Limulus polyphemus, the common horseshoe crab, inhabits coastal estuaries of North America and the Yucatan peninsula. It has been an admirable model for vision research because its relatively simple visual system has marked functional similarities to those of many species, including humans (1). Adult male horseshoe crabs use their lateral eyes to find mates (2), and their ability to see underwater has been studied in the estuaries where mating occurs (3, 4, 5). However, studies of the visually guided behavior of adults in captivity have not been successful (R.B., pers. obs.). On the other hand, juvenile horseshoe crabs do exhibit visually guided behavior under laboratory conditions (6). Here we report a comprehensive investigation into the visually guided behavior of juveniles, both in their natural habitat and in captivity. Between May 15 and June 30, 2002, we studied the behavior of juvenile specimens of Limulus in Nantucket Sound, along the Western shore of the North Monomoy Island Refuge in Chatham, Massachusetts. The sand bars on the tidal flats of the island are a protected natural habitat for both juveniles and adults; the waters are relatively clear and calm. Juveniles were most active near shore at depths ranging from 30–120 cm, usually during the late afternoon and evening. In testing the visually guided behavior of a juvenile horseshoe crab, we followed the procedure of a previous study of such behavior in adults (4). First we determined its direction of travel along the bottom and then placed a small object directly in front of it at a distance of 20–40 cm. We used three cylindrical acrylic plastic objects (diameter 7.6 cm, height 5.2 cm): one was painted matte black, one matte gray, and one remained unpainted. Because these objects were designed for experiments with adults, their size was reduced proportionally for small juveniles (prosomal width range 2–14 cm). We filmed 713 trials, defined as an approach of an animal to an object, on a Sony camcorder (model DCR-PC110 NTSC; underwater housing by Gates Underwater Products) and then analyzed the behavior with a computer. All videotaped trials were analyzed with an eMac computer using iMovie software (copyright 2001–2002 Apple Corporation). We categorized each trial as representing one of three behaviors: “hit,” an animal moves directly toward an object and makes direct contact with it; “pass-by,” an animal passes by an object without changing direction; and “turn-away,” an animal turns away from an object. For turn-away behavior, we measured the distance from the center of the object to the middle of the animal at the point where it turned. For the pass-by behavior, we measured the distance of closest approach to the object. Figure 1 (top) shows percentages for all three behaviors. We performed a similar set of experiments in the laboratory, where we observed the behavior of 10 juveniles in a large trough (130 62 cm; the water 6.5 cm deep). The bottom of the trough was covered with a thin fiberglass mesh screen, giving the crabs traction for walking and righting themselves if overturned. The crabs were maintained in ambient diurnal lighting in a separate trough before being placed in the experimental trough, where we used diffuse overhead illumination to simulate, as closely as possible, an overcast day. The same three objects used in the ocean observations were now placed in the center of the trough and rotated every 15 min. The crab’s behavior was filmed and analyzed, as in the ocean trials. We defined a behavioral event as a crab entering the camera’s field of view (51 74 cm). We filmed 797 trials, mostly in the evening hours when the crabs were most active. Figure 1 (bottom) gives percentages for the laboratory trials.