Skip to main content

Data: Prescribed fire enhances seed removal by ants in a Neotropical savanna

Cite this dataset

Alcolea, Mirela; Durigan, Giselda; Christianini, Alexander (2021). Data: Prescribed fire enhances seed removal by ants in a Neotropical savanna [Dataset]. Dryad.


Seed dispersal and predation by animals often drive plant regeneration. In tropical savannas, such as the Cerrado of Brazil, fire is also a key process in ecosystem dynamics, consuming the lower vegetation strata and killing wildlife, but how fire affects seed-animal interactions is virtually unknown. We investigated the effects of prescribed fires on the removal of diaspores from Miconia rubiginosa and sunflower Helianthus annuus in Cerrado from southeast Brazil. Using plots burned one month or one year before sampling and unburned controls, we assessed the effect of prescribed fire on microhabitat structure and diaspore removal by vertebrates and ants. Covered microhabitats experienced higher seed removal by vertebrates than open microhabitats, but microhabitat features did not influence seed removal by ants. Prescribed fire did not change the total amount of seed removal, and ants were responsible for most removals in burned and control plots. However, fire increased the importance of ants as agents of removal compared to vertebrates. These changes were probably mediated by changes in microhabitat cover. It is likely that species whose seeds are often preyed upon by vertebrates benefit from fire to escape predation, while the opposite would be true for those removed by granivorous ants. By changing microhabitats composition and frequency of seed removal by different agents, fire may create pulses of opportunities for certain plant species to increase their populations and enlarge their spatial distribution, while constraining others. However, how different fire intensities and frequencies influence seed fate of different species is still to be investigated.


Data collection was carried out in a protected area of Cerrado in southeastern Brazil. To investigate the effect of prescribed fire on seed removal we performed removal experiments with seeds of Miconia rubiginosa and sunflower. Sampling was done at three sites > 1.7 km apart. Each site has areas subjected to prescribed burning and controls where we established sampling transects. Each transect contained four seed removal stations 10 m apart from each other. Each removal station received two paired-treatments ca. 30 cm from each other: an open control accessible to all animals and a selective exclosure allowing only ant access to the seeds. We placed 10 seeds of each species per treatment combination. After 24 hr, we counted the number of seeds removed or preyed on at the spot.

To test the impact of fire in microhabitat variables and a cascading effect in seed removal, we measured microhabitat features of canopy and ground layer vegetation around all seed removal stations. To evaluate the density of the ground layer vegetation, we held a graduated pole vertically above ground in four points (north, south, east and west), one meter from where the seeds were placed. We counted the number of vegetation intersections at two levels of the pole: from 0 to 0.5 m and from 0.5 to 1 m from the ground. To estimate canopy and ground layer cover we visually assigned an index for plant cover scored from 0 to 4, as 0 = uncovered, 1 = up to 25% cover, 2 = 25 to 50% cover, 3 = 50 to 75% cover and 4 = over 75 % cover. Canopy cover (percentage) was estimated within a 5 m radius centred at the removal station, while for the ground layer a 2 m radius was considered. Additionally, we counted the total number of trees (>10 cm DBH) and shrubs (woody plants branched up to 1 m above ground) within a 2 m radius from the removal stations.

Usage notes

There are no missing values. Please see Readme.txt file for more details about variables.