Chapter X: The Astronomical Geography of the Known World

We have already examined the broader theories of astronomical geography whereby the relation of the globe to the remainder of the universe was explained. In this chapter we shall speak only of those aspects of astronomical geography which were intimately connected with man’s knowledge of the various parts of the known world, or _oikoumene_, as distinguished from the sphere as a whole.

PHENOMENA RESULTING FROM DIFFERENCES IN LATITUDE

Within the _oikoumene_ the phenomena resulting from varying elevations of the ecliptic in different latitudes were fairly well understood. The facts that there are two summers between the tropics (particularly in India) and that the sun there passes vertically overhead twice a year had been commented on by Pliny and Solinus, whose observations in this connection found their way into Isidore’s _Etymologiae_[1095] and thus to the works of the plagiarists of Isidore in our period. The _De imagine mundi_,[1096] the _Image du monde_,[1097] Gervase of Tilbury,[1098] and John of Holywood[1099] all tell us that the same phenomenon was said to occur in Arabia which lies between the tropics. Similarly the long days and nights of far northern latitudes were described on the authority of Solinus and Isidore. In the _De imagine mundi_,[1100] from which Gervase copies, it is said that in the island of “Chili” (Thule) there are six months of daylight and summer and six of night and winter. Giraldus Cambrensis also quotes Solinus[1101] and Isidore[1102] to the same effect and adds a brief description of how the sun continuously circles around the horizon during the long Arctic day and how its light disappears completely when the luminary departs southward towards the Tropic of Capricorn.[1103]

“CLIMATA”

The ancient geographers had divided the earth’s surface into _climata_, or climates, which, as we have already seen,[1104] were not atmospheric regions but mathematical strips running east and west and bounded by parallels of latitude. Pliny,[1105] for instance, had conceived of seven climates, the first in the latitude of India, where the length of the longest day is fourteen hours, and the seventh in that of the Borysthenes (Dnieper) and of Venetia, Umbria, Milan, and Aquitania, where the longest day is fifteen and three-quarters hours. Martianus Capella[1106] added an eighth climate in the north between the parallel of the Borysthenes and that of the Rhipaean Mountains. Furthermore, he applied names to the strips. It must be added, however, that neither Pliny nor Capella were precise in the data they gave, and neither indicated in degrees the latitude of the parallels which bound their climates.

More definite is the information we find in the two works of Ptolemy. The _Almagest_[1107] and _Geography_[1108] give accounts of the characteristic astronomical phenomena that occur along a series of parallels, thirty-eight in number according to the former, twenty-one according to the latter.[1109] The positions of these were determined by the length of the longest day at each one. Though there is no explicit mention of the older division by climates in the text of either of Ptolemy’s books, such a division not only appears upon the map of the world made by Agathodaemon on the basis of material supplied by Ptolemy but also upon certain of the special regional maps which were probably the work of Ptolemy himself.[1110]

At all events, the conception of the seven or eight climates did not disappear but at a very early period, whether by Ptolemy or not, was correlated with the Ptolemaic parallels.[1111] That is to say, certain of Ptolemy’s parallels were used to designate the imaginary lines marking the centers and bounds of the climates. This practice was adopted by the Arabs and from them transferred to the knowledge of the Christian West in various astronomical treatises. Among the Latin manuscripts of the _Toledo Tables_,[1112] for instance, there are series of astronomical tables for each of the seven climates, according to which the climates occupy the space between latitude 16° N., with a longest day of thirteen hours, and 48° N., with a longest day of sixteen hours. The length of the longest day and the latitude are given for each parallel that bounds the climates. Except that Ptolemy notes minutes as well as degrees and in the _Toledo Tables_ the minutes have in most cases been omitted, the figures correspond essentially with those of the _Almagest_ and _Geography_. Thus: Ptolemy’s eleventh parallel according to the _Almagest_ (or tenth according to the _Geography_) has a longest day of fourteen and a half hours and is at latitude 36°. In the _Tables_ the southern edge of the fourth climate likewise has a longest day of fourteen and a half hours and is at latitude 36°.

Again, in John of Seville’s translation of Al-Farghānī’s _Astronomy_[1113] and in the _De sphaera_[1114] of John of Holywood, who had borrowed from Al-Farghānī in this matter, we find a similar correlation. In both cases the figures of latitude correspond essentially, though with slight divergences in detail, to those of Ptolemy. The boundaries of each climate, however, have here been displaced by one parallel to the south of the parallels used in the _Toledo Tables_ and those which we may presume were the Ptolemaic boundaries of the climates.[1115]

The table, Figure 11 (in the Notes to Ch. X), gives some idea of the relative degree of accuracy of these figures as they were employed in the West during the Middle Ages. But just as in the case of other figures for latitude and longitude, as we shall shortly have occasion to see, this material was not utilized for geographical purposes during our age.

