Data from: Experimental defaunation alters foraging behavior of a small antelope in Kenya
Data files
Nov 04, 2025 version files 222.13 MB
-
Acacia_Spines.csv
16.89 KB
-
Acacia_Tannins.csv
1.29 KB
-
Diet_Composition.csv
7.80 MB
-
Mpala_Plant_Functional_Traits.csv
15.17 KB
-
README.md
8.25 KB
-
UHURU_DikDik_P6_Raw_Sequences.zip
214.29 MB
Abstract
African savannas are experiencing large mammalian herbivore declines and extirpations, the consequences of which will be at least partially dependent on the functional responses of the species that persist. We used a long-term herbivore-exclusion experiment in semi-arid Kenyan savanna to understand how a small browsing antelope (dik-dik, Madoqua guentheri, ~5 kg) responds to the absence of larger herbivores across seasons (e.g., wet vs. dry) and sites (e.g., mesic vs. xeric). We found that dik-dik diets differed significantly from larger, co-occurring mixed-feeders (e.g., elephants: Loxodonta africana; impala: Aepyceros melampus). Dik-dik activity increased significantly in response to large herbivore exclusion, and they foraged selectivity based on plant nutritional properties and avoided Acacia species with strong physical and chemical defenses, especially during dry seasons and in xeric environments. Shifts in selectivity due to competitive release were strongest under resource-scarce conditions, amplifying local impacts of large-herbivore losses on savanna plant communities. Together, these results provide strong evidence that dik-dik exhibit flexible foraging behavior in response to larger herbivores, especially in resource-scarce conditions. If large herbivores are extirpated, subsequent diet shifts by dik-dik may suppress long-term plant diversity in this savanna.
Dataset DOI: 10.5061/dryad.2rbnzs832
Dataset Overview
These datasets and scripts can be used to run the analyses and produce the figures in the associated manuscript (Reed et al. 2025). These data were collected at part of the Ungulate Herbivory Under Rainfall Uncertainty (UHURU) experiment at Mpala Research Centre in Laikipia County, Kenya. UHURU uses size-selective fencing to exclude large herbivores from 1-ha plots on the savanna. UHURU is comprised of three sites across a 20-km climate gradient, with each site containing three replicated blocks of control and exclusion treatments. In this study, we used the South (relatively more mesic) and Central (relatively more Xeric) sites.
Here we report new data on plant functional traits at Mpala. The UHURU plant composition data and the dung count data used in our analyses have previously been published in an UHURU data paper (Alston et al. 2022) and so are not replicated here. The dik-dik, elephant, and impala diet data from outside of the UHURU plots have been previously published in Kartzinel et al. 2019, and the raw sequence reads are stored on Dryad (https://doi.org/10.5061/dryad.c119gm5). The dik-dik diet data from mesoherbivore exclusion plots within the UHURU experiment are published here for the first time. Here we also provide a clean diet composition dataset with the subset of samples used in this manuscript, which can be used for the analyses in the scripts below.
Files and variables
File: Acacia_Tannins.csv
Description: This dataset contains data on leaf tannin concentrations for three Acacia species (N = 10 trees per species) at Mpala Research Centre in Laikipia, Kenya. The data were collected using methods from Ward and Young, 2002.
Variables
- Species: Acacia tree species: A. brevispica, A. mellifera, or A. etbaica
- Tree_Ht: Height of tree in meters
- Branch_Ht: Height of branch from which leaves were collected in meters
- DBH: diameter at breast height of tree from which leaves were collected in centimeters
- UTM_East: Universal Transverse Mercator Easting coordinate for tree location
- UTM_North: Universal Transverse Mercator Northing coordinate for tree location
- Percent_QE_Tannin: leaf percent tannins (% QE)
File: Acacia_Spines.csv
Description: This dataset contains data on spine lengths in the distal 10-cm of branches for three Acacia species (N = 10 spines per branch, 1 branch per tree, and 3 trees per species per site) at Mpala Research Centre in Laikipia, Kenya.
Variables
- Site: South or Central, indicating that the leaves were taken from trees in either the South or Central locations of the UHURU experiments
- Species: Acacia tree species -- A. brevispica, A. mellifera, or A. etbaica
- Tree_ID: unique identifier for each tree
- Spine_ID: unique identifier for each spine
- Spine_Length_mm: length of spine in mm
File: Mpala_Plant_Functional_Traits.csv
Description: This dataset contains data on foliar nutrient concentrations for 211 plant species at Mpala Research Centre. A subset of these data for species that were found in UHURU vegetation surveys and in dik-dik diets (N = 54 species) were included in analysis for the manuscript, but here we provide the full set.
