Behavioral patterns of two fiddler crab species Uca rapax and Uca tangeri in a seminatural mangrove system

Zoos increasingly transform their exhibitions from traditional one‐species enclosures to more natural exhibits, that is, environments that capture part of an ecosystem including a selection of animals and plants that occur there. Thus, enhancing the experience of its human visitors while also allowing its residents to possibly show more natural behavior. In 2017 Royal Burger's Zoo (Arnhem, The Netherlands) created and opened a mangrove‐like environment containing fiddler crabs. Fiddler crabs display a broad range of behaviors, and this research examines which wild‐type behavior and behavioral patterns can be observed on a seminatural mudflat. The behavior shown by Uca rapax and Uca tangeri on the mudflat was counted each hour between 07:00 and 17:00. An asymmetric tidal regime was present in the enclosure including two high water periods. Various known fiddler crab behaviors, including waving and combat, were observed but no copulation. A clear pattern in exposed crabs on the mudflat was found, with low numbers visible in the early morning and the highest numbers present in the early afternoon, while number of visitors did not have a significant effect on this pattern. Interestingly, the highest abundances were not observed around the ebbing tide (07:00–09:00), as observed in the wild, but somewhat later, possibly due to the asymmetric tidal scheme or the interaction of tidal and daily rhythms. This study shows that in captivity, fiddler crabs indeed show a range of natural behaviors which is linked to the tidal and possibly daily rhythm as well.

The objective of the present study is to investigate the behavior of fiddler crabs on a seminatural mudflat in the Royal Burgers' Zoo to assess which types of natural behavior can be observed, whether there are patterns in their activity and the impact of the number of visitors around the mudflat on the visibility of the crabs.

| Study site
The research was performed in the Royal Burgers' Zoo located in Arnhem, The Netherlands. The zoo opened an exhibit called "The Mangrove" in July 2017, which resembles a tropical mangrove ecosystem with a mudflat (210 m 2 ) combined with marine water creeks (93 m 2 ) to house fiddler crabs and Atlantic horseshoe crabs (Limulus polyphemus; Figure 1). Three different water levels were applied in this mangrove system to mimic the tidal system: low, mid (+ 6 cm relative to low tide) and high tide (+ 8 cm relative to the low tide; Figure 2). During low tide, the whole surface of the mud plain was exposed to air and during high tide, some areas did not flood ( Figure 1). In contrast to natural tidal conditions, the tidal scheme in this area was fixed. Furthermore, the changes in water level were realized between 30 and 60 min and thus did not follow a more natural, sinusoidal wavelike flow. During the observational period, the tide lowered from high (+ 8 cm) to mid (+ 6 cm) tide between 07:00 and 08:00 and from mid to low (+ 0 cm) tide between 08:00 and 09:00 until 16:00 after which the water rose to mid tide ( Figure 2). Between 16:00 and 00:00, an 8-hr long period of raised water (i.e., mid or high tide) occurred, while between 05:00 and 09:00 a 4-hr long period of raised water occurred. The tidal scheme is therefore asymmetric.
F I G U R E 1 Overview of the mudflat in the Royal Burgers' Zoo, Arnhem. The dashed line indicates the overhang of the visitors' walking path over a small part of the mudflat. Zone contoured with a red line indicates the zone used for analyzing the influence of visitors on the number of crabs present on the mudflat. Visitors standing at the net fence (= dashed line in parts with overhang) in the red zone were counted. Except for adult (= prop roots are developed) red mangrove (Rhizophora mangle), black mangrove (Avicennia germinans), and mangrove fern (Acrostichum areum), numerous sprouts of R. mangle were present in the area surrounding the adult plants. Crosshatched gray zones roughly indicate areas which were not flooded during high tide

| Abiotic conditions
Water salinity (21.1-29.5‰), pH (7.81-9.18), and temperature (23.7-26°C) were measured (Hach HQ40d Portable Multimeter, Loveland, CO) daily around 11:00 by the animal caretakers and this data are provided in Appendix 1. Air temperature (kept between 24°C and 28°C) was regulated by means of air-conditioning and ventilation at the top of the roof, while relative humidity (± 70%) was controlled by means of water sprayers.

