With the notable exception of bee dances, there are no established examples of multimodal referential signals. The food calls of male fowl, Gallus gallus, are functionally referential and the acoustic component of a multimodal display. However, the specificity of the receiver's response to the visual component (tidbitting) has never been tested. Here we provide the first detailed analysis of tidbitting, and test the hypothesis that these characteristic movements are functionally referential. We conducted a playback experiment with five high-definition video stimuli: Silent tidbit,Matched-frequency motion in the opposite direction, Silent crows, Inactive male and Empty cage. Females searched for food more during Silent tidbitting than under any other condition, suggesting that this visual display specifically predicts the presence of food and hence has similar functional properties to food calls. Silent tidbitting was also singularly effective at evoking approach and close inspection, which may enhance signal memorability. These social responses suggest that the visual component of the display has the unique function of triggering assessment of signaler identity and quality as a potential mate. The acoustic and visual components are hence redundant as a food signal, but synergistic when additional functions are considered. These findings emphasize the perceptual complexity of multimodal displays and provide the first demonstration of multimodal referential signaling in a vertebrate.

Fundamental to any analysis of signal function is the issue of receiver response specificity. Signals produced solely in a particular context can allow receivers to identify the eliciting event, such as the discovery of food or the sudden appearance of a predator. In many cases, this predictive relationship allows adaptive decision making [e.g. food searching or flight(Evans, 1997; Hauser, 1997)]. Signals that evoke such highly specific receiver responses in the absence of contextual cues are considered functionally referential(Evans, 1997). Previous research has demonstrated the existence of acoustic(Evans et al., 1993a; Evans and Evans, 1999; Seyfarth and Cheney, 1990) and visual (Gould and Gould, 1988)referential signals and these now appear to be taxonomically widespread,occurring in species as diverse as non-human primates (e.g. Di Bitetti, 2003; Macedonia, 1990; Seyfarth et al., 1980), birds(Bugnyar et al., 2001; Evans et al., 1993a; Evans and Evans, 2007),suricates (Manser, 2001) and social insects (von Frisch,1993).

Some of the same species that produce referential signals also have multimodal communication. The honeybee, Apis mellifera, waggle dance encodes the direction, distance and profitability of a food source(von Frisch, 1974; Gould and Gould, 1988). During the waggle dance, the direction of the wagging run indicates the direction of food. This visual display is accompanied by vibratory bursts, the duration of which is correlated with the distance to food. The combined display is hence an example of multimodal referential communication.

Multimodal communication has two putative benefits: ensuring that the intended receiver perceives the signal, and maximizing information content(Hebets and Papaj, 2005; Partan and Marler, 1999; Partan and Marler, 2005; Rowe, 1999). Redundant orbackup' signals (Johnstone,1996; Zuk et al.,1993) convey the same information through each modality, thereby increasing the likelihood of signal detection in a noisy environment. By contrast, nonredundant or multiple message' displays(Johnstone, 1996; Møller and Pomiankowski,1993; Zuk et al.,1993) transmit different information via each channel,potentially increasing the rate of information transfer(Candolin, 2003; Partan and Marler, 1999; Partan and Marler, 2005). If one component of a redundant multimodal display is functionally referential,then the other component might be expected to elicit a similarly specific receiver response when produced independently. However, this prediction has never been experimentally tested.

We focus here on the visual component of a food-related audio-visual multimodal display produced by fowl (tidbitting). The acoustic component of this display (food calling) consists of series of pulsatile, cluck-like sounds(Marler et al., 1986a; Stokes and Williams, 1971),which are known to be functionally referential. Audio playbacks are sufficient to evoke substrate-search responses, in the absence of other cues(Evans and Evans, 1999), and the behavior of the hens is mediated specifically by the expectation of finding food (Evans and Evans,2007).

Under natural conditions, discovery of a palatable item by a male in the presence of a hen reliably elicits the multimodal tidbitting display(Evans and Marler, 1994; Marler et al., 1986a; Marler et al., 1986b; Stokes and Williams, 1971). This performance often entices one or more hens to approach the tidbitting male and food-search near him, sometimes taking the food item directly from his mandibles (Gyger and Marler,1988; Marler et al.,1986a; Marler et al.,1986b). In this hierarchically structured system, hens prefer to mate with males that provide food(Pizzari, 2003) and, in the presence of a hen, dominant males respond to a subordinate's food calling and tidbitting display with overt aggression, often chasing him away from the food and then food calling themselves (Stokes and Williams, 1972). On occasions, subordinate males alter the signal by producing only the visual component, suppressing the call. Experimental playbacks demonstrate that these unimodal displays still attract hens to the silently tidbitting male(Smith and Evans, 2008).

In contrast to these detailed analyses of food calling, the visual display has only been described in general terms. Davis and Domm(Davis and Domm, 1943)characterized tidbitting as a repeated, rhythmic motion of the head and neck,including repeated picking up and dropping of the food item. Wood-Gush's(Wood-Gush, 1954) description was similarly brief and he treated the two display components as a single action, whereas Davis and Domm (Davis and Domm, 1943) recognized that these often occur separately and treated them as distinct. Stokes and Williams(Stokes and Williams, 1971)provided the most detailed overview of the tidbitting display, although they did not quantify the specific motions performed or determine if there was any consistency to the sequence order. Many visual displays are stereotyped(Wiley, 1983) and include an alerting component, which engages the attention of the intended receiver(Fleishman, 1988; Rowe, 1999) thereby improving signal detectability in noisy environments. A more precise analysis of the structure of the tidbitting display is an essential prerequisite for further study of signal design.

