Both turtles are permanent residents of the aquarium for a poignant reason. They sustained boat-related injuries that left them with positive buoyancy syndrome, a condition in which trapped gas or lung damage makes it difficult to dive and forage normally. Treated with weights affixed to their shells, they can reach the bottom of their tank, but they spend most of their time in the upper water column. Because they are non-releasable, they will live out their lives under human care, which makes their long-term welfare a genuine scientific and ethical concern. The research team, led by Isabella Barba and Sandra M. Troxell-Smith of Oakland University alongside colleagues from the aquarium, saw in these two co-housed males a rare opportunity to study the social dynamics of captive male green sea turtles, a topic that remains surprisingly understudied.

The experiment was elegantly simple. Over a three-month period from January to April 2026, trained observers who were not involved in the animals’ daily care conducted 108 thirty-minute observations, three times per week. Each observation day included three sessions: one before feeding, one during feeding, and one after. The turtles experienced three feeding conditions in rotating two-week blocks. The baseline condition was their standard target feeding, in which each turtle is fed individually with tongs at a separate, controlled location. The first enrichment was a depth pole, a PVC pole roughly 122 centimeters tall, perforated with holes and stuffed with lettuce, held suspended in the tank by an aquarist and moved side to side so the turtles had to swim up and chase their food. The second enrichment was a pair of floating ice blocks, each containing about 250 grams of frozen vegetables, dropped into the tank so each turtle could engage with one individually.

Observers used two complementary behavioral sampling techniques. Scan sampling, recorded at two-minute intervals, documented species-typical behaviors such as swimming, pacing, breathing, resting, sleeping, feeding, and foraging, along with each turtle’s precise location in the tank using the ZooMonitor program’s heat-mapping function. All-occurrence sampling captured every social interaction as it happened, from approaches and displacements to chases, rams, snaps, and bites, as well as neutral behaviors like gentle contact and close proximity without aggression. In total, the team collected 3,350 raw data points. Inter-observer reliability, tested across six practice sessions, averaged over 95 percent agreement for depth in the tank and interval behaviors, and over 82 percent for space use, giving the findings a solid methodological footing.

The statistical analysis was appropriately rigorous for a two-animal case study. The researchers built linear mixed models with session as a random effect and fixed effects of subject, feeding condition, phase, and time of day, including all two-way interactions, comparing full models against null and simpler models using corrected AIC. Chi-square analyses tested whether behaviors were distributed evenly across the tank’s horizontal and vertical coordinates. The results were striking in one key respect: behavioral variation was driven primarily by feeding phase and time of day rather than by enrichment condition. In other words, what the turtles did depended far more on when food was coming than on how it was delivered.

That feeding-phase rhythm played out in a clear, adaptive sequence. Before feeding, the turtles paced significantly more, swimming back and forth in front of the target feeding area, a classic pattern of anticipatory behavior seen in many captive animals responding to predictable feeding schedules. During feeding, pacing dropped sharply and active engagement took over. After feeding, resting and sleeping surged, with the turtles spending dramatically more time motionless at the bottom of the tank. The researchers interpret this as an energy and digestion cycle consistent with patterns seen in other captive and wild animals. Notably, pacing before feeding is not necessarily a sign of stress or frustration. Since food was regularly delivered as anticipated, the authors suggest it may reflect positive anticipation of imminent food, though they also note that introducing less predictable feeding times could be additionally beneficial.

Individual differences between the two turtles were equally clear. Benson spent significantly more time swimming and breathing, while Carr spent more time at the bottom of the tank and engaged in more resting and sleeping behavior. Both turtles spent more time in the upper portion of the water column, likely influenced in part by their positive buoyancy syndrome. Feeding itself was more frequent in morning sessions than afternoon ones, suggesting a diurnal rhythm and preference. Interestingly, the turtles fed more from the floating ice blocks in the morning than in the afternoon, while the depth pole showed no such time-of-day effect, hinting that different types of enrichment may not benefit turtles equally at different times of day, a practical insight for aquariums structuring feeding activities.

Now for the finding that matters most for aquarium management. Agonistic interactions between the two males were overall very infrequent, which the authors take as evidence that current husbandry practices, including the large, structurally complex tank with rock formations, ledges, and multiple feeding stations, are largely effective at maintaining stable cohabitation. But there was one significant exception. Adverse behaviors were distributed unevenly across phases only for the depth pole condition, occurring most frequently after feeding, with ten adverse events post-feeding compared to just two pre-feeding and three during. No such pattern appeared after baseline or floating ice block feedings. The researchers suspect the reason is spatial: during target feeding, the turtles were fed at separate locations, and during the ice block condition, two blocks were provided, allowing each turtle to separate. The depth pole, by contrast, was a single item both turtles had to interact with simultaneously, potentially creating competition. The authors suggest that using one depth pole per animal, or offering enrichments that ensure food is available in multiple locations, could reduce conflict further. They also note that both enrichments lasted only two to five minutes, and the ice blocks lasted longer on average, which may partly explain the difference.

There was also a quietly positive side effect of the enrichment feedings. Under both enrichment conditions, the turtles were less likely to be out of sight following feeding compared to baseline, suggesting the enrichments engaged them, kept them active, and made them more visible to aquarium visitors. Since enrichment activities are often designed precisely to increase space use and animal visibility, the authors consider the enrichments successful in that regard, even though they produced no additional behavioral benefits above and beyond the standard feeding method. Importantly, the lack of strong condition effects does not mean the enrichment was ineffective; it means the enrichments caused no ill effects and offered no benefits beyond what target feeding already provided, which the authors view as a positive sign that all three feeding conditions supported appropriate behavioral changes.

The study has obvious limitations the authors acknowledge candidly: a sample size of two individuals, a relatively short three-month observation window, and observation periods isolated to pre-scheduled feeding times. A technical failure also meant the space in which aggressive behaviors occurred was not recorded, something future studies should correct. Still, the findings contribute to a very limited literature on male sea turtle cohabitation and carry a hopeful broader message. Under appropriate housing conditions, with effective feeding strategies, sufficient space, and environmental complexity, adult male green sea turtles can be co-housed with relatively low levels of conflict. For the many captive sea turtles that remain in human care for life due to non-releasable medical conditions, that message matters enormously. Evidence-based refinements to enrichment design, feeding schedules, and habitat structure do not just improve the lives of individual animals; they support conservation, rehabilitation efforts, and public education. And for Benson and Carr, two turtles who cannot return to the ocean, the science says they are doing just fine together, one lettuce-filled pole and one frozen vegetable block at a time.

Subject of Research: Lettuce get along: effects of feeding enrichment on green sea turtle behavior and interactions

Article Title: Lettuce get along: effects of feeding enrichment on green sea turtle behavior and interactions

Article References: Barba, I., Vonk, J., Zook, L., Averitt, S., & Troxell-Smith, S. M. (2026). Lettuce get along: effects of feeding enrichment on green sea turtle behavior and interactions. Frontiers in Ethology, 5, Article 1886897. https://doi.org/10.3389/fetho.2026.1886897

Image Credits: AI Generated

DOI: 10.3389/fetho.2026.1886897

Keywords: Lettuce, along, effects, feeding, enrichment, green, turtle, behavior, interactions, scientific research