Skip to main content

Data from: Habitat size thresholds for predators: why damselflies only occur in large bromeliads

Cite this dataset

Srivastava, Diane Sheila et al. (2019). Data from: Habitat size thresholds for predators: why damselflies only occur in large bromeliads [Dataset]. Dryad.


Predators are often more sensitive to habitat size than their prey, and frequently occur in only the largest habitats. Four explanations have been proposed for this pattern: (1) small habitats do not have enough energy to support higher trophic levels; (2) small habitats are less likely to contain particular prey required by specialist predators; (3) small habitats are risky for predators with slow life histories or large body sizes; (4) small habitats are numerically unlikely to be colonized by regionally rare species, such as predators. We critically examine these four hypotheses in relation to the predatory damselfly larva, Mecistogaster modesta Selys. (Pseudostigmatidae), which occurs almost exclusively in bromeliads > 100ml in capacity. We synthesize multiple years of survey data and three manipulative experiments from the Área de Conservación Guanacaste, Costa Rica to conclude that damselflies do not occur in small bromeliads due to their higher risk of desiccation – not because of energetic limitation, trophic specialization, risk of terrestrial predation, or pure numerical effects. These results suggest that recent and predicted declines in precipitation in northwestern Costa Rica may further restrict bromeliad occupancy by damselflies, with cascading consequences for the rest of the aquatic food web.


We have studied, since 1997, the aquatic macroinvertebrates living in bromeliads in the pre-montane rainforest surrounding Estación Biológica Pitilla (Área de Conservación Guanacaste, 10°59'N, 85°26'W, 700 m a.s.l.).

(1) INVERTEBRATE SURVEYS We determined abundances of M. modesta larvae in bromeliads from the secondary forest between September and November in 1997, 2000 and 2002 (n=20, 20 and 18 bromeliads, respectively). Prey communities in these bromeliads were also recorded in 1997 and 2002, but not 2000. Bromeliads were selected to represent a wide range of water capacity: 0.4 -1530 ml in 1997, 35 - 4690 ml in 2000 and 18 - 7132 ml in 2002. In 1997 and 2000, both Werauhia sanguinolenta and Guzmania spp bromeliads were included. In 2002, only W. sanguinolenta bromeliads were examined.

(2) BROMELIAD WATER DEPTHS To examine the effect of bromeliad size on desiccation risk, we measured water depth (± 1 mm) in three leaves (central and two peripheral) of 30 bromeliads every two days from October 2012 to October 2013. Bromeliads represented a range in capacity (12-33ml; 6 bromeliads, 34-67ml: 6 bromeliads, 68-100ml: 10 bromeliads, 100ml – 580 ml: 8 bromeliads).

(3) EFFECTS OF BROMELIAD SIZE ON DAMSELFLY SURVIVAL AND GROWTH In October 2002, we collected 29 M. modesta larvae (10 - 12 mm long, excluding caudal lamellae) and marked these insects by amputating the right middle leg. .Each larva was placed in a leaf axil (one larva per bromeliad), near the center of a bromeliad in the secondary forest (14 Guzmania spp., 15 Werauhia spp., 8 - 2636 ml in volume). Twenty days later, we searched for marked larvae by dismantling each bromeliad leaf by leaf to search and measured bromeliad capacity. To measure larval growth, we weighed larvae to the nearest 0.002g before the experiment (but after amputation) and at the end of the experiment, and calculated specific growth rate the change in mass as a proportion of the initial mass.

(4) EFFECTS OF SPIDERS AND BROMELIAD SIZE ON DAMSELFLY SURVIVAL. We collected twenty bromeliads, suspended them in a comon garden, and placed one spider (Trechaleidae: Cupiennius coccineus F.O. Pickard-Cambridge) in each of twenty of them. After 24 hours, all bromeliads were dismantled leaf by leaf to collect the experimental spiders and damselflies, and damselfly survivorship was recorded. 

(5) EFFECT OF DROUGHT ON DAMSELFLY SURVIVAL To examine the effects of drought on damselfly survival and growth, we manipulated drought length (defined as 0 mm water depth) in artificial microcosms, imposing droughts of 10, 15, 20, 25, 30 days in length. We also had an equal number of controls, which were treated identically, except that they were not subjected to drought. Each microcosm consisted of a 50 ml centrifuge tube sitting in a 207 ml plastic cup, and housed a single damselfly larva. Each drought length treatment or matching control was represented by two microcosms, except for the 30-day drought treatment with six microcosms. Microcosms began with 15ml water, but two days later we removed all water from half of the microcosms by emptying the cups to initiate drought. After completion of each drought length treatment, we added water to the cups to return to a volume of 15 ml inside the tube. In the controls (which were never subjected to drought), 50% of water was replaced every two days to prevent hypoxia. Two days after the completion of the drought, we emptied the mesocosms of that drought treatment, as well as their corresponding controls and checked for damselfly survival.

Usage notes

Please see Readme.txt file


Natural Sciences and Engineering Research Council, Award: Discovery GRants to D.S. Srivastava