Training and enrichment of ectotherms

Introduction

“Many laboratories and zoos are documenting the value of systematic observation and experimentation to test out the best ways to enhance the lives of reptiles in captivity as well as to provide more educational and ecologically appropriate exhibits. A preoccupation with large numbers is not necessary for good science worthy of publication. This reminder is needed to encourage persons who are working with small numbers of large, rare, endangered, if not potentially dangerous, animals to record and publish their endeavors.” (Burghardt, 2013, pp. 295).

Gordon Burghardt, an expert in reptile and amphibian cognition, claimed more than 10 years ago that he could review all the literature in existence on the topic of amphibian cognitive abilities in less than 20 pages (Burghardt, 2013). The claim is likely still close to true. For this reason, it is important to highlight innovative work being done on training and enrichment with reptiles, amphibians (henceforth, herps), and other ectotherms. Similar to work by Monreal-Pawlowsky et al. (2021), the examples in this paper are intended to showcase the cognitive abilities of these species and encourage others to embark upon similar training and enrichment programs. While the examples provided here are single sample case studies or “power studies” (Triana and Pasnak, 1981), they represent a picture of the capacity for cognition in these animals (Kaufman and Kaufman, 2016) and examples such as these are becoming increasingly acceptable as representative of species’ ability in research done in zoos and aquariums (Alligood et al., 2017; Margulis and Westhus, 2008; Saudargas and Drummer, 1996). Every lobster has the same neurological makeup as every other lobster. One of the examples detailed in this paper is that of a lobster named Cupcake obtaining food hidden in a rubber treat toy. We do not argue that every lobster can do this. Our goal is to focus the discussion on why Cupcake, with the same neurological equipment as every other lobster, does. What causes individual variations amongst individuals in a species? What conditions, forces, or selective pressures made for a problem-solving lobster? And, more importantly, how do other zoos, aquariums, and museums develop enrichment and training programs which foster individual differences and specific abilities in their animals?

The importance of training and enrichment

Many research groups have detailed the importance of both training and environmental enrichment for the health and welfare of animals in professional care in zoos, aquariums and other facilities (see, for example, reviews by Alligood and Leighty, 2015; Clark, 2017; Melfi, 2009). Enrichment is rapidly becoming recommended or even mandatory for animals in professional care across the globe (Association of Zoos and Aquariums, 2025; Heidenreich et al., 2023) to enhance general husbandry and provide cognitive stimulation (Westlund, 2014), as well as ensure personnel and animal safety during medical procedures (Fernandez and Martin, 2021). More broadly, there is evidence that the public benefits from seeing training and enrichment – both through a greater understanding of the care provided by facilities, and an increased understanding of the abilities of the animals themselves (Povey and Rios, 2002; Price et al., 2015; Riley and Rose, 2020; Visscher et al., 2009).

Training (i.e. operant conditioning), in and of itself, can be used as a form of enrichment in addition to as a tool for husbandry (Westlund, 2014). Although training is often strongly associated with “intelligence” and cognitive abilities of animals (Burghardt, 2013; De Meester and Baeckens, 2021; Kaufman and Kaufman, 2015), and this perspective is not necessarily bad or incorrect, it may contribute to many ectotherms not receiving training or enrichment in zoos or aquariums (Clark, 2011). Marine mammal trainers are known to boast that anything can be trained given the right reinforcement–and they are right. Pryor (1999) tells a story of a friend who trained scallops to clap on cue. Clapping scallops are fun, but generally training programs are focused on behaviors necessary for routine husbandry and veterinary care. Instead of waiting for lions to wander from one section of their habitat to another, they can be trained to “shift” on cue. Large animals like elephants can be trained to lift their feet for checks and foot maintenance, avoiding unnecessary anesthesia for routine husbandry. Proactive training for medical procedures such as this improves the safety of both the animals and keepers (Fernandez and Martin, 2021). Lastly, operant conditioning may provide an excellent way to learn more about the cognitive abilities of ectotherms – for example, experimental subjects trained this way can then be used in novel item or quantity paradigms, or in match to sample tests (Gatto et al., 2022). Training not only uses an animal’s cognitive abilities, it can also develop them (Howard and Freeman, 2022).

Goals of this paper

Our intention for this paper is twofold. We highlight advances in ectotherm training and enrichment at our facility as well as the behavioral insights these efforts provide for the species studied. Our findings should encourage others within the zoo and aquarium community to think more broadly about the species involved in their enrichment and training programs. We expect that insights derived from our efforts will also help to bridge the communication gap between academic researchers and zoo staff. We aim to bring the advances in ectotherm training and enrichment to the attention of the scientific community, and to encourage others within the zoo and aquarium community to think more broadly about the species involved in their training and enrichment programs (Loh et al., 2018; Riley and Rose, 2020; Rose and Riley, 2019; Skibins and Powell, 2013; Albert et al., 2018). We also hope to showcase the wide variety of options for training and enrichment of ectotherms, which in turn will allow guests to learn about the natural behaviors and adaptations of our animals, keeping them excited to learn, involved in our facilities, and committed to conservation.

Cognitive abilities in ectotherms

As a generalization, it seems that cognition, trainability, and benefit from enrichment are inherently linked. This assumption is both easy to understand and so sweeping as to be difficult to test. Compounding the situation is the lack of cognition research being done on ectotherms in general (Burghardt, 2013; Eagan, 2019). However, the research which has been done shows higher cognitive abilities than are often presumed for these species (Bisazza and Santacà, 2022; Burghardt, 2013; Burmeister, 2022; Gutnick et al., 2020; Liu et al., 2016; Szabo et al., 2021). This aligns with our experiences with training and enrichment at VLM.

What follows is an abbreviated review of literature on the cognitive capabilities of ectotherms. It is by no means exhaustive, but meant to provide a sampling of abilities for which experimental evidence has been presented. Because much of the training and enrichment undertaken with birds and mammals is based on a presumption of cognitive abilities, it is important to frame the abilities of ectotherms similarly.

