Observations of comet C/1652 Y1 recorded in Korean histories

2.1 Historical documents

For the Korean records of comet C/1652 Y1, we consulted four historical documents: Sillok, Silgi, Dilgi, and Bigo. Sillok is the official chronicle of the Joseon dynasty, meticulously documented over 500 years. Meanwhile, Silgi is a daily record maintained by the Seungjeongwon (Royal Secretariat), with extant records dating back to 1623, though it is estimated to have originated at the dynasty’s outset. Hence, Silgi covers half the period of Sillok but is four times greater in volume. Dilgi, akin to Silgi, is a daily log produced by the Sigangwon (Royal Educational Office of the Crown Prince). As Lee (2023) highlighted, Dilgi contains a great number of records on astronomical phenomena, including meteorological ones such as solar and lunar halos and unusual clouds (Bahk et al. 2022). According to the Seoungwan-Ji (Treatise on the Royal Astronomical Bureau) compiled by Ju-Deok Seong in 1818, astronomical observations were reported to both the Seungjeongwon and Sigangwon (Jeon 1974). This could explain the abundance of astronomical records in both Silgi and Dilgi, and why their contents are nearly identical. While Sillok and Silgi underwent one or two compilations (though not comprehensively), Dilgi is known as the original manuscript. However, Dilgi’s chronology is incomplete, as its creation commenced only with the nomination of a crown prince. The series of 13 crown princes is preserved at the Kyujanggak Institute for Korean Studies. Dilgi, like Silgi, was handwritten in Chinese characters; recently, portions of the series have been translated into Korean and published in print. In this research, we reference Hyeonjong’s Dilgi, pertaining to the era when King Hyeonjong was crown prince (i.e., 1650–1659). Lastly, Bigo spans from the Three Kingdoms Period (54 BC–AD 918) to the Joseon dynasty in Korea, cataloging a diverse array of astronomical events. Despite being a frequently cited resource in the study of Korean astronomical records, Bigo is noted for its brevity and occasional inaccuracies.

2.2 Astronomical terminologies

Korean historical astronomy closely mirrors that of China. For an in-depth understanding of the terminology related to astronomical and night-hour systems, refer to the study by Lee (2012, 2023) and Lee et al. (2012). Here, we provided a brief overview of the celestial sphere system and its angular units. The celestial sphere was divided into three yuan (hereafter, enclosure) surrounding the North Pole and four cardinal directions (East, West, South, and North) including each seven Su (i.e., lunar mansion); hence, it comprises a structure of three enclosures and 28 lunar mansions. These enclosures and lunar mansions comprise multiple constellations, including a namesake constellation for each enclosure and lunar mansion. For example, the mao (18th) lunar mansion in the West includes nine constellations, such as yue, juanshe, and tianyin, in addition to the mao constellation itself. Notably, the term xing was employed to denote various entities: constellations (e.g., can xing, i.e., Orion), stars (e.g., hegu da xing, i.e., Altair), and planets (e.g., shui xing, i.e., Mercury). As an example, Figure 1 presents a segment of the Cheonsang-Yeolcha-Bunyaji-Do, a traditional Korean star chart engraved in stone in 1395 (Rufus 1913). This star chart centers on the North Pole, with radial lines demarcating the 28 lunar mansions, and utilizes an equidistant projection method to depict stars. Particularly, a star located at the radial boundary of a lunar mansion is termed suju xing, serving as a determinative star for that mansion. For example, the determinative star for the can (twenty-first) lunar mansion is δ Ori (Figure 1).

Figure 1

A part of Cheonsang-Yeolcha-Bunyaji-Do (source: Kyujanggak Institute for Korean Studies) showing constellations referred to this study: (a) (tian)ce, (b) yujing, (c) canqi, (d) tianjie (Cheonjeol in Korean pronunciation, hereafter refer to tianjie A), (e) tianjie (Cheonga in Korean pronunciation, hereafter refer to tianjie B), (f) mao, (g) juanshe, and (h) daling constellations. The fifteenth, sixteenth, seventeenth, eighteenth, nineteenth, twentieth, and twenty-first indicate the regions of the kui, lou, wei, mao, bi, zi, and can lunar mansions, respectively. The constellation marked with a white circle in the eighteenth lunar mansion is yue constellation consisting of one star. For reference, the red dotted circle represents the circumpolar circle, and the constellations marked with red dotted lines are the beidou (Big Dipper) and can (Orion) constellations on the left- and right-hand sides, respectively.

