Francis Stephenson#


In my research career, spanning some 40 years, I have largely concentrated on the application of early astronomical records (especially from the pre-telescopic period) to the investigation of problems in modern astronomy and geophysics. Despite the low precision of early observations, the lengthy time-span which they cover affords several advantages over more recent data. Examples include the study of very rare events (e.g. galactic supernovae) or long-term trends which cannot be adequately discerned over the short period covered by accurate modern observations (e.g. in Earth's rotation). It seems fair to state that I am widely regarded the leading world authority on this subject, which has become known as "Applied Historical Astronomy". I have contributed extensively to its growth from a somewhat arcane field of study to an important scientific discipline.

My research interests centre round the following areas of Applied Historical Astronomy: (i) Earth's rotation in the past; (ii) historical supernovae; (iii) the past orbit of Halley's Comet; (iv) solar variability. Additionally, I have done much research on (v) oriental star maps and also (vi) the accuracy and reliability of pre-telescopic observations.

I have published extensively on the above topics in books, research papers, conference proceedings and also popular articles - see list of publications below.

Most of the numerous early observations which I have analysed originate from only four cultures: Babylon, China (including Japan and Korea), Europe and the Arab dominions. An important feature of my research is the systematic consultation of historical documents in their original language, either personally (as in the case of Chinese and Latin texts) or with the help of close colleagues (for Babylonian and Arab material). In my view, such a procedure is essential if an objective assessment is to be made of early astronomical records.

I have a sound command of Latin and a good working knowledge of Classical Chinese. In particular, I am in regular touch with colleagues in China, both South Korea and North Korea, and also Japan. I have hosted several Royal Society K.C. Wong research fellows from China. I have made two visits to China under the auspices of the Royal Society, four visits to South Korea, and two to Japan. I have acquired some knowledge of Late Babylonian cuneiform - such as found in astronomical texts - but here I am able to rely extensively on the expertise of colleagues: especially Professor H. Hunger of the University of Vienna and Dr C.B.F. Walker of the British Museum. I have encouraged Dr S.S. Said, formerly of King Saud University, Riyadh, Saudi Arabia and other colleagues possessing an expertise in Classical Arabic, to translate and investigate a wide variety of Arabic texts dealing with eclipses, comets, etc.

A summary of both my past and current research investigations is as follows.

Earth's rotation in the past

Fluctuations in the rate of rotation of the Earth have been detected on time-scales from a fraction of a year to millennia. These are produced by a variety of mechanisms, both tidal and non-tidal. My principal researches in this subject have been made in collaboration with Dr L.V. Morrison and major papers were published in 1984 and 1995 (in addition to many other publications). In the first of these papers (Phil. Trans. Roy. Soc. A, 313 (1984), 47-70) we charted the decade fluctuations (usually attributed to core-mantle coupling) in hitherto unrivalled detail since A.D. 1620, mainly using occultation observations, and evaluated the necessary torques (between 10$^{17}$ and 10$^{18}$ Nm). Our results have inspired other authors to investigate correlations with the westward drift of the Earth's magnetic field. In the same paper we also used ancient and medieval observations (since 700 B.C.) of both solar and lunar eclipses to demonstrate the existence of marked non-tidal variations in the length of the day on a time-scale of millennia, a previously unexpected result.

In our more recent paper - published in (Phil. Trans. Roy. Soc. A, 313 351 (1995), 165-202) - Dr Morrison and I considerably refined our previous results on the millennial time-scale, incorporating many hitherto unused ancient and medieval eclipse data. We deduced that the mean length of the day has increased by 1.7 milliseconds per century (ms/cy) in the last 2700 years (compared with the expected figure due to tidal friction of 2.3 ms/cy) and also demonstrated the existence of non-tidal oscillations (of amplitude some 5 ms and periodicity 1500 years). We showed that our result for the mean non-tidal rate of decrease in the length of the day (i.e. 0.6 ms/cy) is in close accord with current estimates of the effect of post-glacial rebound. In particular, our values for the Earth's rotational clock error and length of the day from ancient times to the modern period are tabulated annually in the Astronomical Almanac.

