Introduction

The main purpose of this book is to investigate in detail long-term variations in the length of the day, or equivalently, changes in the Earth's rate of rotation, using pre-telescopic observations of eclipses. Such variations are mainly produced by lunar and solar tides, but non-tidal mechanisms are also significant. Despite its historical bias, this subject has become an important topic in modern geophysics.

In studying changes in the length of the day which have occurred in recent centuries (since the invention of the telescope), more accurate data than eclipses are available. These include occultations of stars by the Moon up to 1955 and systematic monitoring of the Earth's rotation relative to the atomic time-scale since that date. However, observations of eclipses provide by far the most consistent and reliable source of information on variations in the terrestrial rate of rotation during the pre-telescopic period.

Although numerous early accounts of eclipses are to be found in astronomical works, many others are scattered in a variety of writings – for example the Greek and Latin classics, imperial annals of China and monastic chronicles of Europe. Observations recorded in these works are often extremely crude by modern standards, but the lengthy time-scale covered is highly suited to the detection of long-term trends which are not apparent from more recent data. The investigation of this diverse material adds a whole new dimension to what might otherwise have been a somewhat narrow scientific discipline.

Introduction

In this chapter, all accessible reports of solar eclipses preserved on the Late Babylonian astronomical texts which are of value in studying the Earth's past rotation will be investigated. Compared with the substantial number of observations which must have originally been made, very few Babylonian records of solar eclipses are now extant. Reasons for this are briefly discussed in chapter 4. Reliable observations only range in date from 369 to 136 BC – less than half the period covered by lunar eclipse sightings.

The solar eclipses which form the subject of the present chapter will be divided into four categories: (i) the only known example of a total obscuration of the Sun observed at Babylon (dating from 136 BC); (ii) measurements of the local times of a variety of eclipse contacts; (iii) estimates of solar eclipse magnitude at maximum phase; and (iv) instances when the Sun rose or set whilst obscured. Several of these observations have already been discussed in chapter 3, where they were used to illustrate various analytical techniques. However, they will now be considered in greater detail. At the close of this chapter, a possible allusion to a total solar eclipse recorded in the Religious Chronicle of Babylon will be considered. This event probably took place around 1000 BC – much earlier than the period covered by the Late Babylonian texts.

All eclipse records on the Late Babylonian texts where the date is doubtful will be rejected, as will questionable readings of damaged signs.

Introduction

In chapters 4–13, more than 400 timed and untimed observations of both solar and lunar eclipses from the pre-telescopic period (date range –700 to +1600) have been investigated in detail. ΔT values or limits (depending on whether the observations were timed or untimed) have been derived in almost every case. The fundamental objective of the present chapter is to use these results to obtain the best-fitting ΔT curve to the data and hence to determine changes in the length of the day (LOD) over the historical period. In addition, the various geophysical mechanisms responsible for the observed variations will also be discussed. I am grateful to Dr L. V. Morrison of the Royal Greenwich Observatory for undertaking the data analysis which forms the basis of much of this chapter and producing several of the diagrams. This chapter is essentially an enlargement of section 5 of the paper by Stephenson and Morrison (1995). Although several tens of further historical data have been added (notably medieval Chinese timing – see chapter 9), the basic conclusions obtained in that paper remain unchanged.

All of the ΔT results derived in chapters 4–13 are summarised in tabular form in the Appendices for ready reference. For each individual observation, only the year (– or +) and appropriate ΔT value or limits are tabulated. Appendix A contains timed data and Appendix B untimed material, each being divided into sub-groups depending on the source and type of observation it contains.

Introduction

For the purpose of this chapter, the term ‘medieval’ will be loosely interpreted to mean the period between the close of the classical age in Europe (c. AD 500) and the beginning of the telescopic era. It thus includes the Renaissance. Numerous observations of both solar and lunar eclipses were recorded in Europe during this period, especially after AD 1000. In most cases, the date is accurately reported and the precise place of observation is known. However, with only a few exceptions, times were only crudely estimated to the nearest hour or so. Thus most observations may effectively be regarded as untimed.

As discussed in chapter 3, untimed lunar eclipses are of negligible value for the determination of ΔT (unless it is clearly implied that the Moon rose or set eclipsed). Most of this chapter will be devoted to the investigation of solar obscurations in which the Sun was either totally or very largely covered. In section 11.9, a few careful timings of solar eclipses from the early fourteenth century will also be analysed. Finally, in section 11.10, an eighth century report of an occultation of the planet Jupiter by the eclipsed Moon – a very rare event – will be discussed.

Historical sources

European reports of solar eclipses (as well as lunar obscurations) from the Middle Ages and Renaissance are mainly found in historical works, notably chronicles; only a few observations are reported in astronomical treatises.

Introduction

Arab observations of solar and lunar eclipses which are of value in the study of the Earth's past rotation originate exclusively from the medieval period. These records are found in two main sources: (i) chronicles, and (ii) compendia on astronomy. As might be expected, the observations reported in chronicles are essentially qualitative; measurements of any kind are fairly rare. Eclipses and other celestial phenomena (such as bright comets and meteor showers) were mainly noted on account of their spectacular nature. Yet chronicles contain some important astronomical records. The present chapter will be devoted to eclipses described in these works. Arab compendia on astronomy contain many careful measurements of the times of the various phases for both solar and lunar eclipses. These observations will be discussed in chapter 13.

The Arabic names for the Moon and Sun are respectively al-Qamar and al-Shams. In general, medieval Arab chronicles use the term khusuf al-Qamar for an eclipse of the Moon and kusuf al-Shams for an eclipse of the Sun. These designations are still in use today, both among astronomers and the general public. However, medieval Muslim astronomers mostly used kusuf for both types of eclipse, adding the appropriate term for the Moon or Sun. The word khusuf means ‘sinking’, but in describing a lunar eclipse it came to mean a failing of the Moon's light. By contrast, kusuf means a ‘cut’ – i.e. in the solar (or lunar) limb.

Introduction

The eclipse observations made by medieval Arab astronomers are among the most accurate and reliable data from the whole of the pre-telescopic period. Careful records of both solar and lunar eclipses are contained in a number of compendia – some known as zijes (astronomical handbooks containing various tables along with explanatory text). These include measurements of the times of occurrence and other details such as magnitude estimates. Although the main emphasis in this chapter will be on timed data, solar magnitude estimates and horizon observations of eclipses will also be considered.

Many of the observations discussed below were investigated by Newcomb (1878) and Newton (1970). However, these authors relied on published translations which sometimes contained significant errors, while their own interpretations are occasionally suspect. Furthermore, in neither case was a direct solution made for ΔT.

Sources of data

Most of the accessible eclipse observations by medieval Arab astronomers are contained in a single treatise – the zij compiled by the great Cairo astronomer Ibn Yunus, who died in AD 1009 (his date of birth is unknown). A few eclipses are also recorded in works by al-Battani (who lived between AD 850 and 929) and al-Biruni (AD 973–1048).

Ibn Yunus cites reports of some thirty solar and lunar eclipses from between AD 829 and 1004. His treatise, dedicated to Caliph al-Hakim, is entitled al-Zij al-Kabir al-Hakimi.