• <div style="background-image:url(/live/image/gid/32/width/1600/height/300/crop/1/41839_V14Cover_Lynch_Artwork.2.rev.1520229233.png)"/>


Color and Territoriality are Linked to Aggressive Behavior in Male Betta splendens

Marisol Gelacio, Austin Conn, Gabriela Reid
Department of Biology
Lake Forest College
Lake Forest, Illinois 60045


Male Betta splendens are often associated with their vibrant colors and aggressive behavior, but it remains unclear whether the coloration of an intruder or its territoriality status plays a role in the level of aggression that a male Betta will display. By testing red Betta fish under four separate conditions (red intruder, black intruder, red resident, black resident), we hypothesized that the Betta would display more aggression towards a red challenger, since it has been previously shown that they tend to attack their own kind. We also hypothesized that the resident Betta fish would display more aggression when it was an intruder, since Betta fish are known to be highly territorial. Therefore, if the Betta fish was not the resident in the tank, he would be less aggressive towards the intruder, as they had established a territory before him. Under the four conditions, we measured the number of fin flares and latency from each fish as an indication of aggression. The red color of the challenger proved to be significant in causing more aggression than the black color, while territoriality did not.


We chose to study Betta splendens, which are commonly known for their vibrant colors and aggressive behavior. It was previously confirmed that Betta fish, particularly males, are prone to high levels of aggression and will attack their own kind (Romano et al., 2017). Romano and colleagues used a red robotic fish, mimicking the male Betta fish, to elicit aggressive responses. They found that the Betta fish showed aggression by performing several non-physical displays at first, such as gill flaring and fin spreading, and then physical displays, such as tail biting. There are multiple reasons for male Betta fish to respond in either a non-physical or physical manner towards their own kind, such as fighting for food, to protect their nests and eggs, and when a male of their own kind enters their territory. For our Independent Research Project, we decided to focus on the effect color and territoriality status have on the level of aggression that is displayed by the male Betta fish. Similar to our experimental approach, Niko Tinbergen performed a study on the three-spined Stickleback fish in 1952 to test how color and shape play a role in the level of aggression  displayed by using the color red and four different shaped models (squared, round, flat, and oval shaped). Tinbergen determined that the territorial response and aggressive attack in Stickleback fish was not due to the shape of the fish model. Instead, the fish reacted exclusively to the red color, acting as the sign stimulus (from Biological Principles).

In addition, Tinbergen believed that there were four levels of analysis questions that should always be asked about the natural behavior of any animal, in this case, the aggressive display made by Betta fish: What signals elicit the behavior performed by the animal? (Causation), How does the behavior help the animal survive? (Survival Value), How does the behavior change or develop throughout an animal’s life? (Development), and Why does the behavior arise and evolve in species? (Evolution). The levels of analysis that fit our study were causation (proximate) and survival value (ultimate). We believed that the aggressive behavior that male Betta fish display could be categorized as a causation explanation because it has been previously shown that when there are higher levels of 11-ketotestosterone, which functions as the endogenous androgenic sex hormone (Pretorious, 2016), there are higher levels of aggression, when nesting and territorial status were involved (Dzieweczynski, 2006). We also believed that the aggressive behavior could be explained through a survival value explanation, because it has been previously shown that Betta fish perform a visual threat display (fin flare) when they claim territoriality status to show superiority, which then benefits the Betta fish and increases its fitness (Meliska et al., 1980).

