A Preliminary Study of the Gold Content of Byzantine Coins and a Possible Link to the Supernova of Year AD 1054

Abstract

A series of 11 Byzantine gold coins were investigated, including two examples of an extremely rare type called histamenon “stellatus”, from around the reign of the Byzantine emperor Constantine IX Monomachos (AD 1042–1055). The methods applied were X-ray fluorescence spectroscopy (XRF), specific density measurement (SD), and scanning electron microscopy (SEM). The debasement (decreasing gold content) of the Byzantine nomisma gold coinage during the 11th century was demonstrated. A method combining XRF and SD measurement combined with a graphical presentation/analysis called a ternary plot was also demonstrated. The measured gold content of the 11 coins was corrected for the possible “outwashing” effect and a potential cleaning of ancient gold. A model for the estimation of the gold content of Byzantine histamenon nomisma gold coins from the period AD 1020–1118, based on the specific density (SD), was derived. It was demonstrated that two analyzed histamenon “stellati” coins likely were minted around AD 1054–1055, possibly during the same period as the occurrence of the supernova SN1054, known as the Crab-nebula. It is further discussed if the gold content and size of the stars shown on those coins can be correlated to the visibility of the supernova from June AD 1054 to January AD 1055.

1. Introduction

The devaluation of the Byzantine histamenon nomisma gold coins minted in the Byzantine empire (presumably Constantinople) around and during the reign of emperor Constantine IX Monomachos (AD 1042–1055) was investigated and a basic equation to calculate the gold content via the specific gravity of these coins was developed. The possible age of two extremely rare histamenon gold coins, minted during the reign of the emperor, were investigated, and a possible link between these coins and the appearance of the supernova SN1054, creating the Crab nebula, is discussed. The method applied here focused on the analysis of the gold content of the coins, and it is hypothesized that an analysis of the gold content of the coins can help obtain a more detailed estimate of the actual age of coins from that period, as a devaluation occurred during that period [1,2,3,4]. Due to the reduction in the gold content in the nomisma coins during the reign of Constantine XI, it is hypothesized that a high gold content could be an indicator for the coin being minted early in the period and vice versa. The popular name for this type of coin is histamenon “stellatus”, meaning “the histamenon star coin”, and a popular assumption is that one of the two stars on the obverse side of the coin illustrates a supernova in 1054, (right star), which was not recorded in European sources; the left star is often interpreted as the planet Venus [5]. If the assumption regarding the supernova is correct, it would place the year of minting of the coin to not earlier than July–August 1054 AD. This rare type of coin is in the literature often referenced as “Sear 1831”, after the standard reference book about Byzantine coinage by David R. Sear (see [6], p310, no 1831). Four types of histamenon nomisma gold coins are known from the reign of Constantine IX ([6], Sear 1828–1831), but it is not certain in which order they were issued, so the full time span 1042–1055 is possible (see Figure 1).

The 11 specific coins analyzed in this preliminary study were acquired from personal collections, known to the author. Two extremely rare “stellati” coins, acquired from Jean Elsen et ses fils, s.a, Brussels, BE, on 21 January 2001 and from Stockholms Auktionsverk [7] in 2024 were part of the investigation (see Figure 2). The provenance and handling/cleaning of the coins is not known, and the lack of proper traceable provenance is a general problem with ancient coins.

A period with debasement of Byzantine gold coinage started during the reign of emperor Michael IV the Paphlagonian (AD 1034–41) and accelerated during the following decades [1,2]. The debasement primarily occurred due to the addition of silver and, from the reign of Constantine IX (AD 1042–55) until around AD 1069, the amount of silver in the alloy was increased by approximately 0.4 wt% per year and the purity of the histamenon nomisma fell from approx. 90 wt% to 70 wt% (21 to 17 carats) [2]. Emperor Constantine IX issued four different types of coins during his 14 year long reign [6]. After the reign of Constantine IX, the gold content of the histamenon nomisma stayed in the range of 70–78 wt% until Emperor Romanus IV (AD 1068–1071), where it fell below 70 wt%. For details about the (very complex) political background for the debasement, see [2]. Figure 3 shows the estimated gold content of a series of published gold content values around the reign of Constantine IX found in the literature [2,8]. The rate of debasement is estimated in the literature to be in the range of 0.6–1%/year around year 1050 [9,10]. Despite some deviations and obvious uncertainties in estimating the age of coins, a decline in gold content is clearly visible. It is estimated that for the year 1054, the mean gold content is expected to be in the order of 81–83 wt%.