We have previously shown (Smith and Evans, 2008) that the visual display elicits similar overall levels of food searching to the multimodal display and the acoustic component alone. However, the multimodal and visual display evoke higher levels of inspection – a conspecific recognition behavior(Dawkins, 1995; Guhl and Ortman, 1953) –than the audio alone. Hence the acoustic and visual components of male tidbitting can be classified as perceptually redundant (sensuPartan and Marler, 1999) with regard to food search duration. However, the additional social response suggests that the visual display has an additional function.

These findings are consistent with the idea that the visual display might be a food-related signal with response specificity comparable to that of food calls, in which case tidbitting would have the unusual property of being a multimodal signal in which the components in each modality were functionally referential. However, there are several possible alternative explanations for the hen's food search response. These include a release from vigilance in the presence of an alert companion (Artiss and Martin, 1995) or a general increase in foraging in response to any male movement.

In this study, we tested the specificity of hen responses to tidbitting motion. First, we classified the most common movements and described their typical order over the course of the tidbitting display. We gave males food items in the presence of an unfamiliar hen and scored the types of motor pattern produced and the probability of transitions between them. This defined the gross structure of the visual display and tested whether the temporal sequence was constrained. Results also informed the design of the playbacks that followed.

We then conducted an experiment to test perceptual processing. Stimuli were designed to evaluate a range of motions, with varying levels of spatiotemporal similarity to normal tidbitting. Hens were shown high-definition videos of a male performing four different movement patterns and control footage of an empty cage. We then analyzed video recordings of the hens' food searching and social behavioral responses. If the visual display were functionally referential, we would predict an increase in the food searching behavior specific to normal tidbitting, such that responses to this playback would be greater than those to any other type of motion and to the empty cage control.

### Subjects

Male and female golden Sebright bantam fowl, Gallus gallus(Linnaeus 1758), were the subjects for both experiments. The behavior of this strain closely resembles that of the ancestral form, the red junglefowl, Gallus gallus (Collias and Joos,1953; Collias,1987; Andersson et al.,2001; Schütz and Jensen,2001) from which all domesticated strains have been derived(Fumihito et al., 1994; Fumihito et al., 1996). In particular, Sebrights have not been subjected to artificial selection for rapid growth or egg production.

Birds were housed in 1.0 m×1.0 m×0.6 m (length × width× height) cages in a climate-controlled room maintained at 22°C on a 12 h:12 h day:night cycle and given ad libitum access to high-protein food (Gordon Specialty Feeds, Sydney, Australia) and water in their home cages. Twenty-five males participated in experiment I, which measured the structure of tidbitting. Each male was housed with a single female. Twenty-four different hens participated in experiment II, which tested specificity of response. These hens were house in same-sex pairs.

Recordings for creation of playback stimuli and behavioral testing were conducted in a sound-attenuating chamber (Ampisilence S.p.a., Roberssomero,Italy 2.38 m×2.38 m×2.15 m), which was lined with 10 cm Sonex'foam baffles (Illbruck, Minneapolis, USA) on the side walls and 15 cm baffles on the ceiling to prevent reverberation.

### Experiment I: tidbitting performance

#### Test apparatus

During each trial, the male was confined to a steel-framed wire cage (0.60 m×0.45 m×0.86 m), which had an audience hen cage of similar size abutting the left side wall and a food dispenser on the opposite wall (for details, see Evans and Evans,1999). We monitored tests via a Sony 1450QM color monitor, connected to a Panasonic WV-CL320 video camera in the sound chamber. A Panasonic WJ-810 time-date generator provided a stopwatch' at the bottom of the video display for timing test sessions. All tests were video recorded using a Panasonic AG-7750 VHS-format deck.

#### Design and test procedure

Males were placed in the test cage for 15 min each on three consecutive days to habituate them to the test environment. We placed each male's cage-mate in the adjoining audience cage during these sessions to facilitate habituation. No food was presented.

On the fourth day, the male was again placed in the test cage, but with an unfamiliar hen in the audience cage. Unfamiliar hens were used as audience because males are most likely to display for an unfamiliar female, although the performance does not differ based on familiarity (C.S.E., unpublished). After a 5 min delay, which allowed the male to settle after handling, the food dispenser was activated, delivering five corn kernels onto the floor of his cage.

We recorded the type of movement and position of the male's head in space in every video frame (PAL standard; 40 ms time interval). Movements fell naturally into three discrete classes, which we coded as follows: twitch'(Fig. 1A): a rapid horizontal side-to-side motion of the head with the neck held fully upright; short bob'(Fig. 1B): abrupt vertical movement of the head from a fully upright position to a point halfway above the floor, returning to the upright position; and long bob'(Fig. 1C): vertical movement of the head, plunging through the full arc toward the floor, often picking up the food item with the mandibles, and ending with the head in the upright position. We recorded all transitions between these motor patterns and calculated mean probabilities, which were then summarized in a kinematic plot(Lehner, 1979). Results of this analysis informed the design of the experiment II stimuli.

Fig. 1.

Classification of tidbitting movements. (A) Twitch: side-to-side horizontal movement of the head with the neck upright; (B) short bob: vertical movement of the head from a fully upright position, stopping abruptly when approximately horizontal; and (C) long bob: vertical movement of the head from upright through a full arc towards the floor. This often includes seizing the food item between the mandibles. The small white object on the floor is a chicken dropping.

Fig. 1.

