The SNR of the single beam spectrum is again about 15:1 by the peak-to-peak method. The resolution is half the previous value, and the acquisition time is also half. These two values lead to factors of square root 2 which cancel. All of these spectra so far have been recorded in open lab air on a benchtop. At 20,000 rpm the windage from the disk mirror is sufficient to blow papers off the lab bench; the mechanical vibrations are large enough to rattle hardware and tools off the lab bench. In spite of these disturbances, the SNR is largely unchanged from that observed at 333 scans per second. Certainly, there is more noise immunity than either conventional rapid-scan or step-scan operation allow. It is possible that effect of the turbulence is too small, relative to the detector noise, to be observed.