Aim To examine the response of ring widths of Phyllocladus aspleniifolius to maximum and minimum temperatures and to a zonal index (ZI).Location Tasmania, Australia.Methods Between two and four increment cores were obtained from approximately 20 trees at each of 16 sites of P. aspleniifolius across Tasmania. Four regions were defined and temperature data for climate stations from each of these regions were transformed into dimensionless time series appropriate for each region. The ZI was constructed from Melbourne and Hobart mean sea level pressure (MSLP). Correlations were calculated between climate data and ring widths over most of the 20th century. Data were then segregated into two subperiods to assess whether or not temporal changes in ring width response have occurred over the 20th century.Results The dominant feature of the correlation functions was a negative correlation with temperatures of the previous growing season. Correlation between the ZI and ring width is most significant, and positive, in November and March of the previous growing season. Differences in the shape of the correlation functions, and their strength is apparent over the course of the century. In general, correlations have become weaker and less consistent among the 16 sites for the latter part of the century. Temperatures have generally increased during the 1900s over which time period correlations have been calculated.Main conclusions The temporally changing climatic response in ring widths is consistent with a hypothesis of competition for resources between different parts of a plant. A nonlinear response to climate over a short-time period violates one of the basic assumptions of dendroclimatology, therefore suggesting further investigation of both the physiology of trees used for dendroclimatological purposes and the statistical methods used for climatic reconstruction.