| Fiddler crabs
On the 4th of July 2017, 160 Uca (Afruca) tangeri (Eydoux, 1835) with an unknown male-female ratio, were released on the mudflat. The species originated from Nigeria. Male Uca (Minuca) rapax (Smith, 1870; identified with the use of Crane, 1975;Hopkins & Thurman, 2010) imported from Cuba were released on the 6th (four specimens), 7th (113) and on the 14th (83) of July 2017. No females of U. rapax were released. These two species do not occur together in the wild (Crane, 1975). Experiments started on the 20th of July 2017, thus all crabs had been acclimatized for at least 6 days.

| Data collection
Three days before the start of the observation period (from 17th July 2017 until 19th July 2017), a preliminary investigation was performed to identify which behaviors could be observed on the mudflat. During this preliminary study, the mudflat was scanned every 30-60 min and some individuals were observed for a longer period to determine which behaviors occurred. On the basis of this prior investigation and descriptions of fiddler crab behavior by Crane (1975), a list of behaviors present was constructed (see Appendix 2) and in this current study, we studied behavior related to individual specimens.
The observational period ran from 20th July 2017 until 8th September 2017. Scanning the mudflat with binoculars (Bynolyt Buzzard, Wormerveer, The Netherlands SHR 8 × 42), the total number, species and sex distribution of fiddler crabs present was counted together with their behavior at that moment. As the mudflat was too large to observe from one position, observations were made from different positions at the edge of the exhibit. Observations always started and ended at the same position, and the same sequence of positions was always used.
Observations were performed each hour starting at 07:00 up to and including 17:00 (except 12:00). During each observation day, not all hours were included and a random selection of several hours per day was made to ensure that for each hour at least 10 replicates were available. For a detailed overview, see Appendix 3.
The potential impact of the number of visitors present around the mudflat on the number of crabs visible was tested in a particular zone on the mudflat (i.e., the red zone, Figure 1). This area (ca. 74 m 2 ) was selected because it is rather elongated and thus had a large border at which the visitors could stand and thus possibly influence the crabs' behavior. The number of visitors present at the net fence (Figure 1) of the red zone was recorded before and after the observation in that zone and thereafter averaged.

| Statistical analysis
An initial partial redundancy analysis (pRDA) showed that time of the day accounted for 14.9% (Monte Carlo, F 9 = 13.1, p = 0.002; Appendix 5) and species and sex for 17.6% (Monte Carlo, F 2 = 64.1, p = 0.002; supplements S5) of the observed variation. These results are in agreement with earlier findings of differences in behavior between species and sex (e.g., Crane, 1975) and therefore we chose to analyze the collected behavioral data per species and sex group. Per species and sex group (male U. rapax [mr], male U. tangeri [mt], and female U. tangeri [ft]), percentages of each behavior type observed were calculated. These percentages showed that behaviors "waving," "feeding," and "stationary" were important on the mudflat and the mean number of crabs exhibiting these behaviors was therefore "hour" was added as random effect in the analysis of the effect of the number of visitors. For this latter analysis, all species and sex groups were summed. The variable "day" was added as random effect in both Poisson regression models above. All analyses were performed using SPSS 23 (IBM Corp. Armonk, NY).

| Ethical statement
The Royal Burgers' Zoo is a member of EAZA (European Association of Zoos and Aquaria) and follows the code of Ethics prepared by F I G U R E 2 Daily tidal scheme present on the mudflat in Burgers' Zoo. Arrows indicate the asymmetry in the tidal schema [Color figure can be viewed at wileyonlinelibrary.com] VAN HIMBEECK et al. | 345 EAZA and furthermore follows the Dutch national legislation regarding animal welfare. This study complied with the EAZA code of Ethics and Dutch national legislation, since the studied animals were invertebrates which were not manipulated in any way, and therefore ethical approval was not required according to the Dutch national legislation.
For female U. tangeri "Feeding" was observed in 63% of the observations followed by "Stationary" which was observed for 30% of the observations. "Locomotion" and "Constructing" accounted for approximately 3%, while "Cleaning" was only observed occasionally.
Appendix 4 shows the mean percentages of all the behaviors for male U. rapax, male U. tangeri, and female U. tangeri.

| Number of crabs visible per hour
In general, the number of crabs increased from 07:00 on, until the midday hours, while toward the end of the observational period the number of crabs present on the mudflat decreased again ( Figure 3a,e,i). A Poisson regression indicated that time of the day had a significant effect on the number of visible male U. rapax (F 9,156 = 100.61, p < 0.001), male U. tangeri (F 9,156 = 42.97, p < 0.001), and female U. tangeri (F 9,156 = 49.53, p < 0.001).

