University of Virginia Library

APPENDIX VI

ECLIPSES IN THE REIGN OF EMPEROR WU

i. In Chien-yüan II (the second year of the period Chien-yüan), ii
(the second month), on the day ping-hsü, the first day of the month,
an eclipse of the sun is listed (HS 6: 2b, 27 Cb: 14a; Han-chi 10: 1b)
The "Treatise," ch. 27, adds, "It was 14 degrees in [the constellation]
K'uei." Hoang, Concordance des chronologies néoméniques chinois et
européenne, equates this date with the julian Mar. 21, 139 B.C. But
there was no eclipse on that date.

In the five years between the last preceding correctly recorded eclipse
in 143 B.C. and the next one in 138 B.C., there were 12 eclipses, of which
only one was visible in China.[1] That one occurred on July 8, 141 B.C.,
which date was, according to Hoang, Emperor Ching, Hou III, v, the last
day, yi-ch'ou. The day after yi-ch'ou is ping-yin, which was the first
day of the sixth month; Hoang's calendar might have been one day in
error; and someone, reading a partly illegible list of eclipses which
omitted the year-periods, might have misread "three" as "two," "six"
as "two," and misread or corrected "ping-yin" [OMITTED] to "ping-hsü" [OMITTED], thus
achieving this listing, so that it quite possibly represents a genuine
observation.

The three principal stars of K'uei were then in 344°, 343° and 347°
R.A.; Oppolzer calculates the longitude of the sun at the eclipse of 141
B.C. as 101° = 103° R.A. The heavenly location of the eclipse does
not thus represent any observation.

ii. On Chien-yüan III, ix, ping-tzu, the last day of the month, a second
solar eclipse is recorded (6: 3b; Han-chi 10: 5a). Hoang equates this
date with Nov. 1, 138 B.C., for which Oppolzer calculates his solar eclipse
no. 2545. It must have been quite conspicuous in Ch'ang-an; Oppolzer
charts the path of centrality as passing through the present Urga, Mongolia,
and Korea.

HS 27 Cb: 14a also records the eclipse and adds, "It was 2 degrees in
Wei3." The principal star of Wei3, μ Scr, was then in 219° R.A.; Oppolzer
calculates the sun as in 215° long. = 213° R.A.

iii. On Chien-yüan V, i, chi-szu, the first day of the month, a third
eclipse is recorded (HS 27 Cb: 14a; Han-chi 10: 8b). The "Annals,"
ch. 6, do not list this eclipse. Hoang equates this date with Feb. 16,
136 B.C., but there was no eclipse on that date.

In the four years between the previous eclipse in 138 B.C. and the
next correctly recorded one in 134 B.C., there were nine solar eclipses,
of which two were visible in China.[2] Oppolzer charts the moon's umbra
in the eclipse of Apr. 15, 136 as passing through northern Siberia; calculation
by the method in Neugebauer, Astronomische Chronologie, shows
that in Ch'ang-an this eclipse was invisible, in the present Peiping it
reached a magnitude of 0.07 at 3:26 p.m., local time, and at the present
Ning-hsia, the ancient So-fang, it reached a magnitude of 0.15 at 2:44
p.m., local time. Calculation also shows that the eclipse of Apr. 4, 135
was invisible in Ch'ang-an, but reached a magnitude of 0.08 at 6:32
a.m., on Apr. 5, in the present Peiping.

Apr. 15, 136 was, according to Hoang, Chien-yüan V, iii, the last day,
ting-mao. Ting-mao is the second day before chi-szu. Since Hoang's
calendar may be two days in error, and since "three" in Chinese may
easily be misread as "first," the eclipse we are considering happened on
Apr. 15, 136 with a very high degree of probability. It seems to have
been reported from outside the capital. Yet at the longitude of Ch'ang-an,
the southern limit of visibility for this eclipse was 35° N, only a short
distance north of Ch'ang-an, according to Neugebauer's elements, so
that the eclipse might easily have been reported from somewhere in
central Shensi. Since moreover absolute exactness cannot be claimed for
even the best astronomical computations concerning events two thousand
years ago, it is furthermore possible that this eclipse was actually visible
in Ch'ang-an.

iv. In Yüan-kuang I, ii, on the last day of the month, ping-ch'en,
a fourth eclipse is recorded (HS 27 Cb: 14a; Han-chi 11: 4b). The "Annals"
does not record it. Hoang equates this date with Mar. 25, 134.
But there was no eclipse on that day.