The importance of individual differences in assessing cognition

Much of the purpose behind detailing our experiences with training and enrichment of ectotherms is to advance the idea that single subject can be a valid proxy for the capacity of a species, regardless of whether every individual shows the same level of ability (Kaufman and Kaufman, 2016). However, individual differences within a species should not be completely disregarded, particularly when providing individualized enrichment to animals in professional care, who may have different cognitive needs based on this variation (Carere and Locurto, 2011). Research on personality and individual differences is relatively new – especially when dealing with ectotherms, and while literature reviews show a general consensus for both the existence of individual differences and the relationship of these differences with learning abilities, the direction of the latter relationship has yet to be consistently established (Dougherty and Guillette, 2018; see excellent reviews of literature by Kralj-Fišer and Schuett, 2014 and Lucon-Xiccato et al., 2024).

The majority of studies on individual differences and learning use boldness as the main personality trait (Kralj-Fišer and Schuett, 2014). In reptiles such as the rock agama (Agama planiceps), this manifests as variation in risk taking and anti-predator behavior (Carter et al., 2010, Carter et al., 2012). In the delicate skink (Lampropholis delicata), sociability and exploratory behavior are positively correlated and consistent across gender (Michelangeli et al., 2016); and when an associative learning task is introduced, bold, neophilic individuals made fewer mistakes (Chung et al., 2017). Several studies of fish have provided evidence for personality traits which impact sexual selection (Colleter and Brown, 2011), foraging success (Harcourt et al., 2009), parental care (Barbasch and Buston, 2018; Stein et al., 2016), control of exploratory behavior (Savaşçı et al., 2021; Vinogradov et al., 2022), and predator avoidance (Brown et al., 2014). In spiders, aggressive and docile phenotypes appear to be heritable, and females with aggressive phenotypes are more active in foraging, web building, and defense of the colony (Holbrook et al., 2014; Wright et al., 2014). A similar effect has been found in leaf beetles (Phaedon cochleariae; Müller and Juškauskas, 2018). Several studies have shown the effect of acoustic environment on aggression and dominance in the field cricket (Gryllus integer; DiRienzo et al., 2012; Rudin et al., 2018) and Albers and Reichert (2022) found that in house crickets (Acheta domesticus) latency to emerge from a shelter varied with speed to reach criterion in a learning task. Lastly, cuttlefish (Sepia officinalis) can be distinguished by individual variation in attentiveness to prey, along with the speed and duration of their attack (Zoratto et al., 2018).

Considering variation in personality across individuals when establishing training and enrichment programs has been key to success at VLM. Boldness often encompasses willingness to enter open spaces, sociability, and exploratory behavior – all of which are relevant to visitor learning and experience. Encouraging this is important in zoos, and many of the examples in this paper discuss how observation of individual animals by their keepers led to the keepers fostering bold personalities via adapting training and enrichment to individuals.

Cognitive abilities in reptiles

Of the types of animals discussed in this paper, reptiles have had the most study (Hellmuth et al., 2012), although not close to the extent afforded to birds and mammals – in the academic literature, information on enrichment on birds and mammals is over 15 times more common than reptiles (Howard and Freeman, 2022). There is evidence for cognitive abilities and successful learning in snakes, lizards, and turtles, although testing may be more difficult due to specific sensory abilities and behavioral limitations (Burghardt, 2013). For example, research with reptiles has noted gaze following (the ability to follow the direction of another’s visual focus to obtain information) in tortoises and hognose snakes (Burghardt, 2013; Wilkinson et al., 2010b). Another group of researchers used operant conditioning and shaping (training via small incremental steps) to successfully teach Burmese pythons (Python molurus bitvitattus) to press a button for a food reward (Emer et al., 2015). Still another group used a “lid flipping” task developed for sparrows and presented it to Puerto Rican anoles (Anolis evermanni). Not only were the anoles able to find prey by removing or flipping the lids, they did so at a higher success rate than the birds; particularly given anoles eat only once a day (and thus participated in one trial a day), whereas the sparrows had up to six trials a day (Leal and Powell, 2012). Several other research groups have successfully demonstrated both problem solving and the procedural memory necessary to retain solutions in a variety of lizard species (Cooper et al., 2019, Cooper et al., 2020; De Meester et al., 2011). Chelonians are known to perform visual discrimination tasks, discriminate between quantities, learn via social observation, and follow the gaze of others to gain information (Davis and Burghardt, 2011; Gazzola et al., 2018; Gutnick et al., 2020; Wilkinson et al., 2010a, Wilkinson et al., 2010b). Snakes can also successfully complete both spatial and memory tasks, decreasing the time necessary to find shelter over multiple trials when chemosensory cues are removed (Holtzman, 1998).

Cognitive abilities in amphibians

Maze learning has been used in several species of amphibians. Liu et al. (2016) worked with dart frogs (Dendrobates auratus) in a serial reversal learning task and found that frogs not only used a rule based strategy to learn, but also showed a decrease in errors over time. This is indicative of several processes linked to learning, attention, and other cognitive abilities. Similarly, fire bellied toads (Bombina orientalis) were fairly quick to learn both simple (turn right), reversed (turn left), and complicated mazes – with all subjects reaching criterion in three to four days for simple mazes and the majority reaching criterion in fourteen days for more complicated mazes (Brattstrom, 1990). Tiger salamanders (Ambystoma tigrinum) also learn to navigate t-mazes, starting at one end and making a left or right turn choice for reinforcement. This can be learned as both larvae and adults, and there is evidence that when learned as larvae the ability is retained after metamorphosis (Schwartz and Cogan, 1977). Evidence also supports the idea that salamanders attended to features over geometry in a maze (Crane and Mathis, 2011; Kundey and Phillips, 2019), and that they may have the behavioral flexibility to change the cues they use to more relevant ones as needed (Kundey et al., 2016).

