The writer analysed the s easonal disease, beriberi, by using the means of auto-correlation method and applying the cyclic distribution. The purpose of this study is to find out the type of the seasonal periodicity and the distribution curve of beriberi in the main cities of Japan.
1) Analysis by means of auto-correlation method:
I calculated the auto-correlation coefficient of the d eath rate of beriberi in 5 years from 1926 to 1930. Then I draw correlogramm by using auto-correlation coefficient.
There is the clear period of 12 months which has wide amplitude in four cities in the Kansai district and two cities in the Kanto district. While in the cities in the south Kyushu district and in the Hokkaido district, there exist two periods, one is the period of 12 months which has wide amplitude and the other is the small period of 6 months even though they are not so clear.
2) Means of applying the cyclic distribution: I tried to express the death curve of beriberi b y composition of two periods which have one year normal cyclic distribution. But the phase of the second period slips down by half a year. The first period was supposed to have the cause of outbreak in summer while the second period in winter.
The temporal distribution curve is expressed by
In the case of Osaka City and F ukuoka City, I calculated the value of constants. Next if I calculate, by using the constants, the monthly death rate of beriberi in these two cities as the actual value, I can get the distribution curve shown in the graph, and find out that the actual measurement and calculation coincide well.
Judging from the above-mentioned fact, I can imagine that there will be some reason easy to cause beriberi death even in. winter in some district.
The writer tried the time series analysis of the seasonal disease, beriberi to find out the type, of the seasonal periodicity and the distribution curve of, beriberi by using the means of auto-correlation method and applying the cyclic distribution.