Mud fiddler crabs, Uca pugnax, have a streak of blue coloration located on the front of the carapace above the mouth and centered between the eyes. We documented that this blue streak is absent in juveniles and develops as crabs become sexually mature. By photographing male crabs under controlled conditions in the laboratory, we demonstrated that the brightness of the blue streak (in comparison with the rest of the carapace) is dynamic, and can dim from bright blue to nearly black in fewer than two minutes. We examined blue streak variability in male crabs in response to physical factors (light and temperature) and social context to begin to understand what causes its dynamic response. The blue streak darkens in response to decreased ambient light, but does not respond to changes in temperature. In the field, it is brighter when crabs are roaming on the mudflat or fighting, but darker when crabs are basking or performing waving displays. The highly visual nature of fiddler crabs and the dynamic character of the blue streak suggest that it may communicate information about the state of a crab or its environment.
The coloration patterns of crabs are often dramatic, and are usually a part of species descriptions. In many species, these colors are not static, but have been shown to change depending on lighting conditions, the tidal cycle, and temperature (e.g., Wilkens & Fingerman 1965; Crane 1975; Warner 1977; Palmer 1995; Hemmi et al. 2006; Silbiger & Munguia 2008). More recently, it has become clear that color changes can occur over different time scales, and that the use of color as a visual signal is pervasive among crab species. Changes in color have been shown to be associated with crab size (Detto et al. 2004; Casariego et al. 2011), attracting a mate (reviewed in deRivera & Vehrencamp 2001; see also Baldwin & Johnsen 2009; Todd et al. 2011), stress (reviewed in Zeil & Hemmi 2006), predation threats (Hemmi et al. 2006), and inter-male agonistic behavior (Crane 1975; Takeda 2006; Todd et al. 2011). When bright coloration is present it is often in regions readily visible to conspecifics, such as on the claws (Detto et al. 2004; Detto & Backwell 2009), carapace (Zeil & Hofmann 2001; Hemmi et al. 2006; Detto et al. 2008), or face (Huang et al. 2008; Todd et al. 2011).
Fiddler crabs (members of the genus Uca) are highly visual animals. They respond rapidly to the sight of predators (Layne et al. 1997; Wong et al. 2005), can visually differentiate between familiar and unfamiliar conspecifics (Detto et al. 2006; Zeil & Hemmi 2006), and use visual displays to attract mates (Salmon & Atsaides 1968; Pope 2000; Detto et al. 2006; Zeil & Hemmi 2006; Detto 2007; Detto & Backwell 2009). Fiddler crabs are reported to have a panoramic field of vision (Land & Layne 1995a; Zeil & Zanker 1997; Smolka & Hemmi 2009), and within this field, individuals of Uca pugilator Bosc 1802 can discern wandering crabs at a distance of up to 30 cm, and can identify their sex within a distance of 15 cm (Land & Layne 1995b; Layne et al. 1997).
What a crab perceives is determined by the anatomy and physiology of its visual system and the optical properties of the visual cue. Members of the genus Uca are sensitive to wavelengths of ~400–600 nm (light that appears from violet to orange to the human eye), and electrophysiological data suggest that they may have greater sensitivity in the blue-green and yellow-orange ranges (Horch et al. 2002). Whether fiddler crabs see “in color” is a topic that is still being examined. Behavioral (Hyatt 1975; Detto et al. 2006) and electrophysiological data (Horch et al. 2002) provide strong support for this possibility; the conclusions from microspectrophotometry are less complete (Jordão et al. 2007). Taken together it seems quite likely that fiddler crabs do have some version of color vision, although they may not see color as humans do.
Uca pugnax Smith 1870, the mud fiddler crab, is a well-known saltmarsh species ranging from northern Florida to Cape Cod (Crane 1975). These crabs are primarily dark olive to almost black in color (Meinkoth 1981). The frontal section above the mouth and centered between the eyes of some individuals is blue or blue-green (Fig. 1A,B). This marking, subsequently referred to as the “blue streak,” was described briefly by Crane (1975), but without reference to its biological significance or variation. If the blue streak varies among individuals or on a single individual over time, then it has the potential to convey information. We recorded the presence of the blue streak as crabs mature, and the rate of its change in brightness in adult crabs. If the blue streak has a role in communication, then we would expect it to change depending on circumstance. We examined variation in blue streak brightness in response to physical factors (light and temperature) and social context to begin to understand what causes its dynamic response.
