It’s a Spider-Eat-Spider World: Observations of Nonsexual Cannibalism in the Invasive Jorō Spider Trichonephila clavata

Abstract

Spiders and other arthropods can sometimes consume others of their kind, and this is most often associated with mating activity, whereby females cannibalize males during or after mating, or during mating attempts. Nonsexual cannibalism is less common but may be associated with food availability or territorial aggression. In the Southeastern United States, a non-native orb-weaving spider, Trichonephila clavata (the “jorō spider”), is expanding its range. Prior lab experiments indicated this species to be “shy” compared to other native spiders, based on behavioral reactions to stimuli. Here, we report descriptive observations and photo-documentation of nonsexual cannibalism by this species, including from anecdotal observations, plus findings from controlled pairings of spiders, both in the lab and in natural webs in the field. In the cases where cannibalism was witnessed, it involved one female biting and killing another, typically after a short fight. When two females of a similar size were placed together in a container (n = 25 trials), fights ensued 40% of the time. When females of different sizes were paired (n = 27 trials), fights happened 18% of the time, and the larger females were not always the aggressor. Across all the lab trials (n = 52), six bouts (9%) led to the direct killing of one female. In field trials where two females were placed on an empty web (n = 14 trials), we observed one fight (7%) where the aggressor ended up killing and wrapping the other spider in silk. Given that some of these instances happened away from any web, these observations imply that the aggression is not necessarily an act of territoriality. The intraspecific aggression could arise when females are provoked or stressed, which deserves more study.

1. Introduction

In the Southeastern United States, a non-native orb-weaving spider, Trichonephila clavata (the “jorō” spider), has been expanding its range since it was discovered in 2014 [1]. This species has a native range throughout East Asia, and based on its physiology and knowledge of its climate suitability elsewhere, it is estimated that it will eventually expand beyond the southeast and occupy a wide range within North America [2,3]. In the introduced range, the species now occupies over 40,000 km² [4]. Each fall, webs of this species can be found within a wide variety of habitats and landscapes, including in backyards, on trees, on human structures, and even on street lights over busy roadways (Davis, pers. obs.). Interestingly, this affinity for urbanized environments seems common within this genus of spiders, as both the related T. clavipes, and T. plumipes are known for exploiting urbanized habitats [5,6]. Given this tolerance of disturbance, it is therefore surprising that T. clavata is thought to have a timid personality, based on behavioral reactions to mild negative stimuli [7]. Tests in that study revealed how T. clavata (and T. clavipes) responded to stimuli by freezing for over an hour, while most other orb-weaving species only froze for a few minutes. Those findings implied that these spiders have a “shy” personality, since prolonged thanatosis is a hallmark of timidness [7,8,9,10].

Within the scientific literature on spiders, there are many reports of cannibalism, though the majority refer to cases of “sexual” cannibalism, whereby females attack and/or consume males either during or after mating attempts [11,12,13,14]. Studies of nonsexual cannibalism (one conspecific attacking and eating another, not associated with mating) are fewer [15], but the research does suggest that hunger and body size do influence the behavior in some spiders [16]. In a thorough review of this behavior, Wise [17] pointed out that some research even indicates this behavior to be important in regulating the population sizes of spiders. Other factors that can influence this behavior include habitat complexity and the foraging mode of the species [17].

In this paper, we report instances of nonsexual cannibalism and/or aggression by female jorō spiders in their introduced range, including preliminary observations, plus the results of direct lab and field-based trials. Given the need for biological knowledge of this newly invasive species, we endeavored to provide detailed descriptions of the behavior, plus photo-documentation. Our observations and findings not only add more nuance to the view of this species, which was thought to be “timid” [7], but also provide new insights into the nature of such behavior among spiders and suggest there is more to learn about its function.

2. Methods

Anecdotal observations—The impetus of this project began with multiple, and separate, observations made of apparent cannibalism by female jorō spiders, described as follows. On 17 September 2023 and 19 November 2023, one of the authors (AKD) observed cases where a female T. clavata had apparently captured and eaten another female in her web. The first case was in a web next to the author’s personal residence (Oconee County, GA, USA). It was typical of the many webs in the local area and given the season, i.e., it was approximately 1 m in diameter, and a mature female was in the middle. The female was clutching a smaller female T. clavata when the web was discovered (the attack itself was not witnessed; Figure 1A). The smaller spider had been wrapped in silk and apparently treated as a prey item. The author came back to the web one hour later and the carcass remained hanging in the web. It was removed to obtain a picture (Figure 1B). Close inspection of the remains suggested that the cephalothorax had been consumed and the abdomen was largely intact. The second case was also near the residence, and a female T. clavata was observed wrapped up (and dead) within the web of another female (picture not shown).

