Introduction

Many animals communicate via composite signals emitted through more than one sensory channel (i.e. multimodal signals; Partan and Marler 2005). “Signal” is here defined in a broad sense to include all the modalities used to transmit information (sensu Dawkins 1995), independently of their being specifically elaborated by the sender for communication purposes (but see Wisenden and Stacey 2005). “Stimulus” denotes a detectable change in the internal and external environment that might stimulate behavioural response in an organism (Immelmann and Beer 1989).

When compared to unimodality, multimodality affords new opportunities for communication (Partan and Marler 1999) particularly by enhancing the detection, discrimination, and memorability of signals (Rowe 1999). For instance, the distance at which females of the wolf spider, Schizocosa ocreata, detects mates is higher when vibratory components of male displays are associated with their vision (Scheffer et al. 1996). The production of two or more types of stimulus often refines the ability to perceive a signal against background noise: the crayfish Orconectes propinquus can recognise a predator at a distance when the vision of the latter is combined with the detection of alarm substances emitted by conspecifics (Bouwma and Hazlett 2001). Finally, multimodality may improve the ability of a receiver (e.g. a predator) to learn and remember the association between another animal (e.g. a prey) and some properties of the latter (e.g. unpalatability; Guilford and Dawkins 1991).

The literature is crowded of studies that analyse the diverse sensory channels used by crustacean decapods. The role of olfaction in sex identification and mating has been investigated in several species, including clawed lobsters (e.g. Bushmann and Atema 1997, 2000), crabs (e.g. Bamber and Naylor 1997), and crayfish (e.g. Bechler 1995; Stebbing et al. 2003). The shrimp Palaemonetes pugio uses contact pheromones: individuals have to touch the cuticle of a conspecific with their antennae in order to recognise its sex and reproductive status (e.g. Caskey and Bauer 2005). In several decapod species, antennule ablation (Cowan 1991; Christofferson 1970; Bamber and Naylor 1996), but not blindfolding (Snyder et al. 1993; Bushmann 1999), affects the animal’s ability to identify the other’s sex.

However, notwithstanding the surge of interest in the communication systems of crustaceans (Atema and Steinbach 2007), multimodality has been only rarely analysed in a systematic fashion. An exception is the study by Hughes (1996a) on the shrimp Alpheus heterochaelis. Males of this species respond aggressively to visual stimuli alone, such as an open claw, and do not respond to chemical stimuli alone; but when the two are combined and the odour is released by a female, aggressive responses are suppressed.

Here, we analyse the effects of chemical and/or visual stimuli emitted by conspecifics of either sex on the behaviour of the red swamp crayfish Procambarus clarkii during mating interactions. Males (Ameyaw-Akumfi and Hazlett 1975; Dunham and Oh 1992) and females (Dunham and Oh 1996) of this species use chemical stimuli to identify the sex of a potential mate and even to assess its quality (i.e. body size, Aquiloni and Gherardi 2008a). On the contrary, little is known about the role of visual stimuli during sex identification of this taxon (but see Itagaki and Thorp 1981; Pavey and Fielder 1996; Corotto et al. 1999; Acquistapace et al. 2002; Cronin and Hariyama 2002). Finally, signal redundancy still remains unexplored in crayfish, notwithstanding its recognised central role in assuring the accurate reception and recognition of messages by the numerous species studied so far (Partan and Marler 2005).

We designed an experiment in which crayfish of the two sexes were offered with either chemical or visual stimuli from a potential mate or with the two stimuli combined. We expected that, if sex identification relies on “redundant” signal components, each type of stimulus should elicit equivalent (or enhanced) responses from the receiver regardless of whether the stimuli are presented separately or in combination (Partan and Marler 2005). Conversely, as a result of the combination of “non-redundant” signal components, distinct responses may be still elicited or one component may dominate the other or modulate its effect; alternatively, a new response may emerge.

Materials and methods

Experimental protocol

Form-I males (i.e. reproductive males, cephalothorax length, mean ± SE: 48.5 ± 2.7 mm) and sexually mature females (cephalothorax length: 46.0 ± 2.3 mm) were collected from Massaciuccoli Lake (Tuscany, Italy) and isolated in individual plastic aquaria (25 × 20 × 20 cm; water level: 10 cm). We only used sexually responsive individuals, i.e. crayfish that, once offered with a conspecific of the other sex, displayed pre-copulatory behaviours (Kasuya et al. 1996). Before copulation took place, crayfish (120 males and 120 females) were separated and isolated in individual aquaria as above for 48 h before being used in the experiment.