Crab collection, maintenance, and photography
Members of Uca pugnax were observed and collected on mudflats at Woodland Beach and Lewes, Delaware. Although both male and female crabs exhibited blue streaks, this study focuses on males because we wanted an independent, non-invasive indicator (the display claw) of the maturity of a crab. Crabs that were brought back to the laboratory were housed in opaque plastic bins (34 × 24 cm) containing sloped sand and artificial seawater (20 ppt, a common salinity at these sites). The crabs were exposed to natural light and room temperature (21°C) and provided with fish food flakes if kept more than 48 h. At the end of each laboratory experiment, crabs were returned to their places of origin.
Crabs brought back to the laboratory were photographed within 48 h of collection using a Canon PowerShot A75 camera (Canon, Tokyo, Japan). Each crab was attached to a black felt block (12.7 × 5.1 × 2.5 cm) using a rubber band (Fig. 1A) and photographed at a set distance and with constant lighting from four model EDXP 30–16 daylight fluorescent light bulbs (6500K; Home Depot, Atlanta, GA, USA). Crabs studied in the field were photographed within 30 s of capture (Hemmi et al. 2006) using a Nikon D50 camera; in the field, crab distance from the camera and lighting conditions varied. In both the field and the laboratory, each crab was photographed with the frontal section of its carapace parallel to and level with the camera lens (approximating the viewpoint of an interacting conspecific). A scale bar and standardized color card were included in every image. The resulting digital images were used to determine the brightness of the blue streak and the size of the carapace and display claw of each crab.
We quantified crab size and the brightness of the blue streak and carapace from digital images. Carapace width and claw length were measured using ImageJ software (National Institutes of Health, Bethesda, MD, USA). As it is not known whether crabs see color or to what degree that color corresponds to human visual experience, we have conservatively analyzed our data using only brightness data (ranging from pure black [0%] to pure white [100%]) derived from Adobe Photoshop CS3 HSB (Hue, Saturation, and Brightness; Adobe Systems, Inc., San Jose, CA, USA) analyses of our digital images. To quantify brightness, we first defined sample regions representative of the carapace and blue streak using morphological markers (Fig. 1A). The center of the carapace sample was located along the vertical midline, midway between the posterior edges of the orbits and the most posterior edge of the carapace visible in the picture. The center of the blue streak sample was also located along the vertical midline, but midway between the anterior edge of the frontal region and the most posterior edges of the orbits. Each sample was 0.6 × 0.4 mm in size (no fewer than 54 pixels), and the brightness of the pixels within the sample was averaged (using the “average” filter) and measured in Adobe Photoshop CS3. For the field data collected in May, we used Picture Window Pro 5.0 (Digital Light & Color) and the Macbeth color card (X-Rite, Inc., Grand Rapids, MI, USA) to standardize color across all images prior to analysis.
As digital cameras are designed to record light and color as humans see them, the brightness values obtained from digital photographs may not accurately reflect the brightness of light as perceived by a crab. To address this potential shortcoming, we devised a more “crab-centric” measure of brightness based on the sensitivity of one type of U. pugnax photoreceptor (Jordão et al. 2007). Averaging the spectral sensitivity curves of the blue and green sensors of our camera yielded a curve that approximates the spectral sensitivity of the photoreceptors of U. pugnax. Therefore, we defined our crab-centric brightness measure as the average of the blue and green values of each pixel in our images. While the standard measure of brightness is proportional to the light intensity detected by all three (red, green, and blue) sensors, our crab-centric measure differs in that it omits the light detected by the red sensor. We analyzed all of our images using both the regular brightness measure and this crab-centric brightness measure. Our analysis generally yielded qualitatively similar results for both measures, suggesting that our results are robust to the exact definition of brightness used. Therefore, in the analysis that follows, we present our results using the regular brightness measure. In the few instances in which the two measures differ, we note the differences.
