Long-term follow-up of parkinsonian patients has shown that although levodopa treatment significantly improves the parkinsonian symptoms and the quality of life of parkinsonian patients for several years, various distressing difficulties arise during chronic levodopa treatment, such as the loss of benefit, dyskinesias, on-off phenomena, postural instability and dementia. Clinical, neuropsychological, mortality and post-mortem brain studies indicate that levodopa as a replacement therapy does not modify the progression of the underlying pathology and the natural course of the disease. It seems that levodopa has only a limited period of optimal usefulness in the treatment of Parkinson's disease. However, at present there is no better or more potent therapeutic agent available than levodopa and it is still the primary treatment of Parkinson's disease. It would be reasonable not to begin levodopa treatment in patients with mild symptoms but to withhold levodopa until the severity of symptoms really makes its use necessary. Thus it is possible to get the maximal long functional benefit. Post-mortem brain studies have shown that in Parkinson's disease there is not only a progressive loss of dopaminergic substantia nigra neurons but there are also significant changes in the striatal dopamine receptors. In some patients a denervation supersensitivity seems to develop and in some others a loss of dopamine receptors in the striatum. However, in advanced parkinsonian patients with a deteriorating response to levodopa, there seem to be still enough dopamine receptors in the striatum for drugs stimulating the dopamine receptors directly to improve the parkinsonian disability. Indeed, recent evidence indicates that dopaminergic agonists, such as bromocriptine, seem to be a significant and valuable adjuvant therapy to levodopa in parkinsonian patients with a deteriorating response and/or the on-off phenomena. Although bromocriptine is not completely satisfactory, it is a significant opening to a new mode of treatment. In the future it will be very important to develop more potent and selective dopaminergic agonists affecting only those striatal receptors which are mainly responsible for the parkinsonian symptoms. Then a better therapeutic response is likely to occur and many central side effects can be avoided. Current difficulties in the management of Parkinson's disease greatly depend on the fact that we are dealing with a symptomatic therapy. It is hoped that future research will soon lead to a discovery of the primary cause and consequently to a causal therapy of Parkinson's disease.