In the four years between the last correctly recorded eclipse in 138
B.C. and the next one in 134, only two eclipses were visible in China,
the first of which seems definitely to be the one denoted by the preceding
recording. The other eclipse visible in China, occurring on Apr. 4, 135,
22h.48m.GCT, and visible in China on the morning of Apr. 5, was on
Chien-yüan VI, iii, the last day, jen-hsü, according to Hoang. Jen-hsü
[OMITTED] may easily be misread as ping-ch'en [OMITTED] and "three" may equally
easily be mistaken as "two." Year-periods were not inaugurated until

114 or 113 B.C., so that the mistake of one year is equally intelligible.
The original record probably read the year "six," which was misread
"seven" before it was inserted into the list in the "Treatise," ch. 27.
Hence Apr. 5, 135 B.C. was probably the actual date of this eclipse.

Since the list in the "Treatise" seems to be that of the court astronomers
in Ch'ang-an, it is peculiar that this and the preceding eclipses must
have been entered into that list without having been visible in Ch'ang-an.
The previous eclipse may have actually been visible in the capital; the
southern limit of visibility for the present eclipse, according to Neugebauer's
method, was about 40° N in the longitude of Ch'ang-an; it was
visible at sunrise at T'ai-yüan.

v. In Yüan-kuang I, vii, a day before the last day of the month,
kuei-wei, a fifth eclipse is recorded (6: 5b; 27 Cb: 14a; Han-chi 11: 4b).
The "Treatise" adds, "It was eight degrees in Yi4."

Hoang equates this date with Aug. 19, 134 B.C., for which day Oppolzer
calculates his solar eclipse no. 2555. He calculates the sun in
long. 142° = 143° R.A. Of the stars in Yi4, α Hydrae was then in 136°
R.A., α and γ Crateris in 139° and 145° R.A. There is thus a fair
approximation.

This eclipse must have been quite prominent; Oppolzer charts the
path of totality as passing through the modern Irkutsk and southern
Manchuria.

vi. In Yüan-so II, a sixth eclipse of the sun is recorded. HS 6: 10b
dates it iii, the last day, yi-hai. HS 27: Cb: 14a dates it ii, the last day,
yi-szu, and adds, "It was 3 degrees in Wei4," Han-chi 12: 3a dates it ii;
the last day, yi-hai. Hoang equates the date in the "Annals" with
May 6, 127, and that in the "Treatise" with Apr. 6, 127. He gives no
yi-hai day in the second month, so that the Han-chi's date is impossible.

Oppolzer calculates his solar eclipse no. 2570 for Apr. 6, 127, and
nothing for the other date, so that the recording in the "Treatise" is
correct. It is interesting that the Han-chi gives a dating partly like the
incorrect one in the "Annals" and partly like the correct one in the
"Treatise." Probably in the second century, when the Han-chi was
composed, the text of the "Annals" contained the reading now in the
Han-chi (possibly Pan Ku originally had an incorrect record before him),
and later someone, who knew that such a date was impossible, corrected
the text by changing the month, for the emendation of the month is the
easiest one and gives a seemingly correct result.

This eclipse was visible in the Mediterranean world, especially in
Babylonia. Oppolzer and Ginzel both calculate that the path of totality
passed through northern Sinkiang and just east of Lake Baikal. They

calculate the sun in long. 12° = 14° R.A. The principal star of Wei4,
35 Arietis, was then in 11° R.A.

In the period of 7 years between the preceding eclipse and this one, no
eclipses were visible in China.[3]

vii. In Yüan-so VI, xi, kuei-ch'ou, the last day of the month, a seventh
eclipse is listed (27 Cb: 14a; Han-chi 12: 9a). The "Annals" do not list
it; the Han-chi puts it at the end of the year, on the day kuei-yu. Hoang
gives no kuei-ch'ou day in the eleventh month of that year, but does give
a kuei-yu day as the 19th day of the month, Dec. 14, 124 B.C., at the
beginning, not the end of the year.

