BXTRAIT

DOCUMENTS
DE L’ISLAM MEDIEVAL
S P
DE BECHER.G—HE'

AGTES?-:DE:=iLA,?.TABI3B-='RONDE

fill REG!-Ella-IE SCIBNTIFIQUB

(Pub. 3-'5 mltl

9145.; '1T§‘B

4-@>»* E1
A f ,\‘

INSTITUT ERKH'GQ"lfi"KKBHHDLOGIE ORIBNTALB DU CAIRB

TAII29 -— T991
 courtesy of Stefan Heidemann

Michael L. BATES
AMERICAN NUMISMATIC SOCIETY
New YORK

COINS AND MONEY IN THE ARABIC PAPYRI

A few years ago I began to collect references to Egyptian exchange rates in the
papyri, Geniza documents, and literary texts as data for comparison with the coins.
The texts turned out to be a mine of information on all aspects of the coinage, illustrating
how coins were used as money by the people for whom they were made. The file of
notes was soon extended to include every reference to coins that I could find in the papyri
and the Geniza. At the same time, I was struck by the frequent misunderstandings of
the editors and commentators of these texts when they tried to explicate coin references,
misunderstandings resulting from ignorance of the contemporary circulating coinage of
Egypt in the early middle ages. Admittedly, their ignorance is shared to a large extent
by many numismatists, who have to a too large extent studied coins one by one and
issue by issue, not as elements in a complete monetary system that evolved with the
years. No numismatist has yet tried to summarize Egypt’s early Islamic monetary
history in a coherent narrative that can be used as a reference by non-specialists.
‘ So, the first part of this paper will describe coinage in Egypt from the Arab conquest
in the year 21/641 to the conquest by the Fatimids in 358/969. The Arab conquest of
Egypt brought no change to Egypt’s currency, but the Fatimid conquest was an important
turning point and makes a convenient place to stop. Space does not permit the continu-
ation of the story of Egypt’s monetary system through the F5/timid, Ayyfibid, and
Mamlfikperiods, but one can say that the remarks made here about the operation of
the monetary system apply in general to those later periods also and not only to
Egypt, but to the entire Islamic world and even beyond to medieval Europe and Byzantium.
In fact, the major phenomena of Egypt’s monetary system in the early Islamic centuries,
such as the fluctuating values of coins in different metals, the necessity to weigh coins
in transactions, the autonomy of mints, and the dependence of minting on the willingness
of the general public to bring in coins, not upon the desires of the ruler, were in large
part characteristic of all monetary systems, ancient, medieval, and early modern (except
for East Asia) until the 19th century. The points made here can be supported from the
abundant documentation of the Egyptian papyri and the Geniza papers, but the interpret-
ation of these can be assisted, and I think rightly so, by a general familiarity with monetary
phenomena in the rest of the Islamic world. Egypt had its peculiarities-~and one
~l-l» MICHAEL L. BATES

important point that needs to be remembered is that each country of the Islamic world
had its own currency even when it was part of a wide-spread empire——but there were
broad similarities with the functioning of currency in other Muslim as well as Christian
countries.
The second part of this paper will discuss the use of these coins in the monetary
system, and try to show how the papyri illustrate the system as well as how an understanding
of the system contributes to the reading of the papyri.
The only Byzantine mint captured by the Arabs in the early conquests was in Egypt,
at Alexandria '. Though we know nothing about the transition in the city from Byzantine
to Arab administration, there is reason to believe that the workers at the mint were, in
essence, told to continue with what they were doing. There are, according to Byzantine
numismatists, Byzantine coins from Alexandria at least to very nearly the time of the
Arab conquest '1, and there is a large series of copper coins bearing the abbreviated mint
century name ALEX which do not seem to be Byzantine issues, which come from seventh
archeological contexts in Egypt and almost no where else, and which therefore are almost
certainly issues of the Alexandria mint under Arab rule. These coins, however, mostly
have no indication at all of Arab authority—no inscriptions and no modification of the
crosses that appear on both faces of the coins 3. For this reason, it is likely that the
post-Byzantine series was initiated very soon after the city was taken, virtually as a
continuation of Byzantine minting, for if minting had ceased, only to be initiated much
later, we would expect that such an innovation would have attracted more attention
from the Arab rulers and some emblem of Muslim or Arab sovereignty would have
been placed on the coins. /I propos, it is probably not correct to call these << Arab
imitations >> of Byzantine coins, since, if minting was not interrupted, or only briefly,
the men who made these coins must surely have been the same men who made them
under Byzantine rule, and can hardly be accused of imitating their own work.
These coins, like their Byzantine predecessors, have one or more standing imperial
ligures on the obverse, with no inscriptions. The reverse has the two capital letters I
and B, which together are the Greek numeral 12, because the coins under the Byzantines

