The methods used in a multipractice study about asthma and allergic rhinitis in 1977-78 is described. In total 153 general practices took part comprising 271 physicians. The total practice population was 450.271 persons as calculated from the number of persons insured in groups 1 and 2 of the public health insurance system plus their children. The general practitioners were for the most part younger than the average Danish general practitioners. The age of the practice population was younger than the age of the total Danish population. The region of Zealand was overrepresented, and Copenhagen underrepresented. These differences were, on the whole, relatively small. The course of recording was judged by comparing the doctors' self-estimated recording activity with their actual number of consultations per patient. This self-estimated recording rate was a useful indicator of recording quality. The recording rate seemed to be reasonably good in those two third of the practices, which estimated themselves to have recorded more than 90% of the consultations intended to be recorded. The interpractice variability of recording in multipractice studies is an important methodological problem of general practice research. In this study interaction was found with some practice characteristics, but major sources of variability remain unexplained. In spite of the difficulties mentioned conditions for collecting information about prevalence and course of diseases in general practice in Denmark are very favourable in comparison with most other countries. The reason for that is the very large proportion of the population, which is firmly registered with particular practices, the public personal identification number, the access to relevant registers, and the feasibility of handling substantial amounts of data.