GEOGRAPHICAL COÖRDINATES

At the present time the study of regional geography is largely dependent on a precise knowledge of the geographical coördinates of places. The foremost duty of the explorer is to know where he is from day to day and to find this out by astronomical means, if possible. In classical times and among the Moslems the importance of such observations was not only well understood, but several methods of carrying them through were described by astronomers and geographers, and the latitudes of a great many stations had been determined astronomically. Longitude, on the other hand, long remained a stumbling block, and before the twelfth century, certainly, no systematic attempts to ascertain the longitudes of any large number of places had ever met with success.

A few relics of classical and Moslem study in this field became familiar in the West as a result of the intense interest in Arabic astronomy prevailing in Europe between the tenth and thirteenth centuries.[1116]

Various figures representing the results of Arabic corrections of and additions to the data given in Ptolemy’s _Geography_ found their way into Western astrological tables. The most interesting of these occur in a list of the latitudes and longitudes of some sixty odd cities appended to the Paris manuscript of the _Marseilles Tables_ of Raymond of Marseilles[1117] and also to most of the Latin versions of the _Toledo Tables_.[1118] This list and certain figures scattered through the astrological tables and canons[1119] reveal the results of the reductions made by Al-Khwārizmī and by Az-Zarqalī of Ptolemy’s gross overestimate of the length of the Mediterranean, to which we have referred in a preceding chapter.[1120] The European student of these astrological works might have drawn a by no means contemptible map from the figures to be found in them had he been interested in what these figures could teach him of geography. Figure 6 is a map compiled from the coöordinates given in the Paris manuscript of the _Marseilles Tables_.

At the end of this list of geographical coördinates in many manuscripts additional figures not derived from Moslem sources are given. These show the positions of such points in Europe as London, Hereford, Paris, Toulouse, Barcelona, Marseilles, Novara, Cremona, Florence, and Naples[1121] (see Fig. 12, in Notes to Ch. X). They were undoubtedly determined by observations made during our period or shortly after.

FIG. 6—Sketch map constructed from the list of geographical positions
appended to the Paris manuscript of Raymond of Marseilles’
_Marseilles Tables_. The outline of the coast, arbitrarily indicated
by a shaded band, is shown merely to give some idea of the type of
map that might have been constructed from the data given in the
tables. This may be compared with the Henry of Mayence map (see
above, p. 124) shown in outline in the inset. The original Henry of
Mayence map reveals far greater detail and upon it east (not north,
as in this figure) is at the top.

This list is based on the observations of the eleventh-century Arabic
astronomers Al-Khwārizmī and Az-Zarqalī. Cities and other points
have been plotted according to the coördinates of this list. The
resulting map of the Mediterranean region and the Near East is
remarkable for its comparative accuracy. For a key to the names
represented by the numbers on the diagram and for the figures for
the latitudes and longitudes, see J. K. Wright, _Knowledge of
Latitudes and Longitudes_, 1923, pp. 87–88.
]

METHODS OF FINDING LATITUDE AND LONGITUDE

That such observations were carried out is entirely possible, for there is absolutely no doubt that methods of finding latitudes and longitudes were well understood in theory and were sometimes put to practical use. Rules are given for finding latitude in Az-Zarqalī’s _Canons_, in Plato of Tivoli’s translation of the _Astronomy_ of Al-Battānī, and in many other astronomical and astrological treatises.[1122] Two principal methods were recommended. You may either measure with the astrolabe the altitude of the sun above the horizon at noon at the spring or autumn equinox and find the latitude by subtracting this angle from 90° or you may measure the altitude of the celestial pole above the horizon, which is the same as the latitude. As to longitude, the fact that there are differences in local time between points east and west of each other was recognized and clearly explained by several writers of our age.[1123] The _Marseilles Tables_ give a rule for finding longitude by the observation of eclipses. Roger of Hereford indicates that he himself, by observing an eclipse in 1178, ascertained the positions of Hereford, Marseilles, and Toledo in relation to Arin, the world center of the Moslems.[1124] Gerard of Cremona describes a method of finding longitude by noting the distance of the moon from a given point in the heavens and thereby dispensing with eclipses,[1125] though it is doubtful whether this method was used until the sixteenth century. The lack of accurate instruments for ascertaining time must have rendered it extremely difficult to calculate longitude under any circumstances. Making allowances for this, it is surprising to find how accurate the few coöordinates that have come down to us seem to be, if our interpretation of them is correct.[1126]

The geographical interest of these figures and of investigations of this sort was not appreciated by the majority of the men of our age. The application of astronomical considerations to the problems of navigation was still in its infancy. The purpose of the investigator of the twelfth and early thirteenth century in finding geographical coöordinates was astrological. He wished to make use of them to transpose tables made originally for the meridian and parallel of one station to the meridian and parallel of another. Their influence on the cartography of the age was absolutely _nil_. It is probably safe to make the categorical statement that the maps and geographical treatises of the century and a half preceding the year 1250 were drawn and written with almost complete disregard of any astronomical considerations whatsoever.