l. In Syria, there had been a mint at Antioch, but it was closed before the Persian conquest, some
twenty-five years before the Arabs came. There may have been short-lived emergency mints in Syria
during the Persian occupation. The only other Byzantine mint taken by the Arabs was in Carthage,
captured after 695. i
2. Sec Philip Grierson, Catalogue of the Byzantine Coins in the Dumbarron Oaks Collection and in
the lVlu'!!em0re Collection II, 2, Washington, D.C., 1968, p. 467 for coins attributed to Constans ll.
The Alexandria coppers do not bear the emperor's name and attributions are not completely without
doubt; moreover, the distinction between issues of the Byzantine period and later ones is not at all
clear cut.
3. The lack of any specific Arab characteristic on these coins means that the attribution of some
issues to the Byzantine or to the Arab period is problematic. The rule of thumb at present is : any
issue rejected by specialists in Byzantine coinage must be Arab, but the judgement is evidently subjective.
The attribution of the more common Arab issues is, however, beyond much doubt, supported by
their abundance at Arab sites such as Fustit.
 COINS AND MONEY IN THE ARABIC PAPYRI 45

had a stated theoretical value of 12 nummi ". As far as we know, this theoretical nummi
value was ignored in practice, because the papyri speak of folles and fulzis, units of
roughly the same value as a whole coin. Below these letters in the exergue is the mint
inscription ALEX, with the L and X represented by Greek lambda and chi. These coins
have a distinctive thick fabric, characteristic of Egyptian copper, and are very irregular
in weight, from less than one gram to over eleven grams 5. The Byzantine coinage of
Egypt as well as the Arab-period coinage that followed it was quite different from the
coinage of other Mediterranean countries, and seems to have circulated almost exclusively
in Egypt. ‘ . "
There are quite a few varieties of these very common coins, with no obvious way to
put them in chronological order or assign an absolute date to any given issue 5. The
only issue that stands out is one that must have been late in the series, because it differs
in the mint name: instead of ALEX, it has MA CP 7, which would be pronounced << Masr >>,
the Arab name for Egypt and more specifically for its capital which at that time was
Fustat. This inscription provides two interesting bits of information. First, Egypt’s
mint was moved from Alexandria to Fustat sometime during the period of issue of these
coins. Fustat continued to be the location of Egypt’s main mint, usually its only mint,
for the next several centuries, until the end of the Fatimid period. The other interesting
fact that the spelling of the mint name proves is that the Arabs of Egypt in the seventh
century called the country << Masr », as they do today in colloquial Arabic, not << Misr »
as the classical dictionaries have it. ,
Just as we can only speculate about the beginning of this Byzantine-type coinage, we
are left to speculate about the date of its replacement by Arabic coinage. It is possible
that Arabic copper coinage began in Egypt at the same time as or shortly after the
introduction of Arabic gold and silver coinage in Syria, the metropolitan province of
the caliphate, in 77-79 H. / A.D. 697-699 3. It does not necessarily follow, however,
that Egypt made the transition simultaneously. In the East, that is in Iraq and Iran,
old style copper coins continued to be issued along with the new Arabic type for

4. Under the Byzantines there were also fractions with S for 6 and gamma for 3, but no such coins
attributable to the Arab periodhave been identified.
5. These figures are from Henri Amin Awad, << Seventh Century Arab Imitations of Alexandrian
Dodecanummia», American Numismatic Society Museum Notes, 18, 1972, p. 114. A metrological
study by George C. Miles, << On the Varieties and Accuracy of Eighth Century Arab Coin Weights »,
Eretz-Israel, 7, 1963, p. 86, of Byzantine and Arab-Byzantine 12 nummia coins together (the imitations
were suspected to exist when he wrote, but had not been fully sorted out from the Byzantine issues)
showed a range of 2.54 to 18.93 grams, with a concentration of specimens between 4.80 and 5.60 grams.
6. Awad’s study, cited above, is useful but brief and does not include all varieties nowknown, nor
does it attempt a chronological ordering.
7. Jere L. Bacharach and Henry Amin Awad, << Rare Early Egyptian Islamic Coins and Coin
Weights : The Awad Collection», Journal of the American Research Center in Egypt, 18, 1981, p. 51-56.
8. For a general discussion of Syria’s coinage in the Umayyad period, see Michael L. Bates,
<< History, Geography, and Numismatics in the First Century of Islamic Coinage», Revue suisse de
numismatique, 65, 1986, p. 231-262.
46 MICHAEL L. BATES

