The study aim was to identify the population of live patients in the Lothian area of Scotland, with presenile dementia of various aetiologies, and to describe the clinical profiles of each and the patterns of decline which occur, together with any genetic characterisation possible. Cases were identified using the Lothian Psychiatric Case Register. For the demographic data, the CAMDEX (The Cambridge Examination of Mental Disorders of the Elderly) informant interview was used. The behavioural assessment comprised the CAPE-BRS (Clifton Assessment Procedure for the Elderly, Behavioural Rating Scale), the Cornell Depression Scale and the MOUSEPAD, a new behavioural and psychopathological assessment. Of 290 potential cases, 164 (57%) were excluded. Reasons: Death 50 (31%), Unsuitable 40 (24%), Refused 40 (24%), Untraced 23 (14%), Out of area 11 (7%), Of the 126 (43%) seen, 112 (89%) met DSM3R (Diagnostic and Statistical manual of Mental Disorders, Third Edition, Revised) criteria for dementia. 63 (56%) of the 112 were rated as DSM3R severe type. 80 (72%) of the 112 fulfilled the McKhann criteria for Alzheimer's Disease. Full description of the group is available. The genetic testing included Apolipoprotein E, α-1 antichymotrypsin and Very Low Density Lipoprotein Receptor (VLDL-R) allele typing. Important data concerning the services provided and used by this group of patients and their carers has been collected and can be shared with organisations working to find funding for presenile dementia in the health service. This study will provide a thoroughly documented and clinically well worked-up sample. The analysis of the work will help to identify if subgroups exist, according to the patterns of various clinical features, the rates of decline and genetic variations. This would in turn, give a greater chance to plan appropriately for all those involved in caring for and managing these illnesses.