In dealing with the geochemistry of volcanic rocks, it appears imperative, first, to refine the procedure of rock analysis with respect to both major and minor components, and second, to improve present methods of interpreting analytical data on common rock suites such as basalts, andesites and dacites. In regard to the first problem, the rapid method of silicate analysis as described by Shapiro and Brannock (1956) and by Riley (1958) deserves special attention. With such a method it is possible to estimate the accuracy and precision of data with far greater ease than is possible with a classical one, yet without any loss of quality. For the determination of some components, however, it is felt that the procedures recommended previously lack precision and/or simplicity. In Part I of this paper, new methods are suggested for these components. The sum of Mg, Ca and Mn is titrated with EDTA in the presence of Al, Fe and other metals, using thymolphthalein complexone as indicator at pH 10-10.5. The sum of Al and Fe is determined by the back-titration of excess EDTA with the standard Cu solution. PAN is used as indicator at pH about 4, and tartrate is added as the masking agent of Ti. Na and K are determined by flame photometry using very dilute solutions (1-5ppm Na or K), without the separation of other metals and without the use of the internal standard. Sr is included in the scheme of analysis, and is determined by flame photometry according to the standard addition technique. These procedures are tested for their accuracy using the standard samples G-1 and W-1 (Tables 3, 5, 9, 12 and 15). Procedures for other components are also described in the text in full detail. The system of analytical procedures recommended is shown in Table 16. To investigate the problem of the interpretation of compositional variation, typical rocks from Asama volcano and the surrounding area were selected, namely, andesites and dacites of calc-alkaline type. Twenty-one samples, described in Tables 18 and 19, are analyzed according to the above procedures. Results are shown in Table 20. Variation diagrams are presented in Figs. 3 to 5. Since it may be questioned whether the "trend" seen in the diagrams implies a genetic relationship, a least squares approximation technique has been introduced in order to determine whether the composition of the main components of a rock (F) can be expressed by the linear combination of a selected set of compositions of a magma and the phenocrysts crystallized from it : F(o)=F(1)x(1)+F(2)x(2)+······+F(m)x(m) ······(1) where F(1), F(2), ... denote the compositions of a magma and minerals, and F(o) is the calculated composition of F. These calculations lead to the conclusion that there are at least two series of rocks in Asama (Tables 25 to 30 and Fig. 9), distinguished from each other mainly by their K(2)O content. Those rocks showing features of assimilation (Aramaki, 1963) all belong to the K(2)O-rich series. Contents of minor components such as TiO(2), MnO, P(2)O(5) and SrO are analyzed by the linear regression technique in two ways, for example : TiO(2)(o)=C(1)x(1)+C(2)x(2)+ ······(2) and TiO(2)(o)=aMgO+bFe(o) +cK(2)O ······(3) These methods are found useful in discriminating rocks of different ongm and in distinguishing the characteristic behavior of each component. Results of calculations (Tables 31 to 34) support the conclusion reached by calculations based on the contents of major components.