1. Although dung of mammalian herbivores is an important pathway for nutrient return in savanna ecosystems, differences in dung decomposition rates among species have been little studied. 2. We measured rates of dung deposition and decomposition for various herbivores in a moist Tanzanian savanna, and related differences among species to nutrient concentrations and the activities of soil macrofauna (e.g., different mesh sizes of decomposition bags, or presence and absence of dung beetles). 3. Dung C:N:P stoichiometry varied widely among species, which could in part be explained by differences in feeding strategy (browsers vs. grazers) and digestive physiology (ruminants vs. non-ruminants). Rates of both decomposition and nutrient release were influenced by the C:N:P stoichiometry of dung, with lower relative losses of the least abundant nutrient. Surprisingly, soil macrofauna increased relative losses of the least abundant nutrient, thereby stabilizing the ratio of N loss to P loss. Dung beetles increased rates of N and P release from wildebeest dung significantly and also increased N availability in the soil. 4. We conclude that rates of nutrient return in dung depend not only on where herbivores deposit their dung, but also on its C:N:P stoichiometry, the activity of soil macrofauna, and interactions between these factors. These factors may therefore influence the relative availabilities of N and P in the soil and hence the functioning of savanna ecosystems.
Dung beetle preference
The first excel sheet ("Callibration beetle_weight") contains a table with numbers of dung beetles and their corresponding weights. The second excel sheet ("Dung beetle preference study") contains a table describing the number of dung beetles caught after 12, 24 or 36 hours with dung of different herbivore species. In the third excel sheet ("Dung watercontent") we calculated the water content of dung from different herbivore species: "el" is elephant, "wl" is wildebeest, "zb" is zebra and "bf" is buffalo.
Dung decomposition experiment
The first excel sheet ("all data") is an overview of the dung decomposition experiment containing the relative weight and nutrient (C, N and P) loss together with C, N and P release rates of dung from different herbviore species ("bf" is buffalo, "gf" is giraffe, "rb" is reedbuck, "wb" is waterbuck, "wh" is warthog and "wl" is wildebeest) collected after 41 or 167 days. Dung was placed in dung bags with a fine (mesh.size =1) or coarse (mesh.size = 2) mesh, whereby the latter allowed acces of macrofauna to the dung. Signs of termite and/or dung beetle attack were recorded in the "termites" and "dungbeetles" columns, which were combined to form the "soil.fauna" column. The second excel sheet ("t=0") contains a table with weight and nutrient contents of the dung at the start of the experiment.
Dung decomposition exp.xls
Dung beetle experiment
First excel sheet ("dungbeetle cages") is an overview table of relative biomass loss and C, N and P release rates from dung undergoing different dung beetle and crushing treatments: "nodb.nomash" is no dung beeltes + not crushed, "nodb.mash" is no dung beetles + crushed, "db.nomash" is dung beetles + not crushed and "db.mash" is dung beetles + crushed. Weight and nutrients of the dung at the start of the experiment are indicated by "t0" and these are calculated from data in the second excel sheet ("t0_all species"). Weight and nutrients of the dung when collected (either at day 15, 29 or 47) are indicated by "tx".
Dung deposition per collection time
This excel file consists of several sheets representing different dung collection times (time0-time13). The sheets describe at which site (MAC = Mkwaja area and SAC = Saadani area) dung was collected (in biomass) and which herbivore species deposited that dung. Time0, time6 and time10 were collection times at the start of a field season in which the dungplot was cleared; these data were not taken into account for the dung deposition analyses.
Dung deposition.xls
Dung deposition per guild
In this excel file the dung deposition data of "Dung deposition per collection time.xls" are combined into dung biomass deposited per herbivore feeding strategy and digestive physiology group (noted as column "guild" in the file): "rum.br" is ruminant browser, "nonrum.gr" is non-ruminant grazer and "rum.gr" is ruminant grazer. The dung biomass deposition is combined with the N and P contents of dung in dry and wet periods to estimate annual N and P inputs through dung.
Dung nutrient overview
An overview of dung nutrient concentrations (C, N and P) and ratios of the most common herbivore species of Saadani National Park collected on several occasions during two years. Dung was collected for several experiments (under column "data": see excel file for abbreviations of experiments) and in two areas (under column "location": "SA" is Saadani area, "MK" is Mkwaja area).
Resin bag experiment
Calculated adsorbed N and P values of resin bags placed for 34 days in pots containing only soil, soil and dung, or soil, dung and dung beetles (see column "treatment").