perhaps a quarter of a century after the first Arab dirhams were struck. Another
possible epoch for the introduction of Arab coins in Egypt is the administration of ‘Abd
Allah b. 'Abd al-Malik b. Marwan, 86-90 H./ A.D. 705-709, because his name is the
earliest on the Arab glass weights for coins, or again the administration of Qurra b.
Sarik, 90-96/709-714, under whom the first glass weights for fals appear.
The previous paragraph concerns specifically the actual minting of Arabic coinage in
Egypt. No doubt the new Arabic gold dinars and silver dirhams of Damascus began
to penetrate Egypt soon after their introduction in 697 and 699 respectively, but it does
not necessarily follow that they were immediately made the unique official currency of
Egypt or that their weight standards were immediately adopted 9. The very earliest
Egyptian Arabic glass dinar weights, those of ‘Abd Allah b. 'Abd al-Malik (705-709),
are for the Islamic dinar 1°, but one weight of his successor Qurra (709-714) is for the
Byzantine solidus. Qurra’s stringent instructions in a papyrus letter to one of his
governors to accept dinars only according to the weight standard of the Treasury (‘old
wazn bayt al-mail) “ might indicate that this standard was new when Qurra wrote (the
papyrus is attributed to the years 90-91 by the editor, the beginning of Qurra’s admin-
istration) and not yet fully implemented; or that, while the standard was known earlier,
it was henceforward to be strictly and uniformly applied. Finally, it should be noted
that, whenever the change became official, it is probable that neither Byzantine solidi
nor Byzantine weight standards disappeared overnight from use in private transactions,
especially among the Christian population. 1
The earliest Arab coins of Egypt, by which is meant those with Arabic inscriptions
instead of Byzantine images, have nothing but a short version of the Muslim sahada on
both sides, << There is no god but God alone >> with << Muhammad is the messenger of
God » on the reverse. Very soon, between 734 and 742, an issue appears with the name
of the current governor, and after him appear a series of issues with governors’ names -not
every governor to be sure, because minting seems to have been intermittent-until almost
170 H. / A.D. 786-787, a-year that will be seen to be an important monetary turning point
in other respects. Miles’ classic article of 1958 listed the issues and discussed them in
sequence, but since he wrote, Awad and Bacharach have added a number of new issues
and varieties to the series, while it has also become apparent that some of Miles’ issues
are mere ghosts based on incorrect descriptions by earlier scholars, and many of the new
types that have been published, although interesting, are only accidental variants, not
new issues. What, in this context, is an issue? There is no fixed definition, but it may
be proposed that, since changes in the design and especially the inscriptions of coins

9. It will be seen below that the reformed weight standards of Egypt were different from the weight
standards employed in Damascus, but the new standards in Egypt were analogous to those of Damascus.
10. See Morton, p. 45-46. The three specimens known to him weigh 4.14, 4.22 and 4.22 grams;
another in the ANS weighs, also, 4.22 grams.
11. Adolf Grohmann, Arabic Papyri in the Egyptian Library, Cairo, 1938, no. 149, ll. 19-26
(hereafter APEL).
 COINS AND MONEY IN THE ARABIC PAPYRI 47

come only on the orders of the government, an issue is the entire body of coinage
produced pursuant to such an order, while a variety is the result of engravers’ idiosyn-
cracies, stylistic evolution, accident, or internal regulation of the mint. All the individual
coins of an issue are interchangeable in monetary value, while two separate issues may
or may not be interchangeable. It will be proposed further on that all Egyptian Arabic
coppers until 170 H. were interchangeable, that is, that none were demonetized; the
standard value of the fals was changed by changing the weights used to measure payments.
The headings within the following listing are an attempt to group related issues
together, suggesting an affinity by chronology.

