The main aims of this study were to assess the potential of three indigenous New Zealand tree species, Libocedrus bidwillii, Nothofagus menziesii and N.solandri, for dendroclimatological analysis, to develop tree-ring chronologies from these species and to use the chronologies to reconstruct palaeoclimates. As a prerequisite to chronology development, the annual nature of Nothofagus solandri growth was investigated. Results showed that shoot and radial growth occurred during the summer months only and were closely linked to the course of temperature at this time. Growth ring formation in N.solandri was annual. Thirty-three tree-ring chronologies from the three species were developed from six areas in the South Island; Craigieburn Range, Castle Hill and Flock Hill Basins, Whitcombe Valley, Landsborough Valley, Hollyford Valley and Murchison Mountains. Timberline Nothofagus menziesii and N.solandri chronologies were the most sensitive to environmental factors, assumed to be climatic (having high mean sensitivity and large common variance values), while N.solandri chronologies from bluff sites and subalpine Libocedrus bidwillii chronologies were less sensitive. With increasing altitude N.solandri chronologies became more sensitive, the most sensitive chronologies being developed from trees growing close to, but not forming, the alpine timberline. Climatic analysis of the chronologies showed that the timberline Nothofagus menziesii and N.solandri chronologies were strongly associated with growing season temperature while the three N.solandri chronologies developed at montane bluff sites were strongly associated with growing season rainfall. The Libocedrus bidwillii chronologies were only poorly associated with climate. Two reconstructions of past rainfall (for Amberley and Lake Coleridge) and one of riverflow (Hurunui River) were developed using mainly the rainfall sensitive bluff site Nothofagus solandri chronologies. The reconstructions extend back to 1840 A.D. Variance explained in calibration ranged from 46% to 60% and in verification from 44% to 66%. The reconstructed Hurunui riverflow record suggests that riverflows during the period of modern observations may have been higher than the long term average. This reconstruction and future reconstructions from other rivers presents a potentially important source of information for hydro-electric power generation and in irrigation development planning. New Zealand summer temperature was reconstructed using a grid of seven timberline Nothofagus menziesii and N.solandri chronologies. 59% of the variance was explained in calibrating the reconstruction while 49% of the variance was explained in verifying the calibration with independent data. Based on this, a reconstruction of summer temperature to 1730 A.D. was developed. The three climate reconstructions were interpreted in terms of atmospheric circulation patterns affecting New Zealand and were used to verify other proxy evidence for climatic variation during the last 250 years. In the period 1730 to 1900 A.D. cool summers were common in the 1740's, about 1760 and from 1830 to 1860 and were probably periods of increased southwest to west airflows onto the South Island. Canterbury rainfall, reconstructed as below average in the 1850's, also suggested more persistent westerly airflow at this time. Runs of warm summers occurred more commonly about 1780, from 1790 to 1820 and from 1870 to 1890. Northerly airflows were probably more persistent at these times. It is concluded from the research presented in this thesis that the potential of dendroclimatology as a means to reconstruct past climates in New Zealand is considerable.