Much of my research on Earth's past rotation was undertaken while I was at JPL, Pasadena, during the winters of 1989-90 and 1990-91, having been awarded a one-year Senior Resident Research Associateship by the National Research Council, Washington, D.C. My extensive book entitled "Historical Eclipses and Earth's Rotation" considerably expanded our previous publications on this subject. This book, published by Cambridge University Press in 1997, has received several excellent reviews and is widely regarded as a standard work on the subject.

More recently (since about 2005) I have extended my research to include the following issues: (a) analysis of Babylonian timings of first contacts for both solar and lunar eclipses; (b) the investigation of observations - many previously unused - of solar and lunar eclipses made between AD 300 and 800; and (c) the study of possible archaic references to large solar eclipses (prior to about 700 B.C.). In particular, the last of these extensive investigations has revealed that before about 700 B.C. no solar eclipse observations are of sufficient reliability to provide useful information on variations in the Earth's rate of rotation. Scientifically, this is an inconvenient but unavoidable conclusion.

Historical supernovae

Early observations of "new stars" - giving details of position, changing brightness, etc. - are of particular significance in astrophysics as no outburst of a galactic supernova has been definitely recorded since the invention of the telescope. Although about 200 supernova remnants (SNRs) have been discovered at radio, x-ray and other wavelengths, the ages of most of these can only be crudely estimated. A confident association between a carefully observed temporary star and a SNR thus yields an exact age for the latter and this provides a valuable calibration for theoretical models of the supernova process.

When I commenced my researches in this topic around 1970, only three historical supernovae (appearing in A.D.1054, 1572 and 1604) had been definitely recognised and their remnants identified. Working first on my own and later with Dr D.H. Clark I made many careful historical and astronomical investigations of temporary stars recorded in the pre-telescopic era. Emphasis was placed on those objects which were unusually bright, visible for many months and in low galactic latitude - all characteristics of supernovae. As a result we were able to affirm direct associations between two further supernovae (appearing in A.D. 1006 and 1181) and SNRs of acceptable age. We thus considerably enlarged the sample of SNRs whose precise age was established. The results of this research have since found general acceptance. Our joint book (The Historical Supernovae, Pergamon Press, Oxford, 1977) soon became a standard text on historical supernovae, and has so remained.

In 1998, I began a new phase of research into historical supernovae with Dr D.A. Green (Cambridge University) and our resulting book entitled Historical Supernovae and their Remnants, was published by Oxford University Press in 2002. In this work, we discuss historical supernovae in detail, both historical observations and the present day radio and X-ray remnants. In particular, we have investigated hitherto unknown texts - from both East Asia and the Arab lands - and also developed a variety of techniques for better understanding of the terminology in the early sources. We conclude that over the past millennium, only five galactic supernovae have been definitely recorded - in AD 1006, 1054, 1181, 1572 and 1604. In the previous millennium, possible supernovae ocurred in AD 185, 386 and 393, but in each case there is an element of doubt in the interpretation of the records. We emphasise that apart from these eight objects no other temporary star recorded in history deserves consideration as a potential supernova.

In a paper on this subject ("A reappraisal of some proposed historical supernovae"), published in 2005, we devised a series of selection criteria to use in the investigation of further records of possible supernovae. In this paper we tested these criteria on several further proposed supernovae which had been discussed in recent years and found that in no case were there valid grounds.

Subsequently, in 2009, David Green and I made a detailed study of some 50 early records of "guest stars" recorded in East Asian history. This was entitled "A catalogue of `guest stars' recorded in East Asian history from earliest times to A.D. 1600".