We therefore chose to look at the effect of color and territoriality status in male Betta fish for our Independent Research Project. The research question we aimed to answer was: does the coloration of a challenger and its territoriality status, or a combination of both, play a role in the level of aggression displayed by a male Betta splendens? We divided our hypotheses into two sections, color and territoriality status. For our null hypothesis, we predicted that there would be no difference in the level of aggression the male Betta fish displays toward the lure based on its color and territoriality status. Our two-color hypotheses were that the resident Betta fish would display and stimulate more aggression towards a red lure intruder and the alternative hypothesis was that the Betta fish would display more aggression towards the black lure intruder. It has been previously shown that male Betta fish tend to fight their own kind. Because the red lure mimics the coloration of the experimental fish, we predicted the Betta fish would have a shorter latency time and a higher amount of full flares when he was tested with the red lure rather than the black lure (Baenninger, 1966). Our two territoriality hypotheses were that the intruder male Betta fish would display more aggression towards the red lure resident and the alternative hypothesis was that the intruder Betta fish would display more aggression towards the black lure resident. We made this prediction because it has been previously shown that Betta fish are also known to be highly territorial, especially when they are confined in smaller areas, leading to more confrontation (Forsatkar, 2016). This then led us to believe that if the Betta fish were the intruder in the tank instead of the resident, he would be less aggressive since another fish was there before him and had already established a territory. In addition, to perform our Independent Research Project, our independent variables were the different colored lure models (red or black) used to experiment on the Betta fish, as well as whether the Betta fish was the resident or intruder during the experiment. The dependent variables were the latency of flare made by the Betta fish when it reacted to the lure as well as the level of flare that the Betta displayed (0,1,2).


Fifteen red colored male Betta splendens were purchased from two local pet stores and transferred into tanks held in the Lake Forest College laboratory. The Betta fish were placed in Kritter Keeper tanks, housing them with rocks for environmental enrichment. The smaller tanks were used to increase aggression during testing. Partitions were created using brown cardboard cutouts andplaced in between tanks to eliminate exposure to neighboring Betta fish. Each fish was numbered for testing procedures between one and fifteen using masking tape and a sharpie marker. Three TetraBetta Plus floating mini pellets were fed to the fish routinely every day after the experimentation period to elicit more aggressive behavior. Each of the fifteen fish were tested once a day, and we used a random number generator to randomize the order of testing condition for each fish.  In addition, each fish was tested under 4 conditions (red intruder, black intruder, red resident, black resident) for 5 minutes using Continuous Focal sampling, alternating between condition and fish to avoid habituation. Red and black Booyah Bait Co bass lures were used as the challengers for testing color and territoriality. Pipe cutters were used to remove the hooks from the bass jigs before placing them inside the tank. Lastly, zip-ties were used to shape the lure strings to mimic the fin flare of the male Betta, and a wire was attached onto the lure to hold the model in the fish tank.

Preliminary Observations

We randomly chose a male Betta fish (#2) out of the fifteen to conduct a preliminary observation on and to record its behavior toward the model (Table 1). The red lure was placed in the tank without movement for 10 minutes. We then recorded using Continuous Focal Sampling to observe its behavior and created an Ethogram.

Gelacio Fig 1

Table 1. Preliminary observations on the behavior of the male Betta

Fish #2 towards the red lure.



A random number generator was used to develop the order of experimentation for each fish and each of the conditions (Table 2). There were four conditions: red resident, black resident, red intruder and black intruder. We defined a resident as the individual who had established territory on its specific area, in this case the tank that each Betta splendens was kept in. We defined an intruder as the novel individual who was being introduced to the new territory. When we tested the male Betta fish as the resident, the lure was attached to the lid of the occupied tank so that it remained stationary within the Betta’s territory. A timer was set for five minutes as soon as the lure was placed in the tank. Two of the experimenters recorded the Betta fish behavior toward the lure by using Continuous Focal sampling. This procedure was conducted in the same manner with both the red and black colored lures. The latency of fin flares was recorded and used as the measure of aggression displayed by the Betta fish. The three types of flares recorded were no flare (0), partial flare (1) and full flare (2). Data was analyzed using a two-Way ANOVA, General Linear Mixed Model and a Chi-Square statistical test.

When we tested the male Betta fish as the intruder, the lure was first fastened to the corner of an unoccupied tank, containing the same environmental enrichment factors as the housing fish tanks with rocks. Once the lure was positioned in the tank, the Betta fish was placed into the tank and the time was recorded immediately for five minutes. Two of the experimenters recorded the Betta fish behavior toward the lure by using Continuous Focal sampling. This procedure was conducted in the same manner with both the red and black colored lures acting as the resident. Similar to the resident experimentation, we recorded the latency of fin flares as well as the type of flare each fish performed by categorizing them into the level of flare: no flare (0), partial flare (1), and full flare (2). We analyzed the data by using a two-way ANOVA, General Linear Mixed Model and a Chi-Square statistical test.

