The magnitude of the archaeomagnetic field in Egypt between 3000 ...
Transcript of The magnitude of the archaeomagnetic field in Egypt between 3000 ...
Geophys. J. R. astr. Soc. (1980) 63,45-56
The magnitude of the archaeomagnetic field in Egypt between 3000 and 0 BC
K. P. Games Sub-department of Geophysics, University of Liverpool, Liverpool L69 3BX
Received 1979 December 28; in original form 1979 October 23
Summary. Measurements of the magnitude of the archaeomagnetic field in Egypt using sun-dried adobe bricks are presented. Some results from fired materials are also presented and compared with those from the bricks. The results suggest that the magnetic field in Egypt has fluctuated quite rapidly over 3000 yr. These fluctuations are compared with those obtained by other authors and with records from observatory data. These results demonstrate the necessity of using archaeological material that can be reliably well-dated.
Introduction
Many authors in recent years (e.g. Bucha 1967; Bucha et al. 1970; Burlatskaya etal. 1970; Kitazawa & Kobayashi 1968; Rusakov & Zagniy 1973; Walton 1979; Gunn &Murray 1980; Shaw 1979) have published results obtained from archaeological materials which show fairly rapid fluctuations of the magnitude of the geomagnetic field over several hundreds of years. On the assumption that this high rate of change is a real phenomenon, it now seems clear that one of the most critical factors in archaeomagnetic studies is the accuracy of dating of the samples used. The precision of a I4C date is usually about k 50 yr, and the accuracy of the thermoluminescence method is at present 5-10 per cent. So for samples more than 2000 yr old these methods of dating are not sufficiently accurate, since such errors would not give the resolution needed to pick out the observed rapid rates of change of the geomagnetic field.
This was one of the reasons we decided to turn to Egypt for our archaeological samples, where the chronology that has been worked out is very accurate, certainly in the relative and often in the absolute sense. Most of the material used in this study was sundried adobe brick, which was collected from tombs and buildings which had been reliably dated. The pottery samples used were obtained either from University College, London, or from current excavations in Egypt, and again only samples which could be accurately dated were used.
Methods used to determine the magnitude of the ancient geomagnetic field
The sun-dried adobe bricks were analysed using the method described by the present author (Games 1977). Many of these results have already been presented in thesis form (Games
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
46 K. P. Games
E 10.1
z r t n P
0 0 . 8 1.6 2 . 4 3 . 2 SRM lArn2kg-’x10~5)
Figure 1. The plot of NRM/SRM using the demagnetizing field as parameter for sample E1O.l. The points labelled R were rejected for reasons explained in the text.
1978) where they are analysed and discussed in more detail. An adobe brick becomes magnetized when it is thrown into a wooden mould or former, and we have called this magnetization a Shear Remanent Magnetization (SRM). To determine the magnitude of the ancient geomagnetic field, Ban,, the Natural Remanent Magnetization (NRM) of a 2 cm core of the adobe brick is measured and stepwise demagnetized using alternating field (af) demagnetization. An anhysteretic remanent magnetization, ARM], is given to the sample and this is also measured and af demagnetized. The sample is then crushed, mixed with water, and re-cast in a known laboratory field, Blab, in a way which simulates the original brick-making process. When it is dry, the SRM acquired by the sample is measured and af demagnetized in the same way as the NRM. Then ARM2 is given and measured and af demagnetized in the same way as ARM]. Fig. 1 shows an example of a graph of NRM versus S F W for sample E1O.l using the demagnetizing alternating field as parameter. Since
Figure 2. A plot of the accepted data for E1O.l.
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
The archaeomagnetic field in Egypt 3000-0 BC 47
0 4 0 * 02 -1 '2.0 16 0 ARM121 [Am kg x 10 5,
Figure 3. A plot of ARM(l)/ARM(Z) for sample E1O.l. The sample is rejected if the gradient differs from unity by more than * 5 per cent.
the low coercive force region is susceptible to secondary magnetizations such as viscous remanent magnetization (VRM), the data in this region are rejected on the basis of the stability of the NRM direction. Fig. 2 shows a plot of the accepted data for sample E1O.l. From this graph, the ancient geomagnetic field magnitude, Ban,, is calculated from
Ban, - Lwmlvl --- Blab ASRM where the SRM is acquired in a laboratory field, Blab, of 50 pT. Fig. 3 shows a plot of ARMl (given before the sample is remixed) against ARM2 (given after the resetting process) for the accepted data, which should have a gradient of 1.00 if no alteration of the magnetic properties has taken place during the process of producing the SRM. If this gradient differs from unity by more than 5 per cent, the sample is rejected. In all cases (except one, where insufficient material was available) two determinations of B,, were made by using two cores from each brick, and the error on the mean value of Ban, quoted is half the difference between the two determinations. The average value of this error for all of the samples used is f 4.47 per cent.
