CONSTRUCTION AND INTERPRETATION OF PATH DIAGRAMS FOR HEIGHT AND INCREMENT FROM LOBLOLLY-PINE PROVENANCES TESTS IN TIME-SERIES

Growth is an accumulation process. The total height model is described by the previous height and the last increment. The periodic increment model is described by three causes: (1) the previous height, (2) all prior periodic increments, and (3) other factors or random error. In t his paper, height and increment models at age 5, 10, 15, 20, and 25 ye ars were built by using data from the Southwide Loblolly Pine Provenan ce Test. The regression coefficients from these models were converted into path coefficients and then they were assembled into path diagrams . A synthesized path diagram can show the cause and effect relationshi p in any combinations of heights and periodic increments. Variances an d correlations can be partitioned by tracing separate routes in a path diagram so that important factors may be identified. After constructi on, the path diagram was verified by comparing correlations calculated from the path diagram and those calculated from the data set. They we re identical. For loblolly pine, paths leading to total height are all positive, so there is persistent dominance of tall and fast growing t rees. Paths leading to periodic increment may be positive, negative, o r may be greater than unity, it shows that there may exist a counterba lance between total height and preceding increment in forming subseque nt increment. There seems to be a growth phase shift and a change of d irection in counterbalance about 15 years of age. Applications of path analysis to provenance selection are discussed.