Blue streak presence as a function of body size
Male crabs (n=68) were collected from Woodland Beach on 20 November 2008 and photographed in the laboratory within 24 h of collection. Each photograph was examined to determine the presence or absence of a blue streak and to measure carapace width and display claw length.
Time course of blue streak brightness changes
To track the time course of the blue streak's change in brightness, we took time-lapse photographs through a Wild dissection microscope using a Nikon E995 camera (Nikon Corp., Tokyo, Japan). Male crabs were collected from Lewes on 7 and 22 July 2009 and brought back to the laboratory where they were exposed to sunlight for a minimum of 2 h before beginning a photographic sequence. Each of the 20 crabs was sequentially removed from sunlight and within 2 min was secured under the microscope, held in place between a block of styrofoam and a piece of cardboard with a rubber band. The crab was oriented under the lens such that a section of the blue streak (Fig. 1B) was in focus and well-illuminated with fiber optic lights (the room was otherwise dark). Images were acquired at 5 s intervals for as long as the crab remained in position, in some cases up to 20 min. We quantified the change in the blue streak region by measuring the brightness of a section of the blue streak that remained in focus throughout each photographic sequence (Fig. 1C,D). Note that the exact position of this section (Fig. 1B) varied slightly among crabs, in contrast to all other experiments that used the standardized region (Fig. 1A).
Environmental influences on blue streak brightness
We wanted to test whether the brightness of a crab's blue streak was affected by variation in light and temperature. In addition, we were concerned that blue streak brightness might vary relative to the timing of the tide. To account for these three variables, we defined four treatment groups (light at 21°C and 32°C, and dark at 21°C and 32°C) and conducted an experiment during the 4-h period preceding low tide at the Lewes field site (which occurred at 17:02 on 24 July 2009). Male crabs were collected on 22 July 2009, brought back to the laboratory, and separated into the four treatment groups. Each group consisted of similarly sized crabs (ANOVA, F=0.45, df=[3, 88], p=0.72; mean carapace widths: light at 21°C 1.88±0.11 cm, light at 32°C 1.84±0.20 cm, dark at 21°C 1.85±0.17 cm, dark at 32°C 1.80±0.14 cm) kept in separate bins, each containing a sand bottom covered with ~1.5 cm of artificial seawater (20 ppt). An hour before any data were taken, all four groups were placed outside in full sunlight. Sunlight intensity, measured using a LI-COR Inc. LI-1000 Data Logger (LI-COR Biosciences, Lincoln, NE, USA), fluctuated between 250 μmol s−1 m−2 and 2000 μmol s−1 m−2 during the experiment. We monitored water temperature in all bins with thermometers.
We began by sequentially photographing a single crab from each of the four bins. Each crab was selected haphazardly, photographed under standard conditions (as previously described) within 1 min of removal from the bin, and then released to a holding bin. Photographing one crab from each of the four bins took ~10 min. We repeated this sequence, using previously unphotographed crabs, for 4 h. During the first half of the experiment (~2 h), all four groups of animals remained in the sunlight at warm temperatures (median=31.3°C; interquartile range [IQR]=30.1–31.9°C). In the second half of the experiment, we moved two groups of crabs into a dark room illuminated only by a dim red light bulb that allowed us to continue monitoring temperature and sampling crabs efficiently. Red light (~650 nm) was chosen because members of U. pugnax have maximal sensitivity to wavelengths of 521 nm (green) and limited sensitivity to wavelengths above 600 nm (orange) (Jordão et al. 2007). The bin of one of the groups exposed to these darkened conditions was placed in a warm water bath (median=30.5°C; IQR=30.0–30.9°C) to maintain a temperature close to the median outdoor temperature. The second was cooled (median=20.0°C; IQR=20.0–20.1°C) with small bags of ice placed in the water. Of the two groups of crabs that remained in the sunlight, one was cooled (median=21.2°C; IQR=18.6–24.0°C) with bags of ice, and the other was not (median temperature=35.5°C; IQR=34.9–36.7°C). Immediately after transferring the crabs to their new conditions, we resumed sampling and photographing. At the end of the experiment, each crab had been photographed only once.
