Sequence Dating

The Method of Working.

Some starting connection is need ful, either in a related series of forms or a link to historic times, or some stratification of a site. For instance, a long series of degradations of a globular pot to a cylinder and of a handle into a cord pattern, served to start the interpretation of the later pre historic pottery of Egypt. Or, on the other method, the Badarian settlement was divided by a layer of rock chips, while it was deserted; the pottery from under the chips was certainly then the older.

Having got thus some intimation, by style or by place, which of two groups A and B is the older, the next stage is to group all graves containing A types and none of B, and arrange them in order of the proportion of A types; those with 5, 4, 3, 2, i or o types like A being ranked in that order. Similarly those with B types on the other side away from A. In this way the order is arranged as forms like A, o, 1, 3, 4, 5, 6, then A, A and B types conjoined, B, 6, 5, 4, 3, 2, I, 0 like B. Thus a rough breaking up into many groups in approximate order can be made statistically.

Having a rough classing the next process is to deal with each type individually. It is obvious that if a series were arranged in true order of succession, a disturbance to a false order would be more likely to scatter types than to concentrate them in their range. Hence the shortest possible range of each type is the most probable. Then outlying examples of each type in the series of graves must be examined to see whether that grave could be brought nearer to the other examples of the type, without dis persing any other type which is associated with it. This process may be graphically imagined by supposing that every example of a type is tied to all other examples of it by an elastic cord in ten sion. Then a scattered type will tend to pull itself together, pro vided it does not increase the tension of the cord of another type. The whole process might be solved mechanically by such a model.

Method in Practice.

The way to handle all this material is to take for each grave a slip of card, about Tin. long and tin. high. Rule it into as many spaces as there are kinds of pottery, A, B, C, etc., about half a dozen. Enter in its respective space the number of each type of pot in the graves, in order of numbers. Then place all the cards in a column, so that a hundred grave contents can be reviewed in a height of 25in., laid on a tray. Ten such trays allow of a review of every pot in a thousand graves at once, on a table about six ft. wide. Usually only two or three trays need to be seen at once. Each type is searched in turn, every example noted (by putting a drawing pin point up at each place) ; the earliest and latest are then examined to see what associated types are with them, and if moving these extreme slips will make divergence in other types. This is a process which has to be repeated many times, each time getting a better concentra tion of each type. Where several types found together have to be searched in all their connections, a different pattern of drawing pin can be used to mark each type.

Results.

The eventual result is a classing of, say, i,000 graves in nearly their original order. Each selected grave should contain

at least five different types of pottery. Next the whole series can be divided into numbered groups, modifying the groups so as to divide at the beginning or end of prominent types. When these groups are consecutively numbered, then each type of pot can have its limits in sequence dates ; say from 42 to 48. Some simple types will have a very long range ; other decorated types may have been made by only one generation. The more elaborate a type the more it demands concentration. In mechanical parallel the tension of the elastic cord will be greatest for the peculiar types.

The Sequence Date of Graves.

The different types of pottery in a grave will each give a fresh limitation of the possible range, as in the following cases, from Abydos : B 22b 30-37 W 55 72-78 B 38a 43-66 B 25f 3o-5o R 26 55-8o W 42 62-72 P iia 31-63 L 17c 51-78 R 23c 36-8o P lib 35-71 L 36a 58-8o R 65a 49-68 L 53a R 8o 41-72 L 53c 54-80 Limits 35-37 62-66 These dates of graves serve then to date all the other objects found in the grave. A fresh series of graves will sometimes slightly extend one or two of the ranges already found for a type ; but the graves with sufficient varieties work out with a very small range of uncertainty. The scale of so parts, 30-79, which was adopted for Egypt, is not at all too minute in its subdivision, as the resulting dates work out to two or three units. Twenty-five years after this was done the scale has been continued back to S.D. 20-29 to include the still earlier prehistoric age of Badari.

Precautions.

In forming a corpus, various of the gg numbers should be left unused where a large step lies between types, so as to leave numbers for future discoveries ; similarly in assigning letters for varieties, especially where a large difference exists. In ages where a good dating is already known by records, it is best to make larger groups (as 33-37), and place all the forms in order of date, as this will point to their true relation.

For applying the system of sequences precautions are needed. First, there must be a sufficient series of dated graves, each con taining many varieties of pottery. For an entirely fret country at least soo grave groups are wanted ; for joining on an extension to an existing system a few dozen graves will give a prelimi nary view, useful as a framework for fresh evidence. Secondly, all kinds of evidence must be taken into account, not only forms, but decoration, relation to other forms, descent of a type, peculiar personalities in the work, and not only pottery but all other material of which there may be enough to give an indi cation, as ivory, stone, flintwork, and metal. Thirdly, the histor ical view should have consistency. If there is a large change of types at any point of the dating, this may modify the arrange ment of the slips.

The final aim of archaeology is to have a separate corpus of types of each great period of each country, not only for pottery but for all other classes of objects. Thus the whole of the materi al products of the past can be put in an organized connection, and so give a definite basis for our knowledge of man's ability and purpose. (W. M. F. P.)