OBJECTIVE To determine the best method of preparing synovial fluid specimens for use in quality assurance (QA) surveys designed to assess accuracy of crystal identification. METHODS A previously published method (A) was compared with a new method (B) in the setting of a QA survey. Ten Australian, one New Zealand, and one Hong Kong hospital laboratories took part in the survey. Each laboratory examined six different synovial fluid specimens prepared using method A (first round) and a separate six specimens using method B (second round). In method A, a drop of synovial fluid on a glass slide was surrounded by a rim of Ultramount, sealed with a coverslip, and distributed. The participating laboratory did not need to perform any processing of the specimen before examination. In method B, a capillary tip was filled with synovial fluid, heat sealed, and distributed. The fluid was expelled onto a glass slide in preparation for examination after arrival in the participating laboratory. RESULTS Using method A 36 of 71 (51%) of the specimens were rated as satisfactory, compared with 53 of 61 (87%) of the specimens using method B (Fisher's exact test, p < 0.001). CONCLUSIONS An improved method of preparation of synovial fluid specimens for QA surveys is described. Using the new method it is feasible to perform a synovial fluid QA survey covering a large area (Australasia).