A. ANONYMOUS ISSUES WITH THE SAHADA. »

1. An issue with the sahada inscribed within three concentric circles is extremely
common, and seems likely to be the earliest Arabic coinage of Egypt '2. It is very like
what was probably the earliest Syrian Arab coinage, but is easily distinguished from the
Syrian analogue by its thick fabric, like the Byzantine type coins that preceded it. The
similarity of the design is an argument for putting this Egyptian issue in the same time
span as the Syrian one, that is approximately 700-710, but it is not a conclusive argument.
2. The same inscriptions, with the same epigraphical style and fabric are found in a
square surrounded by an outer circle. This type is also common enough, but it was
overlooked by Miles in his classic article 13. Issues 1 and 2 surely belong close together
in time, and, except for the closer resemblance of issue 1 to the earliest Syrian issues,
there is no obvious reason to make either 1 or 2 the earliest.

B. ISSUES ASSOCIATED WITH AL-QASIM B. 'UBAYD ALLAH.

4. An anonymous issue with the first part of the §a/zada on the obverse, preceded by
the words bism Allah, and on the reverse a six-pointed star in a circle surrounded by the
circular inscription Muhammad rasfil Allah 1". .

12. George C. Miles, << The Early Islamic Bronze Coinage of Egypt », in Centennial Publication of
the American Numismatic Society, ed. Harald Ingholt, New York, 1958, p. 473-474, no. 1 (hereafter
EIBCE). ,
13. Some examples appear in John Walker, A Catalogue of the Muhammadan Coins in the British
Museum, II : A Catalogue of the Arab-Byzantine and Post-Reform Umaiyad Coins, London, 1956,-
p. 205-206, nos 610-615. The issue was recognized as Egyptian by ‘Abd al-Rahman Fahmi,
Fair al-sikka al-'arabiyya, p. 448-454, nos 1378-1518; cf. Bacharach and Awad, 1981, p. 53. Miles
perhaps did not include this issue in his EIBCE because many examples were found in the excavations
of Antioch which he cataloged; nevertheless the attribution to Egypt is supported not only by the
fabric of the coins but also by other archeological finds. Note the number of examples in Fahmi’s
catalogue. In general, Syrian coppers are sometimes found in Egypt at sites like Fustat, where I
worked two seasons. Egyptian coppers are much more rarely found in Syria. Dr. Lutz Ilisch
estimates that Egyptian coppers constitute some 1-3 ‘X, of the hoards he has seen from Syrian sources.
14. EIBCE, p. 474, no. 2.
48 MICHAEL L. BATES

5. An issue with a similar obverse, but with Muhammad rasiil Allah in the reverse
center and amara al-Qasimb. ‘Ubayd Alla/z in the margin; the first datable copper issue
of Muslim Egypt (116-124/734-742) *5.

6. An issue with a reverse identical to no. 4, with a central star, but with the obverse
inscription fals Misr wafin wazn tamaniyat ‘as'ar qirat, that is, << full Egyptian fals weight
of 18 carats»; unfortunately for any immediate metrological conclusion, the weight of
these coins varies from 2.65 to 4.62 grams, much like all other Umayyad fulus of
Egypt *6.

C. THE COINAGE OF ‘ABD AL-MALIK B. MARWAN.

8. An issue bearing the names of the Umayyad caliph Marwan II and 'Abd al-Malik
b. Marwan, 131-132/749-750, the last Umayyad governor of Egypt, is extraordinary
among Egyptian coppers in having been issued at a number of mints, al-Fustat, al-Fayyfim,
al-Iskandariyya (Alexandria), Atrib and Ahnas; perhaps other mint names will be found 17.
The names give some indication of the administrative centers of Egypt in the period,
but the list may notbe complete as these coins are generally rare, with at least one mint
represented by a single specimen. Except for this issue, and coins of A‘smfin around
258-259, there is no evidence for more than one mint in Egypt until the eleventh
century. It seems almost certain that this relative plethora of mints was rtatemporary
phenomenon '8.