Classification of tidbitting movements. (A) Twitch: side-to-side horizontal movement of the head with the neck upright; (B) short bob: vertical movement of the head from a fully upright position, stopping abruptly when approximately horizontal; and (C) long bob: vertical movement of the head from upright through a full arc towards the floor. This often includes seizing the food item between the mandibles. The small white object on the floor is a chicken dropping.

### Experiment II: female response to visual tidbitting display

#### Test apparatus

The test apparatus consisted of a 106 cm Sony high-definition flat panel plasma display (resolution 1080 by 1960 pixels), mounted at floor level next to a 1.2 m×0.30 m×0.5 m cage with a remote-controlled wire door 0.4 m from one end. Previous research has demonstrated the hens can recognize the movements of conspecifics on video screens(McQuoid and Galef, 1993) and that video playback evokes natural anti-predator and social responses(Evans and Marler, 1991; Evans et al., 1993a; Evans et al., 1993b). The HDV format used in this study provides an approximately four-fold increase in resolution over standard-definition digital video and has successfully been used in multimodal playbacks with hens(Smith and Evans, 2008).

Decisions about the overall layout of the test setup were informed by the well-described properties of the fowl visual system. Hens recognize conspecifics using close binocular inspection of the other bird's head and neck region (Guhl and Ortman,1953), but they are myopic in the frontal field and so unable to determine individual identity from distances greater than 30 cm(Dawkins, 1995; Dawkins, 1996; Dawkins and Woodington, 1997). We hence positioned the end of the test cage 30 cm from the plasma display, a distance at which a hen should attempt social recognition by fixating on the screen. Note that this spatial separation was also sufficient to prevent hens from resolving individual pixels, with a concomitant loss of verisimilitude. The long axis of the cage was perpendicular to the plasma display and the remote-controlled door was at the end farthest from it.

#### Stimulus design

We used a 3-CCD high-definition video camera (Sony HDR-FX1) to make high-definition (1920×1080 lines) video recordings of 12 male fowl, Gallus gallus, performing the tidbitting display. During recording,the video signal was monitored on the plasma display later employed for playbacks, allowing us to adjust the camera zoom so that the image of the male was precisely life-sized. Males were confined in a 0.60 m×0.45 m×0.86 m wire cage, 0.8 m from the camera. The audience hen was held in a separate cage, approximately 30 cm from that of the male. After a 10 min acclimation period, four mealworms were delivered from a remote-controlled food hopper mounted above the male's cage. This usually evoked tidbitting and food calling from the male.

We selected four males using the same criteria as in our previous study(Smith and Evans, 2008). Raw footage was then edited using Final Cut Pro 5.1 (Apple Computer) to create five stimuli from each male, each of 15 min duration. These consisted of an empty cage for the first 5 min, to allow the hen to settle, followed by 5 min of a male standing alone, which provided a baseline measurement for food searching in the company of a vigilant companion, then 60 s of one of the five test sequences, followed by 4 min of empty cage. Each successive phase of the test stimuli began and ended with a 0.5 s fade transition to avoid evoking a startle response. The first 10 min and the last 4 min of every trial were hence identical across every condition. Only the 60 s test sequence differed across conditions.

Test sequences were designed systematically to assess the specificity of female responses to male display movements. The five stimuli were Silent tidbitting (normal signal), Matched-frequency motion in the opposite direction, Silent crows, Inactive male and Empty cage. Decisions about design of the stimuli were informed by the results of experiment I, as detailed below.

##### Silent tidbitting

Silent tidbitting consisted of the male performing the full natural display, including twitches, short bobs and long bobs(Fig. 1).

##### Matched-frequency motion in the opposite direction

Analysis of the tibitting movements (experiment I) revealed that the majority (77%) of the movements during a typical tidbitting display include a rapid downward sweep and return upward of the head and neck. To test whether this characteristic, rapid, vertical movement was a crucial signal component,independent of temporal pattern, we created stimuli with Matched-frequency motion in the opposite direction. Each of the Silent tidbitting exemplars was used as a template; we edited spontaneous crowing of the same male so that each upward extension of the neck during the crowing motion corresponded to one of the long, downward sweeping motions in the Silent tidbitting sequence(Fig. 2A,B). Matched pairs of completed stimuli contained movement sequences with similar timing and amplitude of displacement in the vertical plane (range 19–22 movements)but in opposite directions.

##### Silent crows

To assess the role of movement frequency, we created exemplars of Silent crows, in which the male performed four upward neck extensions with 12 s of resting behavior between them; thereby creating a stimulus with a fivefold reduction in the frequency of movements compared to the Silent tidbitting and Matched-frequency motion stimuli. Crowing was chosen as the motor pattern for these two stimuli because it has a similar duration and distance of displacement to the long bob motor pattern in tidbitting. It is also a natural motion performed spontaneously by males. Unlike the acoustic component of crowing, which has a well-established territorial function(Miller, 1978), there is no evidence that the accompanying movement has specific social salience in the absence of vocalizations.

##### Inactive male

To test the effect of an alert companion, we created the Inactive male playbacks, in which the male maintained a stationary alert posture, with occasional spontaneous side-to-side head movements, appearing to scan horizontally around the cage, but without any vertical displacement of the head.

