Delayed aneurysm surgery, once standard practice, is now followed by only a minority of neurosurgeons. We analysed the outcome of such a policy in 400 consecutive patients with ruptured aneurysms treated over a 14-year period. Despite an 'open door' admissions policy, admitting all patients immediately on referral, only 56% arrived within 24 h of the ictus (69% within 72 h). Surgery was generally delayed for 8-10 days in patients in Grades 1 and 2; for higher grade patients often for longer until their condition was stable. Two-hundred-and-eighty-seven patients (72%) underwent surgery, 93% on day 8 or later (78% on day 11 or later). Outcome was assessed at 1 year. For all patients 68% were in Glasgow Outcome Scale Grade 1, while 26% had died. Of the operated patients 88% were in GOS grade 1, while 5% had died (30-day surgical mortality was 3.5%). Fifty-one patients (12.8%) rebled, 30 in the first 10 days. Rebleeding was distributed evenly in time over the first 2 weeks. Eighty-four patients experienced non-haemorrhagic deterioration (NHD) all but 3 within 10 days. NHD peaked at days 4-9. Thirty-three patients died of rebleeding and 16 of NHD, but only 12 of the patients who died from rebleeding were fit for operation at anytime and might have been considered for early surgery. Two of these patients died so soon after admission that surgery could not have been performed leaving 10 patients who might have been saved by early surgery. We review the theoretical advantages of delayed as against early surgery and conclude that it is doubtful whether the timing of surgery has any significant effect on management outcome in line with the conclusions of the Cooperative Study.