In the 5 years from the preceding to the next correctly recorded eclipse
in 122 B.C., there were 11 eclipses, of which 2 were visible in China.[4]
Calculation of the eclipse of Feb. 3, 124 B.C. shows that it was invisible
in Ch'ang-an and the present Peiping, but it reached a magnitude of 0.05
at 2:10 p.m. local time in the ancient So-fang, the present Ning-hsia,
which had recently been conquered. Such a small eclipse in so remote a
spot would hardly have been noticed. The eclipse of Jan. 23, 123 B.C.
was visible in Ch'ang-an and places south and east. It occurred in Yüan-so
VI, xii, the last day, kuei-ch'ou.

The record plainly points to the eclipse of Jan. 23, 123 B.C.; the error
in dating, reading xii as xi, is quite natural.

viii. In Yüan-shou I, v, yi-szu, the last day of the month, an eighth
eclipse is recorded (6: 14a; 27 Cb: 14a; Han-chi 12: 13). The "Treatise"
adds, "It was 6 degrees in [the constellation] Liu. According to the
calculations of Ching Fang [77-37 B.C.] in his Yi-chuan, when, as at
this time, the sun is eclipsed from its right side, his rule says that the
prince will lose a minister."

Hoang equates this date with July 9, 122 B.C. for which Oppolzer
calculates his solar eclipse no. 2582. He charts the path of this annular-total
eclipse as passing through central Shensi, just north of Ch'ang-an.
He calculates the sun as in long. 102° = 103° R.A.; the principal star of
Liu, δ Hydrae, was then in 101° R.A.

ix. In Yüan-ting V, iv, the last day, ting-ch'ou, a ninth eclipse of the
sun is listed (6: 21a; 27 Cb: 14b). The Han-chi does not list this eclipse.

The "Treatise" adds, "It was 23 degrees in [the constellation]
Tung-ching."

Hoang equates this date with June 18, 112 B.C., for which Oppolzer
calculates his solar eclipse no. 2606. He charts the moon's umbra as
passing through Suiyüan and Jehol, and calculates the sun's longitude as
83° = 82° R.A. The principal star of Tung-ch'ing, ν Gemini, was then
in 66° R.A.

In the ten years since the preceding eclipse and down to this eclipse,
there were 23 solar eclipses, of which only one was visible in China.[5]
This eclipse occurred on Aug. 19, 115 B.C. It reached a magnitude of
only 0.28 at 11:21 a.m. local time in Ch'ang-an, so might naturally have
been missed.

x. In Yüan-feng IV, vi, the first day, chi-yu, a tenth eclipse is listed
(27 Cb: 14b). This eclipse is not found in the "Annals" or in the Han-chi.
Hoang equates the date with June 24, 107 B.C., but there was no eclipse
on that date. Chu Wen-hsin, in his Li-tai Jih-shih K'ao, p. 30, suggests
the eclipse of Sept. 19, 107 B.C.

In the six years from the preceding eclipse to the next correctly listed
one in 96 B.C., there were 37 solar eclipses, of which five were visible in
China.[6] These were the eclipses on (1) Apr. 6, 108 B.C., Yüan-feng III,
iii, last day, yi-yu, which reached a magnitude of 0.32 at 3:29 p.m. at
Ch'ang-an; (2) Sept. 19, 107, Yüan-feng, IV, viii, last day, ping-tzu, which
was invisible in Ch'ang-an but reached a magnitude of 0.34 at the present
Peiping at sunrise; (3) July 19, 104 B.C., T'ai-ch'u I, vii, first day,
keng-yin; (4) Dec. 3, 103 B.C., T'ai-ch'u II, xi, first day, jen-tzu, which
reached a magnitude of 0.12 at 1:42 p.m. at Ch'ang-an; and (5) May
17, 101 B.C., T'ai-ch'u IV, iii, the day before the last, jen-yin, which
reached a magnitude of 0.45 at 4:20 p.m. at Ch'ang-an.

