Adequate quantification of weight and stature is essential in order to determine levels of nutritional support and to monitor the effects of nutritional intervention. Traditional anthropometric techniques are difficult to apply in elderly or handicapped patients chair or bed-bound. The purpose of the present study is to elaborate regression equations for the estimation of stature in the italian elderly population from other anthropometric measures that can be more easily determined. We have found a single model valid for both sexes (in which the value of the variable "sex" equals 0 if woman and 1 if man) to predict stature in italian elderly: Stature = 94.87 + 1.58 knee-height - 0.23 age + 4.8 sex. Cross validation on a control sample of 30 males and 54 females yielded pure errors of 3.1 cm for men and 2.74 cm for women.