| Daily patterns in feeding, stationary, and waving
The lowest numbers of crabs which were "Feeding," "Stationary," or "Waving" were observed during the morning hours ( Figure 3). For each of the species and sex groups, the mean number of individuals which were "feeding" increased toward the midday hours, peaked around 13-14 hr and decreased again when the day progressed ( Figure 3b,f,j). After reaching its maximum around 11 hr, "Stationary" appeared to remain relatively constant (Figure 3c,g,k). While only few male U. tangeri individuals were observed waving, waving in male U. rapax was more abundant and a clear maximum was observed between 13-16 hr (Figure 3d,h).

| Effect visitors on the visible number of crabs
The number of visitors around the exhibit did not have a statistically significant impact on the number of fiddler crabs visible on the mudflat (Poisson regression, F 1,167 = 1.647, p = 0.201).

| DISCUSSION
This study shows that there is a clear relationship between the number of crabs visible on the mudflat and time of the day, as the number of visible crabs increased from early in the morning toward the midday hours after which numbers decreased again. This pattern roughly coincided with the tidal movement during the day and was observed for U. rapax males, U. tangeri males, and U. tangeri females.
This response of the crabs on the seminatural mudflat roughly corresponds with observations in the wild where fiddler crabs, in general, withdraw in their burrows during high tide and leave them during the receding tide to perform (social) activities (Barnwell, 1968;Crane, 1975;Murai, Goshima, & Nakasone, 1982;Skov & Hartnoll, 2001;Wolfrath, 1992a), which is further stimulated by their internal tidal rhythm (Barnwell, 1963;Stillman & Barnwell, 2004;Webb & Brown, 1965). It is remarkable, however, that a delay seems to be present in this pattern in the exhibit of the Royal Burger's Zoo, as the highest number of visible crabs were observed a couple of hours after low tide. The observations in the present study, therefore, are not consistent with usual fiddler crab behavior (i.e., emerging from the burrow as the water recedes [e.g., Barnwell, 1968;Crane, 1975;Murai et al., 1982]) and do not confirm earlier research that found that activity was initiated in early ebb and activity peaks could occur before low tide (Barnwell, 1963;Stillman & Barnwell, 2004;Webb & Brown, 1965). An additional inspection on the mud flat in the Royal Burgers' Zoo showed that the drainage of the burrows did not always follow exactly the receding water: Some burrows retained water for approximately 45 min while others drained immediately. Consequently, this heterogeneous drainage pattern seems an unlikely explanation for the observed delay in the activity. This hints toward the fact that, besides the tidal regime, the activity pattern of the fiddler crabs is also governed by another factor. A possible explanation could be that the fiddler crabs possess a daily rhythm. Daily activity patterns have been described in earlier studies (Barnwell, 1966;Stillman & Barnwell, 2004;C. C. Thurman & Broghammer, 2001;Webb & Brown, 1965) and their interaction with tidal rhythms could also play a role in the crabs exhibiting the observed delay in the present study. These daily rhythms (and their interaction with tidal rhythms) can differ widely between fiddler crab species (e.g., Barnwell, 1963;Stillman & Barnwell, 2004; C. C. Thurman & Broghammer, 2001;Webb & Brown, 1965) and this study suggests that also U. rapax and U. tangeri possess a daily rhythm as well. This difference between species is also supported by the results of the multivariate analysis in this study (Appendix 5) which showed that behavior of male U. rapax was twice as much influenced by one of the rhythm components (i.e., daily or tidal) than both U. tangeri sexes. It should be noted that due to the fixed tidal regime F I G U R E 3 Mean individuals (left ordinate) of total visible crabs (a, e, i) and behavior types "Feeding" (b, f, j), "Stationary" (c, g, k), and "Waving" (d, h) per hour (abscissa) for male U. rapax (a, b, c, d), male Uca tangeri (e, f, g, h), and female U. tangeri (i, j, k). Shaded area depicts tidal level (right ordinate). For total visible crabs (a, e, i): error bars indicate 95% CI. Letters indicate significant differences (Poisson regression, p < 0.05). p values were adjusted for multiple comparisons (sequential Bonferroni correction). For behavior types: Error bars indicate standard error of the mean. CI: confidence interval VAN HIMBEECK et al. | 347 on the mudflat in the Royal Burgers' Zoo, it is difficult to uncouple the tidal and daily rhythm. As Crane (1975) states that the internal rhythm of fiddler crabs can be adjusted in a few days, it is likely that the crabs in the present study were adapted to the tidal scheme.
However, the rather abrupt changes between water levels and the asymmetric tidal regime (i.e., a 4-hr and an 8-hr long period of raised water) may interfere with the crabs synchronizing to the tides. As there is a delay of at least 2 hr, it is plausible that the crabs (try to) adjust to the longer high water period or that an intermediate synchronization between the shorter and longer high water period occurs.
Throughout the day, a large proportion of the emerged U. tangeri crabs were feeding. According to Wolfrath (1993), U. tangeri in Portugal started their surface activity by feeding intensively for 1-1.5 hr followed by approximately 1 hr of burrow activities.
Depending on the season, this phase was followed by courtship or migration. The crabs fed roughly up to an hour before the water reached the burrows. This pattern was not observed in the current study for U. tangeri. In general, feeding was the most observed behavior for U. tangeri in the Royal Burgers' Zoo, which has also been observed for other species (Caravello & Cameron, 1991;Tina, Jaroensutasinee, Keeratipattarakarn, & Jaroensutasinee, 2018;Weis & Weis, 2004). The present results show that from the time that the water recedes to the hours after that, U. rapax males initially are either stationary or feeding and that this behavior continues again in the afternoon. The results also show a rise in waving display from 13 to 16 hr. This pattern of the waving display is more in agreement with the activity phases of U. tangeri as observed by Wolfrath (1993).
Many isolated types of behavior as listed by Crane (1975), like cleaning activities and combat, were observed on the seminatural mudflat in the Royal Burgers' Zoo. Waving display, typical male fiddler crab behavior, was observed extensively and most often observed for U. rapax. Waving display differs widely between fiddler crab species (Perez, Rosenberg, & Pie, 2012), and one waving motion of U. rapax is known to be longer than the short motion of U. tangeri (Crane, 1975). As earlier research on U. pugilator found that larger crabs indeed wave more frequently (Hyatt, 1977), differences in body size between the species may also have been of influence.
Interestingly, copulation has not been observed at all during this study. Von Hagen (1962) reported that for U. tangeri surface copulation mainly occurred at night, which could have happened in the present study but was outside the observation period. Also, the animals could have mated in their burrows (Crane, 1975) or no mating at all occurred.
Visitors density had no significant impact on crab abundance, although fiddler crabs are known to seek refuge in their burrows when startled, for example, being approached by a possible predator (Crane, 1975). It has been reported that fiddler crab populations, which are regularly disturbed, can get accustomed to consecutive disturbance (Crane, 1975;Walker, 1972). As fiddler crabs on the studied mudflat are under continuous disturbance by visitors, it seems that the crabs became familiarized with the visitors and do not seem to react to them.