15. EIBCE, p. 477-478, no. 3. Jere L. Bacharach and Henri Amin Awad, << The early Islamic
Bronze Coinage of Egypt: Additions», Near Eastern Numismatics, Iconography, Epigraphy and
History: Studies in Honor of George C. Miles, ed. Dickran K. Kouymjian, Beirut, 1974 (hereafter
EIBCEA), p.» 187-188, no. U.1 is perhaps a transitional issue between EIBCE 2 and EIBCE 3, or
rather a mule (accidental die combination) between them; at any rate, it suggests that the one issue
immediately followed the other. EIBCEA, p. 188, no. U.2, dated 92 H. and said to be the earliest
dated coin of Islamic Egypt, is an example of a rather common issue (John Walker, A Catalogue of the
Muhammadan Coins in the British Museum, II 3 A Catalogue of the Arab-Byzantine and Post-Reform
Umaiyad Coins, London, 1956, p. 290, nos 944-949), an issue which is also represented for the years
91-93 and 95-99. Walker suggests a North African origin for the series probably because the
horizontal inscriptions have words divided between two lines, a frequent occurrence in the Umayyad
Magribi coinage (though not only there). None of these were found at Fustat. It is true, as Awad
says, that the weight and fabric are Egyptian, but this might be attributed to Egyptian influence on a
North African outpost. a
16. EIBCEA, p. 188, no. U.3. s
i 17. Miles, EIBCE, p. 477-480, nos 4-7; mint Ahnas published by Raoul Curiel, <<Monnaie de
bronze d’Ahnas-Misr », Revue numismatique, 6° sér., 10, 1968, p. 131-137. There is also EIBCE, p. 481,
no. 8, another issue of Fayytlm (without al) of uncertain date; more examples have been published
in EIBCEA, p. 189, no. 8 b.
18. The next few copper issues, and nearly all subsequent issues, have the mint name Misr, the
only name found; the copper issues before this extraordinary one probably did not have a mint name
because there was only one mint in the country. I
 COINSAND MONEY IN THE ARABIC PAPYRI 49

D. THE ‘ABBASID RESTORATION OF AL-QASIM’S COINAGE.

9. The first °Abbasid issue of Egypt is very similar in its prominent features to the
earlier coinage of al-Qasim b. 'Ubayd Allah, to the point that the two issues are difficult
to distinguish without reading the governor’s name in the reverse marginal inscription.
This revival of an older type suggests a restoration of an older system or weight standard
in place of the immediately preceding system of ‘Abd al-Malik b. Marwan. Decentra-
lization of minting was perhaps not the only innovation of ‘Abd al-Malik b. Marwan’s
issue. The ‘Abbasid issue has the mint name Misr, the date 133/751, and the name of
the governor ‘Abd al-Malik b. Yazid *9. 1 .

E. THE LAST THICK-FABRIC ISSUES.

10. ‘Abd al-Malik b. Yazid, whose term extended to 141/758, produced another issue
without mint name or date and with the invocation aslahahu Allah 2°, an innovation on
coins which is also found on the glass weights of the 'Abbasid period.

ll. The coppers of Muhammad b. Sa'id, ca. 152-157/769-774, have a design identical
to the previous issue and the. same invocation 2'. a

12. An issue with the name of Matar, ca. 157-159/773-776, differs inarrangement of
the inscriptions but has an epigraphic styleand ornamentation similar to the previous
two issues. Matar also used an invocation, akramahu Allah 2'3. _

13. Two examples of an issue like the preceding two, but with the name Muhammad
b. Sulayman, governor 159-162/776-779, are in the collection of the American Numismatic
Society 23. _

14. The last issue of this series is dated 167 H. and has the name of Ibrahim b.
Salih, with inscriptions much like those on ‘Abbasid dirhams of the era 2";

19. Miles, EIBCE, p. 481-482, nos 9 and 10 (his catalogue entry 10 comprises incompletely described
examples of 9, not a separate issue).
20. Miles, EIBCE, p. 482, no. 11. To be ignored at this point are EIBCE, 12 and 13, which belong
to a later period, and Miles 14, which reproduces a highly implausible 19th-century attribution of a
unique coin to Farama, 147 H.
21. Miles, EIBCE, p. 485-486, no. 15.
22. Miles, EIBCE, p. 486-487, no. 16.
23. Registry numbers 197l.316.228 and 1971.316.229, from the collection of Robert W. Morris who
acquired the coins in Cairo and first identified the governor.
24. EIBCEA, p. 191, no. A.1.
50 MICHAEL L. BATES