The sherds and fired material were analysed using Shaw's method (Shaw 1974). In all cases the thermoremanent magnetization (TRM) was given to the sample in the same direction as the NRM to within about lo", so as to minimize the error due to fabric anisotropy as outlined by Rogers, Fox & Aitken (1979). Also the samples were cooled slowly from 700°C to room temperature in about 7 hr in order to reproduce as nearly as possible the original cooling rate of the NRM. This is to try and minimize the errors due to the effect of cooling rate on the TRM as discussed by Fox & Aitken (1979). The error quoted for these samples is the error on the slope of the best-fitting straight line through the origin. If this was less than 5 per cent, then an error of f 5 per cent is quoted (see GUM & Murray 1979). In Fig. 4, one example of a result for a pottery sample is shown. The left- hand pair of graphs shows all of the data for the NRM versus TRM and ARMl versus ARM2 plots. The right-hand pair of graphs shows the accepted data after rejection of the part of the coercive force spectrum which shows alteration by departing from the straight line of gradient = 1.00 drawn on the plot ARMl versus ARM2.
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
48 K. P. Games
0
> a I I
4
I o /
100 200 300 LOO 0
ARM I21 (Arn2kg~'x10-51 5 10 15 20 0
ARM(21 (Am2 k g ~ ' ~ 1 0 ~ ~ 1
Figure 4. Results for sample (E103). The left-hand pair of graphs shows all of the data for the NRM versus TRM and ARMl versus ARM2 plots. The right-hand pair of graphs shows that accepted data after rejection of the part of the coercive force spectrum which shows alteration by departing from the straight line of gradient = 1.00 drawn on the plot ARMl versus ARMZ.
Results
All of the 56 results quoted in this paper are listed in Tables 1 and 2. The value of B,, determined from each sample is given, and these values are also converted to Virtual Axial Dipole Moment values to enable comparison of results from differing latitudes. A brief description of each site is given along with the archaeologists' estimated age. In all cases, dates were averaged down to decades e.g. E19, Tomb of Karenen, 2133-1991 is rounded off to 2130-1990. As for the First Dynasty dates, the German School would bring the beginning of Dynasty I down to 2 8 0 0 ~ ~ . The dates are based on those in Cambridge Ancient History, revised edition, vol. I , chapter 6 , and on A. H. Gardiner, Egypt of the Pharaohs. An explanation of the basis of the dating is given in the Appendix by Professor Smith .
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
The archueomagnetic field in Egypt 3000-0 BC 49
These results are then plotted as virtual axial dipole moment values against time in Fig. 5. Values derived from fired material are shown by simple crosses and those from adobe bricks by closed circles. Fig. 6 shows a sketch map of Egypt giving the locations of most of the sites listed in Tables 1 and 2.
One of the most interesting features of these results is the fairly close agreement between the values obtained from sun-dried adobe bricks and the values obtained from fired material. Gunn & Murray (1979) have suggested that because of anisotropic effects fired materials will generally yield too great a value of the magnitude of the ancient field. The same effect would also be noticed if the cooling time for the TRM was less than that for the NRM (Fox & Aitken 1979). We have attempted to eliminate both of these sources of error, but the Egyptian fired materials do provide values of virtual dipole moments which are perhaps somewhat greater than the values derived from mud-bricks. Nevertheless, we must treat any possible difference as tentative until we have more data to reach a more definite conclusion. Despite these small differences, the fairly good agreement between the two completely different source materials gives us more confidence in the derived magnitudes of the geo- magnetic field. There are a few points which disagree with the rest of the data, and we cannot with certainty explain these discrepancies. They may be simple errors, but it is also possible that in some cases the adobe bricks in a structure were actually re-used bricks from an earlier building. There are known instances of such re-use.
