Adult crayfish (P. leniusculus) were obtained in September 2005, March 2006, August 2007, and August 2008 from a crayfish dealer (Chris Campbell, Milton-on-Stour, UK). Animals were separated by sex and a maximum of 25 animals were kept in a communal holding tank (91.5 × 30 × 30 cm), containing carbon-filtered water. Males were physically but not chemically separated from females. Crayfish were maintained at 10°C, 10:14 h light/dark cycle (breeding season; September to December) and 14°C, 12:12 h light/dark (March, August). Individuals were never used in more than one experimental trial.
Within the breeding season, prior to social interaction trials individuals were tested for sexual receptivity. Male crayfish were classed as sexually receptive if they tried to turn or mount the female while female crayfish were identified as sexually receptive if glair glands were visible as whitened tissue on the underside of the telson. Unreceptive crayfish were not used in sexual interaction trials.
Urine blocking experiments
Experiments were conducted in November during the breeding season using sexually receptive male and female crayfish. At 1 week prior to the experiments animals were isolated in individual plastic containers (24 × 18 × 8 cm) and stored in a temperature controlled room at 11°C, mimicking the seasonal temperature in the natural environment. Females used were slightly smaller (mean ± standard error (SE) carapace size 35.3 ± 0.5 mm, N = 30) than males (37.1 ± 0.5 mm, N = 30). Animals were blindfolded 24 h prior to the experiment by wrapping opaque plastic (1 × 4 cm) around the eyestalks and rostrum and securing excess material to the carapace using cyanoacrylate glue. Blindfolding served to prevent disturbance of animals during the experiments by the observer introducing stimuli through a syringe.
A urine-blocked female was paired with an unblocked male for 1 h. Female urine (urine treatment, 15 pairs) or filtered tap water (water treatment, 15 pairs) was introduced by a 250 μl Hamilton syringe between the 2 animals in 5 to 8 pulses of 25 to 50 μl separated by at least 10 s when animals were facing each other and aggressively interacting. In a control experiment (N = 15), an unblocked male and an unblocked female were paired for 1 h without introducing other stimuli. Treatment and control experiments were alternated to ensure that all experimental conditions were tested throughout the reproductive season.
At 3 h before experimental treatment, female nephropores were blocked by attaching 1 cm of silicon tubing (1.6 mm diameter, Bio-Rad Laboratories, Hemel Hempstead, UK) to the basal segment of each second antenna covering the nephropore, using cyanoacrylate glue (Zap-a-Gap, Pacer Technologies, Rancho Cucamonga, USA). To ensure the tubing was secure an additional cyanoacrylate layer was applied around the tube and dried using an accelerator fluid (Zip kicker, Pacer). The open end of the tubing was sealed with plasticine and the seal was reinforced by cyanocrylate glue and accelerator. Dye studies were conducted to ensure that the plasticine plug was efficient at blocking urine release from interacting crayfish.
Stimulus urine and water
Urine was collected from the nephropores of receptive females that were not used for the experiments. Females were strapped to a board using elastic bands tied around the board and the crayfish tail and claws. Any water was cleared from the nephropores and the surrounding area so it could not contaminate the urine sample. The urine was then extracted using a vacuum pump connected to Teflon tubing (1.5 mm diameter) and a 1.5 ml collection vial. At least 200 μl was collected from each of 15 different females and kept in a freezer at -67°C. Prior to the experiments, samples were thawed and stored at room temperature for 1 h directly before to the experiment. Stimulus water used in the experiment was filtered through a 25 cm presediment filter followed by a 25 cm activated carbon filter (Pozzani Pure Water, Louth, UK) and stored at room temperature 1 h prior to the experiment.
Individuals were introduced to both sides of a central acrylic divider in an aquarium (30 × 20 × 20 cm). After 10 min acclimation the divider was lifted and animals were allowed to interact for 1 h. Following each experiment, the tanks and dividers were washed thoroughly using carbon-filtered water. Interactions were recorded from the side with a Panasonic camcorder (NV-GS180EB, Panasonic U.K. Ltd., Bracknell, UK) and a Sony DVD recorder (VRD-MC5, Sony U.K. Ltd., Basingstoke, UK) for later analysis. Video recordings were analysed by an independent observer blind to the experimental treatment. The presence of unambiguous male courtship behaviours was noted in each experiment including seizing, turning and mating (mounting with or without spermatophore deposition, see Table 2 for definitions of behaviours).
Urine visualisation experiments
To study social interactions we used 60 intermoult female crayfish (mean ± SE carapace size of 34.7 ± 0.2 mm, mass 29.1 ± 0.6 g) and 60 intermoult male crayfish (mean ± SE carapace size 36.3 ± 0.3 mm, mass 33.8 ± 0.9 g), with intact appendages. Male fights, female fights and reproductive behaviours were studied within the breeding season (October to December) whereas male-female fights were observed out of the breeding season (March, August). To eliminate the effects of body size on social interactions crayfish were size matched. In same-sex fights carapace and chelae length differences were less than 5%, while for mixed-sex experiments animals were matched only for carapace length due to male-females chelae asymmetries (within 5% for mixed-sex fights and 10% for reproductive interactions). At 1 week prior to interactions individual crayfish were isolated in separate 3 l plastic containers (24 × 17.5 × 8 cm). Animals were blindfolded 24 h prior to the experiment by wrapping opaque plastic (1 × 4 cm) around the eyestalks and rostrum and securing excess material to the carapace using cyanoacrylate glue.
