IT has been assumed generally that variations in humidity would not cause errors in knock rating when using the bracketing method. However, the preliminary test results presented indicate that this assumption is not valid for all fuels. Differences in knock ratings of over three octane numbers were found with certain combinations of test and reference fuels when the humidity was varied over the range normally experienced in knock testing.
It was found that the influence of humidity on detonation is not primarily the result of changes in dry air pressure or oxygen concentration, but apparently depends on the nature of the fuel itself.
The results presented are known to be affected somewhat by changes in engine adjustment and bouncing-pin setting. However, for certain fuels the error introduced by humidity changes is considerably greater than the normal experimental error with controlled humidity. Although for some conventional-type fuels humidity control is not required, it is required when testing certain special fuels, leaded gasolines, and octane-heptane blends.
A description is included of the construction of an engine hygrometer and a humidifier both of which have been found convenient to use in routine testing.