The units xiao (small, in literal; 1/4 du), ban (half, in literal; 2/4 du), tai (large, in literal; 3/4 du), qiang (strong, in literal; +1/12 du), and ruo (weak, in literal; −1/12 du) represent fractions of du, an angular unit. For example, 7 du and qiang equals 85/12 du (which is 7 + 1/12 du), and 1 du is roughly equivalent to 0.9856° (= 360°/365.2575 du) before 1654 in Korea. Lastly, we determined the night length (i.e., from evening twilight time to morning twilight time the next day), equally divided into five intervals, and obtained the duration of each interval, named the geng (hereafter referred to as watch). We found that the mean duration of each watch was approximately 2.6 h for the period from December 19, 1652 to January 9, 1653. Assuming the observation time to be the midpoint of the watch recorded in the historical texts, unless indicated otherwise, the uncertainty in time estimation is approximately ± 1.3 h.

2.3 Korean records

According to Pienaar (1948), comet C/1652 Y1 was first sighted by Jan van Riebeeck on the evening of December 17, 1652 in Cape Town, South Africa. As illustrated in Figure 2, this comet was observed by a Dutch on the evening of December 16, 1652 in Recife, Brazil. On the other hand, Silgi records indicate that observations of comet C/1652 Y1 spanned 22 days from December 19, 1652 (the 22nd day in China, according to Kronk (1999)) to January 9, 1653 (the twentieth day in Europe, according to Knobel (1897, 1918)). The absence of Korean records in previous studies, such as the work of Kronk (1999), may be attributed to the comet being termed bai xing (white star, in literal) in Sillok, Silgi, and Dilgi, and classified as a ke xing (guest star, in literal) in Bigo. We provide English translations of the records on comet C/1652 Y1 primarily from Silgi and Dilgi, which offer more detail than Sillok and Bigo, in the Appendix. We also provide Korean original texts, which are written in traditional Chinese characters. These records are presented in consecutive order, starting with the letter A. In the Appendix, dates are converted to the Gregorian calendar following Han (2001), yet the angles and night hours retain their original measurements in du and watch, as documented.

Figure 2

Drawing of comet C/1652 Y1 observed by a Dutch on December 16, 1652 at Recife in Brazil (source: British Museum).

The remarkable characteristics of Korean documentation of comet C/1652 Y1 within each historical source are as follows. Sillok is characterized by a smaller number (14) of records and less detailed content compared to Silgi and Dilgi. Notably, Sillok lacks observation times. For example, Sillok’s record on December 26, 1652 simply notes that the comet moved to 1 du west of the juanshe constellation (A08). Dilgi, on the contrary, does not include any records for the period from January 1 to February 28, 1653; hence, there are no subsequent records regarding comet C/1652 Y1 from January 1, 1653 (A14–A21). In addition, Dilgi has no records of when the comet was unobservable (A04, A05, A09, and A11). Discrepancies are present between the documents; for instance, Sillok and Dilgi report that the comet trespassed the second and fifth stars of the daling constellation, respectively, on December 30, 1652, while Silgi has no record itself on this day (A12). In addition, Dilgi’s account on December 24, 1652 describes the comet’s size (i.e., brightness); as big (in literal) as the shui xing (i.e., Mercury), a detail not recorded in Sillok and Silgi (A06). On the contrary, Silgi’s account on December 31, stating, “At the 5th watch, the white star was still situated within the daling constellation and was 4 du away from the fifth star of the constellation,” is not recorded in Dilgi (A13). The records of Bigo are the most succinct of the four documents, and the last observed date differs, noting January 9, 1653 unlike Silgi, which records the last sighting on the eighth day. Specifically, Bigo merely reports, “On the dinghai day, the eleventh month, during the third year of King Hyojong’s reign (i.e., December 19, 1652), a white star appeared below the tiance constellation.” Its size was as big as the chen xing (also known as Mercury). It then gradually moved toward the northwest and reached the daling constellation, passing through the canqi, tianjie A, and tianjie B constellations (refer to Figure 1), as well as the mao and wei lunar mansions. It disappeared on the wushen day in the twelfth month (i.e., January 9, 1653)” (see Figure 3). Lastly, none of the documents mention observations of tail length, which is in contrast to other European records (e.g., Weigelius and Schiltero 1653; hereafter WS53). The comet C/1652 Y1 might be described as a white star (or classified as a guest star) in Korean documents because its tail was not visible to the naked eye. It is known that telescopes were not used during the Joseon dynasty, although they were first introduced in 1631 (Ahn 2009). In contrast, European observers may have used telescopes, as Arcieri did (Kronk 1999). For reference, the tail of the comet was reported to be at its longest on December 20, 1652 measuring 5° or 6° (Zolotova et al. 2018).