| CONCLUSION
This study confirms that various known typical fiddler crab behaviors (e.g., waving, feeding, and combat) can be observed on a seminatural mudflat. Furthermore, visitors had no influence on the number of visible crabs and consequently they are able to see many of the natural behaviors of these invertebrates. This indicates that the initiative of the zoo to show the general public (part) of the behavioral diversity of fiddler crabs can be considered a success. See Figure A1 here.
F I G U R E A 1 Water temperature (top), pH (middle), and salinity (bottom) measurements during the experimental period. The first day (Day 1) represents 17th July 2017 and the last day (Day 51) 8th September 2017. The dashed line indicates the mean. The peak of pH on Day 34 is probably due to a measurement error. At Day 43 (30th of August) no salinity was measured

APPENDIX 2
See Table A1 here.

APPENDIX 3
See Table A2 here.

Behavior Description
Waving A group (Crane, 1975) of movements of the major cheliped of the male crab Combat A group (Crane, 1975) of physical contacts between two crabs occurring with the physical contact of the major cheliped with the opponent

Feeding
The crab repeatedly scrapes up (with the movement of the dactylus) a bit of substrate in a small cheliped and brings it to the buccal cavity

Locomotion
Wandering of the crab on the mud plain. During this walking, no feeding or other distinct behavior (e.g., waving) is observed

Constructing
The crab visually moves chunks of substrate out of, toward, or around its burrow

OR
The crab uses its pereiopods to put pressure on the mud surrounding the burrow entrance and other mud

Stationary
The crab shows no sign of motion of the pereiopods (including chelipeds). The motion of other appendages (e.g., antennae or maxillipeds) is possible Cleaning The (muddy) crab strikes or scrapes with its minor cheliped over its own carapace, legs, or major cheliped OR The crab rubs multiple (muddy) legs over each other

Bubbling
The crab produces a mass of bubbles at the mouth, which extends over the anterior parts of the crabs Copulation A male and female crab have their ventral sides orientated toward each other in an "embracing" position while the abdomen of both individuals is folded down

Undetermined
The researcher is not able to see which behavior the crab is showing Entrance structures a The crab has built structures (e.g., semi-domes, hoods, mud balls, and pillars) with sediment at the entrance of their burrow Sound production a The crab produces sound or acoustical signals Droving a A group of crabs is moving as a herd a The selected monitoring methodology did not allow for observations of these types of behavior. See Figure A3 here.

