It has been considered that 'exhausted eclipses' (日食旣) were total eclipses. However, modern precise calculations show that a significant fraction of such records are not realized to be total. Thus we doubt that the two concepts are equivalent. Here we investigate the meaning of 'exhausted eclipses' in the east-Asian history. We first find that eclipses of magnitude greater than 0.8 were regarded as 'exhausted eclipses' by a Korean astronomer of the 18th-century Choson dynasty. His notion was based upon the definition of 'exhausted eclipses' in the ephemerides of pre-modern Chinese dynasties. According to those ephemerides, the 'exhausted eclipses', whose magnitude is greater than 0.8, have the first contact at the western part of the solar disk and the fourth contact at the eastern part of the solar disk. A simple geometrical calculation shows that such cases really occur when the magnitude of eclipse is greater than 0.7. We pointed out that such an ancient definition might not be impractical for ancient astronomers, because the uncertainty of eclipse magnitude estimated by ancient Chinese ephemerides was 10% and the human sight has a spatial resolution of 1.2 arcmin, which is approximately one twentieth of the Sun's angular diameter.

In this study, solar eclipse records were investigated during the Three Kingdoms era of ancient Korea using astronomical calculations and numerical simulations. Under the condition that the solar eclipses were actually observed at the well known capitals of the Three Kingdoms, I investigated the probabilities that the optimal observation areas of the Early Silla, Goguryeo, and Baekje records would appear around Chinese continent. I found higher probabilities than those suggested by Park and La (1994), although the numerical values are still low, especially in the case of the Early Silla records. On the other hand, the probability that the optimal observation area of the Later Silla records will be present around South Korea is only 13.6%, although the area shows a good match with the known capital. I also analyzed the number distribution of the eclipse records for the Three Kingdoms (except for the latter Silla's) according to the observers' locations: at the optimal observation areas and at the known capitals. And then I compared with the number distribution of all eclipses observable from those locations. From the χ2-test, I found that the Goguryeo and Baekje records had better representation of their population distributions at the latter regions (χ2=27.93 and 205.5) than at the former ones (χ2=34.19 and 211.5). Therefore, it is difficult to conclude that the observers' locations during the Three Kingdoms period were either near China, as suggested by Park and La, or in the Korean peninsula, solely based on these results. It is thus recommended that more studies are required to confirm the real observers' locations during the Three Kingdoms era.