The situation regarding the debasement was remedied by Emperor Alexius I in AD 1092, by the introduction of a gold hyperpyron of around 20 carats [2].

2. Materials and Methods

A series of 11 available gold coins of histamenon nomisma type from private Scandinavian collections were included in the analysis of the debasement process during/after the reign of Constantine IX. The analyzed coin types are listed in

Table 1. Investigated coins of type histamenon nomisma. A,B refer to the histamenon “stellati” coins investigated (shaded section).

Coin No.	Emperor	Period	Sear Ref.
1	Constine IX, type I	1042–1055	Sear 1828
2	Constine IX, type I	1042–1055	Sear 1828
3	Constine IX, Type II	1042–1055	Sear 1829
4	Constine IX, B, type IV	1042–1055	Sear 1831
5	Constine IX, A, type IV	1042–1055	Sear 1831
6	Constine IX, Type III	1042–1055	Sear 1830
7	Isaac I	1057–1059	Sear 1843
8	Constine X	1059–1067	Sear 1848
9	Romanus IV	1068–1071	Sear 1861
10	Michael VII	1071–1078	Sear 1868
11	Alexius I	1081–1118	Sear 1912

. Every investigation of the ancient coins conducted in this study was performed under the strict rule that the coins were not to be harmed in any way, so any removal of dirt, cleaning with abrasive materials, cutting etc. was strictly prohibited.

In this preliminary study, two measurement methods were combined. The coins listed in Table 1 were analyzed both by the X-ray fluorescence method (XRF) [12,13,14,15,16,17] and the specific density method [18,19,20,21]. A series of scanning-electron-microscopy measurements were additionally performed on one coin.

The use of an XRF instrument enables fast and easy measurements of a long range of elements, where the use of the specific density method gives only the specific density of the sample and not the elemental composition. But the specific density method can be combined with the ternary graphical method [22] by relying on the known element composition. The method presented here is only valid for alloys of gold, silver, and copper, which are the main elements in Byzantine gold coins. A weak point of this method is that trace elements are neglected. The X-ray fluorescence measurement (XRF) technique is thus used in combination with a graphical method (nomogram), employing a ternary plot for Ag-Au-Cu alloys to determine the apparent specific density of the coins analyzed. The real bulk specific densities (SD) of the coins were measured by immersion of the specimen in a liquid and recording the apparent change in weight as well as the dry weight of the sample. The bulk specific density combined with the application of the ternary plot/nomogram method was then used to identify the bulk gold content. The results were then compared to estimate a possible outwashing of copper and silver from the surface, as a higher surface gold content was to be expected. The procedure is illustrated in Figure 4.

Both X-ray fluorescence spectroscopy and the determination of the specific density of a solid sample like a coin are well established methods and are briefly introduced in the following sections.

