London schools operate in an area where crime rates, including violent crime, is statistically more frequent than the average for the whole of England and Wales (Moore and Yeo 2004). Violent crime in the capital increased (though not to a statistically significant extent) between 2002/3 and 2003/4 (Moore and Yeo 2004b). This has led to a significant increase of fear of violent crime in London compared to the average in England and Wales during 2002-2004. This paper is based on a survey commissioned by the National Union of Teachers but was produced with complete academic autonomy by the University of Warwick. The survey questionnaire contained three sections. The first asked respondents about themselves and their schools. The second asked about the incidence of possessing and using different types of weapons on school premises. The final section asked about aspects of school practice and policy related to the control of weapons incidents. In looking at what can be done to protect schools from violent incidents, one can ask whether incidents are primarily due to the availability of weapons or to the characteristics of those using them. If the primary cause of violence is the availability of weapons, schools can best address the problem by ensuring that weapons cannot be brought into the school in the first place. If the primary cause is the psychological or cultural characteristics of the aggressor, schools have to pay more attention to their policies on internal management, counseling and student support. The priority given to a safe environment was reflected in the finding in this survey that support for containment measures, such as searching pupils, was not related to the level of weapon carrying or their use in schools. This suggests that banning weapons was preferred to dealing with them after they had been brought into school.