We appreciate the interesting comments of Dr Gale (2007). By indicating the infectious doses (ID) for Giardia and Cryptosporidium we attempted to highlight the minimal amounts of (oo)cysts reported in human volunteer trials that are able to induce infection, without excluding the possibility that lower IDs achieve the same effect. The paper by DuPont et al. (1995) quoted in our study reports an ID50 of 132 oocysts for Cryptosporidium, with doses of 30 oocysts inducing infection in 20% of subjects (n = 5). A lower ID50 of 83 oocysts has been found by Chappell et al. (2006) in a very recent study. Action level is by definition the concentration of a contaminant which, if exceeded, triggers a treatment or other requirements that a water system must follow. Action levels are widely used, providing water utilities and sanitary authorities with a practical reference limit for preventing the possibility of an outbreak. We agree with Dr Gale that spot sampling tends to underestimate the real prevalence of these micro-organisms in water. In fact, as many as 10 samples may be required to describe the range of (oo)cysts concentrations that a community is experiencing in its water supply (Wallis et al. 1996). However, it is important to take in consideration that the sample processing methodology used in our study is very laborious and time-consuming, allowing us to analyse only four water samples per week, with a time lag of 3 days between the sample collection and the test result. Because of these technical limitations the monitoring of events, where Cryptosporidium and Giardia were detected, was extremely difficult and, as a consequence, inaccurate. It should also be noted that risk assessment must not only considerer the efficiency of water treatment, but also other important factors including specificity and sensitivity of the diagnostic test, viability and infectivity of the (oo)cysts, immunological status of the host, hydrological features, soil conditions, local geology and others. Because risk assessment was not the aim of our study, most of these factors were not analysed and remain to be elucidated in further work. We also agree with Dr Gale that regular water quality monitoring in conventional water treatment facilities (CWTF) is a cornerstone in control programmes, especially considering that these facilities supply drinking water to the majority of the population. Dr Gale states that, in the event of a failure in treatment, residents in towns and cities supplied by CWTF are at high risk of illness because their lower levels of protective immunity. Despite the growing evidence suggesting that strong serological responses may be associated with a reduced risk of illness (Frost et al. 2005), it should be noted that this issue still remains controversial, and more research should be carried out to demonstrate this phenomenon. Given this situation, we strongly believe that special attention should be directed to those drinking water supplies more susceptible to be contaminated with these protozoan parasites, as explained in our study.