The experimental apparatus consisted of glass aquaria (40 × 25 × 25 cm) supplied with an aerator and filled with 12.5 L of well water at the temperature of 20 °C (Fig. 1). Aquaria contained the test crayfish in Area A, a circular glass container chemically isolated from the rest of the apparatus (10 × 4 cm bottom and 22 cm high) placed at the opposite side of the experimenter in Area B, and a black plastic wall. Aquaria were visually isolated from each other and from possible sources of disturbance. All tests were conducted between 0800 and 1400 h during July 2007. Crayfish were kept under an artificial light with a 12L:12D-hour cycle.

Fig. 1
figure 1

Experimental apparatus composed of an aquarium (40 × 25 × 25 cm) with a glass container, C (10 × 4 cm bottom and 22 cm high, area B), on the side opposite the experimenter, and a plastic wall. Single test crayfish was inserted in area A. The arrow indicates the position from which the experimenter observed crayfish

Full size image

Responsive crayfish were randomly assigned to one of three treatments: single chemical stimuli (C), single visual stimuli (V), and chemical plus visual stimuli (CV). These stimuli were produced by source crayfish of either the same (Hom) or the opposite (Het) sex. In each treatment, we tested 20 males and 20 females for both Hom and Het stimuli so that each individual was subject to two tests in two consecutive days in a random sequence (HomC and HetC, HomV and HetV, or HomCV and HetCV).

Twenty trials were run each day, the order of treatments being determined using a random number table. After 24 h acclimation, we observed the behaviour of test crayfish in two consecutive phases of 3-min each (time sufficient for a crayfish to show a clear change in its behaviour as a response to a stimulus; see Acquistapace et al. 2002): (1) a control phase following the injection of 20 ml of well water (control) into the right-hand corner of the aquarium near the experimenter in Area A; during this time, the plastic wall was left in place; and (2) a test phase corresponding to one of the six treatments; this phase started with the removal of the plastic wall. To ascertain that the injection of 20 ml of well water and the removal of the plastic wall did not disturb the test crayfish, we first conducted a preliminary test (20 replicates per sex) that followed the same procedure as above but in the absence of chemical and visual stimuli from any source crayfish.

Following the results of previous studies (e.g. Bushmann and Atema 1997; Simon and Moore 2007), we assumed that sex pheromones were released in the urine. Therefore, we kept single sexually receptive males and females as source crayfish in 2 L of well water for at least 24 h to obtain “whole body water”, i.e. water conditioned by the odour of the whole crayfish body that should also contain the putative sex pheromone. For HomC, HetC, HomCV, and HetCV treatments, 20 mL of whole body water were injected with a syringe into the right-hand corner of the aquarium near the experimenter. In HomC and HetC, the glass container in Area B remained empty. For HomV, HetV, HomCV, and HetCV, the source crayfish was either a male or a female of the same size as the test crayfish that was placed in the glass container (C in Fig. 1). This container was small enough to impede any movement to the source crayfish, thus, reducing the influence of its behaviour on the test crayfish. At the end of each session, crayfish were fed, aquaria were carefully cleaned, and the water was changed.

During the control and the test phases, we measured the time spent by test crayfish in locomotion and in aggressive posture. Locomotion was recorded because its changes in intensity denote the detection by the test crayfish of the stimuli emitted by the source conspecific (Acquistapace et al. 2002). Aggressive posture (consisting in crayfish having their body raised, their chelipeds held off the substratum and parallel to it, or higher, and the abdomen and tail fan extended; Ameyaw-Akumfi and Hazlett 1975) is used here as an indicator of sex identification: P. clarkii males and females (Ameyaw-Akumfi and Hazlett 1975; Dunham and Oh 1992) and Orconectes virilis males (Hazlett 1985) display this posture when they perceive individuals of the same—but not of the other—sex.

Statistical analyses

The data collected in the preliminary test were analysed for significance using a Wilcoxon signed ranks test (statistic: Z; Siegel and Castellan 1988). To compare the control and the test phase, we first computed the difference (in seconds) between the time that test crayfish spent in locomotion or in the aggressive posture during the test phase and the time spent in the same behaviour and posture during the control. Crayfish were then classified into two groups, i.e. (1) individuals that spent, in the test phase, the same time (±15 s) as in the control phase in locomotion or in the aggressive posture and (2) individuals that either increased (>15 s) or decreased (<−15 s) that time. Since time most often increased (see Table 2 for the few exceptions), the analyses were done on pooled data. G tests after Williams’ correction (statistic: G; Siegel and Castellan 1988) were then used to test the significance of the differences between crayfish showing and not showing temporal changes.