Cognitive abilities in fish

In the aquatic realm, guppies (Poecilia reticulata) have been shown to have some ability to discriminate sets of numbers, in addition to flexibility for how this discrimination is done. They are able to discern relative numerosity (i.e. one group is bigger than another), in addition to absolute numerosity. The fish, when trained to identify groups containing a specific number of objects, not only could identify these groups but could generalize the learning to novel situations; picking out groups with the correct number of items when presented along with groups containing unfamiliar numbers of items (Miletto Petrazzini et al., 2015). Further training showed that the fish could discriminate larger and more similar sets of items (4 vs 5) when training was allowed to go on for a longer period of time. Similar abilities have been evidenced in Picasso triggerfish (Rhinecanthus aculeatus), who have shown the ability to use probabilistic information when making choices (Newport et al., 2024). Cleaner fish (Labroides dimidiatus) are able to both discriminate between types of “clients” and develop efficient orders in which to provide their cleaning service, and to delay gratification by eating non-preferred items when rewarded with a preferred item after a delay (Aellen et al., 2021; Truskanov et al., 2021). Mosquitofish (Gambusia holbrooki) also may be able to delay gratification – they show varying degrees of inhibitory control based on boldness, age and sex (females and older fish showed greater control; Vinogradov et al., 2022).

Cognitive abilities in invertebrates

In a study done with the jumping spider Portia spp., one experimental team placed spiders in a high viewing area where a path to both a “correct area” (which held lures) and an “incorrect area” (which held dead leaves) could be viewed. More than 250 naïve spiders were able to navigate down the high tower and to the correct area based only on what they could see from the original height, while only 15 spiders chose the incorrect path (Cross and Jackson, 2014). Bees (Apis spp), have well-demonstrated cognitive abilities in memory and learning, numerosity, and choice/classification (Gatto et al., 2022; Loukola et al., 2017; Mirwan and Kevan, 2014; von Frisch, 1967), and paper wasps (Polistes spp.) show facial recognition (Tibbetts et al., 2019). Recently, six cuttlefish (Sepia officinalis), demonstrated the ability to delay gratification by refraining from eating a non-preferred food in favor of waiting for a preferred one (Schnell et al., 2021). Even dune snails (Theba pisana) appear capable of some degree of numerical discrimination; they show a preference for larger groups of conspecifics when choosing a place to settle for the day, and can make fine-scale discriminations opting to settle near five others instead of four (Bisazza and Gatto, 2021).

Training in ectotherms

The training programs for ectotherms at VLM are often concerned with training behaviors which make husbandry less stressful. “Shifting”, or moving from one habitat or space to another, is one of the most common behaviors. For many ectotherms, which are often prey species, being cornered and/or picked up is extremely stressful (Williams and Beck, 2021). For amphibians, although they are handled with gloves at VLM, there are additional concerns with regard to contact with delicate and semi-permeable skin (Alworth and Harvey, 2007). An additional benefit, and one observed anecdotally many times by keepers, is an increase in boldness and interest in guests on the part of animals who have participated in training. According to keepers, herps who have had the cognitive challenge of training are more likely to approach the front of their habitats and/or show interest in their surroundings.

Why teach an ectotherm to target? Many animals spend most of their day hiding and in dark, damp spaces. They do not enjoy being handled, nor is it especially good for them. Target training - which involves physically moving toward and often touching a marker such as a ball or dot, often on the end of a stick - allows for voluntary shifting from one end of the habitat to the other or into a transport container can be a major stress reducer. This allows for stress free cleaning, and for the animals to be used as ambassador animals without having to be handled. In addition, keepers can facilitate short training sessions during the day to feed the animals in a place visible to guests, creating an opportunity for education and discussion of the importance of training (Anderson et al., 2003; Visscher et al., 2009; Whitehouse-Tedd et al., 2020).

Ectotherm training sessions at VLM take place with animals’ regular diets. As a result, because many of these animals do not receive food daily, there are generally one to three training sessions per week for herps and three to five training sessions per week for fish.

What follows are some examples of successful training projects and outcomes at VLM.

Training reptiles

Snakes

Training snakes, particularly venomous ones, to shift voluntarily between habitats can be an excellent way to reduce both stress for animals and danger for handlers (Fleming and Skurski, 2013; Williams et al., 2022). One facility used scent on target poles and was table to train false water cobras (Hydrodynastes gigas) to both target to the pole and to follow it, thus allowing additional flexibility during shifting and for separation of the animals during feeding time (Hellmuth et al., 2012). This shifting behavior can be extended for veterinary procedures if the snake is trained to enter a tube small enough to serve as a restraint. Small holes in the tube can allow for access to different parts of the snakes body to, for example, remove retained eye caps or inspect and treat injuries (Fleming and Skurski, 2013).

Edna the garter snake (Thamnophis sirtalis sirtalis) had an eating problem. If allowed to, she would eat very rapidly and consequently regurgitate her food. Keepers at VLM used target training as a DRI (Differential Reinforcement of Incompatible) behavior. Keepers would offer Edna food and then prompt her to target elsewhere in her habitat. She could not move forward and regurgitate at the same time, and the forward motion aided in swallowing. Targeting between bites of food slowed down Edna’s eating process and dramatically decreased regurgitation. In addition, having taught the snake to target solved a secondary problem. Snakes are famously hard to find in their habitats and targeting allowed Edna to be moved into a more visible place for both public viewing and keeper access. Edna was particularly attentive to movement, which was conducive to her target training– a little wiggle of the target pole was all that was needed to get her interested in the target and allow her keepers to capture movement toward it. Over time the target was modified from a small ball to a large bottle cap (similar to that of a sports drink) to make it more visible. Interestingly, the back side of the cap had to be wrapped in black duct tape and angled such that it did not reflect on the glass of the enclosure, as she would sometimes attempt to target the reflection. (This was not the only species trained at VLM in which this reaction was noted – snakes in particular display a tendency to target toward reflections and often need their target poles to be modified in this way.) Edna developed cataracts later in life and her keepers were able to track her visual health using her targeting behavior. A persistent decrease in her eagerness to participate over time and reduced accuracy when approaching the target signified that her vision was steadily failing. These signs of decline were important factors in discussions of Edna’s welfare leading up to the eventual decision to humanely euthanize her before her welfare became poor.