Behavioral associations with blue streak brightness
To investigate the possibility that the brightness of the blue streak is associated with behavior, we photographed male crabs engaged in common behaviors and used the images to measure the brightness of each crab's blue streak (as described previously). We collected our data on a mudflat in Lewes during low tide on 26 May 2010 (beginning of the mating season), 30 July 2009 (mid-mating season), and 7 September 2009 (at the end of the mating season) (Crane 1975; Bergey & Weis 2008). The behaviors we observed varied by time of year and included roaming, fighting, displaying, basking, and burrow guarding (defined below, in Results). In the field, after a crab's behavior was confidently identified by two or more observers, the crab was captured and photographed. Each captured crab was rinsed with seawater, and an image focusing on the blue streak was taken within 30 s of capture. After photography, crabs were temporarily released into buckets of seawater to avoid re-sampling.
To compare the size ranges of crabs with and without blue streaks, we used unpaired t-tests (these data were normally distributed, as determined by a Shapiro–Wilk normality test). We compared carapace widths and claw lengths of crabs with blue streaks to those without blue streaks. We adjusted the significance level to 0.025 using a Bonferroni correction to account for multiple comparisons.
We performed two analyses to examine how environmental conditions influence streak brightness. We did a regression analysis of streak brightness versus time of day to evaluate the potential effect of time relative to low tide on blue streak brightness. We used ANOVA to compare whether the treatment groups varied in blue streak or carapace brightness before they were moved to their respective light and temperature treatments. We then analyzed the effects of light and temperature on the brightness of the background carapace and the blue streak. After determining that brightness data within each experimental group were distributed normally (Shapiro–Wilk test), we used two-way ANOVA to test the effects of light and temperature on blue streak and carapace brightness.
To test whether blue streak brightness varied with behavior, we first checked data for each behavioral group for normality with the Shapiro–Wilk test. When the data were distributed normally (September), we used unpaired t-tests to compare brightness between behavioral groups. When the data were not distributed normally (May and July), we used Wilcoxon Signed Rank tests. To account for multiple comparisons between behavioral groups (in May and July), we used a sequential Bonferroni technique (Sokal & Rohlf 1995). All statistical analyses were performed using JMP 7, 8, and 9 (SAS, Cary, NC, USA).
Blue streak presence as a function of body size
Newly settled (data not shown) and juvenile individuals of Uca pugnax lacked the blue streak. It developed in males when crabs had a carapace width ranging 1.2–1.4 cm (Fig. 2). Compared with males without blue streaks, male crabs with blue streaks had significantly wider carapaces (mean width: blue streak present=1.51±0.18 cm, blue streak absent=1.05±0.14 cm; t=11.79, df=68, p<0.01), and significantly longer display claws (mean length: blue streak present=2.13±0.39 cm, blue streak absent=1.02±0.29 cm; t=13.55, df=68, p<0.01).
Time course of blue streak brightness changes
The timing of changes in brightness of the blue streak varied among individuals. In some crabs that had been removed from sunlight and secured under a microscope illuminated by fiber optic lights, the blue streak had begun to darken within as few as 2 min (Fig. 3). Blue streaks observed under the microscope initially appeared as a concentration of bright blue spots interspersed with smaller brown and black spots (Fig. 1C). We observed that when a blue streak decreased in brightness, the size of the blue spots relative to the brown and black spots had diminished (Fig. 1D). Blue streaks that remained constant in brightness did not show qualitative changes in the color or distribution of the colored spots.