The graph in Fig. 5 suggests that the magnetic field in Egypt has fluctuated quite rapidly over the 3000 yr studied. During this period 3000-0 BC there are at least three maxima in the field at approximately 2600, 1350 and ~OOBC, and two minima at about 1900 and 1000 BC. Near the large maximum at 1350 BC, the rate at which the field is changing reaches about 1407 yr-' and this persists for about 300 yr either side of the maximum value. For the rest of the time, the rate of change is considerably smaller, being no greater than about 407 yr-'. This lower rate persisted for about 500 yr at a time. The periods or 'pulse widths' which can be seen here are typically 400-1000 yr and their amplitudes vary by a factor of between 1.5 and 2.5.
These results are consistent with the changes in the archaeomagnetic field found by other authors. For example, Shaw (1979) has found that the field changed by just less than a factor of 2 in 150 yr and Walton (1977) reports periodicities of 100-300 yr with amplitudes varying by factors of up to 1.6. Secular variation with frequencies of between 500 and 5000 yr is suggested by the study of lake sediments (Thompson 1978). It is also interesting to find that over the last 20 or 30 yr, observatory data show that there has been a very large rate of decrease in the strength of the magnetic field centred on Paramaribo observa- tory (5" 49', 304" 47'). The locality in question has shown a rate of change of the field of about 1007 yr-' over an area of some 3000 km2, and this rate has persisted from the time the observatory measurements were first recorded in 1941 up to 1971 (the date of the most recent records at our disposal). Moveover, the same feature seems to spread northwards to Dallas observatory (32' 59', 263' 15') and Tucson observatory (32" 15', 249" 10') where the rate was 407 yr-I for up to 30 yr, and southwards to Huancayo observatory (- 12" 27', 284" 40') where the rate was 567 yr-' for 30 yr. So it would appear that at this locality there are occurring large rates of change in the geomagnetic field today, which are com- parable to those suggested by archaeomagnetic data, though of course the observations do not cover as long a time period. This is another factor which suggests that the changes shown in Fig. 5 could represent plausible features of the ancient magnetic field in Egypt.
With such a rapidly varying magnetic field, it may be possible to use these data in reverse to assign approximate dates to archaeological materials, provided that the period from which the object dates is already known to within a few hundred years.
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
wl
0
Tabl
e I.
Ado
be b
rick
res
ults
.
Cod
e
EO 1
E04
E06
El0
E
l 1
El 2
El 3
El4
El5
El 7
El 8
El9
E20
E2 1
E2 2
E23
E24
E25
Lat
itude
Si
te
("1
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Saqq
ara
29.8
Abu
sir
Des
crip
tion
of si
te
I D
ynas
ty T
omb
No.
3507
I Dyn
asty
Tom
b N
o. 2105
I Dyn
asty
Tom
b N
o. 3036
I Dyn
asty
Tom
b N
o. 3357
Tom
b of
Hes
y (N
o. 2405) 11
1 Dyn
asty
I D
ynas
ty T
omb
No.
2185
I Dyn
asty
Tom
b N
o. 3503
I Dyn
asty
Tom
b N
o. 3504
I Dyn
asty
Tom
b N
o. 3505
Nor
th T
empl
e T
own
(Anu
biei
on)
Sout
h W
all.
Con
stru
cted
not
late
r th
an
Sout
h Te
mpl
e To
wn
(Bub
astie
ion)
Nor
th W
all.
Prob
ably
con
tem
pora
ry
Tom
b of
Kar
enen
. XI
Dyn
asty
Ptol
emy
I1
with
El7
Age
(B
C)
3000-2970
2920-2890
3000-2970
3 100-3070
2690-2620
3070-3040
3070-3040
3040-3000
2950-2920
360-220
360-220
21 30-1990
Ban
c(fi
T)
38.0 f
1.0
40.5 f
1.
0 39.0 f 3.0
35.5 f 2.5
44.0 t 1.0
34.0 c
1.0
35.0 f 1.0
33.5 f 1.0
40.0 f 2.0
54.5 f 4.5
49.5 t 1
.5
31.0 * 1
.0
..
Nor
th T
empl
e T
own.