Bolt was a northern black racer (Coluber constrictor constrictor), a species not frequently displayed in zoos. VLM acquisitioned Bolt with the intention for him to become an ambassador animal, however, it soon became clear that his reactivity to movement and high degree of focus were not a good match for the role. Racers also tend to make poor exhibit animals for the same reasons – they are quick to strike at movement, have high energy levels, repeatedly “surf” or pace along the glass in their habitats, and generally appear very anxious. VLM very much wanted to display this fascinating underrepresented species while maintaining good welfare, and found success by implementing a training program prior to placing Bolt on exhibit. Bolt’s keepers used shift, station, and target training (along with his regular diet for reinforcement) to provide him with more exercise and a predictable feeding routine, which helped quell his undesirable behaviors significantly. It is important to note that Bolt’s “glass surfing” behavior initially increased when his training started as he began to anticipate food when people arrived at his habitat, but after many patient sessions of being reinforced only for calm, non-surfing behaviors, the surfing behavior was successfully extinguished and training of other behaviors could proceed. Shift training allowed Bolt to make use of an enclosed exercise pen on a rolling cart that was wheeled next to his habitat behind the scenes. Initially, keepers would assist him gently in shifting over using their hands as a bridge – a process during which Bolt went from being startled by any human touch, to allowing himself to be helped across the small gap between the habitat and exercise pen. After just a handful of sessions, Bolt began to shift to the pen on his own for a mouse or even just the scent of a mouse. The pen gave him opportunities to stretch out, explore, interact with enrichment, and visually inspect new surroundings based on where the exercise pen was wheeled to. Keepers saw a significant increase in calm behavior and improved focus during training sessions the day of or day after Bolt spent time in the exercise pen. Additional exercise and mental stimulation is known to improve snake welfare (Burghardt, 2013) and the experience with Bolt suggests that it is especially critical to highly energetic and highly interactive species such as racers. Racers in the wild tend to “grab and run” with their prey (Secor, 1995). Although natural and unproblematic in an open area, this behavior can manifest itself as hasty food-striking; rapid locomotion like this (“bolting”) when in an enclosed space poses a risk for injury to the snake or keeper, in addition to a potential for animal escape. To reduce these risks, Bolt was trained to station in order to encourage calm behavior and reduce frantic movement in anticipation of food. While station training, Bolt was only fed at an upside down yellow bowl, first by placing food on it and later by using tongs to hold food just above it. Once Bolt knew where his prey would appear regularly, he stayed near the station and took food with calmer and more precise striking behavior. He also showed a reduction in the duration and intensity of “bolting” after acquiring food. Station training also created a convenient space where Bolt could wait for other training cues and engage with his trainers. When Bolt was finally ready to move to exhibit, keepers witnessed mild regression in his trained behaviors and visibility. This included a tendency to hide in an off-exhibit box connected to the top of the enclosure. When inside the box, Bolt could not see his target, so keepers added an auditory component by attaching a rattle ball to the end of the target pole. This helped draw Bolt’s attention and led to more consistent targeting on exhibit. Bolt has become a highly effective exhibit animal that is frequently visible and calmly watches or interacts with guests as they observe him through the glass. The northern black racer is a species which does not often get to be in the spotlight, and VLM is proud to be one of few institutions to highlight this intelligent species with a high capacity for learning.

Lastly, VLM trained a timber rattlesnake (Crotalus horridus) to voluntarily shift through a tube into an off-exhibit holding box for feeding. This has obvious implications for the safety of his keepers while cleaning the exhibit. This technique is particularly useful in venomous snakes and has been used successfully in other facilities (Williams et al., 2022), however, the rattlesnake at VLM was further trained to shift for non-food reinforcers such as scents and enrichment, which is far less common.

Lizards

Similar to snakes, large lizards and their keepers benefit from voluntary crate training because of both a reduction in stress and a decrease in potential for physical harm to both humans and animals (Fleming and Skurski, 2013). Husbandry training is particularly useful in the case of Komodo Dragons (Varanus komodoensis), and many facilities clip nails, draw blood, provide vaccinations, and do ultrasounds with voluntary training. Komodo dragons have even been trained for application of a mask for anesthesia (Gibson, 2020; Hellmuth et al., 2012).

More than one species of lizard has been trained at VLM – including gila monsters (Heloderma suspectum) and bearded dragons (Pogona spp.) – but the most rapid learner was Rocket the broadheaded skink (Plestiodon laticeps). Rocket had excellent potential to be an ambassador animal upon his arrival at VLM. He took to target training quickly, learning to move to a small stick with a ball at the end from distances of up to two feet and from different elevations when using climbing structures – all within the first two training sessions. Keepers began by wiggling reinforcement next to the ball on the stick, but Rocket quickly learned to associate the target with reinforcement, making the immediate presence of the food item unnecessary. His keepers subsequently trained him to shift into a transport container for use as an ambassador. First, keepers trained him to target reliably through a standalone bulkhead placed in his enclosure. Next, the bulkhead was installed in the side of a plastic box with the lid detached, and Rocket was targeted through the bulkhead into the box. Eventually keepers were able to shift Rocket into the box with the lid attached by targeting and reinforcing through a hole in the lid. Training for this behavior took approximately 4 months, and although Rocket ultimately did not become an ambassador due to external factors, keepers were able to use this shift training to facilitate deep cleans of his enclosure and to move him into the exhibit where he resides today.

Crocodilians

Crate and shift training is advantageous in crocodilians as much as it is with large lizards and venomous snakes (Gibson, 2020). However, in addition to these behaviors, zoos have reported training crocodilians to accept restraint systems for veterinary care, for voluntary weighing and blood draws (Hellmuth et al., 2012), and to participate in harness walks (Gibson, 2020).