2.1. X-Ray Fluorescence Technique

The X-ray fluorescence measurement is an easy to use non-destructive/noncontact measurement method to analyze the composition of materials such as alloys. See [13,14,15,16,17] for a general introduction to the subject. The fluorescence spectrum is a fingerprint of the elemental composition of the test object; however, the X-rays probes only the surface area down to a depth of approximately 100 µm (depending on the acceleration voltage applied) and thus should not be applied to analyze bulk materials thicker than that [17]. It has been reported in the literature that results obtained with the X-Ray fluorescence method in general should be considered with some care, as it is known that if the gold content of ancient object is around or below 80%, an effect called “outwashing” may take place [9,20,22]. The same case applies to cleaned gold coins in general, according to numismatist expert Robert Eberlein from the company DEGUSA (personal communication), as cleaning agents can dissolve/remove minute amounts of silver and copper on the surface of coins, thus enhancing the surface gold content. Trace elements such as Fe, K, Mn will often be present, but usually only in very small quantities. The coins in this analysis were treated as a three-element system consisting of Au-Ag-Cu [22]. The measured gold content, in case outwashing on the surface has occurred, is expected be higher than in the bulk material itself. An indication of outwashing would thus be that the specific density (SD) of the coin would differ (be larger) from the value expected from its measured surface elements. See [18,20] for a detailed treatment of the topic.

The X-ray instrument used in this investigation is a Shimadsu EDX−8100 X-ray fluorescence instrument (Kyoto, Japan) [23], located in a controlled laboratory environment at the Centre for Industrial Electronics, University of Southern Denmark (see Figure 5). The serial number is Q26455700005. The instrument follows a regular calibration schedule. A spot size of 10 mm was applied to ensure averaging over a parge surface. The instrument used all XRF-measurements on the 11 Byzantine coins investigated. All tests were conducted at room temperature (20 °C) and after an instrument warmup time of 1 h. As reference values for the tests, two Danish 10 kr gold coins with known gold content (AD 1908–09, 900‰, 21.6 carat) were analyzed, and the average measured gold content (90.7%) was used as reference for 900‰ [24].

The raw results for one of the tested Byzantine coins are, as an example, shown in

Table 2. Example of a quantitative result report from the XRF instrument. The sample under test was the “stellatus” coin A.

Analyte	Result	3-Sigma	Proc. Calc.	Line	Int. (Cps/uA)
Au	83.938%	0.207	Quan-FP	AuLa	2491.9247
Ag	11.917%	0.118	Quan-FP	AgKa	152.4824
Cu	3.426%	0.025	Quan-FP	CuKa	277.0127
Fe	0.218%	0.006	Quan-FP	FeKa	9.6112
K	0.190%	0.023	Quan-FP	K Ka	0.0877
Cl	0.154%	0.049	Quan-FP	Cl Ka	0.0237
Ca	0.081%	0.012	Quan-FP	CaKa	0.0540
Mn	0.075%	0.011	Quan-FP	MnKa	2.4158

and Figure 6.

The gold content measured in this example was 83.9w% with a 3σ value spread of 0.2, as shown in Table 2. This, however, should be corrected for the instrument offset (0.7%), so the value to be used is 83.2%.

2.2. Bulk Specific Density via XRF Data

The expected specific density (SD) value of a coin composed of Au, Ag, and Cu can as described be determined by a graphical method using a ternary coordinate system designed for Au, Ag, and Cu elements, with given curves for specific density (SD) values [22] (see Figure 7). Each axis has gridlines: Ag going downwards from left; Cu going upwards from below, and Au going horizontally left from the right axis. The intersection gives the specific density of the alloy. The example shown in Figure 7 has 70 wt% Au, 20 wt% Ag, 10 wt% Cu. The resulting SD is 15 g/cm³.

The measured bulk specific density can be used to find the gold content via the application of the inverse use of the ternary plot, based on the previously recorded XRF values.