Eas
tern
Wal
l of
Bes
Cha
mbe
rs. L
ate
F'to
lem
aic-
Rom
an
49.5 f 3.5
Pepi
I T
empl
e. E
arly
VI D
ynas
ty
2340-2300
40.5 f 2.0
Mas
taba
t-el
-Far
aun.
End
of
IV D
ynas
ty
2500-2490
52.0 f 3.0
Tom
b of
Des
i. V
I D
ynas
ty
2 360 -2300
37.5 t 1.0
Tom
b of
Khu
i. V
I D
ynas
ty
2360-2300
43.0 f 2.0
2450-2430
41.0 f 1.0
Mor
tuar
y T
empl
e of
Nef
erir
kere
. 3rd
Kin
g of
V D
ynas
ty
100
BC
-AD
50
VA
DM
X
Am
Z
7.45 + 0.20
7.94 t 0.20
7.64 t 0.59
6.96 f 0.49
8.62 i
0.20
6.66 * 0
.20
6.86 t 0.20
6.56 f 0.20
7.84 t 0.39
10.68 ?
: 0.89
9.70 f 0.29
6.08 f: 0.20
9.70 i: 0.69
7.94 t. 0.39
10.1
9 f 0.59
7.35 t 0.20
8.43 f
0.40
8.04 f 0.19
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
E2 7
E
28
E29
E
30
E33
E38
E
39
E4 0
E
44
E45
E
46
E5 1
E5
2 E
54
E5 7
E6 1
E
64
E7 6
E7
7 E
78
E79
E
90
ER9
29.8
29
.8
29.8
29
.8
29.8
25.4
25
.4
25.4
25
.4
25.4
25
.4
25.4
25
.4
25.4
29
.8
29.8
29
.8
29.8
29
.8
25.4
25
.4
25.4
25
.4
Abu
Ghu
rab
Saqq
ara
Saqq
ara
Saqq
ara
Saqq
ara
The
bes
The
bes
The
bes
The
bes
The
bes
The
bes
The
bes
The
bes
The
bes
Dah
shur
Sa
qqar
a Sa
qqar
a Sa
qqar
a Sa
qqar
a K
arna
k K
arna
k T
hebe
s T
hebe
s
Sola
r Tem
ple
of U
serk
af. 1
st K
ing
of V
Dyn
asty
Sa
cred
Ani
mal
Nec
ropo
lis.
N. S
aqqa
ra. N
orth
Ibi
s G
alle
ries.
Dat
e un
cert
ain
Sacr
ed A
nim
al N
ecro
polis
. N. S
aqqa
ra. T
emp.
Zos
er. T
omb
3518
.111
Dyn
asty
Sa
cred
Ani
mal
Nec
ropo
lis.
N. S
aqqa
ra. P
reci
nct D
in M
othe
r of
Api
s
Sacr
ed A
nim
al N
ecro
polis
. N
. Saq
qara
. Sou
th e
nclo
sure
wal
l of
Tem
ple
Ram
ses
I11 T
empl
e, in
ner g
irdl
e w
all,
Med
inet
Hab
u. X
X D
ynas
ty
Ram
ses
I11 T
empl
e, g
reat
gir
dle
wal
l, M
edin
et H
abu.
XX
Dyn
asty
H
ouse
of
But
eham
un M
edin
et H
abu.
XX
Dyn
asty
C
liff
tom
b of
Dei
r-el
-Bal
ui. X
I D
ynas
ty
Ram
asse
um S
tore
Hou
se. X
IX D
ynas
ty
Tem
ple
of T
uthm
osis
IV. X
VII
I D
ynas
ty
Pylo
n: T
empl
e of
Mon
tuem
hat.
XX
V D
ynas
ty
Pylo
n: T
empl
e of
Pbe
s. X
XV
I D
ynas
ty
Pala
ce o
f A
men
ophi
s 11
1. M
alqa
ta. X
VII
I D
ynas
ty
Tem
p. A
men
emha
t 11
1. X
I1 D
ynas
ty
Lat
e O
ld K
ingd
om-M
iddl
e K
ingd
om to
mbs
T
omb
of U
nas-
Hai
. XI1
Dyn
asty
A
ov A
nubi
eion
Kiln
- u
nfue
d br
ick
from
bas
e G
rain
bin
- c
onte
mpo
rary
with
E75
P X
VII
Dyn
asty
. Kar
nak
Mid
dle
Kin
gdom
. Ses
ostri
s I1
1 T
omb
of S
hesh
onk.
