マイワシ及びカタクチイワシの経年漁獲量の推算

Abstract

In the statistical yearbook compiled by Japanese Ministry of Agriculture and Forestry-called “Norin Tokei” for short-the annual catch of iwashi for the period of 1926 to 1950 are not broken down by species such as sardine, Sardinops melanosticta (T.& S.) and anchovy, En-graulis japonica (HOUTTUYN). For the purpose of estimating the catch during the above period we assumed that a multiple regression equation (2) might represent the relation between the annual catches of anchovy (Y), iwashi (G), and the output of processed iwashi products (P1, P2, …… ), statistics of which are available in “Norin Tokei”. For the statistics of Y during the period under study we modified and made use of anchovy catch in “Norin Tokei” for the years of 1915 to 1925, and of data from investigations conducted by the Imperial Fisheries Experimental Station in 1932, 1949 and 1950 in collaboration with prefectural fisheries agencies (Table 3). The statistics of G, P1, P2, ……were obtained from “Norin Tokei” with no modification (Tables 2 & 3). Applying these statistics in equation (2), we estimated the value of the coefficient of partial regression for each item by method of least squares and then tested the significance of the coefficients. The results of the present study have revealed that of all combinations one combining G (catch of iwashi) and P1 (boiled-dried iwashi) and one for P1 and P2 (dried-salted iwashi) show the highest coefficient of correlation between observed catches of anchovy and the estimate with signi. ficant (5% level) coefficients of partial regression. The equations of regression (6) and (8) were introduced therefrom. With help of equation (6) and G=X+Y, the annual catches of anchovy and sardine includ-ing round herring, Etrumeus micropus (T. & S.), were estimated for 1905-53 and the errors of the estimation were discussed from several points (Section VI and Table 4). Shown in Figures 1 and 2 are confidence intervals of the expected values of anchovy catch, and comparison of estimated catches of sardine with observed catches of sardine and iwashi. In Table 5 the annual catches of sardine and anchovy were tabulated for 1915-50, as intended for the present research, in unit 1, 000 metric tons, including observed catches which were modified by the authors as in Table 3, and estimated ones for years when not observed. Coefficients of multiple correlation between estimated and observed catches are 0.998 for sardine and 0.89 for anchovy (Table 6). The estimated values for sardine in Table 5 may be as much reliable as “Norin Tokei”, but those of anchovy are supposed to be useful only for tracing the annual trend of catch.