General Linear Models for repeated measures (GLMs, statistic: F), followed by Tukey post hoc tests, were used for each sex to analyse differences among treatments (C, V, and CV) and between sexes of the source crayfish (Hom, Het), where treatments were taken as between-subject factors and the sex of the source crayfish as a within-subject factor (Quinn and Keough 2002). Wilcoxon signed ranks tests (statistic: Z) were then used to compare the effects of each treatment between sexes of the source crayfish. Finally, the data from Hom and Het, averaged per test crayfish, were pooled to analyse differences among treatments between sexes of the test crayfish using a Mann–Whitney test (statistic: U; Siegel and Castellan 1988). The level of significance is α = 0.05.

Results

Preliminary test

In both sexes, the time spent in locomotion (males, N = 20, Z = −0.663, p = 0.541; females, N = 20, Z = −1.268, p = 0.216) or in the aggressive posture (males: N = 20, Z = −1.0, p = 1.0; females: N = 20, Z = 0, p = 1) did not change after the injection of well water and the removal of the plastic wall. This denotes that manipulation does not have a detectable effect.

Comparison between control and test phases

In the presence of each type of stimulus (except in the case of visual stimuli when test crayfish were males), the number of crayfish recorded in locomotion significantly increased from the control to the test phase. While visual stimuli alone had no clear effects on male aggressiveness, chemical stimuli (alone or combined with visual stimuli) either increased the number of aggressive males (when released by another male) or kept it at a relatively low value (when released by a female). This suggests that males require chemical stimuli to identify the sex of a conspecific. On the contrary, visual (but not chemical) stimuli from crayfish of either sex or chemical plus visual stimuli from a female increased the number of aggressive females, which pinpoints the importance of vision for females. Interestingly, the combination of the two types of stimulus from a male was associated with a very low number of aggressive females. This might denote that females, in the presence of chemical and visual stimuli, have been able to identify the other’s sex (Tables 1 and 2).

Table 1 Control vs. test phase
Full size table
Table 2 Control vs. test phase
Full size table

Comparisons among treatments and between sexes of the source crayfish

The time spent in locomotion was always independent of the sex of the source crayfish (Fig. 2a, b). Conversely, males and females differed in that responses were stronger (i.e. they showed a more intense locomotion) in the presence of chemical stimuli (alone or combined with visual stimuli) when the test crayfish were males (F 2,57 = 10.152, p = 0.0002; Tukey post hoc test: p < 0.05; Fig. 2a), and in the presence of visual stimuli (alone or combined with visual stimuli) when they were females (F 2,57 = 3.751, p = 0.029; Tukey post hoc test: p < 0.05; Fig. 2b; Table 3).

Fig. 2
figure 2

Differences between the control and the test phases (mean ± SE) in the time spent (in seconds) by P. clarkii males and females in locomotion (A, B) and in the aggressive posture (C, D): comparisons among treatments and sex of the source crayfish (Hom the same sex as the test crayfish; Het the other sex) per sex of the test crayfish. Letters over bars denote the hierarchy among treatments (C, V, and CV); asterisk denotes significant differences at p < 0.001 after a Wilcoxon signed ranks test between Hom and Het. N = 20 per treatment

Full size image
Table 3 Comparisons between sexes of the source crayfish and between sexes of the test crayfish (b) in each treatment
Full size table

Aggressiveness in males (as denoted by the time spent in the aggressive posture; Fig. 2c) did not vary with the type of stimulus (F 2,57 = 0.821, p = 0.445), but, at least in the presence of chemical stimuli (alone or combined with visual stimuli), was significantly more intense when the source crayfish was a male. Females were more aggressive in the presence of visual stimuli (alone or combined with chemical stimuli; F 2,57 = 7.448, p = 0.001; Tukey post hoc test, p < 0.05). However, the intensity of aggressiveness was significantly lower when the two types of stimulus were emitted by a male rather than by a female, as a confirmation that sex identification in females requires the co-occurrence of olfaction and vision.

