To record the change in optical rotation of a nerve fibre associated with excitation, an optical apparatus was constructed using a polarizer, a photo-elastic modulator, an analyser and a lock-in amplifier. The apparatus was calibrated with the sucrose solution as the standard. 2. When a lobster leg nerve was dissected and mounted on the sample stage of the apparatus, stimulation elicited a transient change in the lock-in amplifier output. The signal (here called the 'R-signal') had a rapid time course, formed a peak during the rising phase of the birefringence signal, and often quickly returned to the base line, but sometimes showed a long-lasting later phase. 3. The R-signal arose at about the time when the compound action potential of slowly conducting fibres passed through the window of the chamber for the optical experiment, suggesting that it originates mainly in the smaller fibres. 4. The R-signal reversed its sign when the azimuth of the polarizer was changed by 90 deg, indicating that the R-signal was not due to electrical artifacts. Simultaneously recorded changes in the intensity of the transmitted light had a different time course and an amplitude too small to explain the appearance of the R-signal. 5. When the azimuth of the nerve was changed, the later phase of the R-signal changed its amplitude and direction, but the initial phase of the R-signal was much less influenced, suggesting that the birefringence signal was a component of the later phase. 6. The later phase of the R-signal could be reconstructed as a sum of an R-signal at a different azimuth and the birefringence signal, if the amplitude and direction of the latter were adjusted by multiplication of a factor. 7. Assuming that the nerve is a homogeneous, linearly and circularly birefringent and linearly and circularly dichroic material, the lock-in output was described by mathematical equations. From one of them the birefringence signal could be deduced from a series of R-signals observed at various nerve azimuths. The time course of the calculated birefringence signal agreed well with that of the experimentally recorded birefringence signals. 8. Utilizing the same equations, the contribution of the birefringence change to the R-signal was estimated and subtracted. The remaining part was independent of the nerve azimuth, and could be regarded as representing the time course of the change in optical rotation of the nerve.(ABSTRACT TRUNCATED AT 400 WORDS)