Newt the American alligator (Alligator mississippiensis) holds the record for the largest behavioral repertoire and most successful training history of all ectotherms at the Virginia Living Museum. Among her established voluntary behaviors are targeting, turning around, backing up, entering and exiting the water, moving onto a scale, voluntarily entering a transport carrier, allowing extensive tactile manipulation below the neck, and placing her snout into a specially made PVC frame to facilitate safe application of antibiotic ointment on nose scrapes. Newt’s training began when she was just over a month old and she took to it so well that she began to ignore diets which were placed on the ground in front of her or in enrichment, only eating when food was presented as reinforcement during training sessions. To help Newt remember that she could acquire food independently, keepers used her established target training to teach her to approach free food and foraging enrichment devices. Via approximations, Newt learned to pick up free food and manipulate enrichment devices to obtain food on her own. Newt targets to both a traditional target pole and to a light from a laser pointer which provides flexibility when targeting her across her partially aquatic exhibit. The trusting relationship built between Newt and her keepers during training saved her from more than one potentially life-threatening situation in which she swallowed guests’ belongings which had fallen into her habitat – first a child’s shoe and later a stuffed toy attached to a pacifier. Both times Newt targeted out of the water without hesitation and allowed her trainers and veterinary staff to restrain her and manually retrieve the items from inside her stomach. Perhaps most impressively of all, after both incidents Newt returned to training as if nothing out of the ordinary had happened.

Training amphibians

In general, amphibians can be difficult to display and care for due to their often timid nature and camouflage coloration. Training can be a stress free way to circumvent these natural tendances. For example, dendrobatid frogs have been trained to hop into a small tube or box upon hearing a clicker. Once in the box, fruit flies are released. This provides both reinforcement for the shifting behavior and allows keepers to monitor the diet of individual frogs; something which may be difficulty to do in a species such as this, which live communally and prefer habitats with a large degree of plant cover (Fleming and Skurski, 2013).

Cosmo is a fairly timid Cuban tree frog (Osteopilus septenrionalis) previously housed with a conspecific who was far less timid. When the two were separated, Cosmo continued to fear humans and would exhibit fear responses when interacting with them, even if keepers attempted to carry him gently in cupped (i.e. enclosed) hands. To address this concern, staff used careful training via desensitization to hands. This led to a visible increase in inclination to be picked up and a willingness to be hand fed his entire diet. Cosmo will now sit in a keeper’s hand (thus avoiding having to be cupped) and even investigate keeper fingers without signs of a stress or fear response.

A group of Fowler’s toads (Anaxyrus fowleri) learned to use a shifting platform to move into a transport container. Keepers used a modified silicone colander with one of the sides cut away so the toads could readily access the inside. This group of toads was generally very reactive to humans, jumping at the habitat door in anticipation of feeding, and the shifting platform was conceived partially to eliminate this behavior, which could be dangerous for toads who accidentally jumped out of the habitat. Initially, the platform was placed in the toads habitat and animals who jumped on it were rewarded with a desired food item. After the toads learned to associate the platform with reinforcement, the enclosure was opened and the platform held at the edge of the enclosure. The toads transitioned fairly easily to jumping onto the platform in this location, and were also reinforced. The platform was lined with red tape to make it more visible to the toads from a distance, and to create an additional visual stimulus for toads already on the platform. The directing of the toads’ attention to the platform was due to an anecdotal observation by the keepers that toads which focused on humans while on the platform were more likely to jump off of it than toads which focused on the platform itself.

Keepers at VLM have used operant conditioning to target train a tiger salamander (Ambystoma tigrinum) named Chompers. Keepers originally used a large (three to four inches in diameter) circle as a target for Chompers due the assumption that his eyesight was limited. However, bigger was not better – it was scarier. Chompers froze at the sight of the large object. Keepers corrected this by gradually cutting down the size of the target and eventually converting it to a small ball on a bamboo stick. With patience and consistency from his trainers, Chompers learned to place his snout on the ball to complete the target behavior. The behavior was trained via capturing and shaping – keepers would place the target near Chompers and reinforce when Chompers glanced at it. When he was turning his attention to the target regularly, reinforcement became contingent upon Chompers moving toward the target instead of just looking at it. After this, Chompers was required to make contact with the target to receive reinforcement. During training, Chompers’ keepers noticed the development of what might be called superstitious behavior – a behavior that the animal incorrectly believes is necessary for reinforcement. Chompers always made his final contact with the target with a lunge – he walked to a specific distance from the target and stopped, did a predatory “butt wiggle”, and lunged to close the gap. In other animals this lunging behavior may be undesirable, but given that it had no negative consequences for Chompers, it was not considered an issue. Following the implementation of training, keepers noticed a change in Chompers’ overall behavior. The salamander became quicker to come out of his hiding spots and investigate changes in the habitat around him. He became willing to travel increasing distances to the target pole – up to two full body lengths– and he began displaying what could be called (for lack of a better term) “rearing” behavior, which looks similar to the push up behavior seen in male lizard mating displays. In Chompers’ case, this behavior is believed to be reminiscent of the rearing or periscoping behavior of small mammals, which is used to look around the immediate environment. It is frequently used as a measurement of boldness (Perals et al., 2017; Walsh and Cummins, 1976), and Chompers’ behavior suggests that he became more interested in his immediate surroundings after he began training. This behavior became extremely convenient, as Chompers was on the heavier side and in need of more exercise. Keepers began to capture this rearing behavior as a high target or “push up”, creating a mini salamander calisthenics routine to aid in his weight loss, improve his stretching ability, and encourage him to climb onto taller perches (Figure 1).

In addition to Chompers, two other cave dwelling amphibians who are often difficult for guests to see have been trained to come out of their hiding spots to be fed. A greater siren (Siren lacertina) and a hellbender (Cryptobranchus alleganiensis) emerge in response to an auditory cue (tapping on rocks) at feeding time.

Training fish

In fish, target and station training is most often used during feeding time. Fish which habitually eat too much or too little can be trained to eat at a station away from tank mates. Fish who can be aggressive toward conspecifics or keepers can be trained to touch their rostrum to a target before being fed, thus creating a calmer feeding environment. This is particularly helpful with large rays or sharks who are often fed by hand or encouraged to eat where they are visible to guests (Corwin, 2012). Large fish such as groupers (Epinephelus itajara) can be trained to swim to hands and allow hand feeding, examination of skin, and even to tolerate being lifted out of the water for a skin scrape. One facility reports using broken shells rubbed over the bodies of fish as a successful (secondary) tactile reinforcer (Davis, 2010).

