Can time of day, fish size or learning influence the choice of gobies to stay hidden so they are not eaten or so they can go out and feed?
Deschenaux Solenne and Kerréveur-Lavaud Manon
Abstract
In the coral reef, mutualisms are very common. Understanding behaviours, mechanisms and factors influencing them is essential for proper monitoring and conservation plans. Here we look at the influence that human presence has on goby-shrimp mutualism. To do this we take into account several factors: the size of the gobies, the time it takes them to emerge from their hole after a stimulated threat, the repeatability of the hazard through repeated measurements and the time of the day. The experiment was conducted on 27 associations over 3 days. Statistical analyses revealed that small gobies took significantly longer to emerge after disturbance than medium and large gobies. We hypothesized that small gobies take longer to exit their dens for two main reasons, the first one having two levels. Firstly, small gobies are more likely to be impacted by predation and intra-species competition. Secondly, they need less food to be replete and therefore need to take less risks than bigger gobies concerning feeding. Regarding the time of the day, we observed and deduced that gobies went out quicker in the morning because they did not feed at night. Finally, we did not observe any habituation to humans. Indeed, there was no significant differences in terms of habituation according to the number of trials.
Background
Coral reefs environments are among the most species-rich habitats in the world. Among these species, there are a lot of symbioses and inter-species cooperations and mutualisms (Khoda, 2016). These interactions between species lead to differents advantages and benefices for all the concerned individuals, for instance protection against predators, food supplies or a safe place for breeding (Karplus, 2011; Karplus, 1974; Schmiege and al., 2017).
In this project, we choose to work on the mutualism occurring between gobies and shrimps (Figure 1). In this association, shrimps construct and maintain a burrow that is used as a shelter by both the shrimp and the fish. During the day it is a shelter against predators and during the night a safe place to rest (Karplus, 1974). Sometimes, two fishes can be found in a hole with a shrimp and it have been hypothesised that the hole could also be a safe place to breed (Karplus, 2011). In return, the fish prevent the shrimp, that is blind, if any predator approach (Jaafar, 2012). When the shrimp is outside the hole, it maintains a physical contact through its antenna with the fish. If any predator approach, the reaction of the fish indicates to the shrimp that it is time to go back in the hole. Both gobies and shrimps also provide food supplies for each other. Some species of shrimp feed on the goby's faeces or ectoparasites (Jaafar, 2014) and by moving the sand out and around the hole, the shrimp expose some algae and small crustaceans the fish can eat (Khoda, 2016).
We wanted to know how gobies react to threats, more precisely to human disturbance. To do so, we choose to test whether a goby would get out of its hole quicker or slower after several perceived threats simulated by the experimentator. Such changes in exit time would have interesting implications in gobies survival and thus fitness and could be link with evolution of behaviour related to the approach of a predator. Indeed, this behaviour is a great ecological relevance for the goby. What does the fish gain or lose in terms of fitness to come out of its hole more or less quickly after a danger? Going out quickly will allow him to get more food and therefore increase his fitness. On the other hand, if a predator is still in the area it is riskier and puts him in danger of death. There is therefore a trade-off to those gobies’s behaviour between access to food and predation. We will consequently observe the different performances of gobies and relate it to their sizes and time of day.
Questions:
-
Are the gobies influenced by threats (human activity) and can they get used to the human presence?
-
Are the size and time of the day affecting the goby's exit latency?
Hypotheses:
Considering that predation and human activity in the sea might be factors influencing associations between shrimp and gobies we wanted to know if they influence the behaviour of gobies. We therefore measured the time it takes for a goby to emerge from its hole after having been frightened by a human presence. Through repeated measurements we can calculate whether gobies tend to emerge significantly faster and faster or not. We also want to know if the size and time of day impact this behaviour.
According to this, our first null hypothesis (H01) is that there is no habituation from the gobies concerning human presence. In contrast, our first alternative hypothesis (H11) is that there is a significant habituation of gobies to human presence, in other words, gobies will either faster or slower of their hole. Our second null hypothesis (H02) is that size has no influence on the exit latency of gobies while our second alternative hypothesis (H12) is that size significantly influences this behaviour. Finally, our third null hypothesis (H03) is that the time of the day has no influence on the exit latency of the gobies, unlike our third alternative hypothesis (H13).
Predictions:
We predict that even after a certain number of times the gobies have been scared in the same way, they will be not able to get used to it. Therefore, they will not have a significant tendency to emerge faster from their hole. Our predictions are inspired by the fact that if gobies get used to ever threat around them, they would be killed far too often and therefore would have some problem to survive. We believe that the behaviour of "getting in and staying in the hole" must be a mechanism of protection too effective and stable to be able to change in just a few days or across 5 trials. Regarding the impact of the time of day, we think that in the morning the fish will leave faster than at other times of the day because they do not feed at night. Finally, we think that the largest gobies will come out faster than the small and medium ones, because their size makes them less vulnerable and with fewer predators.