Figure 3

Highlighted text is the accounts on comet C/1652 Y1 recorded in Jeungbo-Munheon-Bigo (source: Kyujanggak Institute for Korean Studies). For translations, refer to the text.

3.1 List of observations

Table 1 summarizes the positions of comet C/1652 Y1 as extracted from Korean historical documents. The first three columns present sequential numbers (No.), the date of the record (Date) in the Gregorian calendar, and the observation hour (Hour) in watch units, respectively. The fourth column provides the reference point used to measure the comet’s position. For the identification of stars in Chinese constellations, as mentioned in the records, with modern stars, we referred to the study by Kim (2020). The fifth column lists the angular distances (Ang. Dist.) in units of du, measured from the reference point. If the direction (Dir.) is specified in relation to the angular distance, it is included in the sixth column. Our inferences based on the analysis are presented in parentheses in columns 4, 5, and 6. The final column notes the corresponding records from the Appendix. The symbol T in the parentheses indicates the record of a fan (trespass, in literal) event. According to the Seoungwan-Ji, such an event is recognized when the separation between two celestial bodies is 0.1 du, as noted in column 5.

The Korean observations offer several distinctive features. First, the initial observation on December 19, 1652 (No. 1) does not provide numerical data, simply notating the comet’s position in the southeast direction, referred to as xun. Second, the observations on December 20, 24, and 30, 1652 (Nos. 2, 3(a), and 10) documented fan events, where comet C/1652 Y1 was reported to trespass the third star of the yujing constellation ( λ , ψ , β Eri, and τ Ori), the upper star in the tianjie B constellation ( υ and ω 2 Tau), and the fifth star of the daling constellation ( η , τ , ι , κ , β , ρ , 16, and 12 Per), respectively; hence, reference points are β Eri, υ Tau, and β Per, respectively, as listed in Table 1. Third, the sole observation reporting the angular distance from the North Pole was made at the first watch on December 24, 1652 (No. 3(c)), which also detailed the comet’s emergence point within the tianjie B constellation (No. 3(d)) and described its brightness as comparable to Mercury. Lastly, records Nos. 6, 7, 8(a), 8(b), and 9 simply identify the juanshe ( σ , ν , ε , ξ , ζ , and ο Per) and daling ( η , τ , ι , κ , β , ρ , 16, and 12 Per) constellations as reference points without indicating specific stars. Conversely, records Nos. 11–16 and 17–21 precisely indicate stars within the daling constellation, such as the fifth and second stars (i.e., β and τ Per, respectively), as reference points.