At the Virginia Living Museum, Voldemoray, a green moray eel (Gymnothorax funebris) was trained to put his head into a length of PVC pipe to promote safe feeding. Moray eels, which often exhibit open-mouthed “scanning” behavior during feeding to help make contact with food, also have fast reflexes and a painful bite. This behavioral combination can make feeding sessions difficult (or even dangerous) for keepers. To address this concern, VLM staff trained this eel to station – a behavior in which an animal goes to a designated place and calmly stays there. In this case, the eel’s station was located such that its head was within a PVC pipe held at the water’s surface. In approximately five sessions Voldemoray learned that food would only be offered if he placed his head inside the tube, and soon after he began keeping his head inside until the session ended and the tube was removed. Once established, this tube stationing behavior allowed VLM keepers to examine Voldemoray’s head and mouth, and veterinarians to take swabs without the need to restrain him. When “Voldy” outgrew his tube, he retained the manners he learned from it and stationed calmly near it during feeding. The generalization seen in this training is an excellent demonstration of how training animals who are reactive or nervous often has the added benefit of reducing general arousal in daily behavior (Laule and Desmond, 1992, Laule and Desmond, 1998; Westlund, 2014).

Training invertebrates

Although behavioral neuroscience has long relied on operant conditioning models of fruit flies (Drosophila melanogaster), sea slugs (Aplysia californica), and pond snails (Lymnaea stagnalis; Brembs, 2003), we have located few published examples of operant conditioning in invertebrates in zoo or husbandry situations. Experimental evidence beyond the species typically used in laboratories has shown that crickets (Gryllus pennsylvanicus) can be trained to navigate a Y-maze (Cyr et al., 2024) and lobsters (Homarus americanus) can be trained to grip a sensor bar with a varying degree of force (Tomina and Takahata, 2010). Honey bee (Apis mellifera) researchers found that, while the majority of bees eventually learned to push a cap aside to uncover a food reward, it was possible to increase the speed of learning by shaping the behavior via small approximations (a technique commonly used in training via operant conditioning). For example, a cap could be set with a small gap underneath, causing the bee to inadvertently push it aside while reaching for the reward (Abramson et al., 2016). The few academic published accounts of zoo training in invertebrates which we have been able to locate where that of a Bahamaian seastar (Oreaster reticulates) which was trained to eat at the surface of its habitat so that tank-mates would not steal it’s diet, and a giant Pacific octopus (Enteroctopus dofleini) who was trained to come to the surface of its tank, crawl into a container for weighing, and then return to its tank (Corwin, 2012).

Pastel is a blue crab (Callinectes sapidus) who receives laser treatments for shell issues. Her keepers had been using a net to remove her from her habitat for treatment. Because netting caused stress for Pastel, she is learning to use a target to enter a basket, which is then used to lift her up for her treatment. Crabs do not have good depth perception, and Pastel was initially reluctant to approach the target from more than three inches away. A diamond shaped pattern on the target provided the necessary visual clarity and currently Pastel will target from as far as nine inches and at multiple angles. The next approximation in Pastel’s training will be the addition of a platform to the target, creating a step up, which will eventually be replaced by the basket.

Enrichment for ectotherms

Environmental enrichment as a tool in zoos began with the pioneering work of Hal Markowitz (Markowitz, 1982; Markowitz and Stevens, 1978) and has steadily become more widespread in the nearly 50 years since. It is now a staple, particularly for zoo-housed birds and mammals, and an enrichment program is required for accreditation by many professional organizations (Association of Zoos and Aquariums, 2025; Heidenreich et al., 2023). Although often used to mitigate self-mutilation or stereotyped behavior in mammals and birds (Mason et al., 2007; Shepherdson et al., 1998), enrichment has a wide variety of applications to all animals in professional care. Enriched animals show signs of increased welfare and cognitive stimulation, display more natural behaviors, and increase public interest and engagement (Alligood and Leighty, 2015; Fernandez and Martin, 2021; Mason et al., 2007; Quirke and O’Riordan, 2013; Riley and Rose, 2020).

However, surveys of peer reviewed and grey (professional) literature calculate that in only three to less than one percent of enrichment studies is the subject an amphibian, fish, or invertebrate (Alligood and Leighty, 2015; Brereton and Rose, 2022; de Azevedo et al., 2007; Kresnye et al., 2022).

Enrichment at VLM is meant to encourage the natural behaviors that species would normally exhibit, in addition to mimicking structures or scenarios that they might encounter in the wild. The resultant behaviors sometimes appear to be play-like, and examining them from this perspective is valuable. Burghardt uses five criteria in order to classify behavior as play (Burghardt, 2005, Burghardt, 2010, Burghardt, 2014), and these are generally accepted by most comparative cognition researchers. Play should be 1) non-functional in the context it is seen in, 2) spontaneous and/or rewarding, 3) different from other useful behaviors – exaggerated in some way, for example, 4) repeated, but not repetitive in a stereotypic or anxious manner, and 5) be seen during times of low stress. These criteria become very useful when working with ectotherms, whose behavior and reactions are not as easily recognizable or interpretable as that of mammals and birds.

The next section discusses examples of enrichment at VLM, supplemented by specifics in the accompanying tables and Supplementary Material.

Enrichment in reptiles

There is relatively little information on environmental enrichment in captive snakes and reptiles in general (Alligood and Leighty, 2015; Burghardt, 2013; Hoehfurtner et al., 2021; Melfi, 2009), but what has been published shows very positive results (for an excellent review, see Howard and Freeman, 2022). This includes a noted increase in behavioral diversity – a distinct benefit for facilities looking to interest guests in reptiles (Uwakaneme et al., 2004). When provided with the option, snakes strongly preferred an enriched enclosure, showed changes in behavior indicative of improved welfare and increased problem solving abilities, and displayed a larger variety of species typical behaviors (Almli and Burghardt, 2006; Burghardt et al., 1996; Hoehfurtner et al., 2021; Hollandt et al., 2021; Nagabaskaran et al., 2022). Like snakes, turtles have also shown a distinct preference for enriched environments when living under professional care (Case et al., 2005). Nile soft-shelled turtles (Trionyx triunguis) in particular have been known to play with floating objects and even engage in tug-of-war behaviors with keepers and their hoses (Burghardt, 2015). Lizards have also displayed cognitive problem-solving abilities – in one test, varanid monitors (Varanus albigularis) quickly learned to manipulate a clear feeding tube with hinged doors to obtain prey, and maintained correct responding between trials, while decreasing incorrect or futile responses (Manrod et al., 2008). Varanids in particular appear particularly suited to enrichment due to their cognitive abilities and interest in play (Cooper et al., 2019, Cooper et al., 2020; Howard and Freeman, 2022). They have been observed engaging in play in a variety of situations with and without objects, in the wild and under professional care.