##### Empty cage

The Empty cage video was identical to the other stimuli in every respect,except that the male was absent. We standardized the five test sequences for several parameters that seemed likely to influence hen responses. To prevent potential interactions between food calls and the motor patterns, which were the focus of the study, all playbacks were silent. Mealworms appeared and remained on the video screen during all five 60 s test sequences to control for the presence of a preferred food item. Finally, to reduce the possibility of social facilitation of food search caused by observation of a feeding companion (Zajonc, 1965; McQuoid and Galef, 1993), the males were not shown consuming the mealworms. Results from experiment I revealed no obligatory transitions between the three motor patterns(Fig. 3), suggesting that temporal sequence is unlikely to be an important aspect of display structure;hence we did not manipulate transition patterns in experiment II.

Fig. 2.

Images of Silent tidbitting (A) and Matched motion (B) stimuli. Each long bob in the Silent tidbitting exemplar was paired with a crow by the same male in the Matched motion exemplar.

Fig. 2.

Images of Silent tidbitting (A) and Matched motion (B) stimuli. Each long bob in the Silent tidbitting exemplar was paired with a crow by the same male in the Matched motion exemplar.

#### Test procedure

Hens were individually placed in the test cage for four 15 min periods, at intervals of 48 h, to acclimate them to the apparatus and sound chamber. At the start of each session, the wire door was closed and the hen was confined to the end section of the cage, preventing her from approaching the plasma display. During each session the wire door was remotely released once and the empty cage video, without mealworms, played on the plasma screen. By the fourth acclimation session, all hens readily emerged and walked the length of the cage after the door opened and none exhibited signs of disturbance such as wing-flapping or crouching.

We used a within-subjects design in which each hen was first assigned one of the four male exemplars and then a unique random sequence of the five treatments (test sequences). To ensure that the video male was unfamiliar to the hen, we applied the constraint that she should have had no social contact with the real male for at least six months prior to the experiment. Hens were tested at the same time of day to minimize diel variation in behavior. The inter-trial interval was 48 h. Each trial began with the hen behind the closed wire door in the section of the cage farthest from the plasma screen. This standardized the hen's minimum distance from the video male at the beginning of the 60 s stimulus. For the first 10 min, the door remained closed while the empty cage and then the male video played. The test sequence (i.e. one of the four movement types or Empty cage control) then began and the remote control door was opened, allowing the hen to approach the video male.

Fig. 3.

Kinematic plot of motor pattern frequency and transition probabilities during tidbitting display. Diameter of circles is proportional to the frequency of occurrence of each motor pattern (twitch: 23%; short bob: 40%;long bob: 37%). Width of connecting bars is proportional to frequency of transitions between the motor patterns. Bars of same shading add to 100% of transitions. Specific transition frequencies were as follows: twitch to short bob (78.8%), to long bob (21.2%); short bob to twitch (63%), to long bob(37%); and long bob to twitch (15%), to short bob (85%).

Fig. 3.

Kinematic plot of motor pattern frequency and transition probabilities during tidbitting display. Diameter of circles is proportional to the frequency of occurrence of each motor pattern (twitch: 23%; short bob: 40%;long bob: 37%). Width of connecting bars is proportional to frequency of transitions between the motor patterns. Bars of same shading add to 100% of transitions. Specific transition frequencies were as follows: twitch to short bob (78.8%), to long bob (21.2%); short bob to twitch (63%), to long bob(37%); and long bob to twitch (15%), to short bob (85%).

We monitored the tests using a CCD camera (Panasonic WV-CL320) connected to a Sony color monitor (Sony PVM-1450QM). The video output was converted into MPEG-2 format using a Miglia-EvolutionTV and saved for later analysis. Behaviors of interest were scored using JWatcher Video 1.0(Blumstein et al., 2006), which reads the time-code of the video file to permit single-frame resolution. We measured the duration of food searching, which is characterized by distinctive close binocular fixation of the substrate(Evans and Evans, 1999; Evans and Evans, 2007) and two social responses: time spent in close proximity to video male, indicated by approach to within 0.1 m of the end of the cage closest to the screen, and inspection behavior, which was characterized by the hen stretching her neck towards the flat panel monitor at the height of the male's head, exactly as hens scrutinize other flock members (Guhl and Ortman, 1953).

Fig. 4.

Food search duration. Time spent food searching during 60 s playbacks(values are means ± s.e.m.). Scores adjusted for individual baseline level (see text for details). Different letters indicate significant differences (P<0.05) as determined by repeated-measures ANOVA,followed by post-hoc Tukey's HSD. Inset shows characteristic binocular close inspection of substrate during food search.

Fig. 4.

Food search duration. Time spent food searching during 60 s playbacks(values are means ± s.e.m.). Scores adjusted for individual baseline level (see text for details). Different letters indicate significant differences (P<0.05) as determined by repeated-measures ANOVA,followed by post-hoc Tukey's HSD. Inset shows characteristic binocular close inspection of substrate during food search.

There is considerable individual variation among hens in time allocated to foraging. To increase the sensitivity of our response measure, we used a difference score, adjusting responses to playback with a corresponding estimate of the spontaneous level of food searching(Evans and Evans, 1999; Evans and Evans, 2007). This was calculated by measuring each hen's total duration of food searching during the 60 s test sequence and then subtracting a baseline estimate, which was obtained by averaging two samples: her time spent food searching during the first 60 s after the video male appeared (min 6 of 15 min trial period) and that during the last 60 s before the test playback began (min 9). This approach also further controlled for day-to-day variation.

Tests for overall treatment effects were conducted with repeated measures ANOVAs (SPSS 15.0.6 for Windows), using male exemplar as a blocking factor. Exemplar was never significant, so all data were pooled for further analysis. Significant differences were further explored using Tukey's honestly significant difference (HSD) to conduct multiple pair-wise comparisons; this maintained the overall alpha level at the nominated value of 0.05.

