Stress resulting from an investigator handling or observing subjects can affect the results of field studies. However, effects of such stressors are rarely investigated. In a series of field experiments in southern Florida we examined whether display behavior of territorial male lizards (Anolis sagrei) was affected by length of time between handling and observation, length of time that an observer was present and length of time that subjects were held in bags prior to release in the field. Proportion of headbob displays that were bobbing displays did not differ between periods 1-2 h after release and one day later, nor did it differ between periods 0-1 h and 1-2 h after the first appearance of an observer. Similarly, this proportion did not differ between animals previously held in bags for one night and those held for two nights. Our results contrast with other studies that have shown strong effects of several social factors on the proportion of bobbing displays. Thus, the factors we studied appear unimportant in affecting the display behavior of A. sagrei. A number of studies have examined the social significance of variation in display behavior in territo- rial males of the lizard Anolis sagrei. These studies have shown that the relative frequencies of two forms of headbob displays (see descriptions below) vary mark- edly with residence time, the signaler's location within its home range and familiarity of potential rivals (Paterson, 1999; McMann, 2000; McMann and Paterson, 2003). However, such studies may induce stress in lizards indirectly through observer effects or directly through capture and handling effects. Although it is obviously possible to induce anthropogenic stress in almost any animal, including lizards (Moore et al., 1991), it is unclear whether such stress may have confounded effects of other factors on display behavior in previous work on A. sagrei. In this study, we examined behavioral responses of male A. sagrei to variation in time between release and observation, as well as variation in time held in captivity. Our choice of types and levels of treatments were designed to be similar to between-treatment variation in these variables in previous work (McMann, 2000). However, the results of this study provide general information about effects of human disturbance on behavior of A. sagrei in the field. Sugerman (1990) found no effect of an observer on behavior of A. sagrei in a laboratory setting, but patterns of behavior observed in recent field studies (Paterson, 1999, 2002; McMann, 2000; McMann and Paterson, 2003) would have been impossible in the small containers in which the lizards were held. The role of observer effects is important for studies of animal behavior (Macfarlane and King, 2002) but has been often neglected in studies of lizards and other taxa. The highly variable headbob displays that appear to mediate social interactions in A. sagrei can be classified as either ''nodding displays'' or ''bobbing displays'' (McMann, 2000). A nodding display is a volley of continuous rapid dorso-ventral oscillations of the front portion of the lizard's body, whereas a bobbing display is a more prolonged volley of discontinuous oscillations of the front part of the body. Because the proportion of headbob displays that are bobbing displays appears important in communication in this species (Paterson, 1999; McMann, 2000; McMann and Paterson, 2003), we used this display characteristic as a dependent variable in the current study.