Captive rearing reduces the sensitivity of Acartia tonsa copepods to predator cues

Captive populations are often subject to different selective pressures than their wild counterparts, which could result in trait differences between these populations. This study investigates the effect of captive rearing on the swimming behavior and escape responses of Acartia tonsa, a species of marine copepod zooplankton that use hydromechanical signal detection to aid in finding food, locating mates, and avoiding predation. As captive populations of A. tonsa experience reduced interspecific predation and higher population densities compared to wild populations, it was hypothesized that these differences may drive adaptive evolution of swimming behavior in this species. Several components of routine swimming were compared (swimming speed, number of hops, distance of hops, frequency of hops) for groups of captive-reared and wild-caught A. tonsa, revealing that wild-caught copepods swim faster and hop more frequently than captive-reared copepods. However, when the escape responses of the captive-reared and wild-caught populations were compared using an artificial predator mimic, no significant differences were found in the number of sequential hops performed during the escape response, the maximum velocity of the response, or the total distance traveled during the response. Although the escape responses performed by the captive-reared and wild-caught copepods were similar, the captive-reared individuals often showed no response to the artificial predator mimic (34% of individuals responded to the predator mimic), whereas wild individuals almost always showed a response (96% responded). This suggests that captive rearing may have resulted in reduced sensitivity to hydromechanical signals in captive copepods compared to wild copepods, as responding to these signals in a predator-free captive environment would impose an unnecessary energy cost. This study offers new insight into how captive-rearing may impact copepod populations and provides evidence of how predator-driven evolution and density dependent selection may influence the behavior of copepod species.

Methods
Wild A. tonsa were collected from Ballast Point Park, Tampa Bay using a 30 cm diameter, 150 µm mesh zooplankton net. Captive copepods were obtained from Aquaculture Nursery Farms, based in Charlotte County, Florida. 

Copepod swimming behavior was recorded using an Edgertronic high speed imaging camera with a Nikon AF micro NIKKOR 105 mm lens. Each water sample contained approximately 100 copepods. Videos of the copepods swimming were recorded for 9 seconds at a resolution of 1280 x 1024 pixels, with 6400 ISO sensitivity, 1/10000 shutter speed, and 60 frames per second. 14 videos were recorded for the captive populations and 30 videos were recorded for the wild populations.

To collect the escape response data, an artificial predator mimic was generated using a piezoelectric actuator. This was generated using a BK Precision 4030 10MHz pulse generator. The piezoelectric actuator was used to rapidly move a narrow aluminum rod in a vertical orientation to create a strong, radiating hydromechanical signal to trigger escape behavior in copepods (methodology from Gemmell et al., 2013). Video recordings spanned immediately before and after this predator cue was administered, with a pre trigger frame rate of 500 frames per second and a post trigger rate of 500 frames per second. 31 videos were recorded for the captive populations and 21 videos were recorded for the wild populations, with approximately 100 copepods per water sample.

The video recordings were analyzed using ImageJ (version 1.54i) image analysis software, using the tracking plugin to track and record the movement of the copepods. The video was analyzed frame by frame, with each frame representing 0.02 seconds of the total recording. 200 copepods were randomly selected for tracking (50 per group – captive routine swimming, wild routine swimming, captive escape response, wild escape response). For the routine swimming videos, a random coordinate generator was used to select the copepods that were tracked (the copepod closest to the random coordinates was tracked), and for the escape response data the copepods within several body lengths to the piezoelectric needle were selected to be tracked. As we were working with two-dimensional video recordings, we only tracked copepods that remained clear and in focus (copepod swimming is three-dimensional, so loss of focus indicates that the copepod has moved out of the X-Y focal plane).