The improvement of civil registration reguires continuous study rather than periodic efforts. More and better statistics, however, are urgently required to formulate development programs and planning. Data obtainable from the civil registration are usually marred by errors of omission which are difficult to correct. This study aimed at finding out the problems occuring when a set of crude birth registration data in a rural area is used. Data Sources of this study are: 1) For birth registration: government birth registration records obtained from myun office and other government offices. 2) For the actual number of births: birth and child records from the Kang Wha Community Health Project The study area is Sunwon Myun and Naega Myun in Kang Wha Gun, Gyunggido. The referrance period for the accumulated data is one full year: Jan. 1st 1975-Dec. 31st 1975 Major findings are as follows: If the number of registered births is compared with the actual number of births which occured in the target area, the former is far greater than the latter. The general assumption usually is, that the actual number of births exceeds the registered number of birth in Korea. The observation from this specific study in this specific target area, shows the opporsite trend. The number of births which were registered is 550. This is more than two times as much as the number of births which actually occured during the year of 1975 in the study area namely 256. The difference comes mainly from the fact that many cases of births from other areas were registered in the target area. In other words birth is not registered where it occured but where the permanent residence adress is. Among 550 births registered in the target area 66% did not occur in the target area. Only one third of all registered births were registered within the legal period for birth registration which is 2 weeks. 34% of the registered births actually occured in 1974, but were registered in 1975. In 55% of the cases a difference was observed between the actual date of birth and the registered date of birth. From the 256 births which occured in the target area, only 153 births (59%) were registered at the myun office and the remaining 130 births (41%) were not resistered there in the year of study. 6% of the 550 cases listed as registered have no seperate registration sheets. Nevertheless, they definitly have been registered in the birth list at the myun office. 3% of the 550 cases are not recorded in this list but have a separate registration sheet at the myun office. In conclusion, birth registration data have many errors and problems. Their usefulness as. a source for vital and other statistics should be reconsidered. A series of sound methological studies will be necessary to establish their actual usefulness. A continuous and permanent compulsory system of birth recording is needed.