At Naqada and Diospolis Parva (Hu), Petrie dug thousands of graves of the Predynastic cultures, from before writing and before the kings. Nothing in them was dated. How could he put them in order?
Key idea
Seriation puts things in order by their changing styles. Petrie listed the pottery types in each grave and arranged the graves so that each type appeared in one unbroken run. He numbered the result as sequence dates (SD 30, 31, … 80). They give an order, earlier and later, but not years.
Why it works
Styles come in, are popular for a while, and fade out. A pot type made for fifty years will appear in graves from those fifty years, not scattered across centuries. So if graves are in the right order, each type's graves should sit together. The order that keeps every type together is most likely the order of time. Which end is earlier has to come from somewhere else: Petrie used a jar whose wavy handles shrank, over time, into a painted wavy line.
| Grave | Type A | Type B | Type C | Type D |
|---|---|---|---|---|
| 1 | ✓ | ✓ | ||
| 2 | ✓ | ✓ | ||
| 3 | ✓ | ✓ | ||
| 4 | ✓ |
Worked example: Finding the order
Using the table above, put the graves in a sequence in which every pot type appears in one unbroken run.
- Grave 4 has only A, and grave 1 has A and B, so 4 and 1 sit together: 4, 1.
- Grave 3 has B and C, so it follows 1 (which shares B): 4, 1, 3.
- Grave 2 has C and D, so it follows 3 (which shares C): 4, 1, 3, 2.
- Check: A runs 4–1, B runs 1–3, C runs 3–2, D is only in 2. Every run is unbroken.
Answer: 4, 1, 3, 2 (or the reverse, 2, 3, 1, 4).
Worked example: Which way is time?
Grave 4 contains jars with real wavy handles; grave 2 contains jars with only a painted wavy line. Which end of the sequence is earlier?
- Petrie found that the handles came first and shrank over time into a decoration.
- So the grave with real handles, 4, is at the early end.
- The sequence, earliest first, is 4, 1, 3, 2.
Answer: Grave 4 is earliest.
Common mistake
Taking sequence dates as years. SD 40 is earlier than SD 60, but the gap is not 20 years, or any number. Petrie started at 30 to leave room for earlier cultures not yet found. Absolute dates came later, from radiocarbon.
Petrie published the method in 1899 and used it in Diospolis Parva (1901). It was one of the first uses of seriation anywhere, and the same idea, grouping things by shared features, is used today with statistics and computers.