Gelacio Fig 2

Table 2. A random number generator was used to randomize the four testing

     conditions (red or black resident, red or black intruder) for each male Betta.

     The numbers indicate the order in which each condition was performed.


We recorded the latency of fin flares displayed by the Betta splendens during the procedure for each condition as seen in Figure 1. Figure 1 depicts the average time in seconds for the Betta splendens to project their fins as a sign of aggression in each condition tested. The average latency for the red resident condition was 140.5 seconds and the black resident was 223.6 seconds. The average latency for the red intruder condition was 193.5 seconds and the black intruder condition was 227.8 seconds. The average latency for red was lower than for black as seen in Table 1. The p-value was 0.178 and the F-statistic was 1.700. Table 1 displays the results of the General Linear Mixed Model analysis as F-statistics and P-values of ownership (resident or intruder condition), color (red or black lure) and ownership and color together.

Table 2 displays the results of the statistical analysis Chi-square on categorical data, such as the type of flare. We categorized the level of flare into three categories: no flare (0), partial flare (1) and full flare (2) and the averages for each level are seen in Figure 2. Table 2 shows the variation in response of individual identity with repeated measurements. There is a significant effect of identity due to individuals’ responses being vastly different. There is an effect of color, no effect of ownership and no interaction between ownership and color as seen in Table 2.

Gelacio Fig 3

Figure 1. Average Latency of the thirteen male Betta fish to flare their

fins as a sign of aggression under the four conditions.


Table 3. Generalized Linear Mixed Model Analysis of ownership, color, and both.

Gelacio Fig 4

Gelacio Fig 5

Figure 2. Average level of flare (0,1,2) produced by the thirteen male

   Betta fish under the four conditions.


Table 4. Chi-Square Statistical Analysis of the type of flare made by the male Betta

   fish based on territoriality, color, and in conjunction.

Gelacio Fig 6


After our testing period was completed, we were able to reject part of our null hypothesis. There was a significant difference in the level of aggression shown by the male Betta splendens, based on the color of the challenger. However, there was no significant change in the level of aggression due to difference in territoriality status, whether the male Betta fish was the resident or intruder. In addition, our hypothesis that the male Betta fish would show more aggression towards a red novel intruder rather than a black intruder was supported. We hypothesized that the male Betta splendens would be more aggressive towards the red lure, since it has been previously shown that Betta fish tend to fight their own kind. (Baenninger, 1966). We also made this prediction due to all of the experimental Betta fish having a similar red coloration to the red lure.

While the color hypothesis was supported, the territoriality status hypothesis was not. A possible explanation could be that the males may need further external stimuli to exhibit aggressive behavior in regard to their territorial status. Previous studies have shown that male Betta fish will show higher levels of territoriality only if two specific factors are present at the same time. The study suggests that the factors of a male audience being present and both the resident and intruder Bettas not having a nest is crucial. The study supported that the aggressive behavior of male Bettas can be dependent on the sex of the audience, their reproductive state, and resource possession (Dzieweczynski et al. 2006). There was no audience present or resource to influence aggression during the interaction between the Betta fish and the lures in our study. Also, the reproductive status of all of the fish were the same and there were no females present. Another previous study showed that Betta fish will show more aggression when defending their home tanks, although in this 1984 study the fish had visual and tactile cues that differentiated their home tank from other’s. The fish also responded to the chemical cues of other Betta fish (Bronstein, 1984). In our study, all tanks were visually the same, had no defining tactile cues, and a lure was used which did not exude any chemical cues. Therefore, the Betta fish may not have been able to differentiate between their own tank and the tank that we used for the intruder testing.

We do not believe that habituation played a role in the level of aggression that the Bettas showed in later tests, due to previous studies regarding Betta aggressiveness towards strangers and neighbors. The first study showed that Betta fish will show similar levels of aggression towards all intruders, no matter how habituated they had become. The experimenters suggested that this was likely due to all Bettas posing the same level of threat (Alyan, 2010). Another study suggested that Betta fish do not show habituation, but rather a strategy of avoidance after abandoning an hour of failed attacks (Rhoad, 1975.) For our future studies, we would like to extend our testing time, perhaps longer than five minutes, as previous studies have shown that aggressive behavior can continue for over an hour.

