Abstract: This paper reports on the results of a study of a diallel crosses system of six parents (including selves) of Chinese fir. Materials for tests came from a breeding population of plus tree clones in Fujian. Characters measured in the study were height (H), diameter at the root collar (D. R. C.) , total number of branches (T. N. B.), number of whorls of branches (N. W. B.) and length of the longest branch (L. L. B.) on a seedling. Heterosis of hybrid progenies was calculated in terms of 100% (Fi-best parent)/best parent, which is similar to the formula used by Bahman (1975) and Pashall (1975), et al. and selfing depression was estimated with reference to the method used by Kheradnam, et al. (1975) Open-pollination progenies were used as genetic controls. Using Griffing’s (1956) Method I, Modelll (fixed model), estimates of general and specific combining ability and reciprocal effects were obtained. Relationships between the superior effect of the crosses and the combining ability effect of parents, between the selfing effect and the general combining ability of parents were analysed. Patterns for genetic controlling of characters of parents were also studied. With the genetic components of the variances estimated by Griffing’s (1956) method I, model II (random model), family and individual heritabilities in both broad and narrow senses were estimated respectively. The major results of this study are as follows:(1) 23.23% of all the crosses exceeded their best parents in height-growth heterosis. The most significant one was superior to its best parents by 25%. This indicates that for the selection of the second generation, it is possible to choose some crosses with outstanding superior effects from the combinations of parents sieved in the first generation seed orchard.(2) The level of heterosis of crosses were associated with the magnitude and direction of the effect value of the combining ability of their parents, (a) The rank of the specific combining ability of combinations of parents did not directly correspond with the order of the superior effects of seedling height; (b) Combinations of parents with higher general combining ability did not necessarily form superior effects;