Hirata and Whiton [J. Acoust. Soc. Am. 118, 1647–1660 (2005)] found that the durational ratio of stop closure to word (C/W ratio) is an invariant parameter for disyllabic words that differentiates voiceless single and geminate stops in Standard Japanese, a mora-timing language. The present study examines whether C/W ratio is an invariant parameter for voiced as well as voiceless stop quantity distinction in disyllabic words in Kagoshima Japanese, a syllable-timed dialect in Japanese. A production experiment was conducted in which three native speakers of Kagoshima Japanese (ranging from 54 to 86 in age) uttered in carrier phrases 27 disyllabic words, including 4 minimal pairs for voiceless stop quantity (e.g., popo versus poppo) and 4 minimal pairs for voiced stop quantity (e.g., kogo versus koggo). The rest of them were filler words. C/W ratio and five other possible parameters were calculated based on duration of each segment. Results of statistical analysis [classification accuracy, Akaike's information criterion (AIC), and Bayesian information criterion (BIC)] indicated that C/W ratio most accurately distinguishes single and geminate stops in Kagoshima Japanese as well as in Standard Japanese. [Work supported by Grant-in-Aid for Scientific Research, JSPS.]