Early studies [J. Acoust. Soc. Am. 25, 313 (1953); J. Acoust. Soc. Am. 25, 322 (1953)] related noise output to fan horsepower and indicated the usefulness of this relation in predicting noise in ventilating systems. More comprehensive measurements yielded a simpler, more accurate equation relating noise and rated horsepower for large centrifugal fans in actual air conditioning installations [J. Acoust. SOC. Am. 27, 217 (1955)]. The present paper shows that recent laboratory measurements of noise from small centrifugal fans [Noise Control 1, No. 6, 26 (November, 1955)] is consistent with the relation found for large fans. In addition, empirical equations derived for the small fans expressing total noise as the sum of two parts, one arising from air flow and one from static head, can be simplified with improved fit to the data Pf = C1V5 and PH = C2H3(WH/WR)5, where Pf and PH are noise power resulting from flow and static head, respectively; C1 and C2 are constants; V is a linear velocity equal to the volume flow of discharge air divided by the area of the outer cylindrical surface of the rotor; H is static head; and WR and WH are widths of rotor and housing, respectively. These equations indicate that an optimum relation exists between air speed and pressure head which will minimize noise for a given horsepower delivered from a fan of a given size. Comparisons of these relations to subsequent data for large fans will be made.