Assuming that a strip contains many ultimate units continuously and timber volumes of whole strip consist of unit volume, we can calculate serial correlation coefficients and obtain correlograms. From the data of Shiragadake experimental forest, concerning each strips running along three directions, serial correlation coefficents of stand timber volumes were calculated. From these results of sample correlation coefficents, we can recognize that there are not autocorrelation of stand timber volume series in such forest. Then we can consider p = 0 in the variance formula from systematic sampling, and for systematic sampling we can use random sampling formula instead of systematic formula. In this report based on the numerical data from actual forest volumes, systematic sampling error percents are compared with random sampling error percents. According to this contrast table we can find that the informations from systematic sampling are more satisfactory than random sampling.