Experimental design
The aim of our study was to know if gobies and shrimp associations (although we particularly considered the goby) is influenced by human presence and whether they can habituate to it. This experiment was conducted in the Red Sea, in the National Park of Ras Mohamed, Egypt (Figure 2). So, to carry out this experiment, we observed and measured whether the gobies and shrimps have gotten used to us. To do that, we found and marked with numbered stones 27 independent associations including only one goby and one shrimp. Although it was the hole that was marked and not the fish and shrimp, we considered that we were experimenting with the same individuals every day. Therefore, we considered that each marked association was the equivalent of one individual. Moreover, we were unable to identify and differentiate between goby and shrimp species. We did our experiment in five days with two days of observation and marking of the 27 individuals and three days of measurements, the morning between 6 and 8 am, at noon between 11 and 1 pm and in the evening between 4 and 6 pm. Every gobies were sample once a day and to avoid carry over effect, we randomized the time of the day the data were collected. Our experience consisted in scaring the goby so that it enters its hole with the shrimp, and then we measured how long it took for the goby to come out thanks to a chronometer. We made repeated measurements, five times in a row at the three times of the day. Between each measurement a waiting time of 2 minutes after the gobies has come out has been respected. In case the goby was not out after 10 minutes of waiting, we dropped the measurement. To avoid certain biases, we have also standardized the frightening approach. Indeed, to scary them we approached the hole with a fin then made a small horizontal movement not to lift the sand. Finally, we also noted the size of each goby according to whether it was rather little, medium or big to know if this factor has an influence on our studied behaviour.
Statiscal analyses
All analyses were conducted using the R software version 3.5.1.
In order to study the performance of gobies to go out of their shelter after a threat, we performed a generalize mixed effects models with non-normal error (GLMER) with Poisson distribution using the package lme4. We compared the AIC of different models and choosed the one which was the lowest. This model allowed us to analyze in the same time the effect of the size, of the time of the day and of the trial number on our response variable. The response variable was the duration the fish took to get out of his shelter after the stimulus (in seconds). The explanatory variables we studied with this model were the number of trials, the time of the day which was categorical, either morning, midday or afternoon, and the size of the fish which was also categorical, either small, medium or big. To avoid pseudo-replication errors, we also considered the identity of gobies as a random factor.
This analyze allowed us to deal with the fact that the trials (or stimuli) were nested in the timeday and not independent and to consider, also, that the size of the fish was dependent of the identity of the fish.
Figure 3: R boxplot of the latency related to the size of the goby.
Results
The only factor that had a significant level was the size with the small gobies taking significantly more time to go out of their shelter after a stimulus (estimate: 1,149; Std. Error: 0,318; z-value: 3,605, p-value: <0,001) than medium or large gobies. The size medium only approached significance (estimate: 0,512; Std. Error: 0,265; z-value: 1,931, p-value: 0,053) but still, medium gobies tend to take a bit more time than big fishes to go out of their hide (Figure 3).
Figure 4: R boxplot of the latency related to the time of the day.
Both morning and midday levels of the time of the day approached significance with gobies tending to go out faster in the morning (estimate: -0.288 ; Std. Error: 0.159 ; z-value: -1.810, p-value: 0.0704) and slower at midday (estimate: 0.304 ; Std. Error: 0.159 ; z-value: -1.910, p-value: 0.0561 ) compared to the end of the day (considered as afternoon)(Figure 4).
Finally, considering the number of trials, none of them was significant and only the trial 3 approached significance (estimate: -0.379; Std. Error: 0.205; z-value: -1.847, p-value: 0.0647), fishes tending to take less time to go out than during the first trial (Figure 5).
The identity of the gobies and therefore the inter-individual variation account for 13,5% of the variation (Intercept: 0.13506) in the latency between the stimulus and the apparition of the fish. The size, which was dependent to the identity of the fish, account for 19,2% of the total variation (Intercept: 0.19221). Our intercept was significantly different from 0 (estimate: 4.332, Std error: 0.276, z-value: 15.680, p-value: < 2e-16).
Figure 5: R boxplot of the latency related to the trial.
