During the time between May of 1997 and September of 1998, for sixteen consecutive months, each month broke the previous monthly all-time record high temperature. Using autoregressive intervention moving average (ARMA) models in a series of Monte Carlo experiments the probability of such an event was analyzed for various rates of temperature change. The string of record-breaking global temperatures could not be readily explained by the best fit linear increase of temperature since the late 1970's (2 ¿C/Century), although the event was not implausible (probabilities slightly less than 5%). The 1997--98 event could signal yet another change point in the rate of global temperature increase, but the warming rate over the past few decades is already comparable to that projected during the 21st Century based on IPCC business as usual scenarios of anthropogenic climate change (Kattenberg et al., 1996). ¿ 2000 American Geophysical Union