1. Census effectivities of bird species so far not treated in my previous papers are analysed by quantification method, type 1, of Hayashi (1961). 2. This method predicts
Yi from
X1,
X2, …
Xm. Where
Yi (called outsider) is the census effectivity of
i species given with quantitative value and
X1,
X2, …
Xm (called items) are given with categorized ranks of each
X property, such as, call note, song, body size, etc.. A value (called score) is given to each categorized rank by computer calculation so as to minimize
NΣi=1(
Yi-
Yi)
2 where
Yi is observed and
Yi is estimated (by
Yi=
X1+
X2+…+
Xm) values of the census effectivity for
i species. 3. In this study, four items and several categories in each item were selected as follows; 1) call: the synthetical value of strength and frequency of call note, 2) song: frequency of song syllables, 3) vegetation: forest profile which a bird inhabits or in which it was recorded in census, 4) size: the synthetical value of body size, home range size and velocity of bird's movement. 4. Category ranking of items was decided for each bird species based on these categorized items. As an exception, a vegetation category in which a bird species was mostly recorded by strong song was always given the first rank. In categorizing size item, three ranks for 50m and two for 25m observation radii were set. In call and song categories of some species, lower rank were selected from analytical base. 5. Computer analyses of quantification, type 1, were made and the results are shown by 50m and 25m observation radii. Observed
Yi and estimated
Yi census effectivities, with standard deviation σ of the difference are tabled. Observed census effectivities were calculated under standard conditions, which are: about 150 minutes after sunrise, fine weather and census speed 1.5km per hour. Scores given to each category rank are tabled with multi-correlations coefficient Ro between
Yi and
Yi. 6. In each cases, Ro was very significant (ca. 98%), and σ was rather small. Namely, 96% (=Ro
2) of variation of
Yi can be explained by these items and categories. Suppose that samples of
Yi-
Yi fit normal distribution, then 95.4% of samples should fall between 0+2σ and 0-2σ. Namely,
Yi can be estimated significantly with ±9.14% (50m) or ±12.56% (25m) differences. In conclusion, application of quantification theory to estimation of census effectivity is considered
View full abstract