The past orbit of Halley's Comet

The apparition of Halley's Comet in 1985/6 provided ample incentive for research into its past history. My interest was largely inspired by a translation of East Asian records of Halley's Comet going back to earliest times which Dr K.K.C. Yau and I undertook in 1985 at the request of the editor of {\it{Journal of the British Interplanetary Society}. This demonstrated to me that despite the careful work of Dr D.K. Yeomans of JPL (along with Professor T. Kiang of Dunsink Observatory, Republic of Ireland) the more ancient history of Halley's Comet (i.e. before 12 B.C.) was most uncertain. As a result, I requested Professor H. Hunger of Vienna University to make a search of selected Late Babylonian astronomical diaries in the British Museum. This led to our discovery (with Dr Yau) of Babylonian observations of Halley's Comet in both 87 and 164 B.C. (Nature, 314 (1985), 587-592). This discovery of the earliest reliable sightings of Halley's Comet (there is a brief Chinese record in 240 B.C.) attracted widespread attention. As a result of these our researches, every apparition of Comet Halley since 240 B.C. has at least one identifiable observation.

In these same studies, various orbital solutions based on numerical integration were tested. Of these, the calculations by D.K. Yeomans and T. Kiang were shown to agree remarkably well with observation. I have since shown that owing to the paucity of good cometary records before 240 B.C., no reliable sightings of Halley's Comet at this archaic period can be found. However, the fact that the motion of Halley's Comet can be confidently traced back over more than 2200 years provides convincing evidence that it has made no "close encounters" with stray asteroids, etc. or suffered anomalous mass-loss from any other source in that same period.

A matter of current concern to me is the analysis of medieval East Asian observations of Halley's Comet. When comparision is made with the results of numerical integration, these reveal that for four successive apparitions (from A.D. 989 to 1222), perihelion passage of the comet was delayed by approximately 3 days. The most likely cause is unusual outgassing of the nucleus, causing a rocket effect.

Solar Variability

In principle, sunspot observations provide the most viable index of solar variability in the historical past. However, during the whole of the pre-telescopic period there are only about 150 known sightings (nearly all from China and Korea). As well as producing the first comprehensive catalogue of these observations to be published in a Western language (jointly with Dr D.H. Clark) - and a subsequent revision of this paper (with Dr K.K.C. Yau) - have also discussed in depth the utility of these early records, which are particularly challenging because of their uniqueness.

My principal research paper on solar variability was published in Phil. Trans. Roy. Soc. A, 330 (1990), 499-512, as part of a symposium on "The Earth's Climate and Variability of the Sun over Recent Millennia" held at the Royal Society in February 1989. The main conclusions of my research were threefold: (a) Telescopic observations of sunspots provide a reliable index of solar activity since about A.D. 1715. (b) The apparent scarcity of such observations in the 17th century (e.g. during the Maunder Minimum) is only partly due to diminished solar activity; the relative inaccessibility of astronomical data of any kind at this early period must also be taken into account. (c) Although the temporal distribution of recorded sunspot sightings in the pre-telescopic period (largely from China and Korea) shows considerable variation, most of these features would appear to result from the incompleteness of the observational record rather than indicating real trends in solar activity. In general the most effective usage of early astronomical records involves analysis of measurements - not simply frequency trends.

Recently, Dr D.M. Willis of Rutherford Appleton Laboratory and I published a paper on conjugate auroral sightings made in the late 18th century by both British seamen north of Australia and East Asian astronomers (in China and Japan). This is the earliest known example of conjugate auroral sightings. More recent joint work -- much of it already published -- includes investigations of concurrent sightings of aurorae in China, Korea and Japan; the earliest drawing of sunspots (from England in AD 1128); independent sunspot (from England) and auroral (from Korea) evidence for an intense recurrent geomagnetic storm in AD 1128"; identification of possible intense historical geomagnetic storms using combined sunspot and auroral observations from East Asia.

My studies (partly with Dr Willis) of Korean records of sunspots and aurorae have focused largely on the data in such works as the {\em Koryo-sa}, {\em Choson Wangjo Sillok} and {\em Chungbo Munhon Pigo}. I have lately studied in detail the numerous auroral sightings in the earliest surviving volume of the {\em Sungjongwon Ilgi} (the first few years of King Injo). One of my aims was to verify the reliability of these records by investigating the accuracy of parallel reports of lunar and planetary observations.

