Recent studies on the measurements of magnetization at high temperatures are reviewed. Use of a second gradient magnetometer is the main advance in instrumentation. Direct observation of how remanent magnetization is acquired and/ or lost by changing temperature provides strong evidence about the origin of thermoremanent magnetization. According to such observation, Neel's single domain theory seems to be correct. Neel's multi domain theory (which is originally a two domain theory, but is frequently applied to grains with many domains), however, cannot explain the behavior of magnetization at high temperatures. The behavior of magnetization in pseudo single domain (psd) grains is strongly dependent on the grain size. It seems that the heterogeneity in the magnetic grains and the screening effect are the main effects which characterize the psd moment.It should be emphasized that at high temperatures, there are still many interesting phenomena which have not been adequately described and/ or explained. Some of these interesting phenomena are shown as a guide for future studies.The high temperature measurement of magnetization is undoubtedly important for understanding TRM (thermoremanentrnagnetization), because TRM is acquired at high temperatures. In the 1950's and 1960's, when astatic magnetometers were the main instruments for the measurements of magnetization, there were many studies which reported measurements of magnetization at high temperatures and important advancements were made (e.g., NAGATA and OZIMA, 1955). In the 1970's and early 1980's, however, relatively few such studies were done. This is probably because of the difficulty in using spinner magnetometers and SQUID magnetometers at high temperatures (COLLINSON, 1983). Although some spinner magnetometers can be used at high temperatures, it is difficult to control the atmosphere, and induced magnetization cannot be measured. For rock magnetic purposes, it is important to control the atmosphere so that the magnetic grains are not altered during high temperature measurements. SQUID magnetometers of ordinary design cannot be used with electronic furnaces (CoLLINSON,1983). Such disadvantage has prevented the use of these modern instruments for high temperature measurements. Recently, however, second gradient SQUID magnetometers were put to use for high