A ''curious reduction'' in the normalized growth rate found in recent
experiments on compressible mixing layers is explained using linear st
ability analysis. The maximum amplification rate from the stability an
alysis agrees well with the growth rate measured in different experime
nts. No anomalous reduction is observed in the normalized maximum ampl
ification rate, suggesting that the model of growth rate for incompres
sible mixing layers used to normalize the compressible mixing layer da
ta is inaccurate for extreme density ratios.