Department of Physics, Chemistry and Biology (IFM)

Study area

The study was conducted during the dry season, between 14 th of June and 10 th of November 2010, in the vicinity of the village Santa Rosa in the southern end of the Cordillera del Condor, northwestern Peru. The area is dominated by grazing grounds, coffee and banana plantations followed by agricultural fields for diverse crops (maize, cereals, root crops) leaving small areas with remnant forests. Rock walls are common elements in the landscape mosaic. Both remnant forests and rock walls are threatened by uncontrolled slash and burn agriculture, as well as the increasing deforestation. None of the areas are close to (<500m) running or standing water (bigger than the water cavities represented in the bromeliads), making the frogs strongly connected to the bromeliads for survival.

The frogs were present in two different Bromeliaceae species, both still unidentified. The first species, named Bromeliad Species 1 had a mean width of 72 cm (range 25-187 cm) and a mean height of 42 cm (range 10-89 cm) (Fig.2). The Bromeliad Species 1 were recognized by their red and green coloured and wider leaves, compared to the Bromeliad Species 2, which was the second species (Fig.3). The Bromeliad Species 2 had a mean width of 142 cm (range 50-280 cm) and a mean height of 85 cm (range 30-150 cm). This species had thin, long and green leaves with white stripes on the leaves’ undersides. 

Occupancy of tadpoles, juvenile and adult frogs

To estimate the occupancy of tadpoles, juvenile and adult frogs, randomly selected bromeliad specimens with a minimum distance of 10 m from each other were checked in each site. Pellet and Schmidt (2005) showed that multiple (up to six) visits to each site were needed to establish presence or absence for all of the four amphibian species investigated, with 95 % confidence. Therefore, six visits were set as the minimum limit of visiting times to get a good estimation of the occupancy of each bromeliad. The presence was then determined to the different recordings of presence (for adults, juveniles and tadpoles together): presence ≥1 time, presence ≥3 times and presence all 6 times. Additionally recordings of lifestages: tadpole presence, juvenile presence and adult presence. The order of visits to each bromeliad was randomized. The visiting to bromeliads were performed during hours 05:00 am - 07:00 pm. Every individual bromeliad measured was marked with an individual ID number, by a permanent marker pen at first visit, to be recognized for the coming five visits. At all visits the bromeliads were searched for tadpoles, juveniles and adults. All encountered frogs that were inside or <20 cm outside the bromeliads were directly related to the individually measured bromeliad.

In search for adult and juvenile frogs, firstly the bromeliad was only observed for detection; secondly all leaves of the bromeliad were searched through. The search for tadpoles during the day was done using only the sun light, or in darker weather conditions as well as at morning or night with a special handmade LED-lamp in the water cavities. Every cavity of water that was reachable was searched through.

The frogs were counted, captured, photographed and measured on the snout-vent length (SVL) and on the width in the widest part during resting position, to nearest millimeter. For all areas this was done only during the first visit to each bromeliad, and then during following visits, the individual frogs were only counted, not captured. But in the A-area this was a method performed at each visit, to be able to track individual dispersals of the frogs by photo recognition. Therefore, every encounter was recorded with date, time and location. After the first capture, photos were taken from distance to avoid disturbing and inducing additional stress if the individual was recognized as an earlier capture. Captured individuals were always released at the same spot of catchment. Individuals measuring ≤17 mm were classified as juveniles and individuals measuring ≥ 18 mm were classified as adults (modified after Excidobates measurements by Twomey & Brown 2008). Calling individuals were assumed to be males. Detected males were therefore grouped as “males” and the rest of the adult frogs grouped as “adults” containing unsexed females and males.

Habitat utilization

All bromeliads visited to investigate the occupancy of adult frogs, juveniles and tadpoles were measured to gather information of the importance of bromeliad characteristics for the frog species occupancy.

A number of variables for evaluating the quality of bromeliads and its surroundings were recorded: the size of the bromeliads, depth of the interfoliar water cavities in the bromeliads, sun exposure (%) received by the bromeliads, the bromeliad species, bromeliad density and finally the bromeliads’ altitude.

The sizes of the bromeliads were measured as width (cm) with a regular field tape. At first visit to each bromeliad sun exposure was estimated by eye in percent taking into account shadings and location. The water cavities at the base of the bromeliads were measured in centimeters by dipping down a stick in the water cavity in its deepest part. The bromeliad density was measured by estimating or counting (depending on the location) all nearby bromeliads in a 5m radius from the individually measured bromeliads. The altitude (m.a.s.l.) of each bromeliad was monitored with a Garmin GPS 60CSx.

In total 281 bromeliads were sampled, ending up with a total of 1686 visits to these bromeliads.

Behavioural recordings

To gain information about the dispersal, activity, breeding and feeding behaviour of the E. mysteriosus behavioural recordings were performed (Table 2). Ad libitum sampling was used, making it possible to perform behavioural recordings simultaneously with the investigation of habitat quality variables and occupancy of frogs, since it is performed without systematic constrains. But also to be able to catch the behaviours of interest. Continuous recording was applied at every sampling session.

For recording individual frogs’ dispersal all encounters of frogs were photographed, date and location noted. The activity was continuously recorded during field hours. The feeding behaviour was observed 20 times during different field visits. In addition, the tadpole release (breeding) behaviours were observed during six different times (by six different adults) in the two tadpole-areas (the A- and D-area). In the A-area all the three observations were done in the afternoon (~05:00 pm). In the D-area two observations were done in the forenoon (~10:00 am) and one in the afternoon (04:30 pm).

Data analysis and statistics

All data were analyzed using STATISTICA for windows version 9.1. Between the habitat quality variables standard correlation analysis (matrices) was used with correlation coefficients. The relationship between presence of frogs/tadpoles and the different habitat quality variables were analyzed with Generalized Linear Models (GLM; binomial and logit- link). For behavioural data descriptive statistics were applied.