3.2 Comparison with known orbital path

To scrutinize the Korean records of comet C/1652 Y1, we compared them with modern computational results, using the astronomical algorithms of Meeus (1998) and the DE406 ephemeris of Standish et al. (1997). We disregarded the value of ∆ T (the difference between terrestrial time and universal time) because it was nominal (approximately 46 s in 1652, as per Morrison and Stephenson (2004)) compared to the estimated observational time uncertainty of approximately 1.3 h detailed in Section 2.1. Factoring in the geographic context, we adopted the coordinates of Hanyang (contemporary Seoul), specifically 37° 33′ 59″ N latitude and 126° 58′ 58″ E longitude. We calculated the magnitude of Mercury on December 24, 1652 using Hilton’s (2005) formula and derived a brightness of −0.9 magnitude. For reference, the magnitudes of comet C/1652 Y1 were 4 and 5 on January 1 and 3, 1653, respectively (Kronk 1999).

We calculated the trajectories of comet C/1652 Y1 using Marsden’s (1983) orbital elements initially set to the B1950.0 equinox and compared them with the comet positions summarized in Table 1. Recently, several studies, such as Neuhäuser et al. (2021) and Martínez et al. (2022), showed that Keplerian orbits for comets can be determined just from historical observations using the software find_orb (https://projectpluto.com/find_orb.htm). We also attempted to obtain a Keplerian orbit from Korean observations using the same software. However, we could not obtain reasonable orbital elements, presumably because of uncertainties in the observation hour, measured angular separation, direction, etc. Hence, as an alternative approach, we obtained the orbital elements giving minimum residuals with Korean observations by adjusting each element in the study by Marsden’s (1983) orbital elements by up to 20%. Table 2 summarizes the orbital elements determined by Marsden (1983) and in this study. In the table, symbols T P , e, q, Ω , ω , and i represent the time of perihelion passage, eccentricity, perihelion distance, argument of perihelion, longitude of the ascending node, and inclination, respectively, all reduced to the equinox of J2000.0. Figure 4 maps the orbit of comet C/1652 Y1 from December 19, 1652 to January 5, 1653 within the equatorial coordinate framework. In the figure, the black solid and blue dotted lines represent the Chinese and modern constellation outlines, respectively, while the black rectangles are the comet positions on the dates noted in Table 1 but expressed in units of universal time (UT). For example, the midpoint of the first watch on December 25, 1652 (No. 05) is 19.2619 h in Korean standard time (i.e., UT + 9 h). This corresponds to approximately the 25.4th day in UT, as illustrated in Figure 4. The Chinese constellations are (a) tiance, (b) yujing, (c) canqi, (d) tianjie A, (e) tianjie B, (f) mao, (g) juanshe, and (h) daling. We aligned all coordinates with the J2000.0 standard equinox, and the proper motions of the stars were corrected to the date of the perihelion passage, 1652 November 13.153 UT, referencing the Yale Bright Star Catalog.

Figure 4

Orbital path of comet C/1652 Y1 calculated on the basis of the orbital elements of Marsden (1983). Horizontal and vertical axes are right ascension and declination in the equinox of J2000.0. Black rectangle symbols in the path are the positions of comet C/1652 on dates given in Table 1 but notated in units of UT. The Chinese constellations (black solid line) are (a) tiance, (b) yujing, (c) canqi, (d) tianjie A, (e) tianjie B, (f) mao, (g) juanshe, and (h) daling.

The record on December 19, 1652 states that at night the first watch, a white star was spotted below the tiance constellation, located at xun (No. 1). The term xun corresponds to a cardinal direction, specifically southeast within the octagonal compass system, or at an azimuth range of 112.5–157.5° when measured clockwise from true north. We determined the azimuth and altitude of the comet at the midpoint of the first watch (i.e., the 1.5th watch) on that date by transforming the apparent right ascension and declination obtained from orbital elements of Marsden (1983). The azimuth was approximately 119.6°, which corroborated the historical account, and the altitude was just above the horizon at approximately 0.7°. For reference, the azimuth and altitude of the comet at the end of the first watch are 132.7° and 14.0°, respectively. Directions associated with stars or constellations are deduced to correspond to positions on the celestial sphere as follows: north is directed toward the North Pole, and east refers to the clockwise rotation relative to each reference star or constellation, as depicted in Figure 1.

