High levels of fear of humans may influence the productivity of laying hens and there is evidence from laboratory-based studies of tier effects within battery-cage systems on fear of humans and productivity of laying hens. The present study examined some of the factors that may influence the commercial laying hen's fear of humans. In Part 1 of this study extensive behavioural data from a previous study at commercial farms were re-examined to study the effects of tier level on the behavioural responses of commercial laying hens to humans. Significant effects of tier level were found in multi-tier (three- or four-tier) battery-cage systems on the behavioural responses of commercial birds to humans. On the basis of the withdrawal responses of birds to an approaching experimenter, fear of humans may decline from moderate levels among birds housed in Tier 1 (bottom) to Tier 3, in both systems and then sharply increase from Tier 3 to Tier 4, in the four-tier system, reaching the highest level recorded. For example, birds in Tier 3 (top tier) in the three-tier system were more frequently (P<0.01) observed in the front 20 cm of the cage as the experimenter approached than birds in Tier 1 (bottom tier), however in the four-tier system birds in Tier 4 (top) were observed less frequently (P<0.05) in this area as the experimenter approached than birds in Tier 1. In a series of paired comparisons in Part 2 of the study, a number of aspects of human approach were classified within the paired comparisons as more or less threatening to birds on the basis of the birds' withdrawal responses. The most striking effect was the marked withdrawal (P<0.01) of birds when the experimenter stood 30 rather than 60 cm from the cage. Furthermore, there was more withdrawal (P<0.05) when the experimenter positioned himself so that his eyes were 30 cm below rather than 30 cm above the height of the bird's eyes. A greater (P<0.05) number of bouts of crouching or escape attempts were observed when the experimenter placed his hands rather than his face 30 cm in front of the cage. The determination of stimulus-response relationships in Part 2 of this study, will be useful in identifying those features of human-approach behaviour which may regulate the bird's fear of humans. With this type of information, it may be possible to reduce the commercial bird's fear of humans in order to improve bird productivity and welfare by manipulating these human factors.