Background: Pain in people with advanced cancer is prevalent. When a stable dose of opioids is established, people still experience episodic breakthrough pain for which dosing of an immediate release opioid is usually a proportion of the total daily dose. Methods: This multi-site, double blind, randomized trial tested three dose proportions (1/6, 1/8, 1/12 of total daily dose) in two blocks, each block with three dose proportions in random order (6 numbered bottles in total). When participants required opioid breakthrough doses and it was their first breakthrough dose for that study day, they took the next numbered bottle rather than their usual breakthrough dose. (Subsequent doses on that day reverted to their usual dose.). Results: Eighty-five people were randomized in this study of whom 81 took at least one dose and 73 (90%) took at least block one (one of each dose proportion). No dose was found to be optimal at 30��min with approximately one-third of participants showing maximal reduction with each dose proportion. Median time to pain relief was 120��min. There were no differences in harms: drowsiness, confusion, nausea or vomiting at 30, 60 or 120��min. Conclusions: This adequately powered study did not show any difference with three dose proportions for reduction in pain intensity, time to pain relief, pain control on the subsequent day nor any difference in harms. From first principles, this suggests 1/12 the 24 hourly dose should be used as the lowest dose that delivers benefit. Future studies should include a placebo arm. Significance: Despite the widespread use of immediate release morphine solution for breakthrough cancer pain, the ideal dose derived from background dose has not been determined in an adequately powered randomized, double-blind, crossover, dose ranging study. This study tested three dose levels in people with advanced cancer. Given no differences in time to onset, level of analgesia achieved, nor side effects, the lowest dose tested (1/12th of the daily dose) should be used.