height-growth and site-index curves were developed for estimating site quality of trembling aspen {populus tremuloides michx.) in north central ontario. these curves were developed from stem analysis data of dominant and codominant, uninjured aspen trees obtained from 89 plots covering a wide range of site quality in north central ontario. the actual height-growth patterns were modelled using several non-linear biological growth models: chapman - richards function, modified weibull function, monserud logistic function and an expansion of the chapman - richard function. in addition, a new height-growth model was developed using a similar approach to that of cieszewski and bella. height-growth patterns of aspen varied with level of site-index. height growth curves show an almost linear growth pattern for poor sites (si < 16 m) to a highly curvilinear pattern on good sites (si > 24 m). medium sites (s116-24 m) show a rapid linear surge of height growth before 40 years followed by a slowing curvilinear pattern. height-growth curves, site-index curves and a site-index prediction equation were constructed from trembling aspen stem analysis data. goodness of fit tests were computed using a modified chi-square test. in addition, the accuracy of the height-growth curves, site-index curves and site-index prediction equation were tested using independent stem analysis data from 19 plots supplied by the ontario ministry of natural resources. comparisons with the independent data source shows close agreement; the 95% site index error prediction interval for the site-index curves and the site-index prediction equation are 0.19 ± 1.37 and 0.21 ± 1.35 respectively.