Since the installation in April 1970, the McKinney type mass spectrometer has maintained a good condition and about 5000 isotopic analyses of oxygen, carbon and sulfer has been cerried out so far. The ion source filament was changed to spiral one from single wire. Consequently three times sensitivity and the longer life of filament were obtained. Gradual changes in the ratio of potential divider (β) and the sensitivity (|△ E(0)(2)|/E(0)(1)) for the difference of β have been observed (Table 1). Because these changes are opposite in trend but similar in degree, it is supposed that these changes are caused by decrease of feed-back resister value in the balancing system. Although there remain some problems in the corrections for the difference in observed isotopic tatio between two cases that a same CO(2) is fed to ion source from the ST side of inlet system on the one hand and from the X side on the another, the reproducibility of the measurement of isotopic ratio for long period is
about ±O.1‰ (Table 2, Fig. 1). δ(18)O values of standard waters, MSA-1 and MSA-2, measured during a period from May to Dec. 1970 and in Sep. 1971 yielded good agreements implying the reproducibility of the measurement to be constant (Table 3). Both calculated δ(18)O values relative to a given sample based on each observed δ(18)O value relative to two different samples agreed each other. It is concluded that the linearity of observed a-value holds over a range of 30‰ (Table 4).