Section 3.5 on page 331 contains the following sentence:

Fundamental frequency was then measured at 10 ms intervals beginning 40 ms before release of closure to 100 ms following, the range in which CF0 effects are generally expected.

This sentence is partially incorrect. A more accurate description of the procedure is as follows:

Fundamental frequency was then measured at 10 ms intervals beginning 40 ms before the release of closure. For voiceless obstruents, especially when produced in isolation, these values were typically undefined. F0 was also measured at 10 ms intervals over the first 100 ms of the following vowel, beginning at the onset of glottal pulsing, the range in which CF0 effects are generally expected.

(hat tip to Alif Silpachai)