Nine of the 20 crabs observed produced a sufficient number of focused images for quantitative analysis. Of these nine individuals, four demonstrated a clear decrease in blue streak brightness (Fig. 3). This pattern consisted of a rapid decrease in brightness, followed by a period of relatively stable brightness. In the greatest change we observed, the blue streak's brightness decreased by about a third of its value during the first 110 s of observation. Due to the lag (~2 min) in getting a crab secured and positioned under the microscope, these values should be considered as conservative. The remaining five individuals displayed relatively stable blue streak brightness for the entire 12–20 min period of observation. However, because the brightness of the blue streak can change rapidly, it is possible that these individuals underwent the majority of their change before we began recording.
Environmental influences on blue streak brightness
We verified that experimental groups did not differ in either blue streak brightness (ANOVA, n=40, F=0.10, df=[3,36], p=0.96) or carapace brightness (ANOVA, n=40, F=1.40, df=[3,36], p=0.26) before they were moved to their respective light and temperature treatments.
In addition, no correlation between blue streak brightness and carapace brightness existed before we imposed experimental treatments (R=0.1189, n=40, p=04.48). Using regression analysis of blue streak brightness versus time of day, we were able to rule out the potential effect of time relative to low tide on blue streak brightness (r2=0.002, F=0.07, df=39, p=0.80).
Blue streak brightness was strongly affected by light, but not temperature (Fig. 4; Table 1). Crabs moved from sunlight into darkness displayed darker blue streaks than did crabs maintained in sunlit conditions. Meanwhile, crabs kept in sunlight, but cooled to 21°C, did not show significant changes in blue streak brightness. Carapace brightness was not affected by temperature. When we analyzed the effect of light on carapace brightness using traditional brightness values, there was no significant change; however, if we used the crab-centric measure of brightness, there was a slight but significant darkening of the carapace when the crabs were subjected to dark conditions (Table 1). There was no significant interaction between light and temperature (Table 1).
Table 1. Effects of light and temperature on the brightness of the blue streak and carapace based on blue-green crab-centric brightness measurements. Asterisks indicate statistical significance. Standard brightness measurements gave similar results, except that no statistically significant change in carapace color was observed in response to light.
Sum of squares
Behavioral associations with blue streak brightness
At the end of May 2010 (the beginning of the mating season), we observed three behaviors: roaming (a crab moving across the mudflat), fighting (where two males were locking display claws), and displaying (where we witnessed a sequence of three or more waves of the large display claw from a particular male immediately prior to his collection). Roaming and fighting crabs displayed blue streaks of similar brightness that were significantly brighter than those of displaying crabs (Fig. 5).
At the end of July 2009 (the middle of the mating season: Crane 1975; Christy 1978; Bergey & Weis 2008), we also observed crabs roaming, fighting, and displaying. In addition, we observed basking (where a crab holds its carapace and claws off of the substratum and remains motionless for an extended time with its dorsum oriented toward the sun). As in May 2010, roaming crabs displayed blue streaks that were significantly brighter than the blue streaks of displaying crabs. Crabs engaged in fighting displayed a range of brightness values that overlapped those of both roaming and displaying crabs and were not significantly different than either. The blue streaks of basking crabs had the darkest median value and were significantly darker than those of fighting and roaming crabs, but not different from those of displaying crabs (Fig. 5).
Toward the end of the breeding season, in September 2009, we observed roaming and burrow-guarding behaviors (a crab standing at a burrow entrance), but did not see any fighting or displaying. There was no significant difference in brightness (t=1.30, df=29, p=0.21) between the blue streaks of roaming and burrow-guarding crabs.
Repeating the above analysis using the crab-centric brightness measure yielded qualitatively similar results. The only difference observed was that, using the crab-centric brightness measure, basking crabs (July 2009) were not significantly different from roaming (Z=−0.98, df=1, p=0.32) or fighting crabs (Z=−0.65, df=1, p=0.52).
In all cases, behavioral groups did not differ in either carapace width (ANOVA, May: F=0.04, df=[2,48], p=0.97; July: F=1.25, df=[3,48], p=0.30; September: F=1.20, df=[1,28], p=0.283) or claw length (ANOVA, May: F=0.63, df=[2,48], p=0.54; July: F=1.732, df=[3,48], p=0.173; September: F=1.068, df=[1,28], p=0.310). There was no difference in crab size across dates (ANOVA, carapace width: F=2.07, df=[2,133], p=0.13; claw length: F=0.93, df=[2,133], p=0.13).
