This paper analyzes the tonal behavior of an adaptive second-order sigma-delta modulator, which was developed and published by the same authors. Idle channel tones, caused by non-white quantization error, is not desirable in applications where the human ear is the end receiver. Besides their relatively small magnitude tones in the baseband, most sigma-delta modulators produce high-powered tones near fs/2. It is a more serious problem because the clock noise near fs/2 can couple these tones down into the baseband. Various simulations show that the more randomized nature of the aforementioned adaptive architecture makes it more advantageous in tonal behavior, particularly attractive in that it significantly reduces the dominant tone near fs/2, which can not be reduced by dithering in a standard second order single-bit modulator. With comparison to the standard second-order sigma-delta modulators, the results are illustrated in both frequency and time domains.