### Experiment I

We analyzed the average frequency of each motor pattern in the tidbitting display and the transition probabilities between them(Fig. 3). Short bob' and long bob' combined accounted for the majority of the movements (40% and 37%,respectively). Although the individual motor patterns appear stereotyped and are readily distinguishable, the sequences are relatively unconstrained. Over the course of the display, males switched continually between the three motor patterns. Transitions between twitch' and short bob' were common and ashort bob' was most frequently followed by a long bob'(Fig. 3). Transitions betweenlong bob' and twitch' were less frequent.

### Experiment II

#### Food search duration

Analysis of food searching duration, adjusted for baseline rate, revealed that Silent tidbitting playbacks evoked a significantly higher response than any of the other test sequences (Fig. 4). Responses to Inactive male, Silent crows and Matched motion in the opposite direction did not differ significantly from one another, despite considerable variation in the frequency of movement depicted in these three stimulus types. In addition, Silent crows and Matched motion were not significantly different from Empty cage (F4,88=9.61, P<0.0001; Tukey's HSD, P<0.05).

Fig. 5.

Close approach to video male. Time spent within 0.1 m of the video male during 60 s test stimulus (values are means ± s.e.m.). Different letters indicate significant difference (P<0.05) as determined by repeated-measures ANOVA, followed by post-hoc Tukey's HSD. Inset shows hen standing in region closest to plasma screen (white line added for illustration).

Fig. 5.

Close approach to video male. Time spent within 0.1 m of the video male during 60 s test stimulus (values are means ± s.e.m.). Different letters indicate significant difference (P<0.05) as determined by repeated-measures ANOVA, followed by post-hoc Tukey's HSD. Inset shows hen standing in region closest to plasma screen (white line added for illustration).

#### Proximity to video male

We measured approach to the video male by scoring the time each hen spent within 0.1 m of the end of the cage closest to the plasma screen. Mauchly's test of sphericity was significant (P<0.05), so we applied a Huynh-Feldt correction (ϵ=0.85). The overall treatment (stimulus type)effect was highly significant (F3,75=7.52, P<0.001). Post-hoc tests revealed that hens spent significantly more time close to the simulated male in the Silent tidbitting playbacks than in any other treatment. There were no significant differences among the other four stimuli (Tukey's HSD, P<0.05; Fig. 5).

#### Inspection

As a complementary measure of social response, we measured the amount of time hens spent in inspection behavior(Fig. 6). Mauchly's test of sphericity was significant (P<0.05), so we applied a Huynh-Feldt correction (ϵ=0.78). The overall treatment effect was highly significant(F3,68=40.04, P<0.001). Post-hoctests revealed that females spent significantly more time inspecting the male on the plasma screen during the Silent tidbitting sequences than in any of the other playbacks. None of the other treatments were significantly different from one another (Tukey's HSD, P<0.05).

Fig. 6.

Inspection of video male. Time hens engaged in binocular fixation directed toward the video male during 60 s test stimulus (values are means ±s.e.m.). Different letters indicate significant difference(P<0.05) as determined by repeated-measures ANOVA, followed by post-hoc Tukey's HSD. Inset shows hen inspecting video male.

Fig. 6.

Inspection of video male. Time hens engaged in binocular fixation directed toward the video male during 60 s test stimulus (values are means ±s.e.m.). Different letters indicate significant difference(P<0.05) as determined by repeated-measures ANOVA, followed by post-hoc Tukey's HSD. Inset shows hen inspecting video male.

### Structure of the tidbitting display

The visual component of the tidbitting display is made up of three distinct motor patterns: twitch, short bob and long bob(Fig. 1). Displays are primarily composed of the short bob and long bob motor patterns, with relatively fewer twitches (Fig. 3). Over the course of the displays, the males continuously transitioned between the three motor patterns. These motor pattern sequences are highly variable, with no indication of temporal stereotypy(Fig. 3), suggesting that none of them has evolved specifically to have an alerting function (i.e. design for particular efficacy in engaging a visual grasp reflex), as in some other dynamic visual signals (Peters and Evans,2003). One possible explanation for this may be that the displays contain far more motion in the vertical plane than in the horizontal. Such signal design should generate a large sweep area in the visual field of potential receivers that is robust to variation in angle of view and therefore should be consistently conspicuous (Marr,1982; Peters et al.,2002; Peters and Evans,2007). A second possibility is that the visual display is typically accompanied by food calls, which are easy to localize and audible over a considerable distance (Stokes and Williams, 1971), thereby relaxing selection for a specialized initial movement. A third possibility is that the wattles, which move considerably, swinging from side to side, during each type of motor pattern and produce a large proportion of the apparent motion during the display(C.L.S., unpublished), may so enhance detectability that no constraint on movement sequence is necessary. Planned experiments will test these predictions.

### Referential signaling

We tested whether the food search response evoked by visual tidbitting is specifically dependent on the spatiotemporal characteristics of the display,including frequency and direction. In its most restrictive form, this hypothesis requires that hens respond significantly more to a silently tidbitting male than to any other motion type. Comparisons of food-searching duration reveal precisely the predicted pattern of responses(Fig. 4). It is particularly striking that Matched control movements at the same frequency as tidbitting,but in the opposite direction, were much less effective(Fig. 4). This implies that the downward direction of motion is an important component of the signal.