XX
II D
ynas
ty
XV
III
Dyn
asty
- e
xact
orig
in u
nkno
wn
Com
plex
. inf
. La
te X
XV
I-X
XV
II D
ynas
ty
prec
inct
. in
f. X
XV
III-
XX
X
Dyn
asty
2490
-244
0
2690
-265
0 60
0-45
0
400-
200
400-
360
1200
-117
0 12
00-1
170
11
50-1
050
2130
-199
0 13
00-1
240
690-
650
1420
- 1
380
1430
-142
0
650-
610
1840
-179
0 23
00 -2
000
1980
-192
0 20
0-0
300-
200
1650
-157
0
940-
900
1560
-1 3
20
1880
-184
0
42.0
? 3
.0
46.5
t 3
.5
48.5
? 2
.5
36.0
t 1
.0
40.5
t 2
.0
43.5
f 0
.5
38.0
f 2
.0
30.0
f 3
.0
37.0
f 0
.5
65.0
f 4
.0
52.5
f 1
.5
40.5
f 2
.5
55.0
f 1
.0
56.0
f 2
.0
22.5
f 1
.5
35.5
f 1
.5
47.5
f 0
.5
59.0
f 1
.0
47.5
f 1
.5
37.0
f 2
.0
37.5
? 0
.5
31.5
t 0
.5
43.0
f 1
.0
8.23
t 0
.59
9.11
f 0
.69
9.50
f 0
.49
7.05
t 0
.20
7.94
t 0
.39
9.03
f 0
.10
7.89
f 0
.42
6.23
t 0
.62
7.68
t 0
.11
13.4
9 f
0.83
10
.90 f 0
.32
11.4
2 t
0.21
11
.62 t
0.42
4.
41 t 0
.43
6.96
t 0
.29
$ 9.
31 t 0
.10
11.5
6f 0
.20
(D
9.31
2 0
.32
$
7.68
f 0
.40 2
7.78
t 0
.10
6.54
t 0
.10
9'
8.87
t 0
.20
2 2 z 9
8.41
t 0
.52
(D
P,
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
P
.% T
able
2. P
otte
ry a
nd fi
red
bric
k re
sults
.
Cod
e L
atitu
de
E04P
29
.8
E75P
29
.8
E76P
29
.8
E83P
25
.4
E85P
25
.4
E86P
29
.8
E94P
29
.8
E95P
29
.8
E102
P 29
.5
E103
P 26
.2
E105
P 29
.5
E106
P 27
.4
E107
P 27
.4
E121
P 26
.8
Site
Saqq
ara
Saqq
ara
Saqq
ara
Kar
nak
The
bes
Dah
shur
Dah
shur
D
ahsh
W
Ger
zeh
Aby
dos
Kah
un
Am
arna
A
mar
na
Qua
el K
ebir
Des
crip
tion
of s
ite
Sher
d fo
und
in fi
rst D
ynas
ty, b
rick
E04.
Inf
. 1st
Dyn
asty
Po
t fro
m fl
oor
AU
M (
wat
er ja
r) P
tole
mai
c A
OV
Kiln
. Fired b
rick.
Pto
lem
aic
Sher
d fr
om fo
unda
tion
depo
sit,
Tuth
mos
is 1
. XV
III
Dyn
asty
K
iln b
rick.
Men
kepa
re. X
XI D
ynas
ty
Sher
ds - te
mp.
Sne
feru
. N D
ynas
ty
Sher
ds. X
I1 D
ynas
ty
Sher
ds. X
I11 D
ynas
ty
Ger
zean
she
rd. H
emam
ieh
G: 2
61c,
2' 6
" Pre
dyna
stic
Sh
erd
from
Royal T
omb
of D
jer.
I Dyn
asty
W
orkm
en's
villa
ge. X
I1 D
ynas
ty
Sher
d w
ith p
otm
ark.
Akh
enat
en.
XV
III D
ynas
ty
Pink w
are
sher
d, c
olle
cted
by
Mrs
V. H
anke
y, A
khen
aten
.
