The aim of the study was to examine whether the two main sources of new notifications of addicts were accurate and were drawing from similar populations. The research compared the clinical, social and criminal characteristics of a representative sample of men notified by each of these sources and examined the relationship between their licit opiate use and criminal histories during a 5-year followup period. From the total notifications during the year, a random sample of 95 men designated by the hospital and 117 designated by penal institutions was drawn. For both groups, Home Office records were analyzed with regard to information on renotifications, prescriptions for opiates, and deaths during the 5 years following first notification. It was found that 95 percent of the hospital sample but only 30 percent of the penal sample were addicted to opiates at some time during the followup, as indicated by the official records. Three explanations for this highly significant difference between the samples may be suggested. Institutionalization may have sured some penal cases of addiction, or users were able to obtain supplies of opiates from illicit sources during followup, or they were not actually addicted to opiates when first notified, although they might have misused the drugs. Additional evidence fails to support the first two explanations. It is therefore concluded that notification may depend as much upon the setting and circumstances in which doctors see their patients as on their clinical judgment of a notifiable case of addiction. Three tables and 16 references are included in the article. (Author abstract modified)