2.3. Bulk Specific Density via Volume/Weight Estimation

The standard method of obtaining the specific density of a material is by applying the well-known principle of Archimedes to determine the volume of the sample; see [9,20] for detailed descriptions and uncertainty discussion. The method required a precise balance to measure the weight of the sample and the weight loss. The analytical balance used was a Mettler Toledo MS204TS/00 with four digits, serial no. B948663196 (Greifensee, Switzerland) [25]. The instrument was within its current calibration period. The apparent weight loss during the immersion of the sample in a liquid was used to calculate the volume (see Figure 8). The samples in this investigation were submerged in a liquid, namely ultrapure water (conductivity < 0.06 µS/cm) with a specific density SD_liquid of 0.9982 g/cm³ @ 20 °C/1 atm [26]. The basket holding the sample was made of thin isolated copper wire with a submerged volume V_wire of 0.41 mm³. The recorded weight loss M_loss, due to the updrift of the submerged volume, was used to calculate the total displaced volume. The net sample volume Vsample was calculated by subtracting the basket volume V_wire. See Equation (1).

$$V_{sample}={\displaystyle \frac{M_{loss}}{{SD}_{liquid}}}-V_{wire}$$

(1)

The bulk specific density SD_sample is calculated via the calculated volume V_sample Equation (1) and the measured weight of the sample M_sample. See Equation (2).

$${SD}_{Sample}={\displaystyle \frac{M_{sample}}{V_{sample}}}$$

(2)

The SD_sampl calculated in this example is 16.53 g/cm³. For further information regarding the accuracy of the volumetric measurement technique, see [18].

2.4. Bulk Specific Density Estimation via XRF Data

The measured XRF gold content can, as addressed above, be influenced by the phenomenon of outwashing. In the event that the specific density determined by the XRF values for Au-Ag-Cu is higher than the bulk specific density of the coins, it is likely the case that outwashing has occurred, and the measured gold content should be corrected as described below. The calculated bulk specific density SD_sample (Equation (2)) is used as an input to the ternary plot system, together with the Au, Ag, and Cu values from the XRF analysis, as described in detail in the next section. The method is illustrated by using data obtained for one of the coins under testing. The volume V_sample was estimated to 0.2664 mm³ and the weight M_sample measured to 4.4049 g. The weight loss was measured to 0.2668 g, resulting in a bulk specific density of 16.53 g/cm³ via Equation (2).

Under the assumption of having a system with three types of metallic elements, the specific density of the system can be estimated via the nomogram shown in Figure 9. The intersection of the measured XRF values (blue intersection) is approximately at 16.8 g/cm³, which is higher than determined by the specific density method, thus indicating outwashing. The correct intersection in the ternary plot must be in on the bulk SD 16.53 g/cm³ line (purple area). The bulk silver and copper values must thus be increased (to compensate for the outwashing), meaning that their constant-value curves move the grid lines (orange arrows) downward. The resulting gold content (dotted orange line) represents a value of approx. 82.8 wt%. This corresponds to the listed values in [21].

A condition for the use of the ternary system is that the other elements present are negligible. The values for Fe, Ca, Mn etc. in this example are all at or below the 0.2 wt% level and are considered negligible.

2.5. SEM Analysis

The use of a scanning electron microscope (SEM) to analyze the gold content was investigated. Access was granted to the SEM in the cleanroom at the Mads Clausen Institute at the University of Southern Denmark and technical support was kindly given for the analysis of one coin. The coin selected for test was coin 4 in Table 1 (Constantine IX, type IV, B). The SEM used was of type Hitachi S-4800 (serial number HI-9135-0002; Tokyo, Japan) equipped with Energy-dispersive X-ray (EDX) material analysis for quantitative analysis of samples providing detailed information of a sample’s chemical composition [27]. The main test parameters were acceleration voltage 20.0 kV, working distance 15,000–15,500 µm, and amplification 25–250× (see bellow). Before insertion in the instrument, the sample was degreased with isopropanol (a requirement) and mounted inside a fixture in the vacuum enclosure of the instrument. In all, 13 measurements were made at different areas of the sample.

3. Results

3.1. Specific Density (SD) Estimation via XRF/Ternary Plot

The coins listed in Table 1 were analyzed both by the XRF method and the specific density method/ternary plot method. The use of the specific density method combined with the ternary graphical method rely on the known element composition, and the method presented is only intended for alloys of gold, silver, and copper, as trace elements are neglected.