Nub
ian
sher
d. X
VI D
ynas
ty
XV
III D
ynas
ty
42.0
f 2
.0
49.0
t 2
.5
300 -
200
100
BC
-100
AD
51
.0 f 2
.6
1520
-1 5
00
55.0
f 5
.0
1080
-940
22
.0 f 1
.0
3100
-290
0
56.0
f 3
.0
51.0
k 2
.5
2620
-260
0 1
1840
-1 7
90
33.0
f 4
.0
1750
-165
0 37
.0 f 1
.9
U.C
. No.
19
678
4105
-344
5 31
.0 f 1
.6
1734
7 30
70-3
040
34.0
t 1
.7
1858
5 19
90-1
780
31.0
t 2
.0
2434
8 13
70-1
350
79
.0 f 4
.0
2504
5 13
70-1
350
64
.0 f 6
.2
1680
-1 5
70
39.0
f 2
.0
Q
VA
DM
X
8 Q -.
10zz
A m
z
8.23
f 0
.39
9.60
f 0
.49
10.0
0 f 0
.51
11.4
2 f 1
.0
4.57
t 0
.20
10.9
7 f
0.62
10
.00 f 0
.50
6.46
t 0
.79
7.25
? 0
.37
6.10
+ 0.
32
6.98
k 0
.35
6.10
t 0
.39
15.9
8 f
0.81
12
.95 f 1
.24
7.95
f 0
.39
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
The archaeomagnetic field in Egypt 3000-0 BC 53
t I f + t ,
I + ti
+ + ' + $: i t +
+ +
3000 2 000 1000
AGE IYEARS B C I
Figure 5. A plot of Virtual Axial Dipole Moment values against time. Values derived from fired material are shown by simple crosses, and those from adobe bricks by closed circles.
EGYPT
*Aswan 211"
A.EY
2 SAQQARA 3 DAHSHUR .4 GERZEH 5 AMARNA 6 ABYDOS 7 THEBES 8 KARNAK
i ABUSIR,ABU GHURAB
Figure 6. A sketch map of Egypt showing the location of most of the sites listed in Tables 1 and 2.
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
54 K . P. Games
Conclusions
These results show that values of the magnitude of the archaeomagnetic field can be obtained from sun-dried adobe bricks and fired material which are in fairly good agreement with each other. This fact, together with the accuracy of the dating of Egyptian material, suggests that the fluctuations observed in the magnitude of the archaeomagnetic field in Egypt are real. The results also demonstrate the necessity of using such well-dated material in archaeomagnetic studies of geomagnetic field strength. There are still some gaps in the curve (Fig. 5) where there exists little or no data, or where more data is needed to define the curve more precisely. Until this is done, it is a little too soon to try to ‘date’ archaeological samples using this curve, but in the near future application of the curve may be of some assistance to archaeologists.
Acknowledgments
I would like to thank Professor R. L. Wilson for his helpful discussions about this project and for his help in preparing this paper; Dr John Shaw who accompanied me on my first visit to Egypt, and to Professor A. F. Shore for his help in planning the trip; the Egyptian Antiquities Department, in particular Dr G. Moukhtar, for their permission to visit the sites and collect the material; Dr Ali-el-Khouli for his help and cooperation in the collection of samples; Professor H. S. Smith for his help and advice in organizing my visits to Egypt, for his invaluable help in the collection of the samples at Saqqara, for checking the dates of the samples listed in this paper, and for his interest and cooperation in the whole project. My thanks are also due to the following Egyptologists for their help while visiting their sites and for the samples they provided: Dr J . and Dr H. Jacquet, Dr D. Arnold, Dr P. Munro, Dr Kaiser and to Dr Colin Hope for his help with our collecting at Luxor. I would also like to thank Mrs B. Adams for her help in obtaining sherds from the University College Collection.
Finally, I wish to thank Mrs B. Bridges for measuring the fired material and the staff of the Geophysics Department for their help and support.
This work was supported by a NERC grant no. GR/3/2396.
References
Bucha, V., 1967. Intensity of the Earth’s magnetic field during archaeological times in Czechoslovakia, Archaeometry, 10, 12 - 22.
