W'
cided, the ratios W in column 11 would have been unity in all cases, and the deviations from unity which the numbers in this column show, indicate directly the discrepancies between the two meter results. These discrepancies are more plainly shown in columns 12 and 13, from which we learn that, according to the corrected readings, when meter 327 was in the upper aperture, it indicated, at low velocities, with warm air, 29 per cent. more than 136 did in the lower aperture; while with cold air, it indicated 30 per cent. less. But, when their relative positions were exchanged, and 327 was placed in the lower aperture, it indicated with warm ah', 30 per cent. less than 136 did in the upper aperture, but with cold air, 27 per cent. more.
The conditions in which tile meters were placed were, as nearly as could be ascertained, alike in both apertures, except in respect to temperature ; and all the conditions were certainly as carefully observed in these experiments as they would be in most experimental uses of this kind of instrument. The results of these experiments then seem to show that the indications of the two meters differed so widely, and in such a manner, as to destroy confidence in their usefulness when they are employed under circumstances at all similar to those described.
The variations in the meter indications do not seem to follow any easily recognized law.