the Lorris forest has the data on plant species dispersed by both vector via epizoochory, with 2 taxa (Betula pendula and Galium sp.) in common for the two vectors.
When comparing the species richness in the pool dispersed by red deer to that of wild boar, the number of species dispersed by red deer is not significantly different from the number of species dispersed by wild boar (endozoochory: p=0.13; epizoochory: p=1) (Fig.3d,e). Furthermore, the level of complementarity between ungulate vectors does not significantly differ among dispersal modes (p=0.57) (Fig.3f).
Comparison between dispersal modes or between dispersal vectors
The results of the comparison of plant traits between dispersal vectors or between dispersal modes are presented in Table 2, and are visualized in Fig.4. Concerning the species dispersed by red deer, habitat, the presence of hooked appendage, diaspore releasing height, diaspore mass, diaspore length and seed bank longevity index (SLI) are significantly different between plant species dispersed via different dispersal modes (p<0.05). Based on fig.4, plant species dispersed exclusively via epizoochory are often released at greater height and have longer, heavier diaspore with more chance of possessing a hooked appendage. Differently, plant species dispersed exclusively via endozoochory tend to have higher seed bank longevity.
Regarding the plant species dispersed by wild boar, diaspore shape (Vs), diaspore length, the presence of hooked or of flat appendage differ significantly between plant species dispersed via different dispersal modes (p < 0.05). Hence, among species dispersed by wild boar, the species dispersed exclusively via epizoochory are more likely to have hooked appendage, flat appendage and long diaspores. On the other hand, species dispersed exclusively via endozoochory have smaller diaspore shape (Vs) value, indicating that these species have rounder shaped diaspores. Among species dispersed via endozoochory, plant species dispersed exclusively by red deer have significantly higher diaspore length than species dispersed exclusively by wild boar or by both vectors (p=0.02). This means that plant species that are dispersed exclusively by red deer tend to have longer diaspore than those dispersed by wild boar or by both vectors. No plant attributes differ between red deer and wild boar for epizoochory.
Correlation between plant traits, dispersal modes and dispersal vectors
The PCA plots in Fig.5 show us the correlation between plant traits, dispersal modes and dispersal vectors. The angle between the vector variables indicates the correlation between these variables. The two first axes of the PCA explained respectively 19.09% and 12.15% of the variance of the dataset. The relationships observed from the PCA plots were later confirmed by the Pearson correlation tests, the correlation coefficients and the p-values of these tests are shown in table 3.
Complementarity endozoochory – epizoochory
We observe a gradient in traits related to dispersal modes from endozoochory to epizoochory. Habitat, SLI and the Ellenberg light index are positively correlated to endozoochory while negatively correlated to epizoochory. Plant species living in open spaces with high light requirement and high seed bank longevity are thus more likely to be dispersed via endozoochory. Contrarily, diaspore length and diaspore releasing height are negatively correlated to endozoochory and positively correlated to epizoochory. Therefore, plant species with longer diaspores or with diaspores released at great height are more likely to be dispersed via epizoochory. These results are proven significant based on the results from table 3.
There are other plant traits with significant correlation to either of the dispersal modes, despite not being “obvious” enough in the PCA plots. Diaspore shape (Vs) and diaspore balloon shape both correlate negatively to endozoochory, indicating that rounded diaspore are more likely dispersed via endozoochory. The presence of mucilage and the absence of appendage also correlate positively to endozoochory, which means that diaspores with mucilage or with no appendage are more likely to be dispersed via this mode. On the contrary, the presence of elongated, flat or hook appendages correlates positively to epizoochory. Thus, plant species which have diaspores with an appendage tend to be dispersed by epizoochory.
Complementarity endozoochory-epizoochory
The number of species dispersed via endozoochory is not different to that of epizoochory. This result is in agreement with previous studies (Benthien et al., 2016 for sheep and goats; Petersen et al., 2019 for red deer). However, these results are not directly comparable since the diaspores on an animal’s fur or hooves probably represents much more extended time of plant–animal interactions than do diaspores released in animal feces. Furthermore, the collection of feces left behind is relatively easy compared to investigating fur or hooves, this can eventually lead to the false impression that endozoochory is a more important seed dispersal pathway than is epizoochory (Petersen et al., 2019).
The endozoochorous dispersal partly shares the plant species with the epizoochorous dispersal. Furthermore, red deer does not disperse more species via both mechanisms than wild boar, which means that we cannot tell which species has a higher complementarity between dispersal modes. When we analysed the dataset altogether, it is possible that endozoochory counterbalanced epizoochory for red deer and conversely, epizoochory counterbalanced endozoochory for wild boar. Using study sites as sampling units, our dataset was probably too small, which leads to one variable overshadows the effect of another. However, red deer tends to disperse more species via endozoochory than via epizoochory. Contrarily, wild boar tends to disperse more species via epizoochory than via endozoochory. These trends hint that red deer might be more efficient in endozoochorous transfer and wild boar more efficient in epizoochorous transfer.
Table of contents :
I. INTRODUCTION
II. MATERIALS AND METHODS
II.1. Study sites
II.2. Data collection
II.2.a. Data from COSTAUD project
II.2.b. Data from previous reports
II.3. Data analysis
A. Taxonomic complementarity
B. Functional complementarity
III. RESULTS
A. Taxonomic complementarity
Complementarity endozoochory-epizoochory
Complementarity red deer – wild boar
B. Functional complementarity
Comparison between dispersal modes or between dispersal vectors
Correlation between plant traits, dispersal modes and dispersal vectors
IV. DISCUSSION
A. Taxonomic complementarity
A1. Complementarity endozoochory-epizoochory
A2. Complementarity red deer – wild boar
B. Functional complementarity
B1. Complementarity endozoochory – epizoochory
B2. Complementarity red deer – wild boar
C. Methodological constraints
V. Conclusion
SWOT Analysis