Urine visualisation procedure
Fluorescein was used to visualise urine release following the methods developed by Breithaupt and Eger . A 0.3% sodium fluorescein solution (dose 9 to 10 μg/g body mass) was injected into the pericardium region of crayfish 3 to 4 h prior to experiments using a 250 μl syringe (Hamilton Bonaduz AG, Bonaduz, Switzerland) and a 45-gauge needle (BD Microlance™, Drogheda, Ireland). After injection the hole was sealed using plasticine and tape to avoid haemolymph loss, and crayfish were fed on defrosted prawns. The technique was successful in visualising urine in all individuals (N = 120; see Figure 2 for an example of urine visualisation in two males; see also Additional file 1 for original data used to perform the analysis).
Additional file 1: Movie1. This movie shows an example of male-female interaction prior to mating including female urine release. The movie shows male 'initiation', 'seizing' and the beginning of 'turning'. Female responds with aggressive behaviour (claw pushing, claw lock) accompanied by urine release and defensive behaviours (tail flip) prior to submitting to the male (receptive) at the end of the movie clip. (MOV 8 MB)
Interactions took place in a glass aquarium (40 × 20 × 20 cm) adapted for filming fluorescein release by covering the walls with black opaque lining. Light from a 250 W slide projector was reflected from the top into the tank by a mirror (44 × 20 cm). Interactions were filmed with a camcorder (Sony Hi8, CCD-VX1E or Panasonic NV-GS180EB) from a front view only. Interactions started after a 30 min acclimation period, where animals were physically and chemically isolated by an opaque divider. Fights were recorded for 30 min after the divider was lifted; reproductive interactions were recorded until 5 min after mating ended (defined as when an animal dismounted and mating behaviour did not reoccur after 5 min). Following each experiment the tanks and dividers were washed thoroughly using carbon-filtered water.
Urine release analysis
Recorded interactions were analysed using a behavioural software package (The Observer V. 5.0, Noldus, Leesburg, VA, USA). The timing of urine release and the behaviour of each crayfish was analysed in 15 male-male, 15 female-female and 15 male-female fights and 15 male-female reproductive interactions. The release of stained urine was recorded for both individuals during the acclimation and experimental period. A measure of urine output was determined for each individual from the time spent releasing urine (% of time spent releasing urine in relation to the total time animals were in contact).
Throughout the experimental phase both crayfish were assigned a behavioural score using a mutually exclusive scale, incorporating agonistic (adapted from ) and reproductive behaviours (adapted from ) (Table 2). For each interaction dominant animals were identified as those that initiated fights and showed high levels of aggression (grasping of opponents body with the claws, unrestrained aggression). Subordinate animals displayed submissive behaviours (avoidance, fleeing) and did not reengage in fights.
Female crayfish were scored during reproductive interactions to assess their motivation towards mating. Females were scored as: (1) receptive, where females stretch their claws out in front of their body, which is lowered towards the substrate, aiding the male in mounting and spermatophore deposition; (2) submissive, where the female flees or avoids the male (levels -2, -1. Table 2); (3) aggressive, where the female displays characteristic aggressive behaviours found in fighting (levels 1 to 5; Table 2), or, resists mating attempts by the male by pushing, boxing or clamping onto the male using chelae/pereopods; or (4) separate, where the female was not in contact with the male.
Male responses to female urine (in urine blocking experiments) were analysed using Fisher exact tests. All other data were tested for normality (Kolmogorov-Smirnov test) and homogeneity (Levene test) prior to parametric analysis data. If parametric test assumptions were violated equivalent non-parametric analysis was performed. All data on the percentage urine release duration were arcsine square root transformed. Two-way ANOVA was used to investigate influence of behavioural state (dominant, subordinate, reproductive interaction) and sex on duration of urine release. Post hoc interaction effects were analysed using the Tukey HSD test (significance level P < 0.05). Paired t tests were used to compare urine output between the two interactants in each of the four treatments (Figure 1) since P values for these planned comparisons were not available from the two-way ANOVA. Comparisons of female urine release during different behaviours in reproductive interactions were analysed using a non-parametric Friedman ANOVA with post hoc Tukey tests.
To identify animals displaying high levels of aggression, an aggression index was assigned to each animal by calculating the proportion of time spent displaying aggressive behaviours at levels 4 and 5 (Table 2) in relation to all aggressive behaviours (levels 3a to 5). Pearson correlations were performed to assess if animals displaying a high aggression index also showed a high level of urine release. For correlation analysis we removed one outlier value of a male with a high aggression index that did not release any urine at all. Since all other 29 males had released at least some urine we suspect that urine visualisation was not successful in that 1 male. One-way ANOVA was used to compare the aggression index of animals between male fights, female fights and mixed-sex fights. A Tukey HSD test was used for post hoc comparisons.