Hudson & Holbrook (1982) found lower mean fundamental vocal frequencies in Negro males and females than others had found for whites. As is well known (and found by them), males display a lower frequency (deeper voice) than do females, and puberty deepens the male voice. This deepening is generally attributed to testosterone. The deeper Negro voice may reflect the influence of higher testosterone levels at puberty or prenatally.
More recent studies have failed to replicate the result cited by Miller. A study of "vocal productions of 44 Euro-American and 40 African-American elderly speakers" showed "Euro-American elderly speakers did not differ significantly from African-American elderly speakers on the measurements of all the selected acoustic parameters of voice [including fundamental frequency]". A study on age, height, and weight-matched samples of black and white adults detected no "significant mean differences between the African American and White speakers" in fundamental frequency. A study on vocal samples of 50 black and 50 white men found "no significant differences in the mean fundamental frequency or formant structure of the voice samples". Though the difference is not statistically significant, the whites in this study registered a lower mean fundamental frequency than the blacks (107.55 Hz vs. 108.85 Hz).
The study cited by Miller compared data on blacks collected by the authors to data on whites collected by others, a potential source of error. The study's claim is not sustained by more recent, direct black-white comparisons, and can not be used to argue in favor of black-white differences in masculinization.
 Physiol Behav. 2007 Nov 29 [Epub ahead of print] The relationship between testosterone and vocal frequencies in human males. Evans S, Neave N, Wakelin D, Hamilton C.
 Percept Mot Skills. 2000 Dec;91(3 Pt 1):951-8. Effects of race and sex on acoustic features of voice analysis. Xue SA, Fucci D.
 J Voice. 1997 Dec;11(4):410-6. Aerodynamic and acoustic characteristics of the adult African American voice. Sapienza CM.
 J Speech Hear Res. 1994 Aug;37(4):738-45. Speaker race identification from acoustic cues in the vocal signal. Walton JH, Orlikoff RF.