目的 根据中国南方1901例土家族（男性858例，女性1043例）成人7项头面部骨性指标（头最大长、头最大宽、面宽、下颌角间宽、容貌面高、形态面高、头水平围）进行身高的推测。 方法  本研究采用倍增因数、线性回归、多元回归及多元逐步回归，对土家族成年人中身高与7项头面部测量指标之间的关系进行探讨并对几种估测方法进行可靠性比较。结果  相关分析结果显示，土家族成人中，7项头面部测量指标均与身高有显著相关性。4种身高估测方法中，线性回归最为可靠。4种身高估测方法中，预估可靠性依次为线性回归，多元回归，多元逐步回归及倍增因数。结论  头面部测量指标能够用于身高的估测。在本研究中，头水平围与身高相关性最大，因此，根据头水平围建立的身高估测方程最为准确。

Objective  To conduct stature estimation from seven facial measurements of 1901 Tujia adults (858 males, 1043 females) in Southern China aged from 18 to 75 years. Methods  The present study applied variable statistical method, such as multiplication factors, linear regression, multiple regression as well as multiple stepwise regression to explore the relationship between stature and cephalic-facial measurements and the reliability of the estimation method of stature in Tujia. Results  All the facial measurements had significant coefficient with stature according to Pearson’s analysis. Linear regression was the best method for stature estimation according to cephalic-facial measurements in Tujia among the four method. Overall, the range of the four method based on the reliability was linear regression, multiple regression, multiple stepwise regression and multiplication factors. Conclusion  Cephalic-facial measurements can be used to establish stature estimation equation. Horizontal head circumference has the highest correlation in the present study so that the estimating equation based on horizontal head circumference is most accurate.