The example in Section 2 showed a single measurement giving a gold content of 83.9 wt% with a 3σ value spread of 0.2 wt%. The coin was in total analyzed eight times with an X-ray aperture opening of 10 mm, and each side of the coin was measured four times. The differences between the sides of the coins were in the order of 0.2 wt%. The averaged measured gold content was 83.6 wt% with a standard deviation of 0.3 wt%. The instrument offset was, as described in Section 2, determined to be 0.7 wt%. The resulting gold content of that coin therefore is considered to be 82.9 wt%. A similar analysis for the other “stellatus” coin (no.4, B) in Table 1 resulted in a gold content of 84.0 wt%.

The other nine Byzantine coins listed in Table 1 were analyzed following the same procedures as described in Section 2. The measured XRF gold, silver, and copper surface values are listed in

Table 3. Measured main element composition of the tested coin via XRF, the corrected Au values, and the estimated corrected specific density (Tern. SD) via the tertiary plot method and corrected for the instrument offset. A, B = histamenon “stellatus” (shaded section).

Coin no.	Emperor	Period [AD]	Au [wt%]	Ag [wt%]	Cu [wt%]	Au (corr.) [wt%]	Tern. SD [g/cm3]
1	Const. IX	1042–1055	92.7	5.2	1.0	92.0	17.93
2	Const. IX	1042–1055	95.5	1.9	1.6	94.8	17.90
3	Const. IX	1042–1055	91.0	6.8	1.4	90.3	17.54
4	Const. IX, B	1042–1055	84.7	11.8	2.7	84.0	16.97
5	Const. IX, A	1042–1055	83.6	11.9	3.4	82.9	16.54
6	Const. IX	1042–1055	93.5	5.0	1.0	92.8	17.88
7	Isaac I	1057–1059	78.5	16.5	3.7	76.8	16.15
8	Const. X	1059–1067	82.9	14.0	2.4	82.2	16.42
9	Romanus IV	1068–1071	76.8	19.8	2.6	76.1	15.73
10	Michael VII	1071–1078	75.4	20.7	1.9	74.7	15.63
11	Alexius I	1081–1118	88.9	7.8	2.9	88.2	17.31

, together with the estimated specific density values derived from the ternary plots (Tern. SD). The correct gold content was, due to the instrument offset, expected to be 0.7% lower. The content of the elements silver and copper therefore must be increased, but the ratio is uncertain, indicating a weakness by the XRF method, if no rigorous calibration sequences are performed with calibration grade gold, silver, and copper samples (not available for this preliminary study).

3.2. Specific Density (SD) and Gold Content Estimation via Volume, Weight and Ternary Plots

All analyzed samples were submerged into ultrapure water and the volumes were determined via the method described in Section 2. The weight was measured and the bulk specific density calculated via Equation (2). These bulk specific density values were used as input to the ternary plots, as shown in the example in Figure 8. The intersection of the gold, silver, and copper lines must meet at the line for the bulk specific density, giving bounds on the possible variation. The combination of these boundaries with the values obtained by XRF enabled the estimation of the actual gold content, as described in Section 2. The resulting values for the coins investigated are listed in

Table 4. Measured volume and weight of the tested coin, the bulk specific densities, and the estimated gold content via ternary plot analysis.

Coin No.	Emperor	Period	Vsample [mm3]	Msample [g]	Bulk SD [g/cm3]	Au [wt%]
1	Const. IX	1042–1055	232.4	4.2631	18.3	93.4
2	Const. IX	1042–1055	245.0	4.3461	17.9	91.3
3	Const. IX	1042–1055	250.8	4.4011	17.6	89.7
4	Const. IX, B	1042–1055	259.8	4.4053	17.0	84.0
5	Const. IX, A	1042–1055	266.4	4.4047	16.5	82.8
6	Const. IX	1042–1055	242.0	4.3279	17.9	91.8
7	Isaac I	1057–1059	275.1	4.4432	16.2	78.3
8	Const. X	1059–1067	268.9	4.4139	16.4	80.5
9	Romanus IV	1068–1071	276.9	4.3557	15.7	75.4
10	Michael VII	1071–1078	279.7	4.3727	15.6	73.6
11	Alexius I	1081–1118	250.0	4.3270	17.3	88.1

below. See Section 3.4 for a graphical presentation of the data. For a general in-depth analysis of estimating the composition of ancient coins and the specific density methods, see [8,9,10,11,12,13,14,15,16].