The same 1975 study suggests that an identical level of effectiveness is shown in stationary models, moving models, and a mirror (Rhoad, 1975). In this case, our stationary model was ideal for this experiment. This experiment was important as it challenged previous studies of challenger color and how coloration is linked to Betta splendens aggression. Although Betta fish are generally understood, not many scientific studies regarding the connection between color and aggressive behavior have been done. This study adds to the previous knowledge of how color and territory affect the aggression of the Betta splendens.

Future directions relating to this study could include examining the response of the Betta fish to additional colors such as blue or green. Blue Betta splendens should be tested in order to discover if they are more aggressive towards the same color as themselves or neon colors in general. Different model sizes could also be tested, along with changing the shapes of the challenger, similar to Niko Tinbergen’s experiment on Stickleback fish. By changing the environment of the Betta fish, it can also be used to further understand what leads to territoriality aggressiveness. In addition, chemical cues from our model would be helpful in signaling to our subjects that they are intruders to the lure. Lastly, for future directions, we could test Betta fish through a larger sample size, as we were only able to include data from thirteen Bettas, in which we excluded two fish that died during our experimental time. We would also like to increase the acclimation time of the Betta fish, increase observance time, and eliminate stress from the transfer of the fish from tank to tank. Eliminating stress could potentially be accomplished by creating open passageways to other tanks that already contain the lure, rather than removing the fish with a net each time.


*Ethical Note

No fish were harmed during experimentation, and each individual found permanent homes post experimentation.



We would like to thank Professor Barbosa, Professor Houde, and the Lake Forest College Biology department for their resources, knowledge, time, and support.



Alyan, S. (2010). Male Betta splendens are equally aggressive toward neighbors and strangers.

Journal of Ichthyology, 50(11), 1066-1069.


Baenninger, R. (1966). Waning of aggressive motivation in Betta splendens. Psychonomic 

Science, 4(7), 241-242.


Behavioral Ecology. (2016, August 19). Retrieved April 26, 2018, from



Bronstein, P. M. (1984). Agonistic and reproductive interactions in Betta splendens. Journal of

Comparative Psychology, 98(4), 421-431.


Dzieweczynski, T. L., Eklund, A. C., & Rowland, W. J. (2006). Male 11-ketotestosterone levels

change as a result of being watched in Siamese fighting fish, Betta splendens. General

and comparative endocrinology, 147(2), 184-189.


Forsatkar, Mohammad; Nematollahi, Mohammad; Bron, Culum (August 2016). “Male Siamese

Fighting Fish Use Gill Flaring As The First Display Towards Territorial Intruders”.

Journal of Ethology. 35: 51–59. doi:10.1007/s10164-016-0489.


Meliska, C. J., Meliska, J. A., & Peeke, H. V. (1980). Threat displays and combat aggression in

Betta splendens following visual exposure to conspecifics and one-way mirrors.

Behavioral and Neural Biology, 28(4), 473-486.


Pretorius, E., Arlt, W., & Storbeck, K. H. (2017). A new dawn for androgens: novel lessons

from 11-oxygenated C19 steroids. Molecular and cellular endocrinology, 441, 76-85.


Rhoad, Karen D., et al. “Aggression and Avoidance By Betta Splendens toward Natural and

Artificial Stimuli.” Animal Learning & Behavior, vol. 3, no. 3, 1975, pp. 271–276.,



Romano, D., Benelli, G., Donati, E., Remorini, D., Canale, A., & Stefanini, C. (2017). Multiple

cues produced by a robotic fish modulate aggressive behaviour in Siamese fighting

fishes. Scientific Reports, 7(1), 4667.


Eukaryon is published by students at Lake Forest College, who are solely responsible for its content. The views expressed in Eukaryon do not necessarily reflect those of the College.

Articles published within Eukaryon should not be cited in bibliographies. Material contained herein should be treated as personal communication and should be cited as such only with the consent of the author.