In this first phase of coinage in Egypt under the Muslims, only copper coins were
issued. Whether these are the seventh century Byzantine type coins or the eighth
century Arabic coins, their fabric is similar--thick and dumpy-—-and the coins are very
irregular in weight, with no clearly defined peak weight but with a concentration of
specimens between 3.5 and 515 grams. Although these were the only coins struck in
Egypt 95, it is clear that foreign coins of gold and silver circulated in Egypt. The papyri
refer frequently to transactions in gold solidi or dinars and sometimes to transactions
in silver dirhams. Also, the Egyptian glass weights for coins include many examples
labeled << dinar>> or << dirham ». In the seventh century, the gold coins would
have been Byzantine solidi; in the eighth century, Islamic dinars, but probably solidi
continued to circulate as well”. References to silver coins would be expected only
in the eighth century 97; these would have been Umayyad and 'Abbasid dirhams
from other parts of the caliphate. Evidently Egypt enjoyed a favorable balance of
payments sufficient to provide it with enough gold and silver coinage to maintain
circulation.
The year 170 H. / 786-787 marks a major turning point in Egypt’s monetary history,
for in that year appear the first dinars that can be identified as Egyptian. From this
time on the word to keep in mind about Egyptian coinage is gold. Immediately the
Egyptian mint became the principal producer of gold coinage in the eastern Mediter-
ranean, and continued to be one of the most important until the early twentieth
century.
The Egyptian dinars from 170 until the early 3rd century higri can be identified by
the names of governors of Egypt on them. Until the l90’s, only the ism, or personal
name, of the governor appears; for example, the dinars of 170-171 have the name ‘Ali.
0

25. It has been hypothesized (notably by Andrew S. Ehrenkreutz, << Studies in the Monetary History
of the Near East in the Middle Ages : The Standard of Fineness of Some Types of Dinars», Journal
of the Economic and Social History of_ the Orient, 2, 1959, p. 128-161) that some Umayyad and
'Abbasid dinars bearing no mint name were struck in Egypt before 170, but this idea cannot be subs-
tantiated. The Umayyad dinars without mint name form a coherent unified series that must have come
from a single mint, probably in Damascus or at the residence of the caliph (see Bates, << History»,
p. 257-259). The 'Abbasid dinars without indication of mint from 132 to 170 H. show minor variations
that may enable them to be assigned to different mints, but as yet it has not been possible to identify
any of them as the predecessors from Egypt of the dinars (beginning 170) with Egyptian governor’s
names.
26. They seem to have circulated with official sanction for some time, because at least one glass
weight with the name of the governor Qurra b. Sarik (90-96/709-714) weighs 4.48 grams, the weight
standard of the solidus, not the dinar (American Numismatic Society 1974.268.6; Paul Balog, Umayyad,
'/ibbasid and Tulunid Glass Weights and Vessel Stamps, ANS, Numismatic Studies, 13,New York,
1976, p. 43-44, no. 3; the weight given by Balog is wrong). I
27. For example, a Greek papyrus of 90 H. / 708-709, P. Lond. IV, 1338, cited by Adolf Grohmann,
Ein/Fihrung und Chrestomathie zur arabischen Papyruskunde, I: Ein/iihrung, Prague, 1954, p. 212, n. 7,
referring to a <<mi1iaresion». No earlier references to silver coins in the Arab period are known
to me. _
 COINS AND MONEY. IN THE ARABIC PAPYRI 51

This could be anyone; but taking the sequence of names as a whole, it is obvious that
they correspond to the sequence of governors of Egypt. The list is as follows :

Name. Dates on coins.

‘Ali . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170-171.
Mftsa . . . . . . . . . . . . . . . . . . . . . . . . . . 171-172.
‘Umar . . . . . . . . . . . . . . . . . . . . . . . . . 172-174.
Dawfid . . . . . . . . . . . . . . . . . . . . . . . . . 174.
Mfisa .. . . . . . . . . . . . . . . . . . . . . . . . . . 175.
Ibrahim . . . . . . . . . . . . . . . . . . . . . . . . 176.
\-:~ ,.