3.3. SEM Analysis of “Stellatus” Coin B

The histamenon “stellatus” coin B (coin no. 4) was also investigated by scanning electron microscopy (SEM) with element analysis capability at the Mads Clausen Institute at the University of Southern Denmark (Figure 10). A series of 13 measurements were made, 12 around the surface and one inside an apparently recent scratch/grove on the surface. The surface gold content values measured were in the range of 92–96 wt%, quite higher than for the XRF measurement, but those measurements were obtained with a measurement spot diameter of 10 mm, resulting in an averaged surface value.

The SEM technique generate spot data and is expected to be very dependent on the location of the actual measurement spot. See

Table 5. SEM surface scan data from histamenon “stellatus” coin B. 13 measurements were performed. Results are rounded to 1 decimal. The shaded area (no. 12) highlights the data from a scratch on the surface of the coin.

Meas. No.	Location	Au [wt%]	Ag [wt%]	Cu [wt%]
1	Surface	91.6	5.9	1.5
2	Surface	65.5	0.5	3.0
3	Surface	93.2	4.7	2.1
4	Surface	95.4	3.1	1.6
5	Surface	95.6	3.6	0.9
6	Surface	94.6	4.3	1.2
7	Surface	95.4	3.5	1.2
8	Surface	94.0	4.8	1.3
9	Surface	95.4	4.5	1.7
10	Surface	89.4	4.6	1.2
11	Surface	86.4	11.3	2.3
12	Inside scratch	85.0	10.8	2.3
13	Surface	92.7	6.2	1.2

, Figure 11 (measured values) and Figure 12 (SEM spectrum). The surface measurements show an outlier (no. 2) and in general values in the 92–96 wt% area for the Au content, indicating quite a large variation over the surface area. The gold value measured inside a scratch/grove on the surface (no. 12) showed a value of 85.0 wt% (see Figure 12). Measurement no. 2 is considered an outlier/polluted due to the exceptionally large difference to the other measurements performed, and the emission spectrum from that measurement showed a carbon content of 65%, indicating a polluted area. The penetration depth of the SEM with 15–20 kV acceleration voltage is estimated to be in the range of a few microns, making the measurements very sensitive to surface conditions.

The 85.0 wt% gold content obtained by measurement 12 from inside the scratch is applied in this study. It is, however, not certain that the value represents the bulk specific density of the coin, as it is not known if the scratch reached a sufficient depth, where outwashing has not occurred.

3.4. Comparison of Results

The gold values obtained via the XRF-based method and the specific gravity method are shown in Figure 13. It can be observed that, for most coins, the XRF-based gold content is higher than the estimated bulk gold content. This can indicate that outwashing has occurred on the surfaces of some of the coins. For two coins, both methods gave the same results, indicating that outwashing might not have occurred, as the bulk specific density value is not (or only to a very low degree) affected by a potential outwashing of gold from the surface. Therefore, this method is considered to be more generally appliable for determining the gold content of ancient coins. The method, however, is more difficult to apply as the measurement of the volume of the sample requires a complicated measurement setup. The XRF method further has the advantage that can be applied to estimate the elemental content of the sample and the presence of trace elements. The determined gold wt% values are presented in Figure 13, together with a 2. order curve fitted to the bulk values to estimate the gold content of coins from that period. The obtained SEM-value for “stellatus” coin B is also shown (*).

$$Au=-0.916{SD}^2+38.4SD-303$$

(3)

where Au is the gold content in wt% and SD is the specific density of the coin in g/cm³.