Our results, in combination with previously published data, suggest that the blue streak could serve as an intraspecific signal. The blue streak's color lies within the range of wavelengths perceptible to individuals of Uca pugnax (Horch et al. 2002; Jordão et al. 2007), and its frontal location would make it visible to nearby conspecifics facing one another. Its range of brightness values compared with those of the background carapace would also make it stand out. In addition, the blue streak varies in brightness over multiple timescales: it first appears with the maturation of the crab, after which variation in its brightness can cause it to range from pale blue to nearly black (Fig. 4) within a few minutes, depending on an individual crab's behaviors and environmental conditions. Thus, information could be encoded by blue streak brightness, the duration of a display, or the rate of its brightness or color change.
Previous studies of color change in Uca suggest that variation in the blue streak could be associated with stress or courtship. Color change in response to stress has been observed in adults of Uca vomeris McNeill 1920 (Zeil & Hofmann 2001; Hemmi et al. 2006) and Uca capricornis Crane1975 (Detto et al. 2008). In members of these species, carapace color patterns can fade within 15–20 min in response to capture and handling (Zeil & Hofmann 2001; Hemmi et al. 2006; Detto et al. 2008). These results are consistent with our photomicroscopic observations, which showed that the blue streak could darken within minutes when an individual of U. pugnax was captured, taken out of the sunlight, and secured under a microscope. However, as we have also shown that the blue streak darkens after a decrease in sunlight (Fig. 4), it is not possible to know whether the response in this case was due to the effects of stress or changes in lighting.
Courtship rituals in crabs are also known to involve color changes. In some crab species, males prefer females displaying brighter coloration (Baldwin & Johnsen 2009; Todd et al. 2011), and in many species of Uca, males whiten their claws while waving (reviewed in deRivera & Vehrencamp 2001; see also Crane 1975; Detto & Backwell 2009). Given the known responses of crabs to color, and instances of color change during courtship behavior, a reasonable conjecture might be that the blue streak is associated with mating. Indeed, it becomes prominent only as crabs reached sexual maturity (at a carapace size of ~1.2–1.4 cm: Fig. 2), a size range that roughly corresponds to the onset of sexual maturity in other species of Uca (Christy & Salmon 1984; Castiglioni & Negreiros-Fransozo 2006). However, arguing against the blue streak as a mating display is the fact that during claw waving displays the blue streak is very dark, nearly black (Fig. 5). One explanation might be that a darkened carapace allows for more contrast with the whiter display claw, but in that case, why do crabs develop the ability to brighten the blue streak as they mature? Perhaps when it is brighter it is a signal that the individual is specifically not engaging in a sexual display. Our data confirm that during the reproductive season, roaming crabs exhibit blue streaks that are brighter than the blue streaks of displaying crabs (Fig. 5). Crabs observed after the reproductive season have bright blue streaks that do not vary with the behaviors we observed (Figs. 2, 5). Though the specific function of the blue streak remains unclear, these results support its potential as a social signal.
Uca pugnax is a well-studied species that is considered an important ecological engineer of salt marshes (Bertness 1985). Despite the large number of publications on this species, ours is the first to report that its characteristic blue streak develops with maturation and has the capacity for rapid change. This change is influenced by light and is associated with at least some behaviors, and as such the blue streak reflects information about the individual's current state and its experience of the environment.
We appreciate the helpful conversations we had about this project and manuscript with Julia Berthet, Brian Clark, Julie Hagelin, Emily Hager, Sara Kim, Sönke Johnsen, Michael Peshkin, Kathy Strandburg, Liz Vallen, and Steve Wang. Two reviewers provided us with excellent and helpful suggestions. Jose-Luis Machado and Colin Purrington kindly lent equipment. August Merz, Seth Merz, and Jocelyne Noveral were of critical assistance in the field. Funding from HHMI and the family of Walter Kemp supported this work.