APPENDIX 5 Materials and Methods
Multivariate analyses of the arcsine square root transformed percentages of counts were performed to recognize patterns in behavior using Canoco5 (Braak & Šmilauer, 2012). Counts of observed behavior of each of the three species and sex groups were analyzed. Initial Detrended Correspondence Analysis showed that the gradients were rather short (<3 SD) and therefore the linear response method was regarded appropriate to use (Braak & Šmilauer, 2012). Redundancy analysis is a method of constrained ordination which reduces the dimensions of a data matrix (= ordination) and requires "(some of) the ordination axes to be a linear combination of explanatory variables" (= constrained;Braak & Šmilauer, 2012). By defining a variable as covariate (= covariable) its effect can be accounted for before analysis, and this variable is left out (i.e., partialled out) in the actual analysis (Lepš & Šmilauer, 2003). The resulting analysis is called a partial RDA (pRDA). Such a pRDA was performed on all observations together with hour and species & sex as explanatory variables, and separate pRDAs on each of the 3 species and sex groups with hour as explanatory variable. As the number of crabs of each species and sex was not equal, percentages of each behavior per observation were calculated for each of the species and sex groups. To account for the effect of the nuisance variable "day", this variable was included as covariate (= covariable) in all pRDAs (Braak & Šmilauer, 2012). Significance of explanatory variables was tested by means of Monte Carlo permutation tests on all constrained axes. Figure   Relative to species and sex groups, two general behavioral groups (i.e., groups of behavior arrows on the triplot which point in roughly the same direction) seem to be present. The first group, consisting of "Constructing" and "Feeding" point toward the left lower corner, in the direction of both U. tangeri sexes. Thus, these behaviors were observed the most for both U. tangeri sexes and showed a slightly higher association with the female crabs. All of the other behaviors (e.g., "Locomotion," "Waving," and "Stationary") make up the second group and are pointing to the right lower corner, in the direction of the male U. rapax and away from the U. tangeri sexes. Thus, behaviors of the latter group were associated more with U. rapax individuals. As male U. tangeri is almost located in the origin of the plot (toward U. rapax, relative to female U. tangeri), it is also slightly more correlated with behavior shown by male U. rapax, than female crabs. p = 0.002) of the variation for male U. rapax, male U. tangeri, and female U. tangeri respectively.

Male U. rapax
A partial RDA (covariate: day) was performed on only male U.
rapax individualss as explanatory variable (Figure A5 a). Again, the early morning hours are situated on a nearly vertical line in the right side of the triplot while the remaining hours are located more to the left. However, a spread along the first axes of these later hours can be observed. Hours 13, 14, 15, and 16 are clustered in the far left, while hours 10, 11, and 17 more in the middle of the plot. Being stationary is pointing towards the right bottom corners in the direction of the early hours. "Feeding" is pointing towards the left bottom corner. "Waving" is pointing almost horizontally in the direction of the afternoon hours. Thus, during the hours 13-16, "Waving" and "Feeding" was observed in greater amounts relative to the other hours, while the least was noted in the early morning hours. Intermediate levels (relative to the other hours) of these behaviors seemed to occur at hours 10, 11, and 17. Being stationary was correlated more with the early morning hours (right-hand side of the graph).
Male U. tangeri Figure A5 b shows the result of a partial RDA (covariate: day) performed on male U. tangeri individuals with time as explanatory variable. Hours from 10 to and including 17 are more or less clustered together on the left early morning hours 7, 8, and 9 are located to the right, with some distance between them along the horizontal axis. All arrows-except for "Stationary"-Point toward the left side of the triplot. "Feeding" points almost horizontally towards the left, which indicates an increasing intensity of feeding from 7 hr till midday, while for the other, later hours no large variation was observed. "Waving," "Locomotion," "Combat," "Cleaning," and "Constructing" are also associated more with the later hours. The arrow of "Stationary" points down, indicating relatively less variation among the hours.