This experiment also demonstrates that the hens' increased food searching is not simply a consequence of a decrease in vigilance caused by the presence of an alert companion (Artiss and Martin,1995). Although the difference between the Inactive male and Empty cage is consistent with such an effect, the fourfold increase in food searching to the Silent tidbitting display over that recorded with the Inactive male reveals an additional response specific to the tidbitting signal(Fig. 4). We conclude the food searching response to a silently tidbitting male is not caused by a change in the foraging/vigilance tradeoff (Artiss and Martin, 1995). In addition, the increase in food searching response to Silent tidbitting cannot be attributed to social facilitation(Zajonc, 1965; McQuoid and Galef, 1993)because the playback male was shown signaling, rather than searching the substrate. We conclude that the tidbitting display is sufficient to evoke foraging behavior and that responses to these movements have specificity similar to that previously demonstrated for food calls(Evans and Evans, 1999).

Tidbitting movements. therefore have all the characteristics of a functionally referential visual signal. When combined with food calling, as in the majority of displays, this constitutes the first experimental demonstration of multimodal referential signaling in a non-human vertebrate.

### Social responses

For males, achieving close proximity to a hen is an important factor in mating success (Graves et al.,1985). The playback experiment reveals that tidbitting can play an important role: hens spent more time standing close to the silent tidbitting male than they did during any other treatment(Fig. 5). This visual display was uniquely effective, both relative to other movements and in absolute terms; none of the other video male stimulus evoked significantly more close approach than the Empty cage control sequence. Hens, therefore, responded very specifically to the visual display, and not simply to rapid motion or to the presence of a simulated companion. It has been suggested that one of the functions of the food calling might be to lure the intended receiver close to the signaler (Stokes and Williams,1971) and many descriptions of multimodal tidbitting have noted that the hen typically approaches the calling male(Collias and Collias, 1996; Stokes and Williams, 1971; Wood-Gush, 1954). Our experimental results confirm that the visual display alone is sufficient to evoke this social response.

Tidbitting may also improve memorability of both the signal and signaler,an important aspect of visual signal design(Guilford and Dawkins, 1991)that has primarily been explored in the context of aposematic coloration(Halpin et al., 2008). Hens spent approximately four times longer inspecting the silent tidbitting male than any of the other male stimuli, none of which differed from Empty cage(Fig. 6). Close inspection, as indicated by the hen binocularly fixating on the video male's head, is hence uniquely triggered by tidbitting. Previous research(Dawkins, 1995; Dawkins, 1996; Hodos, 1993) has implicated binocular fixation with the frontal field as a critical process in conspecific recognition and has established that hens use the area around the eye to identify flock mates (Dawkins,1996). Hens prefer males that tidbit more(Pizzari, 2003; Zuk et al., 1993), so it seems likely that the inspection of the male's head evoked by tidbitting may facilitate formation of an association between the appearance of a particular male and provisioning with food. Mating does not usually occur immediately after tidbitting (Stokes and Williams,1971), so females must retain some memory of individual male tidbitting performance for preference subsequently to be expressed.

Although the visual and acoustic components of this multimodal display are redundant (sensuPartan and Marler, 1999), with regard to predicting food availability(Smith and Evans, 2008), the visual display has a synergistic effect on social responses that the sounds do not (Evans and Evans, 1999),increasing the time spent in close proximity and inspecting the male. This contrast presents a challenge for theoretical models of multimodal signaling,since categorization will be sensitive to the function(s) of interest.

### Unimodal production of the tidbitting signal: functional implications

Short-term changes in signal structure can reduce potential costs(Ryan et al., 1982). Conspicuous signals attract predators(Bayly and Evans, 2003; Roberts et al., 2007),parasites (Bernal et al., 2006)and competitors (Stokes and Williams,1971). In fowl, dominant males obtain the majority of matings(Pizzari, 2003) and are aggressive towards subordinate males that tidbit, often displacing them and taking the food item (Stokes and Williams,1971). However, Johnsen et al.(Johnsen et al., 2001)observed that subordinate males were able to tidbit as frequently as alphas when the subordinate male could display out of sight of the alpha. In flocks living under naturalistic conditions, we have observed that subordinate males tidbit without perceptible food calling. This behavior created additional mating opportunities by attracting nearby hens, which approached and food searched near him (C.L.S., unpublished). If the silent tidbitting signal is less conspicuous than the multimodal display, then this behavior may reflect a tradeoff by subordinates between the benefit of attracting females and the social cost of increased conspicuousness to a dominant male. Planned studies will test the conspicuousness of unimodal and multimodal signals and the frequency of their occurrence as a function of social context.

These experiments comply with the Principles of animal care', publication No. 86-23, revised 1985, of the National Institute of Health, and the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (NHMRC, 1997). We thank R. Miller and C. Jude for bird care, and R. Marshall for veterinary support. We also thank A. Taylor for assistance with the statistical analysis. This research was supported by a grant to C.S.E. from the Australian Research Council.

