The diagnosis of ALS is often delayed, probably largely because the disease is relatively uncommon. A practical approach to diagnosis, with a simple algorithm for investigation, might improve diagnostic sensitivity in the early stages of the disease. It is clear that the first step in the diagnostic algorithm must be referral to a neurologist, which implies recognition by the primary care physician that the illness is neurologic in origin. It is the role of the neurologist to make the diagnosis, to confirm it beyond all reasonable doubt, to exclude other treatable conditions, and to inform the patient and the family of the diagnosis in a timely, honest, and compassionate way. Definition of different disease stages allows a model of the process of care to be formulated and indicates the time points at which intervention may be possible. There are two pathways to diagnosis of ALS, which overlap to a considerable extent. Diagnosis by positive criteria is limited by the absence of a specific diagnostic test, whereas diagnosis by exclusion of other disorders may introduce delay that can be tempered only by efficiency in the investigative pathway. The imperative necessity for diagnostic review is perhaps the most important aspect of the diagnosis of ALS, in view of the increasing recognition of the need for early diagnosis, and when an effective therapy becomes available, this process of review will become more important, albeit more difficult.