Otoacoustic emissions have historically been measured with an acoustical probe assembly hermetically sealed in the ear canal, imposing in most cases a limited stimulus bandwidth. A physically open recording system should afford the possibility of a greater stimulus bandwidth but the change in acoustical load may affect the magnitude of otoacoustic emissions obtained. Here it is reported that the authors have measured in the guinea pig transient-evoked otoacoustic emissions extending in frequency to 20 kHz and cubic distortion tone otoacoustic emissions forf2=4737and 8096 Hz with a physically open sound system. To address the effect of acoustical load provided by a physically open versus hermetically sealed system, the authors compared the amplitude of electrically evoked otoacoustic emissions recorded from a guinea pig in each case. The change in acoustical load in the ear canal introduced by the change in recording setup did not appear to make a substantial difference to the magnitude of otoacoustic emissions measured. A physically open recording system provides a good alternative to traditional acoustical probe assemblies sealed in the ear canal for the laboratory measurement of acoustically evoked otoacoustic emissions, with the advantage of permitting a greater stimulus bandwidth.