A quantitative polymerase chain reaction (qPCR) was validated for the detection of Mycoplasma synoviae (PCR equivalents of colony-forming units [CFU eq.]) in chicken joint specimens with time and compared with direct mycoplasma culture. Joint specimens were obtained from 70 layer pullets inoculated intravenously with M. synoviae at 6 weeks of age. Ten control birds were injected intra-articularly with Freund's complete adjuvant. Macroscopic joint lesions were observed in 54 infected birds, of which 11 showed positive M. synoviae culture. The specificity of direct mycoplasma culture was high (100%, 95% confidence interval [CI] = 74 to 100), but its sensitivity low (16%, 95% CI = 8 to 26). Most positive results were obtained during the first 2 weeks after onset of joint swelling using synovial fluid. The qPCR was positive in 26 of 28 synovial fluid samples and in 51 of 70 joint swabs. The sterile joint samples obtained from Freund's complete adjuvant-injected birds were negative in the mycoplasma culture. The specificity and sensitivity of the qPCR for synovial fluid samples were 100% (95% CI = 65 to 100) and 93% (95% CI = 77 to 99); for joint swabs they were 100% (95% CI = 74 to 100) and 73% (95% CI = 61 to 83), respectively. Positive qPCR results (10(0.3) (to) (4.6) CFU eq./ml) were found until the end of the experiment (12 weeks post inoculation). At the end of the study, eight out of 16 joint swabs from birds without macroscopic joints lesions were positive in the qPCR (10(2.0) (to) (2.8) CFU eq./ml). Under the conditions of this study, the sensitivity of the qPCR was higher than that of direct mycoplasma culture (P< 0.0001) during the acute, subacute and chronic stages of arthritis.