Bucha, V., Taylor, R. E., Berger, R. & Haury, E. W., 1970. Geomagnetic intensity changes during the past 3000 years in the western hemisphere, Science, 168,111-114.
Burlatskaya, S. P., Nachasova, I. E., Nechaeva, T. B., Rusakov, 0. M., Zagniy, G. F., Tarhov, E. N. & Tchelidze, Z. A., 1970. Archaeomagnetic research in the USSR: Recent results and spectral analysis, Archaeometry. 12,1 ,73-85 .
Fox, J . M. W. & Aitken, M. J., 1979. The effect on TRM of rate of cooling and its relevance to the determination of the ancient field magnitude, Geophys. J. R . asfr. SOC., 57, 268.
Games, K. P., 1977. The magnitude of the palaeomagnetic field: a new non-thermal, non-detrital method using sun-dried bricks, Geophys. J. R . astr. Soc., 48, 315-329.
Games, K. P., 1978. A new method for determining the magnitude of the Archaeomagnetic field, PhD thesis, University of Liverpool.
Gunn, N. M. & Murray, A. S., 1980. Geomagnetic field magnitude variations in Peru derived from archaeological ceramics dated by thermoluminescence, Geophys. J. R . asfr. SOC., 62, 345- 366.
Kitazawa, K. & Kobayashi, K., 1968. Intensity variation of the geomagnetic field during the past 4000 years in South America, J. Geomagn. Geoelectr., 20, 7-19.
Rogers, J. , Fox, J. M. W. & Aitken, M. J., 1979. Magnetic anisotropy in ancient pottery, Nature, 277, 644- 646.
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
The archaeomagnetic field in Egypt 3000- 0 BC 55
Rusakov, 0. M. & Zagniy, G . F., 1973. Intensity of the geomagnetic fieid in the Ukraine and Moldavia
Shaw, J . , 1974. A new method for determining the magnitude of the palaeomagnetic field, Geophys. J.
Shaw, J. , 1979. Rapid changes in the magnitude of the archaeomagnetic field, Geophys. J. R . as@. SOC.,
Thompson, R., 1978. European palaeomagnetic secular variation 13000-0 BP, Pol. Arch. Hydrobiol.,
Walton, D., 1979. The geomagnetic intensity in Athens between 2000 BC and AD 200, Nature, 277,
during the past 6,000 years, Archaeomerry, 15,275-285.
R. astr. Soc., 39,133-141.
58,107-116.
25,414-418.
643- 644.
Appendix
Apart from their geophysical interest, Dr Games’s results, if confirmed, will clearly have important archaeological applications. Valuable as radiocarbon dating has been, from the point of view of ‘historical archaeology’ it suffers from three disadvantages: (i) it can rarely be certain how long a period elapsed between the cutting of ancient timber and its use in a particular structure or its deposition as charcoal in a particular stratum; (ii) the standard errors given in determinations of radiocarbon dates are often too great to make them par- ticularly useful in historical archaeological contexts: (iii) the necessity, originally demon- strated by the comparison of a long sequence of radiocarbon determinations of carbon specimens from sites in Egypt with their ‘historical’ dates, to use the S u e s Bristlecone-pine calibration for dates later than 5 0 0 0 s ~ has introduced an extra uncertainty. Thermo- luminescent and magnetic data of potsherds likewise suffer from the disadvantage that an unknown time interval must in most cases have elapsed between the firing of the pottery vessel and the deposition of the sherd in the stratum in which it is found. Mud-brick, however, is a material which is virtually always freshly made and sun-dried for the building of individual structures; and even in those cases where mud-bricks have been re-used from earlier structures, this should normally be observable by the archaeologist from the condition of such bricks, their sizes, and the manner in which they have been employed. Mud-brick, in areas where it was or is a basic building material, is almost ubiquitous and almost indestructible. Nothing could be simpler to provide and be less subject to erratic documentation than the requirement of a single half-brick hewn from the interior of a wall.