Ga‘far . . . . . . . . . . . . . . . . . . . . . . . . . 176-186.
I;lalid . . . . . . . . . . . . . . . . . . . . . . . . . . 187. D
No name . . . . . . . . . . . . . . . . . . . . . . . 187-188.
al-Halifa . . . . . . . . . . . . . . . . . . . . . . . 189-194.
al-l;lalifa/al-Amin . . . . . . . . . . . . . . . . 195.
al-I;lalifa/al-imam . . . . . . . . . . . . . . . . 196.
al-I;Ialifa/al-Ma’miin/'Abbad . . . . . . 196-198.
al-Imam/al-Ma’mi‘1n/al-Fadl . . . . . . 198.
al-Imam/al-Ma’mf1n/al-‘Abbas . . . . . 198.
al-Imam/al-Ma’mf1n/lid-Muttalib ... 19s-199.
lillah/DEVI-Riyasatayn/li-Muttalib ' - a 199.
'

At dubious points in the chain, such as the issues with Ga‘far (the Barmakid wazir),
which could have been issued anywhere, the continuity is established by die links, that
is, by the existence of coins with the name Ibrahim and coins with the name Ga'far
struck from the same obverse die *8. Ga'far was in fact the nominal governor of Egypt,
though he never went there and administered the country through lieutenants. Once
the sequence has been established, the distinctive criterion for the identification of
Egyptian dinar issues is obvious: all dinars with a single name below the reverse field
inscriptions (170-194, except 187-188), or with a title above and below the reverse field
(195-196), or as the latter with a third name below the obverse field (196-199), are from
Egypt 29.

28. The die link sequence was worked out by a student in the ANS Graduate ‘Summer Seminar
some years ago. Unfortunately the student’s paper was never completed and the student has left
academia with the notes.
29. The remaining ‘Abbasid dinars of the period are mostly to be assigned to Madinat al-Salam
(Baghdad), though there are some evidently from other mints. The Aglabid dinars of Ifriqiya take the
Egyptian design, adding the single word galaba above the reverse field. Until 198, no ‘Abbasid dinar
has the actual name of the mint city, and many after that date also lack the name. The general
reform of about 215 included the universal use of mint names on dinars, except those of the Aglabids
who were no longer under caliphal control.
52 MICHAEL L. BATES

After 199, nearly all Egyptian dinars have the mint name Misr or the full name of an
Egyptian governor. In 215 Egypt adopted the new anonymous style ordered by al-Ma’miin,
always with a mint name. The caliphs after al-Ma’mfin had their own names on the
dinars, with the name of the wali 'alzd if one had been proclaimed. As a mark of special
autonomy, the names of the Tfilfinid and llfsidid governors also appear on coins of their
administration, but the general design of Egyptian gold coinage was not otherwise
changed.
From about the end of al-Ma’mDn’s reign until the Fatimid conquest, then, Egypt
had a stable and apparently abundant gold coinage, while silver dirhams, though not
prohibited, were not struck in great quantities. In the early 3rd/9th century some debase-
ment of the gold coinage began, culminating in the 320’s with a precipitous decline in
fineness. When Muhammad b. Tugg the lljaidid began issuing dinars with his own
name in 331, the fineness of the Egyptian dinar was restored to a very high level that
was maintained by his successor 3°, but nevertheless problems with the currency remained.
When Gawhar, the Fatimid general, arrived on the west bank of the Nile in 358
he promised a delegation from Fustat that he would restore the coinage, specifying a
change in the design, the raising of its fineness to the level of al-Mansiir’s coinage,
and the cessation of alloyingin it as the three qualities necessary for a sound
coinage 3'. r
Probably minting of silver dirhams in Egypt was introduced in the same year, 170,
but this is not certain as the earliest known dirham of Misr is dated 171 /787-788 3'2.
In general, however, dirhams of Egypt from 170 until the Fatimid conquest are exceedingly
rare today. For most years, no dirhams are known at all, and for many years
only one or two examples survive--in contrast to Egyptian dinars, which are quite
common.
The copper coinage of Egypt after 170 is quite different in fabric from that of the
preceding years, being rather small (12-13 mm), thin, and neatly inscribed. There are
issues with the names Mahffiz [b. Sulayman] and Salih [b. Muslim], 186-187 33; with
l