A second incident of interest occurred while gathering video footage of T. clavata for an unrelated project. For this, two of the authors had collected two mature females from webs in Oconee County, GA, USA, and transported them to a lab 20 min away at the University of Georgia campus in Athens, GA, USA, where they were placed in a glass aquarium designed for macro-photography. The aquarium (90 × 45 × 45 cm) walls and floor were black to provide contrast with the spiders and enhance their visibility. No other materials (i.e., sticks, vegetation, leaves, etc.) were provided at the time. A video camera was positioned in front of the aquarium to gather closeup footage of the spiders. After this relocation to the aquarium, the two spiders each explored the container for 10 min, depositing lead lines as they travelled. At some point, they each came into contact with the other and appeared to pause. They remained in close proximity for 5 min.

Suddenly, the larger of the two spiders lunged at the smaller one, traversing approximately 5 cm. This escalated into a physical fight between the two spiders that lasted less than 5 s. The fight ended with the smaller spider clutching the larger one in the cephalothorax region and apparently subduing it with venom (Figure 2A). The larger spider did not move after that. The smaller spider retained this grip for at least 45 min thereafter. The observation was stopped before the smaller spider loosened its grip, so the exact time is unknown. At no time did the smaller spider wrap the larger one in silk. This continued until, apparently, the smaller spider was finished and then moved away. From examining the carcass of the larger spider (Figure 2B), we could see that only the cephalothorax had been consumed. This resembled the carcass from the observation of cannibalism in the wild (Figure 1B).

A final incident of cannibalism was witnessed by three of the authors when conducting other, unrelated research on this spider [18]. Specifically, the authors had been collecting female jorō spiders from webs in the field in Clarke County, GA, USA, during fall 2023 and temporarily holding each in a 50 mL Falcon tube for weighing purposes. After weighing, the spiders were released on the ground. When two of the spiders happened to be released at the same time, and near each other, they began fighting (see Figure 3), and one ended up killing the other.

Laboratory trials—The preliminary observations above led us to conduct the following trials to ascertain how often this behavior occurs, in both lab and natural settings. The lab trials were conducted during fall 2024, using field-collected females from areas in Clarke and Oconee County, GA, USA. Females were removed from webs and held individually in 50 mL Falcon tubes for transport back to the lab. We first weighed each spider with an electronic balance. Then, trials were conducted by placing two females into a 1 L plastic container (see Figure 4A) and then observing them from a short distance away for 1 h. The observer recorded any incidences of aggression and/or direct killing (cannibalism) during that time. We conducted two versions of these trials, one where we only used females of similar body mass (n = 25 trials) and other trials where females of mismatched body sizes were used (n = 27 trials). By comparing the outcome of these two scenarios, we had hoped to understand if the aggression was due simply to overall body size (i.e., if the larger females tended to kill the smaller ones).

Field trials—Since the lab-based trials involved a very unnatural scenario for these spiders, we endeavored to conduct a small number of field-based trials for comparison. On 12 November 2024, three of the authors visited a nature preserve in Oconee County, GA, USA, where there was an abundance of large jorō spiderwebs. The authors first collected a number of mature female spiders to use in the trials by pulling down their webs and placing the females each in 50 mL Falcon tubes. Then, the authors randomly selected other, intact webs (that were at least 1 m wide) to use for the trials (one different web for each trial). The spider(s) within the web was removed, so that the web was empty for the trials. Then, the authors placed two random females from the Falcon tubes into the empty web and watched the spiders for one hour, recording any instances of aggression and/or cannibalism. This scenario was intended to replicate the natural conditions that the spiders would be in, including being outdoors and on a web (though not their own).

3. Results

Laboratory trials—In the pairings of similarly sized females, there were fights observed in 10 trials (40%). Not all the fights ended up with the direct killing of one spider, but there was often damage to one or both spiders, usually in the form of one or more legs being damaged or lost. In the trials of mismatched body sizes, fights were witnessed in five trials (18%). The differences between these frequencies approached significance in a Chi-square test (ꭓ² = 2.92, p = 0.087). Collectively, of the 52 total lab-based trials, 6 led to obvious cannibalism (11.5%), where one spider killed the other. These involved one spider biting and securing the other until it succumbed (see Figure 4B).