Many of the snakes, lizards, and turtles at VLM interact with “snuffle mats” made of felt strands tied together. Often these are used to hide treats for dogs, however, the reptiles appear to benefit from both the tactile quality and the ability to hide within the mat (see Table 1, row 1). Ball pits with plastic balls are often used as “entry level” reptile enrichment – they are large, colorful, shiny, and easy for reptiles to move out of the way, and because the balls float they can even be used with aquatic turtles (see Table 1, row 2). Balls encourage reptiles to examine new things in their environment, and to maneuver and hide amongst them. Many reptiles at VLM will readily use pocket-shaped items such as oven mitts, bags, and boots as hiding places, and will explore boxes filled with items of various textures (referred to as “Boxes O’ Stuff”). Snakes use wooden peg boards and fire hose chains for climbing enrichment (see Table 1, row 3), which provide tactile stimulation, exercise, and an excellent opportunity for guests to observe how snakes move. Keepers offer seasonal fresh produce, herbs, whole flowers, shed mammal fur, and molted feathers as sensory enrichment, with turkey feathers being a particular favorite (see Table 1, row 4). Newt the alligator is particularly fond of practicing “death rolls” with her toys (see Table 1, row 5). Reptiles of all species at VLM have proven to be receptive to a range of enrichment types, with several individuals interacting more enthusiastically than some of VLM’s birds and mammals.

Table 1. Reptiles.

Enrichment in amphibians

Our searches did not find any published literature detailing enrichment in amphibians. However, VLM has found them to be receptive to many of the same types of enrichment as reptiles, provided that materials are safe and non-abrasive for their delicate skin.

Frogs and salamanders make use of foraging devices such as puzzle feeders and silicon foraging mats (see Table 2, rows 1 and 2), with frogs showing enough persistence to use feeders with variable reinforcement. Fully aquatic amphibians such as newts benefit from sinking foraging items such as small wheels filled with gel diets (see Table 2, row 3). Enrichment can be highly effective for eliciting other natural behaviors in amphibians, for example, climbing structures and hanging water cups can encourage climbing and soaking in arboreal species such as tree frogs (see Table 2, row 4). VLM’s pair of eastern hellbenders (Cryptobranchus alleganiensis) have been especially receptive to enrichment and participate in a variety of activities such as foraging along chain feeders with novel food, investigating when the water pump in their exhibit is set to different flow strengths, and perhaps most charismatically, actively rubbing their bodies across artificial turf to aid in shedding (see Table 2, row 5).

Table 2. Amphibians.

Enrichment in fish

Applying Burghardt’s criteria for play provides several examples of behaviors which could be classified as play in fish. One of the most common is “batting” of objects such as balls or weighted thermometers which return to upright positions when knocked over (Burghardt, 2015). Habitat change can also provide significant enrichment - both as a positive novel stimulus and a negative stressor (something keepers should be attuned to). For example, research into habitat substrate color with gilthead seabream (Sparus aurata) shows that blue or red-brown substrate (as opposed to green) leads to a less stressful social organization and a decrease in aggressive behaviors, which can be empirically quantified in the measurement of neurotransmitters (Batzina et al., 2014; Batzina and Karakatsouli, 2012). In addition, structural complexity in tanks decreased aggression and resource monopolization, and increased proliferation of brain cells in zebra danios (Danio rerio, Basquill and Grant, 1998; von Krogh et al., 2010). Structural complexity has also been shown to impact welfare in a variety of other fish species, although there is significant variation by species and measurement parameter (Näslund and Johnsson, 2016).

At VLM, keepers place diets in puzzle balls or attach them to plastic chains and weave them in and out of structures in the habitat to encourage fish to forage and problem solve. Species as diverse as short nosed sturgeon (Acipenser brevirostrum), coral dwelling fish (i.e. Tangs, Acanthuridae spp.), greater siren (Siren lacertina; a paedomorphic amphibian), and American eels (Anguilla rostrata) make use of puzzle balls, floating chains, and PVC pipe puzzles (see Table 3, row 1 and Supplementary Material, Videos 2, 3). Seahorses (Hippocampus spp) and pipefish (Syngnathinae spp) can be enriched with a variety of materials and textures to use as anchors in their habitats (see Table 3, rows 2 and 3). Seahorses will also solve puzzles made up of balls and shells in which the ball must be pushed away to reveal reinforcement in the shell (see Supplementary Material, Video 4). Often seen in marine mammal enrichment, PVC pipes with holes for food are excellent enrichment for fish and shark species as well, creating similar foraging challenges (see Table 3, row 4). Chain catsharks (Scyliorhinus retifer), a small benthic shark species, often make use of elevated platforms attached to the side of the habitat or floating hammocks made of pool noodles and fire hose (see Table 3, row 5). Juvenile oyster toadfish (Opsanus tau), gobies (Gobiidae spp.) and blennies (Blennioidei spp.) shelter in and display natural territorial behaviors over environmental enrichment such as PVC tubes and 3D printed hides. Structure-hugging species such as Atlantic tripletail (Lobotes surinamensis) and orange filefish (Aluterus schoepfii) weave in and out of floating driftwood, vegetation, and fake seaweed made of recycled canvas banners attached to buoys. Virginia native flounder species (Paralichthyidae spp.) sample sunken bins containing substrates of different textures and colors, and change their pigmentation to match their surroundings. Young Atlantic spadefish (Chaetodipterus faber) and other pelagic fishes consistently swim into the flow of adjustable powerheads and nibble at bubbles produced by sprinklers. Changes to habitat are often the most successful non-food-based enrichment opportunities for fish at VLM.

