The spectrum computed from this data have a SNR of about 15:1 by the peak-to-peak method. The spectral features due to water vapor and carbon dioxide can be seen clearly. The performance is well below theoretical predictions. Part of the problem is caused by the pulse-width-modulated motor controller which generates considerable electromagnetic radiation in spite of careful shielding and grounding. The interferometer alignment is not yet optimized for compactness so beam divergence leads to some loss of energy. When these minor problems are solved the effects of turbulence on the noise level will be easier to discern. If necessary to avoid turbulence, the system will be operated in vacuum. In vacuum, the performance will ultimately be limited only by detector noise.