Variables
- Date_Collected: Date that the leaf samples were collected, ranging from August to September 2018
- Plant_Family: plant family
- Species: species of plant
- DikDik_Diet_ASV: ASV in the dik-dik diet data that matches to the plant taxon
- Protein_per: Foliar crude protein (%)
- Digestibility_NGCD_per: Foliar neutral cellulase plus gammanase digestibility (%)
- NDF_per: Foliar neutral detergent fiber (%)
File: UHURU_DikDik_P6_Raw_Sequences.zip
Description: This folder contains raw Illumina sequence read data (.fastq files) for all dik-dik diet samples in mesoherbivore exclusion ("dik-dik-only") plots, which had not been previously published.
File: Diet_Composition.csv
Description: This dataset contains data on diet composition for dik-dik, elephant, and impala at Mpala Research Centre. The raw sequence reads for control samples were previously published in Kartzinel et al. (2019). The dik-dik diet samples from mesoherbivore exclusion plots ("dik-dik-only") are published here for the first time.
Variables
- ASV: amplicon sequence variant
- Sequence_Reads: sequence count for that ASV and sample
- Sample_ID: identifier for fecal sample
- Herbivore_Common_Name: commone name of the herbivore species (DikDik, Elephant, or Impala)
- Herbivore_Latin_Name: Latin binomial of the herbivore species (Madoqua guentheri, Loxodonta africana, or Aepyceros melampus)
- Sample_Date: date the sample was collected (month-day-year)
- Rain_90d_Prior: cumulative mm over rain over the 90 days preceding sample collection
- Season: wet or dry
- Longitude: longitude where sample was collected in decimal degrees
- Latitude: latitude where sample was collected in decimal degrees
- Site: xeric or mesic region of Mpala near (<8km from) the UHURU sites
- Treatment: UHURU experimental treatment. Meso excludes all herbivores larger than dik-dik (~5kg). Control plots are unfenced and allow access by all local animals. Blanks represent samples that were collected outside the UHURU plots in environments similar to control plots.
- Block: Block within a site within the UHURU experiment (e.x., C1 for Central site, plot 1)
- Taxonomic_Rank: rank to which plant taxon was identified (e.g., family, genus, species)
- Family: taxonomic family to which sequence was identified
- Genus: taxonomic genus to which sequence was identified
- Scientific_Name: Most specific name (family, genus, species) to which sequence was identified
- Sequence: DNA sequence
- Potential_Species_List: For sequences that could have come from multiple species within a genus, we include a list of all the potential species associated with that sequence that are found at Mpala
Scripts
File: Forage_Quality_and_Selectivity.Rmd
Description: R markdown script to run analyses and produce figures associated with plant forage quality and dik-dik diet selectivity. Uses Diet_Composition.csv, Acacia_Spines.csv, Acacia_Tannins.csv, and Mpala_Plant_Functional_Traits.csv.
File: Dung_Analysis.R
Description: R script to run analyses and produce figures associated with dik-dik dung counts. Uses dung survey data from Alston et al. 2022 ("DUNG_SURVEYS_2009-2019.csv").
File: Diet_Composition.R
Description: R script to run analyses and produce figures associated with dik-dik, elephant, and impala diet composition. Uses Diet_Composition.csv.
File: Herbivore_Impacts_On_Acacias.Rmd
Description: R Markdown script to run analyses and produce figures associated with the impacts of herbivores on Acacia sapling abundance. Uses tree census data from Alston et al. 2022 ("TREE_CENSUS_SUMMARY_2009-2019.csv").
References
Alston, J. M., Reed, C. G., Khasoha, L. M., Brown, B. R. P., Busienei, G., Carlson, N., Coverdale, T. C., Dudenhoeffer, M., Dyck, M. A., Ekeno, J., Hassan, A. A., Hohbein, R., Jakopak, R. P., Kimiti, B., Kurukura, S., Lokeny, P., Louthan, A. M., Musila, S., Musili, P. M., Tindall, T., Weiner, S., Kartzinel, T. R., Pringle, R. M., & Goheen, J. R. 2022. Ecological consequences of large herbivore exclusion in an African savanna: 12 years of data from the UHURU experiment. Ecology, 103, e3649. https://doi.org/10.1002/ecy.3649
Kartzinel, T. R., Hsing, J. C., Musili, P. M., & Pringle, R. M. 2019. Covariation of diet and gut microbiome in African megafauna. *Proceedings of the National Academy of Sciences, *116, 23588-23593. https://doi.org/10.1073/pnas.1905666116
Ward, D. & Young, T. P. 2002. Effects of large mammalian herbivores and ant symbionts on condensed tannins. Journal of Chemical Ecology, 28, 921-937. https://doi.org/10.1023/A:1015249431942
