HE Radcliffe Meteorological Station, Oxford, possesses the longest series T of meteorological observations at the same site in the British Isles. Records of rainfall, sufficiently accurate and regular to enable monthly totals to be derived, go back to 1815 (Smith 1968). Although most of these observations have been published in the form of daily results up to 1935 (Radcliffe Observations) they have attracted less attention than the comparable series of daily and monthly temperature records. The astronomers who maintained these observations until 1935 were extremely meticulous in their methods of observation. Any changes of instruments, known instrumental errors and minor changes of site and exposure were recorded and, whenever necessary, corrections were applied and overlapping records from two different instruments or exposures were maintained. Since 1935 when the maintenance of this station became the responsibility of the School of Geography no further changes of site or exposure have been made and there is no reason to suspect that new buildings on the adjacent Radcliffe Infirmary site have affected the rainfall record.This paper examines the annual, seasonal and monthly fluctuations of rainfall at Oxford since 1815 in an attempt to determine whether there have been longterm variations in the total amount of precipitation at Oxford and changes in the distribution of rainfall throughout the year. Studies of climatic fluctuation in the British Isles have largely concentrated on temperature and snowfall as exemplified by the work of Lamb (1966 (b)) and Manley (1939Manley ( , 1940Manley ( , 1959 and also on changes in atmospheric circulation and weather type (Lamb 1950(Lamb , 1965(Lamb , 1972). Rainfall fluctuations have been studied both regionally and in terms of a secular trend by Gregory (1954Gregory ( , 1955Gregory ( , 1956 but his work was confined to the period since 1881, using the former British standard period 1881-1915 and the standard 30-year period 1901-30, both of which happen to be in some respects unrepresentative of earlier and later periods and of the long period means. A very extensive series of papers on British rainfall was published by Glasspoole and in some of these he attempted to illustrate and determine annual and monthly secular fluctuations for the British Isles as a whole and for England and Wales, Scotland and Ireland separately (1921, 1922, 1926, 1928 (a) and (b), 1933, 1941). In some papers Glasspoole took the record back to the year 1727, using such scattered and discontinuous observations as were available for the early period (1 928 (b) and Nicholas and Glasspoole 193 I). Glasspoole suggested 2