There is ample contemporary evidence that most meteorological thermometers in Australia were not exposed in Stevenson screens until very late in the nineteenth century, and in many places not until well into the twentieth century. There is also evidence, from a long-running comparison at Adelaide, that mean temperatures in a Stevenson screen are lower than in an open stand in Australian conditions. Thus, there are strong grounds for expecting that nineteenth century, and some early twentieth century, Australian temperatures are biased warm, relative to modem exposures.

No abstract