Predictions and effects of these theories were analyzed using data for 100 urban neighborhoods selected from Seattle, Washington, census tracts in 1990. Data were also obtained from the Seattle Police Department and a 1990 victimization survey. Baseline recursive equations indicated informal surveillance was inversely associated with robbery and stranger assault. In contrast, burglary rates were not affected by informal surveillance, but burglary had a positive effect on informal surveillance when robbery and stranger assault were controlled. Simultaneous equations showed that robbery and stranger assault had a moderately strong inverse effect on informal surveillance, and that this effect was mediated by resident perceptions of risk. When risk perception was controlled, informal surveillance had an inverse effect on robbery and stranger assault. The latter analysis also indicated burglary increased surveillance, suggesting that some types of crime served positive functions. The results lend support to systemic community decline and functions of crime theory, and they suggest that the relationship between informal surveillance and crime is complex. Implications for community crime research are discussed. Appendixes contain a correlation matrix and information on variables of non-recursive burglary models. 57 references, 5 footnotes, 7 tables, and 2 figures