The Geographical Survey Institute purchased a La Coste and Romberg Geoditic Gravity-Meter Model G-29 in October, 1962. The test observations were carried out at Wakkanai, Asahigawa, Abashiri and Chitose in the Hokkaido District via Mizusawa and Sapporo in November, 1962. Another observations were made at Kyoto, Hiroshima, Kochi, Fukuoka and Kumamoto in Western Japan in January, 1963. At each station, the gravity difference from Tokyo (G. S. I.) were determined with a caribration table that was furnished by the maker. During these field work also the test observations were tried at Kakioka, Kawagoe, Hakone and so on in the neighbourhood of Tokyo. The purpose of these observations were to check the accuracy of the reading value and the reliability of the maker's calibration table and, finally, to determine the most reliable gravity value at each station. The results are; 1) Reading accuracy obtained from several observations at one station by one or several observers is less than ±0.02 mgals when the meter is set quietly at that station. 2) Reading accuracy resultant from the irregularity of the drift of the meter is less than ±0.06 mgals when the meter is transported by railway or automobile along the ratherlarge distance. 3) Mean normal drift at the gravity room in G. S. I. is +0.80 mgal per month. This. is rather large as compared with another La Coste and Romberg gravity meter but the difference of the drift according to the condition that whether the meter is set quietly at the station or transported can not be recognized. 4) The calibration factor determined by the maker's experiment in the room is ratheraccurate and the calibration error is less than 0.1%. This was mainly confirmed by the comparisons of the gravimeter results with those of the G. S. I. pendulum apparatus at; Sapporo, Mizusawa, Kakioka, Hakone, Kyoto, Fukuoka and Kumamoto.
