Study site and experimental set-up
Experiments were conducted between 15 June and 31 July 2015 and 2019 on a 50m section of river Biała Nida (Fig. 1), in southern Poland. To reduce the possible influence of weather on damselfly behaviours, the studies were performed during warm, dry and not windy weather (Golab et al., 2013; Tynkkynen et al., 2004). Both the riparian and floating vegetation was prepared for the experiments so that the composition and the spatial structure was homogenous. The size of all floating vegetation rafts (patches), that are defended by males as their territories and used by females as oviposition sites, was similar/comparable among patches (details about the study site and preparation of the river vegetation is described in Golab et al., 2013, Fig. 1). These conditions minimised any microclimate differences between the studied territories. Also, predation in the study area was controlled/limited to a minimum. The only C. splendens predators at the study river section were birds, which were not hunting damselflies during the study due to the constant presence of 2-3 observers in the studied river section. Amphibians are other possible predators, but they were not recorded at the study area (Golab and Sniegula, pers. obs.).
Data collections were run between 1000 and 1600 h CEST. First, allC. splendens individuals present at the studied section of the river were caught with entomological hand-net and individually marked with a three-digit numbers (white marking pen). Then, five randomly chosen mature males and five females were caught and glued to the fishing-line (Tynkkynen et al., 2008, Fig. 2), placed in a cool-box to prevent energy expenditure, and stored until experiment begins. Experimental bouts were preceded by a 10 min observation of studied territories (patches) in order to assign resident males to their territories. The age of males was assigned to 4 age categories (1 - immature, 2- mature with soft wings, 3- mature without visible wing wear, and 4-mature with some wing damages; Golab pers. obs.). Only males from category 3 were chosen for further studies, since the age can influence male territorial behaviour (Corbet, 1999; Tynkkynen et al., 2009) and could influence a males response to our experimental treatments.
For each resident male three types of experiments were run: (1) courtship experiment – female attached to a fishing line was presented to a focal resident male for 2 min, (2) aggressiveness – a male attached to a fishing line was presented to a resident male for 2 min and (3) boldness – a bird decoy was moved toward a resident male (simulated predator attack) until he flew away and latency to return to his territory was measured (Tab. 1). In every experiment we measured three traits (described in Tab. 1). Each experiment was run twice: on the original patches (morning trial) and repeated on the patches manipulated by sinking ca. 25% of each vegetation patch using a ballast (afternoon trial) (Fig. 1B). The minimum time between the two rounds of the experiment was 1 hour. Males and females on the fishing line were replaced by new ones every 20 min, in order to avoid exhaustion or rejection signals in case of females (Tynkkynen et al., 2008). The patch manipulation aimed at measuring the traits across two situations and times (Dingemanse and Reale, 2005; Sih et al., 2004).