“…They have been described variously as tumours diagnosed (i) during admission for the first primary or (ii) within varying intervals from the first primary, namely, 1, 6 or 12 months. For statistical purposes they have been: (i) included in the analysis (Robbins & Berg, 1974); (ii) included on an arbitrary basis (Berg, 1967); (iii) assessed separately and excluded from the final comparisons between observed and expected numbers (Schottenfeld & Berg, 1971;Veronesi et al, 1974); (iv) eliminated automatically with all other cases diagnosed within 5 years of the first primary (Greenberg, 1963;Schoenberg et al, 1969); and (v) deliberately excluded (Schoenberg & Christine, 1974) on the grounds that it was impossible to calculate an expected number for this group. These differing approaches to coincidental tumours, in addition to invalidating comparisons between surveys, suggest that there is some debate concerning the validity of their inclusion in an analysis.…”