The choice of Egypt as a field from which to obtain a sequence of test samples had two obvious advantages. First, mud-brick has been in use in Egypt from the fourth millennium BC to the present day. Egypt should be capable with the advance in modern scientific excavation of supplying carbon, pottery and mud-brick samples from a series of controlled contexts with ‘historical’ dating evidence, spread evenly both in point of location and time sequence, which would allow thorough testing and correlation, magnetic and radiocarbon systems of dating. That the series which the Egypt Exploration Society and other archaeo- logical missions in Egypt have, with the generous permission and cooperation of the Egyptian Antiquities Service, helped Dr Games to collect is chronologically incomplete and not at present evenly spaced through history has been due to the fact that excavation is uneven in its incidence and progress, and that sites quickly become sanded up in Egypt. Time and careful planning would be required to produce the perfect trial sequence; but the potential value of t h s method to archaeologists is certainly such as to justify the most rigorous testing.
Secondly, the chronology of ancient Egyptian history is reasonably well established back to the beginning of the third millennium BC. All reigns back to 71 5 BC (beginning of Dynasty XXV) are futed by synchronisms with other cultures or by exact information within a
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018
56 K. P. Games
maximum limit of error of 2 yr, and in general precisely. Between 1100 and 715 BC the exact dating of individual reigns and even in certain cases the order of rulers is much less certain, and for this reason no samples of this period (Dynasties XXI-XXIV) have been used in this initial study: but, even in this period, at no point is an error of more than f 25 yr at all probable. The dates of accessions of the Ramesside Kings of Dynasties XIX-XX though uncertain in detail, can be fixed reasonably closely by monumental and documentary evidence from the accession date of Ramesses 11, c. 1290 BC, down to that of &messes 111 in 1 1 8 2 ~ c , which are well-fixed by Near Eastern synchronisms. Within the preceding XVIIIth Dynasty, two ancient observations of the heliacal rising of Sirius (Sothis) allow astronomically calculated fixed dates for the second king, Amenophis I (9th year = 1541 BC
f 6 yr) and (partially) for the sixth, the conqueror Tuthmosis 111, though arguments about whether the observations were made in the latitude of Thebes or that of Memphis cause a possible 25 yr shift in the calculation. Nevertheless, the succession and length of reigns of the XVIIIth Dynasty is well fixed within this limit of error. For the preceding Second Intermediate Period (Dynasties XIII-XVII), dates and sequence of royal reigns are often very uncertain, and no dates of this period have been used in this initial survey. But an ancient observation of the heliacal rising of Sirius in the reign of Sesostris 111 of the XIIth Dynasty (9th year = 1872 BC t 6 yr) fixes the preceding XIIth Dynasty, for which the lengths of all reigns are known, to the period 199 1 - 1786 BC * 6 yr. Thus the uncertainties of the Second Intermediate Period are confined within the period 1786-1670 BC, without effect on earlier periods. Previous to 1991 BC, the chronology is much less certain. However, a summary in the Turin papyrus, our most reliable ancient document, gives a total of 955 yr for the period from the beginning of Egyptian Dynastic history (Dynasty I) to the end of the Memphite Old Kingdom (Dynasty VEII). This leaves an uncertainty concerning the duration of the First Intermediate Period (Dynasties IX-XI), but a combination of monumental, archaeological and stylistic evidence suggests strongly that they can hardly have exceeded a period of more than 150 yr. Thls would give a date of c. 21 50 BC for the end of Dynasty VIII, and thence of c. 3100 for the beginning of Dynasty I. These dates have been challenged (a) by scholars wishing to reduce the length of the First Intermediate Period; (b) by those wishing to lower the Turin Papyrus figure on the grounds that the papyrus, a XIXth Dynasty document, may have duplicated or extended reigns. They would advocate a date of 2900 BC for the beginning of Dynasty I: but even if these arguments are allowed 3000 BC k 100 yr is a reasonably soundly attested date for the beginning of Dynasty I. Within the third millennium, the succession of individual rulers is fairly well established, and in many cases lengths of reigns can be estimated on a combination of king lists and monu- mental sources. An error of f 25 yr in Dynasty VI mounting cumulatively to f 100 yr in Dynasty I should be allowed, but the relation of individual reigns is better established than this figure would suggest. The Egyptian Dynastic sequence provides a very sound test of Dr Games’s thesis, even though not scientifically demonstrable in the strictest sense.
H. S. SMITH, Department of Egyptology, University College London, Cower Street, London WClE 6BT
Downloaded from https://academic.oup.com/gji/article-abstract/63/1/45/667923by gueston 15 March 2018