It can be observed that the second order best fit to the specific density (SD) values quite match the values obtained by applying the XRF-method. Both “stellati” coins are flat and not cup-shaped, which is noteworthy as the shape changed during the reign of Constantine IX [8,28].

A series of coins from around the reign of Constantine IX Monomachos (AD 1042–1055) were analyzed by a combination of X-ray fluorescence, the measurement of specific density, and the use of a tertiary plot method to estimate the gold content of the coins. The methods have comparable results, but the specific density method is considered the most reliable and applicable. Both the XRF and the SEM methods rely on clean and unaltered surfaces, which rarely is the case with ancient coins, as outwashing of silver and copper from surfaces is most likely to occur when gold coins have been buried in soil for centuries.

4. Relation to the Supernova SN1054

On 4 July AD 1054, a new bright object was visible in the sky; a star had turned into a supernova and the Crab nebula was born [29]. This event has been linked to the presence of one of the stars on the “stellati” coins of Constantine IX Monomachos. See [5,11,30] for a detailed discussion about this possibility. The idea presented in this paper is that it might be possible to link the known debasement of histamenon nomisma coins from the period around Constantine IX Monomachos to the appearance and intensity of the supernova SN1054.

4.1. Historical Records of the Supernova SN1054

The appearance of the supernova SN1054 on 4 July 1054 is not known to have been recorded in any European sources, but only in Chinese and Japanese records, see [5,30]. The event, recorded in the Chinese Songhshi and the Wenxian Tongao texts, describe that the new star appeared on the fifth lunar month of the first year of the Zhihe reign period [4 July AD 1054]. The date of the last reported sighting was on 6 April AD 1056. A ”guest star” was recorded in Japan and documented in the Meigetsuki text, which was compiled mid−13th century: “The new start appeared after the middle ten-day period of the (fourth?) lunar month in the second year of the Tenki reign period” [20–29 (?) June AD 1054]. A brief reference by Al-Muktar Ibn Butlan was identified in AD 1978, compiled by Ibn Abi Usaybi’a around AD 1242: “One of the well-known epidemics of our time is that which occurred when the spectacular (athari) star (kawab) appeared … in the year AH446” [12 April AD 1054–1 April AD 1055]. Ibn Butlan lived in Constantinople/Cairo around that time.

It is surprising that apparently no convincing written records exist from Europe. The reason for this is not known, but several theories have been issued, from no interest in celestial phenomenon (doubtful), over long-term cloudy weather all over Europe (very unlikely), to the Great Schism, the break of communion between the Catholic Church and the Eastern Orthodox Church in AD 1054. The appearance of the star could be interpreted as a disturbance of the astronomical status quo and hence God’s heaven as told by the Church. See [5,11] for detailed discussions of this subject.

The possibility that one of the stars on the “stellati” coins actually shows the birth of the Crab nebula in AD 1054 via a supernova explosion is by no means proven, but it has been speculated that the extra star on the right might represent that event [5,11,30].

4.2. Supernova SN1054 Visibility

The supernova of AD 1054 was visible from circa 4 July 1054 to ca. 6 April 1056. A possible relationship between the relative sizes of the two stars shown on the “stellati” coins was proposed by Filipovich et. al. [5,11,30], hypothesizing that the actual size of the stars on the “stellati” coins might represent a deliberate attempt to reflect the declining magnitude of the supernova SN 1054 over time. The supernova was first visible during daytime on 4 July 1054 (indicating a magnitude of −4.0), reaching a maximum magnitude of around −4.3 around 16 July 1054, outshining even Venus (magnitude −4.2) during daytime. The supernova was visible during daytime for around 23 days, and the daytime recognition ended on 27 July 1054. The last recorded sighting was 6 April 1056 (at night, magnitude + 5.5) [29]. Constantine IX passed away 11 January 1055. See Figure 14 for a sketch of the absolute apparent magnitude of the supernova SN1054.

