Five-day-old chicks were trained in a delayed-response task requiring them to rejoin, after a delay of 10, 30, 60, 120 or 180 s, an object hidden behind one of two identical screens. In each of the 20 consecutive trials per chick, the object could be made to disappear randomly behind either screen. Two different goal objects were used: an artificial social companion or a palatable food item. In experiment 1, the screen behind which the goal had disappeared remained visible throughout the delay (the chick was confined behind a transparent partition). Under this condition, with both goal objects chicks proved capable of remembering and choosing the correct screen above chance level at all retention intervals; however, chicks were better at finding the social object than the food item. In experiment 2, chicks were confined behind an opaque partition that prevented them seeing either the correct or the incorrect screen during the delay. Chicks still scored well above chance level at all retention intervals even when they could not see the screens during the delay, although there was a slight decline in performance. Once again, the performance with the social attractor was better than with the food item, but overall above chance for both goal objects. These findings thus reveal that domestic chicks have striking abilities to maintain working memories of the location of biologically attractive objects that parallel or outperform those of most mammals, including primates.