Video recordings and ethological analysis of behavioural categories
During acoustic-visual recordings, a second hydrophone (HTI-96-Min, High Tech Inc., Long Beach, MS, USA, sensitivity: -201 dB re. 1V µPa-1, frequency response 2 Hz to 30 kHz), placed less than three cm from the nest opening, was connected directly to a video camcorder (Canon Legria FS200, 41x digital zoom, 25 frames/sec, Tokyo, Japan) to directly synchronise acoustic and visual signals into a uniform dataset (.mod format) for subsequent analysis. By including or excluding different females into a male’s territory, we could control the experimental variables of interest under standardised conditions and study multimodal communication in males. The camcorder was mounted on a stand and positioned approx. 40 cm from the front of the experimental glass tanks. Quantitative ethological analysis of the courtship, pre-spawning and spawning phases was performed by calculating the number of sounds co-occurring with each behavioural category. Courtship behaviour began when the females entered the male territory at a distance of <5 cm from the male’s nest, while the pre-spawning phase was observed when the ripe female entered the male’s nest. Spawning was observed beginning with the female turning upside-down in the nest numerous times in short succession (assumingoviposition ) and started circling the ceiling. Male behaviours and the associated sound emissions were observed in four soniferous males during seven recording sessions and analysed using Solomon Coder (ver. beta 19.08.02). Again, the ripe females were chosen for the recording sessions according to two indicators of their readiness for spawning (belly and eye coloration). Behavioural categories (and behavioural acts per category) expressed by the males were classified and scored according to our observations and the literature (Amorim & Neves, 2007, 2008; Amorim et al., 2013; Malavasi et al., 2009). In Solomon Coder, two datasets were analysed separately and then compared. Firstly, behavioural coding was performed by re-watching the video recordings with sound production involving five females (N = 5; 37.23 ± 4.29 mm LS; range: 31.62 – 43.76 mm) and documenting the frequency (n min-1) and duration (in sec) of behavioural categories, calculating their total occurrence and percentage. In addition, we analysed eight video recordings (two per male) containing the behaviours of the same four tested males, but performed with six different females (N = 6; 37.23 ± 4.29; range: 31.62 – 43.76 mm) when they did not co-occur with sound production (i.e., males were mute for the entire recording period). By having these two datasets, we investigated the differences in frequencies of behavioural categories in males when they engaged in sound production and when they did not. In total, we used a total of eight males in our experiments, but four were unresponsive (i.e. , did not perform courtship behaviour or sounds), resulting in insufficient data for their further analyses. We analysed videos using Solomon Coder for the following behavioural parameters: male behaviour rate (the total number of behavioural acts per min) (1) co-occurring with sounds or (2) not co-occurring with sounds; number of times a female entered the male’s nest accompanied with sounds (3) or (4) without sounds; (5) total behaviours (number of behavioural categories per video recording).