Field trials—A total of 14 field trials were conducted. Since the lab-based trials above indicated no noticeable effect of body size (i.e., the larger spider did not tend to dominate the smaller or vice versa), we did not a priori select females of similar body size for these trials. Of the 14 trials, we observed one case (7%) where a fight ensued between the females, where one female ended up killing the other and then wrapping it in silk (see Figure 5). In the other instances, either nothing happened or one or both of the spiders left the web.

4. Discussion

The collective observations here, including the preliminary sightings, the lab trials, and the field-based trials, indicate that jorō spiders are capable of a surprising degree of intraspecific aggression, although the controlled trials suggest that instances of direct cannibalism are not common (less than 10% of interactions). Nonsexual cannibalism is not well-studied within the literature on Trichonephila spiders and so it is not clear if this is a “normal” behavior for this spider species. In theory, orb-weaving spiders, at least solitary species, would not normally encounter other spiders off a web to begin with, and so some of our “observations” here may be unrealistic. Nevertheless, we did witness direct cannibalism in our field trials, and we anecdotally witnessed at least two instances of jorō spiders being prey within other (unmanipulated) jorō spiderwebs. Thus, while uncommon, this behavior does occur within the species.

What makes this behavior interesting is that prior research in our lab with T. clavata indicated that spiders of this genus appear exceptionally “shy”, based on observations of their behavioral reactions to mild negative stimuli (air puffs) [7]. The observations here suggest these spiders are capable of a surprising level of intraspecific aggression. Consider that in the lab-based trials, or in the aquarium incident, neither spider could have been considered a “territory-holder” in the environment we put them in, nor were there any prey items to fight over, and so there should not have been any reason for a direct physical fight, and seemingly without provocation. Moreover, a few of these fights escalated until one spider actually consumed the other (or at least the cephalothorax), and all without wrapping it in silk. From these observations, we conclude that females of this species can be hostile to other females.

The fact that there were more fights when the females were of equal size (though this trend was only marginally significant) could suggest that fights happen more when the spiders perceive a potentially equally matched competitor. Conversely, the fact that there were fewer fights when there was a size discrepancy was interesting in itself. Most observations of sexual cannibalism within spiders involve females consuming much smaller males [11,17]. Further, prior work with spiders that do not have mismatched sexual size showed that nonsexual cannibalism occurs more often when there are size differences between conspecifics [16]. Further, spiders with a smaller body size are more likely to have their webs taken over by larger spiders [19]. Thus, it would have been logical to expect that large T. clavata females would simply attack smaller females more often. However, in cases of sexual cannibalism, and also nonsexual cannibalism, it is thought that the female (or larger spider) derives nutrition from consuming the smaller spider [11,16,17,20], possibly benefitting her reproductive success. Moreover, hunger has been shown to be a factor that influences cannibalism tendencies [16], which is something that we did not track, as all the spiders tested were collected from the field.

The level of aggression we observed is also at odds with the tendency of spiders of this genus to form “communal” webs, which we have observed in Georgia. Communal webs of the genus Trichonephila can be very large, with sometimes dozens of females, each with their own web that is connected to others on all sides, sometimes within 10 cm of each other (Davis, pers. obs.). Such aggregations are known within members of this genus, though they are not technically “social” aggregations [21]. The factors that lead to this behavior are not yet clear; there is evidence that some communal species form such webs in areas with high food abundance [22,23], though this may not be the only explanation [21]. Of interest here is the apparent tolerance of each female on these communal webs to be so close to others, when our observations here suggest this can (sometimes) lead to aggression. One observational study of T. clavipes did reveal how females on communal webs can sometimes interact with their nearest neighbor in an aggressive manner by communicating via “rapid exchange of web jerks” [23], yet no physical altercations were observed. This could be interpreted to mean that the web itself is needed to facilitate communication between rival females, which prevents escalation of the interactions. Clearly, the nature of the communal web behavior within this genus of spiders, and especially how these spiders “get along with each other”, deserves further study.

A final possible factor to consider with regard to our captive observations is that the initial capture of the spiders and placement into the novel environment caused a degree of physiological stress to the spiders, which can manifest as aggression [24]. In fact, the frequency of aggression in both lab-based trials (40% and 18%) was higher than that observed in the field (7%), which is consistent with this idea. Future work could be performed to evaluate if this is indeed the case, perhaps by replicating the captive conditions described here and/or exposing spiders to standardized stressors. Such investigations could help to elucidate the true personality profile of this invasive species in the southeast.