Discussion
Even though our interests were directed toward the gobies's performances to go out of their shelter after a stimulated threat, those performances also impacted the associated shrimp whose survival depend on the fish alerting it, who's blind, from the danger. Those performances are mostly about finding a trade-off between going out from the shelter to eat and hidden from being eaten. Bigger gobies tend to associate with bigger shrimps who dig bigger holes than small shrimp (Jaafar et al. 2012) but then, the dyad will also have more needs in food. In our analyses, we found that small gobies significantly take more time to go out of their shelter after being threatened than medium or large gobies. There might be at least two possible explanations, that can both explain part of the latency. Karplus et al. (2011) found that signals emitted to alert the shrimp from a danger were almost always made when big fishes were approaching, independently from their diet. Therefore, we hypothesized that small gobies will perceive the danger bigger than might do larger fish and therefore they will be more cautious and wait more before going out to be sure the danger has passed. Also, considering the competition intra-gobies for a hole and a partner, and knowing that bigger gobies tend to displace smaller one (Karplus 1074), the latter must gain to stay hidden in the hole, so its shelter is not taking apart from him by a larger conspecific he will not be able to displace in return. Secondly, the small gobies being satiated faster, with less food, they will not need to take as much risks as bigger gobies and waiting a bit longer to go out will not be as costly for them.
However, our dataset being quite small (only 27 different individuals) and the stimulated threat being standard (our fins), we would need to do more experiences, this time with different sizes of “threat” to see if it impacts the latency. To complement this and test the second part of the trade-off, so called the feeding part, we should test the gobies after they already ate or after a certain time of starvation. This opening to future experiments is also supported by our analyses of the effect of the time of the day, even though it was not significant. Both morning and midday approached significance and gobies tended to go out of their shelter faster on the morning, therefore after a night without eating.
Considering the short-term learning, nothing was significant, and we therefore consider there was no learning from the fish across the five trials. We firstly thought that the gobies would notice that the threat was not real and learn that it was not a predator. Therefore, we expected them to go out of their shelter faster and faster across the trials. However, after considering the results and even though they are not significant, because gobies took more time to go out of the shelter on trial 5 compared to trial 1, we would expect them to understand that the threat is still around and therefore to stay hidden longer. However, five trials might not be enough for them to learn anything and the experiment could be improved by doing more trials. In this case, we also would reconsider our predictions.
Another limit to our experimental setup is that we considered shrimps to burrowed only holes. However, some species are able to dig tunnels instead (Karplus 1974). That might explain why some gobbies took more than 10 minutes to go out after the stimulus and must be taken into account if more studies were performed on this gobie/shrimp association.
Conclusion
In this study, we analyzed whether different factors would affect the latency of a goby to go out of its shelter after being threathened. We found that small gobies took significantly more time than bigger gobies while the time of the day did not affect significantly the latency. There were not any learning accross the 5 trials also. Because there are more and more tourism around coral reefs and that palms movements are considered as a threat by gobies, we can expect that constant mouvements aroud the fishes might affect their feeding habits and the time they spend into their hole. The absence of learning might emphasize that as every new movement is consider as a single threat (or the same remaining in the area). More studies concerning the perception of the threat and whether different types of threats or stimuli affect the latency should be performed to obtain a more comprehensive vision to take into account into management plans.
References
Freinschlag, M., & Patzner, R.A. (2012). Shrimp-gobies in the southern Gulf of Aqaba (Red Sea). Zoology in the Middle East, 55(1), Doi:10.1080/09397140.2012.10648916
Jaafar, Z., & Dexiang, C. (2014). Goby and shrimp associations: more than meets the eye. Coral Reefs, 33(3), 863. Doi: 10.1007/s00338-014-1143-4
Jaafar, Z., & Hou, Z. (2012). Partner choice in Gobiid fish Myersina macrostoma living in association with the alpheid shrimp Alpheus rapax. Symbiosis, 56(3), 121-127. Doi: 10.1007/s13199-012-0166-2
Karplus, R., & Szlep, M.T. (1974). The Burrows of Alpheid Shrimp Associated with Gobiid Fish in the Northern Red Sea. Marine Biology, 24 (3), 259-268. Doi: 10.1007/BF00391901
Karplus, R., & Tompson, A.R. (2011). The Partnership between Gobiid Fishes and Burrowing Alpheid Shrimps. The Biology of Gobies, 559-607. Doi: 10.1201/b11397-29
Khoda, M., & Yamanouchi, H., & Hirata, T., & Satoh, S., & Ota, K. (2016). A novel aspect of goby–shrimp symbiosis: gobies provide droppings in their burrows as vital food for their partner shrimps. Marine biology, 164:22. Doi: 10.1007/s00227-016-3060-2
Figure 2: National Park of Ras Mohamed, Egypt
https://www.google.com/maps
Figure 1: Goby and Shrimp mutualism.
https://www.leisurepro.com/blog/wp-content/uploads/2009/12/goby-shrimp.jpg
References Images
Figure 1 : https://www.leisurepro.com/blog/wp-content/uploads/2009/12/goby-shrimp.jpg
Figure 2: https://www.google.com/maps