Feather pecking remains a serious problem in commercial egg production. It has been argued that feather pecking occurs as a result of misdirected pecking, so a possible solution would be to increase the likelihood that such pecking was targeted at another object in the environment rather than to the feathers of conspecifics. Chickens of various strains and ages will readily peck at a device consisting of strands of white string but it is not yet known if pecking at that device would substitute for pecking at conspecifics. Therefore, the effects of providing string devices on feather pecking in an experimental situation (Experiment 1) and on feather condition under commercial conditions (Experiment 2) were examined. In experiment 1, 300 chicks of a high-feather pecking strain of white leghorn-type layers were housed in groups of five in litter-floor pens. The 60 pens were randomly allocated to one of five treatments: devices incorporated in the chicks’ pens continuously from 1 day of age till the end of the experiment at 57 days; devices presented for 4 h per day from 1 day of age; first presented from 22 days of age; first presented from 52 days of age; and finally, devices never presented. Feather pecking was virtually eliminated when the devices remained in the pens from 1 day of age or when they were presented for 4 h per day. Feather pecking was most pronounced among birds that had never received the device whereas its introduction at 22 or 52 days of age yielded intermediate results. This orderly pattern of more pecking at feathers when the device was added at later ages was significant (P < 0.005). In experiment 2, 768 Lohmann LSL laying chickens were housed in rearing cages and 720 were transferred in groups of three to conventional laying chickens when 16 weeks old. The birds were allocated to one of four treatments: devices present from 1 day of age; presented for 24 h every 4 weeks; continuously present from 16 weeks of age; and finally, devices never presented. At 35 weeks of age, hens with access to the device had significantly better plumage condition than those that had never received the device (P < 0.05).In conclusion, the addition of a simple string device to the hens of non-beak-trimmed high-feather-pecking birds decreased feather pecking behaviour (Experiment 1), and to the cages of non-beak-trimmed commercial layers significantly improved feather condition (Experiment 2).