Minimal European recommendations for the prevention of the development of four major tumour types by mass screening are proposed by a panel of experts in this issue of the European Journal of Cancer (pp. 1473± 1478). These recommendations are summarised in Table 1 with additional epidemiological and cost-eectiveness data. Brief reference is also given to the generally neglected aspects of logistics and importantly, as screening is only one means of lowering cancer mortality, competing innovations. Improved `normal' access to diagnostic care and/or other emerging diagnostic techniques, as well as the large-scale application of medical treatments could also lower mortality rates. The inclusion of some of the experts in these screening projects improves the advice given, but can also limit the focus. Although mass screening for cancer as proposed in the statement oers an important contribution to early detection and cancer control, its eects remain modest. Firstly, the cumulative incidence and, in particular, the mortality rates observed in the unscreened relevant age groups remain low. It is the individual risks that may be high. Moreover, the age-speci®c relative mortality for each of the four cancer sites described is usually below 5%, with the exception of breast cancer in the age group of 35 to 55 years, where it is approximately 20%, declining to 5% by the age of 70 years. The most eective results appear to be obtained by endoscopy for colorectal cancer, although this has only been examined observationally. Cervical cancer screening, if carried out systematically (as in Finland [1]), could have a marked eect on both incidence and mortality rates. However, in most European countries to date, either pathologists, gynaecologists or GPs screen in isolation instead of being guided in a collaborative eort with and by epidemiologists. Therefore, although the experts stress the importance of organisation and discipline they are rather unspeci®c in their recommendations for successful screening procedures that must also remain free from bureaucracy. A successful approach has proved to be a mixture of `top down' thinking, budgeting, assessment and training with `bottom up' involvement and commitment. Moreover, in the report in this issue I detected little awareness of all the de®nition problems and the, partly related, shortage of pathologists in the various countries referred to. Importantly, not all the advice proposed by the committee of experts is based on solid evidence; for example, the cervical cancer screening that is proposed for ages 20±29 years, unless this is being suggested for high risk individuals only. Breast cancer screening for the age group of 40 to 49 years and the use of sigmoidoscopy for colorectal screening are two other examples where supporting evidence is not yet very strong. Prostate cancer screening was declared experimental for the next 10 years when in reality it is already being implemented by certain urologists in several countries. They also do not put forward arguments of other experts who opposed even research of prostate screening [2] on the basis that it is not possible to distinguish the aggressive from the non-aggressive tumours and address the subsequent problems associated with over-treatment. If the countries are rated in a league table according to their present mass screening programmes, Finland, Sweden and The Netherlands would top the table while Germany, Spain, Portugal and Belgium would be at the bottom of such a list, the latter group because of too much or too little screening activity. Using such comparative studies, European countries can learn from each other on how to improve their own programmes. It also becomes obvious from the report that cancer screening is a rather complex set of medical activities. It seems particularly important to screen appropriately, as screening is aimed at a predominantly healthy population who generally perceive screening to be uncontroversial with obvious bene®ts. Thus, nothing can or should go wrong, i.e. false-positives and negatives should be minimal. This contradiction is not addressed enough in the report. A further description, perhaps as an appendix, of the underlying biological, medical,