Oriental Star Maps

The positions of many comets and "new stars" (novae and supernovae) were carefully recorded by East Asian astronomers in relation to the traditional constellations. These star groupings differ markedly from the their occidental equivalents. I have undertaken an extensive investigation into the history of celestial cartography in both China and Korea. The principal results of this research were published by the University of Chicago Press in 1994 (in vol. II, part 2 of the {\it{History of Cartography}). This article was the first comprehensive discussion of Chinese and Korean star maps to appear in a Western language.

Partly as the result of visits to both China and South Korea - I have built up an extensive collection of high quality photographs of early oriental star maps. My studies of these reveal that the accuracy of star positions is very variable from chart to chart - ranging from about 1 degree to more than 5 degrees. Comparison of star charts of widely differing dates has also revealed that for many individual constellations containing fainter stars the recognised configurations have varied considerably down the centuries. This remarks is particularly true in the 17th century when the Jesuits intoduced Western mapping techniques to China.


The dating of eclipses recorded in ancient Greek and Latin writings attracted much attention during the 19th and early 20th centuries. However, after the extensive work of John K. Fotheringham between about 1910 and 1920, little further progress was made for more than fifty years. Using the results of my studies of Earth's past rotation with Dr Morrison, I have been able to compute the local circumstances of ancient eclipses with much improved precision. This led to my renewed interest in chronological studies of eclipses. Much of this work is published in by 1997 book on {\em Historical Eclipses and Earth's Rotation}. However, detailed investigations of eclipses recorded by Herodotus and Thucydides have more recently been published jointly with my former research Student Dr L.J. Fatoohi.

Until recently, studies of the Chinese calendar were very much neglected by Western scholars. Professor Liu Baolin of Nanjing, who spent a year at Durham, made a careful study with me of both the history and operational rules of the traditional Chinese calendar; this work was published in a series of papers.

Accuracy and reliability of pre-telescopic observations

Many ancient and medieval astronomers timed the various phases of both solar and lunar eclipses. These observations enable the accuracy with which time was measured in the past to be investigated in some detail. Both alone and in collaboration with Dr S.S. Said, Dr Fatoohi and another former research student Dr J.M. Steele, I have made the first comprehensive investigation of this subject, working with Babylonian, Chinese and Arab and European observations. This project, the results of which have been published in a series of papers, has significantly improved existing knowledge of time measurement in the historical past.

Other projects

In a paper published in 1992, Dr W.S. Rada (a frequent visitor to Durham) and I compiled and translated -- giving exact Julian dates where possible -- some 30 medieval Arab records of meteor showers. Scarcely any of these observations had been previously published. The main motive in making this compilation was to encourage further research in meteor showers orbits by making hitherto unused historical data generally available. It is well established that on account of orbital perturbations individual meteor showers vary considerably in extent from year to year and there is some evidence for periodicity. Also a gradual change in the date of a shower can often be detected down the centuries, only part of which is due to the discrepancy between the sidereal and tropical years.

The late Professor Xu Zhentao (who spent a year working with me at Durham University on a Royal Society K.C. Wong Fellow) and I also studied and published the various astronomical references on the oracle bones of the Shang Dynasty in China (ca 1300 - 1100 B.C.). These include allusions to eclipses, comets, stars and planets. Professor Xu had specialised in the study of the ancient Shang script.

Any further pages in alphabetic order of their title as created by you.

Just click at "Create new page", then type a short title and click OK, then add information on the empty page presented to you (including maybe a picture from your harddisk or a pdf-file by using the "Upload" Button) and finally click at "Save". Data available yet!

Imprint Privacy policy « This page (revision-4) was last updated on Thursday, 17. March 2011, 10:15 by Kaiser Dana
  • operated by