4.3. Star Size Relationship to Magnitude

The left star on the “stellatus” coin A analyzed is 3 mm in diameter and the right star is 3.5 mm in diameter, so the diameter is about 17% larger, hypothetically indicating a supernova brighter than Venus. The “stellatus” coin B analyzed is 2.8 mm in diameter and the right star is 3.6 mm in diameter, hypothetically indicating an even brighter star. For a detailed discussion about star size, see [18,21].

Filipovich et al. identified 36 “stellati” coins and analyzed the size of the stars on the coins based on photograph from auction houses etc. (the two “stellati” coins analyzed in this article were not included in their analysis) [29]. The authors hypothesize that, while there are no exact historical records, it is likely that the coins intentionally were produced in order of decreasing size of the left star. The size of the stars therefore would represent the actual magnitude of the supernova (see [11] Table 1). The present study analyzed the ratio of the size of the two “stellati” coins A and B. The sorted ratios between the left and right stars on all 36 known coins are shown in Figure 15 (red dots) as well the ratio values of the “stellati” coins A and B investigated here (blue stars). The blue stars show the gold content of coins A and B. It can be observed that the “stellati” coins A and B published here have the largest star size ratios and, combined with the fact that their physical shape is flat and not curved (scyphate), this may indicate that they likely are older than those published by Filipovich et al., in [29].

The values give an indication of a relation between the ratio of the size of the stars and the gold content (highest gold content for the coin with the largest star size ratio), but with only two data sets available, there is no basis to conclude that there is a relationship between the star size ratio and the magnitude of the supernova. Future research aims to investigate this by gaining access to other “stellati” coins for analysis. In case the hypothesis describe above is correct, it would place the production of the analyzed “stellati” coins A and B at the beginning of the supernova observation period, as both apparently have a larger size ratio than all 36 identified “stellati” coins investigated in [29].

5. Conclusions

The decrease in the gold content of histamenon nomisma coins minted around/during the reign of Constantine IX was verified and a model for the estimation of the gold content of this type of coins, based on the specific density of the coins, from that period was developed. The gold content thus can be estimated via the bulk specific density of the coins, potentially eliminating the need for expensive XRF og SEM equipment for the determination of the gold content of those coins.

The analysis further resulted in estimating the gold contents of two extremely rare ”stellati” coins (Constantine IX, type IV, Sear 1831) and linking them to the supernova of 1054. After adjustment for surface gold enrichment, due to the anticipated outwashing of silver and copper from the surfaces, the gold content was estimated to 82.9 wt% and 84.0 wt% respectively. A SEM analysis of the second “stellatus” coin resulted in a slightly higher averaged estimated gold content of 85.0 wt%. The estimation of the gold content of the histamenon “stellatus” type coins A and B likely place the production of them in the same period as the supernova of AD 1054. This dating does not contradict the popular statement that one of the stars on the coin depicts the supernova appearing in AD 1054. The analysis of the possible link to the supernova SN1054 has so far been limited by the access to only 2 “stellati” coin.

The estimated gold content for those two coins indicate their time of production late in the reign of Constantine IX (around AD 1054–1055) and not at the same time, as their gold contents show a difference in the order of 1 wt%. Their shapes are flat, further indicating a rather early minting period, as the shape of the histamenon nomisma coins changed from flat to cup-shaped (scyphate) during the reign of Constantine IX. All other known histamenon “stellati” coins are cup-shaped, indicating a late time of production. The reason for this change in shape from flat to cup-shaped is not known [28] and both shapes can have coexisted for some time.

The extension of this preliminary analysis depends on access to other “stellati” coins, preferably both flat (if more unpublished samples exist) and cup-shaped samples, as well as optimizing the measurement techniques. This is ongoing. Future work should include the development of a low-cost portable device for field use, capable of easily measuring the specific density of a coin.