Comparison between sexes of the test crayfish

Significant inter-sexual differences were found when test crayfish were offered with chemical and visual stimuli but not with the two stimuli combined. The time spent by females in locomotion and in the aggressive posture was shorter in the chemical treatment but longer in the visual treatment (Fig. 3, Table 3).

Fig. 3
figure 3

Differences between the control and the test phases (mean ± SE) in the time spent (in seconds) by P. clarkii males and females in locomotion and in the aggressive posture: comparisons between sexes per treatment. One and two asterisks denote significant differences at p < 0.01 and p < 0.001, respectively, after Mann–Whitney tests. N = 20 per treatment

Full size image

Discussion

This study shows that, to identify the sex of a conspecific, P. clarkii females rely on chemical and visual stimuli as non-redundant signal components, olfaction and vision acting in concert to evoke an adaptive response. We found, in fact, that visual (and not chemical) stimuli alone of either a male or a female conspecific elicit the aggressive behaviour of females. It is the combination of chemical and visual stimuli that suppresses female aggressiveness, but only when the source conspecific is a male.

Our results also confirm previous studies (Ameyaw-Akumfi and Hazlett 1975; Dunham and Oh 1992, 1996) showing that crayfish males identify a female through the detection of chemical stimuli alone: P. clarkii males exhibit a lower aggressiveness when smelling a female rather than a male. In other decapod species, odours were also found to be used by males to assess the receptivity (in crayfish: Villanelli and Gherardi 1998) or the ovigerous status of females (in fiddler crabs: Goshima et al. 1996). Taken together, these results might suggest that in P. clarkii, similarly to the crabs Carcinus maenas and Telmessus cheiragonus (Bamber and Naylor 1996; Kamio et al. 2000), the male is the sex involved in mate search. The exclusive use of olfaction by P. clarkii searching males is not surprising given that (1) chemical stimuli may be detected at a longer distance than visual stimuli and (2) this species lives in turbid water conditions (Gherardi 2006). Conversely, in crayfish species that inhabit clearer waters such as Austropotamobius pallipes, searching males use both olfaction and vision (Acquistapace et al. 2002); visual stimuli might provide male crayfish with additional information about, for instance, the size of the encountered female as an index of her fecundity (Nobblitt et al. 1995). The importance of odours in mate search is confirmed by the relatively few cases of decapods, in which sexual roles are inverted, such as Homarus americanus (Bushmann and Atema 1997). In this species, it is the female that, at about 5 m, identifies the other sex by odour; in the American lobster, odours are also used by the sheltering males to identify the sex of the approaching individual and to decide of whether to accept it (if a female) or to drive it out of the shelter (if a male intruder; Bushmann and Atema 2000).

Sexes differed in their responses also in the presence of visual stimuli alone. In fact, the vision of a conspecific had a slight effect on male behaviour, whereas females were found to respond to visual stimuli from either sex. The social life of females, thus, mostly relies on vision. This confirms previous evidence showing that, during mate selection (exceptions are the American lobster and the rock shrimp; Bushmann and Atema 1997, 2000; Díaz and Thiel 2004), females use visual stimuli (e.g. body or chela size) as indices of male quality (see, e.g. Forbes et al. 1992; Villanelli and Gherardi 1998; Sneddon et al. 2003; Galeotti et al. 2006; Gherardi et al. 2006; Aquiloni and Gherardi 2008a) and even eavesdrop on fighting males before choosing dominant mates (Aquiloni et al. 2008).

However, visual stimuli are not sufficient for P. clarkii females to identify the sex of a conspecific and/or to elicit their disposability to mate: only when they are combined with the odour of a male is female aggressiveness suppressed. A likely advantage offered by the non-redundant combination of odour and vision in this sex is to enhance the female ability to discriminate potential mates in the crowded social context experienced during mating period (Gherardi et al. 1999). Thanks to the enhanced discrimination ability, females might also avoid mating with low-quality mates, i.e. small-sized males (Aquiloni and Gherardi 2008b). Indeed, the reproductive investment of P. clarkii females is relatively high (Aquiloni and Gherardi 2008c) in terms of time (parental care may last 3 months; Huner and Barr 1991) and energy (each clutch is composed of 200 eggs on average).

To the best of our knowledge, sex identification in female P. clarkii is one of the few examples of non-redundancy in a bimodal communication system of invertebrates in which one signal component (odour) modulates the “message” of another (vision; Hughes 1996a, b).