Andersson, M., Nordin, E. and Jensen, P.(
2001
). Domestication effects on foraging strategies in fowl.
Appl. Anim. Behav. Sci.
72
,
51
-62.
Artiss, T. and Martin, K. (
1995
). Male vigilance in white-tailed ptarmigan, Lagopus leucurus: mate guarding or predator detection?
Anim. Behav.
49
,
1249
-1258.
Bayly, K. L. and Evans, C. S. (
2003
). Dynamic changes in alarm call structure: a strategy for reducing conspicuousness to avian predators?
Behaviour
140
,
353
-369.
Bernal, X., Rand, A. S. and Ryan, M. J. (
2006
). Acoustic preferences and localization performance of blood-sucking flies(Corethrella Coquillett) to túngara frog calls
Behav. Ecol.
17
,
709
-715.
Blumstein, D., Evans, C. S. and Daniel, J. C.(
2006
).
JWatcher 1.0
. http://www.jwatcher.ucla.edu/.
Bugnyar, T., Kijne, M. and Kotrschal, K.(
2001
). Food calling in ravens: are yells referential signals?.
Anim. Behav.
61
,
949
-958.
Candolin, U. (
2003
). The use of multiple cues in mate choice.
Biol. Rev.
78
,
575
-595.
Collias, N. E. (
1987
). The vocal repertoire of the red junglefowl: a spectrographic classification and the code of communication.
Condor
89
,
510
-524.
Collias, N. E. and Collias, E. C. (
1996
). Social organization of red junglefowl, Gallus gallus, population related to evolution theory.
Anim. Behav.
51
,
1337
-1354.
Collias, N. E. and Joos, M. (
1953
). The spectrographic analysis of sound signals of the domestic fowl.
Behaviour
5
,
175
-188.
Davis, D. E. and Domm, L. V. (
1943
). The influence of hormones on the sexual behavior of domestic fowl. In
Essays in Biology
, pp.
171
-181. Berkeley, CA: University of California Press.
Dawkins, M. S. (
1995
). How do hens view other hens? The use of lateral and binocular visual fields in social recognition.
Behaviour
132
,
591
-606.
Dawkins, M. S. (
1996
). Distance and social recognition in hens: implications for the use of photographs as social stimuli.
Behaviour
133
,
663
-680.
Dawkins, M. S. and Woodington, A. (
1997
). Distance and the presentation of visual stimuli to birds.
Anim. Behav.
54
,
1019
-1025.
Di Bitetti, M. S. (
2003
). Food-associated calls of tufted capuchin monkeys (Cebus apella nigritus) are functionally referential signals.
Behaviour
140
,
565
-592.
Evans, C. S. (
1997
). Referential signals.
Perspect. Ethol.
12
,
99
-143.
Evans, C. S. and Evans, L. (
1999
). Chicken food calls are functionally referential.
Anim. Behav.
58
,
307
-319.
Evans, C. S. and Evans, L. (
2007
). Representational signaling in birds.
Biol. Lett.
3
,
8
-11.
Evans, C. S. and Marler, P. (
1991
). On the use of video images as social stimuli in birds: audience effects on alarm calling.
Anim. Behav.
41
,
17
-26.
Evans, C. S. and Marler, P. (
1994
). Food calling and audience effect in male chickens, Gallus gallus: their relationships to food availability, courtship and social facilitation.
Anim. Behav.
47
,
1159
-1170.
Evans, C. S., Evans, L. and Marler, P. (
1993a
). On the meaning of alarm cells: functional reference in an avian vocal system.
Anim. Behav.
46
,
23
-38.
Evans, C. S., Macedonia, J. M. and Marler, P.(
1993b
). Effects of apparent size and speed on the response of chickens, Gallus gallus, to computer-generated simulations of aerial predators.
Anim. Behav.
46
,
1
-11.
Fleishman, L. J. (
1988
). Sensory influence on physical design of a visual display.
Anim. Behav.
36
,
1420
-1424.
Fumihito, A., Miyake, T., Sumi, S. I., Takada, M., Ohno, S. and Kondo, N. (
1994
). One subspecies of the red junglefowl(Gallus gallus gallus) suffices as the matriarchic ancestor of all domestic breeds.
91
,
12505
-12509.
Fumihito. A., Miyake, T., Takada, M., Shingu, R., Endo, T.,Gojobori, T., Kondo, N. and Ohno, S. (
1996
). Monophyletic origin and unique dispersal patterns of domestic fowls.
93
,
6792
-6795.
Gould, J. L. and Gould, C. G. (
1988
). Evolution of the dance. In
The Honey Bee
, pp.
101
-123. New York: Scientific American Library.
Graves, H. B., Hable, C. P. and Jenkins, T. H.(
1985
). Sexual selection in gallus: effects of morphology and dominance on female spatial behavior.
Behav. Processes
11
,
189
-197.
Guhl, A. M. and Ortman, L. L. (
1953
). Visual patterns in the recognition of individuals among chickens.
Condor
55
,
287
-298.
Guilford, T. and Dawkins, M. S. (
1991
). Receiver psychology and the evolution of animal signals.
Anim. Behav.
42
,
1
-14.
Gyger, M. and Marler, P. (
1988
). Food calling in the domestic fowl, Gallus gallus: the role of external referents and deception.
Anim. Behav.
36
,
358
-365.
Halpin, C., Skelhorn, J. and Rowe, C. (
2008
). Naïve predators and selection for rare conspicuous defended prey: the initial evolution of aposematism revisited.
Anim. Behav.
75
,
771
-781.
Hauser, M. D. (
1997
).
The Evolution Of Communication
. Cambridge, MA: The MIT Press.
Hebets, E. I. and Papaj, D. R. (
2005
). Complex signal function: developing a framework of testable hypotheses.
