stem analysis is a common forest mensurational technique used to gain individual tree information for various growth attributes. interest in stem analysis has been renewed with the availability of computer technology, and an increased emphasis on forest growth and yield research. this thesis deals with two main areas of concern. the first is the need for a new computer algorithm capable of processing stem analysis data produced by annual ring measurement equipment. the development and application of two new algorithms, duffno and stem, are discussed. duffno's main functions are; to aid in data verification, and to produce the duff-nolan sequences for the ring width data. stem's main function is to calculate and produce tabular and graphical output of the growth attributes. able the second area of concern involves stem analysis sampling techniques. nine trees were sectioned intensively to obtain true volume estimates, which were used as control values. these were compared statistically against volume estimates derived from sub-samples of the disc data. reliable volume estimates, within 10 percent of control values at a confidence level of 95 percent, were obtained from three basic sampling methods. these were referred to as the "uniform section length" method, the "form class" method, and romberg's method. recommendations for further research are offered.