Table 3. Fish.

Enrichment in invertebrates

By now, many people are familiar with the interactive behaviors of octopuses. Videos of the cephalopods solving problems, exploring toys, and attempting to escape their habitats have appeared on a variety of internet platforms (Finn et al., 2009; Kuba et al., 2014; Richter et al., 2016; Zylinski, 2015). At VLM, we have had similar success with enrichment for many types of invertebrates.

For example, Cupcake the American lobster (Homarus americanus) loved dog toys. Keepers would place her diet in rubber puzzle balls designed for pets to provide her with food acquisition challenges (see Table 4, row 1). Initially, the toys were simple and merely required her to reach in and pull food out, but over time keepers introduced her to progressively more challenging items that required more complex sequences of behaviors. At rest, a slit in one type of rubber ball was too narrow for Cupcake to reach in to access the food. Upon receiving the ball for the first time, she explored it with her pedipalps, after which she transferred it to her front claws and rolled it as she had learned to do with simpler items. Over the course of the day, Cupcake learned through trial and error to use her large claws to squeeze the toy perpendicular to the slit to widen the hole, and to hold the squeezed position long enough to retrieve the reward with her smaller claws (see Supplementary Material, Videos 1a-e). From then on Cupcake always performed the squeeze and grab motions when offered that toy – rarely attempting to remove the treat from the toy without first enlarging the slit. It was evident that she retained information learned during previous foraging challenges and used it to inform her behavioral choices when posed with similar challenges in the future. As a group, decapods seem to be particularly adept at foraging enrichment which requires manipulation of an object, or to being housed in a complex and changing environment (see Table 4, rows 2 and 3). VLM staff have witnessed similar complex manipulation and interaction in various species of crabs, crayfish, and shrimp.

Table 4. Invertebrates.

Enrichment for insects at VLM tends to be in the form of new foods or new environmental features. Insects are given a variety of substrates, platforms, and places to climb in their habitats. For example, decomposers such as cave crickets, millipedes, snails, and isopods (Porcellio spp., Polygryidae spp., Narceus spp.) will gather on snake skin to explore new scents and textures. Water bugs (Lethocerus americanus) in particular are attracted to plastic mesh, especially when it is layered together or made into tubes and placed in water (see Table 4, row 4). Spiders (grouped here but not insects) such as tarantulas interact with large sheets of crinkly paper – walking back and forth across the surface and then hiding under the layers – or sit in hides suction cupped to the walls of the habitat.

Echinoderms such as sea stars (Asteroidea spp) and sea urchins (Echinoidea spp) are extremely receptive to foraging enrichment. Sea urchins will forage in “slow feeder” dog bowls, on different textures, and on a variety of novel foods (see Table 4, row 5). Recently, keepers provided sea urchins with brussels sprouts for the first time and had to break up “fights” as the urchins stole the sprouts back and forth from each other. Sea stars will pursue and hang on to floating enrichment containing their diets, sometimes piling on in such large numbers that the items sink to the bottom of the tank (see Table 4, row 6).

Invertebrates which forage by prying open food are enriched by live clams, but this behavior can also be simulated with clam strips in rubber balls or plastic easter eggs (see Table 4, rows 7 and 8). Predictably, decapods benefit from live or simulated clams, but perhaps more notable is the benefit to predatory mollusks like knobbed whelks (Busycon carica) which have the opportunity to engage in natural hunting behavior - using their muscular foot to envelop and pry open these items.

Conclusion

Much of what happens in zoos, aquariums, and museums is done on a far smaller and less controlled scale than in an experimental laboratory environment. We make no attempt to dismiss this shortcoming. However, we have been thrilled to see that in recent years more and more of the academic research community has come to see the value of what can be learned from animals under professional care in zoological institutions, despite the lack of a strict empirical environment (Alligood et al., 2017; Burghardt, 2016; Hopper, 2017; Margulis and Westhus, 2008).

We at the Virginia Living Museum in Newport News, Virginia strive to provide training and enrichment for all living residents in our care whenever possible, regardless of species. We believe our success to be noteworthy, and we hope that our work has demonstrated to the scientific community that ectotherms possess far more impressive cognitive abilities than is typically assumed. Many of our species of invertebrates, fishes, reptiles, and amphibians which receive training and enrichment do not normally receive purposeful cognitive stimulation under human care; highlighting our successful results is therefore vital to focus attention on cognition in these animals. A single animal at VLM showing the ability to respond to training or solve an enrichment puzzle may only provide a sample size of one, however, that animal is fundamentally similar to others of its species and can provide a valuable model for training and enrichment (J. C. Kaufman and Kaufman, 2016). It is our hope that by increasing awareness of the success of cognitive training and enrichment in zoos, aquariums, and museums, academic researchers will be inspired to explore the inherent “why” and “how” questions behind these examples.

In a practical sense, our goal with presenting these examples is also to encourage other facilities to expand their ectotherm training and enrichment programs (see Table 5). Keepers are a valuable asset for understanding the personalities of their charges, and tailoring enrichment and training to foster individual strengths in ectotherms will strengthen the public’s positive perception of these species. Caretakers do not need to shy away from cognitive based enrichment and training, and will be excited by outcomes similar to what would be expected when working with more charismatic species. This in turn will increase guest enthusiasm and move us further end goal of conservation – particularly with some less well known or loved species.

Table 5. Summary of enrichment and training provided, behavioral results, and recommendations.

Note: for each species, information involving training is in gray shaded cells while enrichment cells are not shaded.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fetho.2026.1799441/full#supplementary-material

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Supplementary material Supplementary material at the source

Footnotes

Edited by: Matteo Chincarini, University of Teramo, Italy

Reviewed by: Scott Nunes, University of San Francisco, United States

Yann Henaut, ECOSUR Conservation and Biodiversity, Mexico