Behav. Ecol. Sociobiol.
57
,
197
-214.
Hodos, W. (
1993
). The visual capabilities of birds. In
Vision, Brain and Behavior in Birds
(ed. by H. P. Zeigler and H. J. Bischof), pp.
63
-76. Cambridge, MA: MIT Press.
Johnsen, T. S., Zuk, M. and Fessler, E. A.(
2001
). Social dominance, male behaviour and mating in mixed-sex flocks of red jungle fowl.
Behaviour
138
,
1
-18.
Johnstone, R. A. (
1996
). Multiple displays in animal communication: backup signals and multiple messages'.
Philos. Trans. R. Soc. Lond. B Biol. Sci.
351
,
329
-338.
Lehner, P. N. (
1979
).
Handbook of Ethological Methods
, pp.
316
-319. New York: Graland STPM Press.
Macedonia, J. M. (
1990
). What is communicated in the antipredator calls of lemurs: evidence from antipredator call playbacks to ring tailed and ruffled lemurs.
Ethology
86
,
177
-190.
Manser, M. B. (
2001
). The acoustic structure of suricates' alarm calls varies with predator type and level of response urgency.
Proc. Biol. Sci.
268
,
2315
-2324.
Marler, P., Dufty, A. and Pickert, R. (
1986a
). Vocal communication in the domestic chicken. I. Does the sender communicate information about the quality of a food referent to a receiver?
Anim. Behav.
34
,
188
-193.
Marler, P., Dufty, A. and Pickert, R. (
1986b
). Vocal communication in the domestic chicken. II. Is a sender sensitive to the presence and nature of a receiver?
Anim. Behav
,
34
,
194
-198.
Marr, D. (
1982
).
Vision: A Computational Investigation into the Human Representation and Processing of Visual Information
. San Francisco: W. H. Freeman.
McQuoid, L. M. and Galef, B. G., Jr (
1993
). Social stimuli influencing feeding behaviour of Burmese red junglefowl: a video analysis.
Anim. Behav.
46
,
13
-22.
Miller, D. (
1978
). Species typical and individually distinctive acoustic features of crow calls of red junglefowl.
Z. Tierpsychol.
47
,
182
-193.
Møller, A. P. and Pomiankowski, A.(
1993
). Why have birds got multiple sexual ornaments?
Behav. Ecol. Sociobiol.
32
,
167
-176.
Partan, S. and Marler, P. (
1999
). Communication goes multimodal.
Science
283
,
1272
-1273.
Partan, S. R. and Marler, P. (
2005
). Issues in the classification of multimodal communication signals.
Am. Nat.
166
,
231
-245.
Peters, R. A. and Evans, C. S. (
2003
). Introductory tail-flick of the Jacky dragon visual display: signal efficacy depends upon duration.
J. Exp. Biol.
206
,
4293
-4307.
Peters, R. A. and Evans, C. S. (
2007
). Active space of a movement-based signal: response to Jacky dragon (Amphibolurus muricatus) display is sensitive to distance, but independent of orientation.
J. Exp. Biol.
210
,
395
-402.
Peters, R. A., Clifford, C. W. G. and Evans, C. S.(
2002
). Measuring the structure of dynamic visual signals.
Anim. Behav.
64
,
131
-146.
Pizzari, T. (
2003
). Food, vigilance, and sperm:the role of male direct benefits in the evolution of female preference in a polygamous bird.
Behav. Ecol.
14
,
593
-601.
Roberts, J. A., Taylor, P. W. and Uetz, G. W.(
2007
). Consequences of complex signaling: predator detection of multimodal cues.
Behav. Ecol.
18
,
236
-240.
Rowe, C. (
1999
). Receiver psychology and the evolution of multicomponent signals.
Anim. Behav.
58
,
921
-931.
Ryan, M. J., Tuttle, M. D. and Rand, A. S.(
1982
Am. Nat.
119
,
136
-139.
Schütz, K. E. and Jensen, P. (
2001
). Effects of resource allocation on behavioural strategies: a comparison of red junglefowl (Gallus gallus) and two domesticated breeds of poultry.
Ethology
107
,
753
-765.
Seyfarth, R. M. and Cheney, D. L. (
1990
). The assessment by vervet monkeys of their own and another species' alarm calls.
Anim. Behav.
40
,
754
-764.
Seyfarth, R. M., Cheney, D. L. and Marler, P.(
1980
). Monkey responses to three different alarm calls: evidence of predator classification and semantic communication.
Science
210
,
801
-803.
Smith, C. L. and Evans, C. S. (
2008
). Multimodal signaling in fowl, Gallus gallus.
J. Exp. Biol.
211
,
2052
-2057.
Stokes, A. W. and Williams, H. W. (
1971
). Courtship feeding in Gallinaceous birds.
Auk
88
,
543
-559.
Stokes, A. W. and Williams, H. W. (
1972
). Courtship feeding in Gallinaceous birds.
Auk
89
,
177
-180.
von Frisch, K. (
1974
). Decoding the language of the bee.
Science
185
,
663
-668.
von Frisch, K. (
1993
).
The Dance Language and Orientation of Bees
, pp.
301
-304. New York: Harvard University Press.
Wiley, R. H. (
1983
). The evolution of communication: information and manipulation. In
Communication
(ed. T. R. Halliday and P. J. B. Slater), pp.
156
-189. Oxford: Blackwell.
Wood-Gush, D. G. M. (
1954
). The courtship of Brown Leghorn Cock.
Anim. Behav.
2
,
95
-102.
Zajonc, R. B. (
1965
). Social facilitation.
Science
149
,
269
-274.
Zuk, M., Ligon, J. D. and Thornhill, R. (
1993
). Effects of experimental manipulation of male secondary sex characters on female mate preference in red jungle fowl.
Anim. Behav.
44
,
999
-1006.