Of these eclipses, the one of Apr. 6, 108 is the most plausible one, for
chi[OMITTED]-yu may easily be mistaken for yi[OMITTED]-yu. If Hoang gauged incorrectly
the number of days in a month, yi-yu might have been the first day of the
fourth month. The eclipse of July 19, 104 was on the right day of the
month; chi-ch'ou [OMITTED] (which may easily be misread for chi-yu [OMITTED]) is the

day before keng-yin; but the year-period, year, and month present difficulties.[7]
The day of the eclipse of Sept. 19, 107, ping-tzu, presents
insuperable difficulties, and the eclipse was invisible in the capital, which
seems to be the locality indicated for the list in the "Treatise." In all
probability Apr. 6, 108 B.C. was the actual date of this eclipse.

It is interesting that from the beginning of Emperor Ching's reign in
154 B.C. down to the eclipse of June 18, 112 B.C., every eclipse that could
reasonably be expected to be recorded was actually listed, while here
within six years at least four eclipses seem to have been missed. Why
should this failure have occurred just before the calendar was rectified?

xi. In T'ai-shih I, i, yi-szu, the last day of the month, an eleventh
eclipse is recorded (HS 27 Cb: 14b; Han-chi 15: 1a). The "Annals" do
not list this eclipse.

According to the correction of Hoang's calendar in Chavannes, Documents
chinois, p. 71 (cf. n. 35.6), this date was Feb. 22, 96 B.C. Oppolzer
calculates his solar eclipse no. 2644 for Feb. 23, 96 B.C. and charts it as
visible in China. This mistake of one day may well have been that of
Hoang's calculations.

Chavannes found, on a tablet discovered by Stein in a Han watch-tower
in the desert, a record giving the cyclical day for the first day of the
twelfth month in T'ai-shih I, which shows that the intercalary month was
added at the end of T'ai-shih I, not in T'ien-han IV, as Hoang has it.
This change discovered by Chavannes gives a date correct within one
day for this eclipse. Neither Hoang, in his Catalogue des éclipses de soleil
et de lune nor Chu Wen-hsin, in his Li-tai Jih-shih K'ao, seems to have
noticed this necessary correction in Hoang's calendar. Hoang concludes
that no eclipse corresponds to this listing! This unexpected tallying
shows the essential correctness both of Han recordings and (within a limit
of about three days) of Hoang's calendar.

xii. In T'ai-shih IV, x, chia-yin, the last day, a twelfth eclipse is listed
(HS 6: 36b; 27 Cb: 14b; Han-chi 15: 3a). The "Treatise" adds, "It
was 19 degrees in [the constellation] Tou."

Hoang equates this day with Dec. 12, 93 B.C., for which Oppolzer
calculates his solar eclipse no. 2652 and charts the moon's umbra as
passing through Suiyüan and Chahar. He calculates the sun's longitude

as 258° = 257° R.A. The principal star of Tou, φ Sagitarii, was then
in 249° R.A. This eclipse was also visible in Athens, Memphis, and
Babylon.

Hoang lists this day, however, as the first day of the eleventh month, so
that one day probably needs to be added to some preceding month.

In the three years since the preceding eclipse, there were 7 solar
eclipses, none of which was visible in China.[8]

xiii. In Cheng-ho IV, viii, hsin-yu, the last day, a thirteenth eclipse is
listed (HS 6: 38a; 27 Cb: 14b; Han-chi 15: 11a. The Han-chi reads,
"the seventh month.") The "Treatise" adds, "It was partial, like a
hook, 2 degrees in [the constellation] K'ang. In the late afternoon [3-5
p.m.], the lower part of the sun was eclipsed from the northwest. In the
late afternoon, the eclipse was [also] over."

Hoang equates this date with Sept. 29, 89 B.C., for which Oppolzer
calculates his solar eclipse no. 2661. Calculation shows that this eclipse
reached a magnitude of 0.85 at 3:41 p.m., local time at Ch'ang-an; that
it commenced at 2:21 p.m. and ended at 4.53 p.m. The sun's longitude
was 183° = 183° R.A.; the principal star of K'ang, κ Virginis, was then
in 186° R.A. There was thus a close checking (except in longitude)
between the recording and calculation.

In the four years since the preceding eclipse, there were 8 solar eclipses,
of which only one, that of Oct. 11, 90 B.C. was visible in China.[9] Calculation
shows that this one reached a magnitude of 0.17 at sunrise in
Ch'ang-an.