The account on December 24, 1652 indicates that the comet was located “below” and “above” the tianjie B constellation within the mao lunar mansion, as per Dilgi (and also Sillok) and Silgi, respectively (A06). The data represented in Figure 4 supports the description provided by Silgi, which places the comet “above” the tianjie B constellation. However, it is difficult to verify the observation from December 24, which reports the comet’s angular distance from the point of emergence as 43 du (No. 3(d)), because the reference point is ambiguous. The record on December 28 states that the comet had moved to 3 du east of the daling constellation according to Silgi (and also Sillok), and 2 du according to Dilgi, at the fifth watch (refer to A10). Upon analysis, the positioning at 3 du aligns more closely with the orbital path as outlined in Table 1 than 2 du positioning does, as demonstrated in Figure 5. In addition, the record on December 30 describes the comet as trespassing against the second and fifth stars of the daling constellation ( τ and β Per) in Sillok and Dilgi, respectively (refer to A12). Subsequent discussion confirms that the comet’s path indeed intersected with β Per, as indicated in Table 1.

Figure 5

Differences of angular separation between observation values (O) recorded in Korean historical documents and calculated ones (C) using orbital elements of Marsden (1983). Horizontal axis represents sequential numbers given in Table 1 and vertical axis represents O–C. Black thick vertical lines are O–C values for the duration of each watch and red cross symbols are mean values in each O–C.

Regarding the trespass events, on December 20, 1652 (No. 2), the angular separation between the comet and the third star of the yujing constellation, β Eri, was calculated at 3.38° ± 0.22 over the course of the fourth watch. However, the angular separation from the second star of the constellation (i.e., the third star when counted in reverse order), ψ Eri, was 0.87° ± 0.32. Therefore, it is likely that the third star of the yujing constellation indicates ψ Eri, as noted in Table 1 with the parentheses. Instances of stars being enumerated in reverse order within a constellation are found in other historical Korean documents (Lee 2023). On December 24, 1652 (No. 3(a)), the comet’s angular distance from the upper star in the tianjie B constellation, υ Tau, was 5.82° ± 0.08 over the duration of the first watch, and from the lower star, ω 2 Tau, it was 4.17° ± 0.16. Notably, the star nearest to the comet at that time was 37 Tau from the yue constellation, with an angular separation of 0.89° ± 0.15. This star is situated above the tianjie B constellation (as indicated by the white circle in Figure 1). Therefore, the record from December 24, 1652 should be interpreted as referring to “the upward star of the tianjie B constellation.” The angular separation between the comet and the fifth star of the daling constellation, β Per, on December 30, 1652 (No. 10), was 0.70° ± 0.04 over the duration of the first watch. If we presume that the comet trespassed 37 Tau and β Per at the first watch on December 24 and 30, respectively, the average angular separation for these occurrences is approximately 0.82°. This figure aligns more closely with Lee’s (2023) finding of approximately 1.0°, rather than 0.1 du explained in the Seoungwan-Ji.

In analyzing the differences between observed values (O) in Korean historical documents and calculated values (C) using Marsden’s (1983) orbital elements, we tabulated the O–C values for all relevant observations. These are illustrated in Figure 5, where the horizontal axis is indexed by the sequence numbers from Table 1, and the vertical axis represents the O–C values. The bold black vertical lines denote the range of O–C values calculated over each watch, in increments of 0.001 d, provided the comet and reference star were above the horizon. The red crosses mark the average of these O–C values for each observation. It is noteworthy that O–C values are positive when measured from the fifth star (Nos. 11–16) and negative from the second star (Nos. 17–21) of the daling constellation. The observations from January 5, 1653 (No. 21) and December 25, 1652 (No. 5) display the most substantial deviations at −4.09° ± 0.02 and +3.51° ± 0.12, respectively. These significant discrepancies could be attributed to potential misidentifications of the second star in the daling constellation, transcription errors within the historical documents, or inaccuracies within the assumed orbital elements provided by Marsden (1983).