There is now widespread agreement that serial brain MRI is useful in monitoring treatments designed to modify the course of multiple sclerosis. It has been less clear whether gadolinium enhancement is needed. We therefore compared the relative sensitivity of long repetition time (TR) spin echo (SE) and gadolinium enhanced short TR SE sequences in detecting active lesions. A blind analysis of the two sequences was performed in 26 untreated patients with early relapsing--remitting (19) or secondary progressive (seven) multiple sclerosis who underwent monthly MRI on four occasions (one baseline and three follow-up). Active lesions were defined as either new or enlarged lesions on long TR SE, or new or persistent enhancing lesions on short TR SE. In one patient there were 144 active lesions, all of which were seen with enhancement on short TR SE, but only 17 were seen on long TR SE. Amongst the remaining 25 cases, a total of 106 active lesions were seen: 68 (64%) were seen only with enhancement on short TR SE, 16 (15%) were seen only on long TR SE, while 22 (21%) were active on both sequences. We conclude that gadolinium enhancement markedly increases the sensitivity of monthly brain MRI in monitoring the treatment of relapsing--remitting or secondary progressive multiple sclerosis. With this frequency of scanning, a post contrast short TR SE sequence is the most sensitive method for detecting active lesions. The smaller yet still substantial incidence of active lesions seen only on the long TR SE sequence suggests that it should also be obtained.