30. For the fineness of Egyptian dinars from al-Ma’mt'm’s reign to the Fatimid conquest, see
Andrew S. Ehrenkreutz, << Studies in the Monetary History of the Near East in the Middle Ages, II:
The Standard of Fineness of Western and Eastern Dinars Before the Crusades», Journal of the
Economic and Social History of the Orient, 6, 1963, p. 148-153.
31. Maqrizi, Itti'a;: al-lzunafa' fi alibar al-hulafa’, p. 104. The Fatimid al-Mansftr introduced a
markedly handsomer coinage than that of his predecessors in North Africa, and though his coinage
does not seem substantially higher in fineness, it is possible that he regulated its alloy more strictly.
See Ehrenkreutz, << Studies», 1959, p. 257. It is nevertheless surprising to find his name in this
context, as the Mu'izzi dinar of the reigning caliph in 358, with its radically different concentric
design, is the one always mentioned as the normal coinage in documents and texts of the
period. 2
32. In the collection of Tiibingen University; with the name Musa, as on dinars of the later part of
the same year. 1
33. EIBCE, p. 489-492, nos 17-22.
 COINS AND MONEY IN THE ARABIC PAPYRI 53

the name Gabir (ca. 195) 3°; with al-Sari b. al-Hakam (200-203) plus various other
oflicials 35; and with Abfi Ishaq b. al-Ra§id and ‘Isa b. Mansfir (215-227) 3°.
With this last issue copper coinage in Egypt ceases, and about the same time it stops
being issued in Syria and the rest of the ‘Abbasid caliphate. No one has any idea why
this happened, but it is closely connected in time with the establishment of the new
style coinage of 215/830-831 (in Egypt; a few years earlier in the eastern caliphate) and
it therefore seems likely that copper was prohibited as part of the general coinage
reform of al-Ma’mfin’s late years and the reign of al-Watiq. Copper coinage was
resumed in Egypt briefly under the first Tfilanid Ahmad at the beginning of his admin-
istration with an issue bearing the mint name Misr and the dates 258 and 259 (872-873) 37.
Also from the Tftlfinid period are some very rare issues, one bearing the mint name
Madinat Asmiin, that seem all to come from the same hoard 38. Copper was not issued
again in Egypt until 622/1225-1226 in the reign of the Ayyiibid al-Kamil 3°.
Aside from the coins, an important and unique feature of Egypt’s currency system
is the glass weights for coins. As far as I know, glass was not used to make weights
anywhere other than in Egypt, except very occasionally in Egypt’s immediate neighbors
under Egyptian influence "°. Coin weights, of course, existed almost everywhere before
the 19th century, but only in Egypt and its neighbors have they been made of glass.
Roman and Byzantine glass coin weights from Egypt are fairly common, and there
are a few that were made, probably unoflicially, during the Arab-Byzantine seventh
century. But the Arab glass weights, which begin quite suddenly in the early eighth
century, are extraordinary for their precision and for their extensive and legible inscriptions,
which name one or more olficials under whose authority the object was made, state the
purpose of the weight, sometimes give the date, and usually add one or more pious
phrases. Along with the coin weights there are also heavier commodity weights and

34. EIBCE, p. 483-484, nos 12-13; Miles identified these as coins of Gabir b. ‘Ubayd, an ‘amil in
133-141, rejecting an attribution to Gabir b. al-As‘at, 195, with the comment << the style of the coin
does not suit the [latter date] ». On the contrary, the style (and fabric) of these coins is very like those of
the late 2nd and early 3rd centuries Higra. It is hard to understand Miles’s comment, except that he
had seen only one specimen of the issue. EIBCE 12 and 13 are varying descriptions of the same issue,
which was quite common in the Fustat excavations.
35. EIBCE, p. 493-495, nos 23-26.
36. EIBCE, p. 495-496, no. 27.
37. EIBCE, nos 29-31.
38. EIBCE, nos 32-37.
39. Paul Balog, The Coinage of the Ayyublds, Royal Numismatic Society Special Publication, no. 12,
London, 1980, p. 159-160. A very few copper coins have been attributed to the Fatimids, but these
are almost certainly debased dirhams that have acquired a thin outer layer of copper from corrosion
in the ground or in the cleaning process. In any case, there is no reason to attribute any of these to
Egypt (cf. Paul Balog, <<Monnaies islamiques rares fatimites et ayoubites», Bulletin de l’1nstitiit
d'Egypte, 36, 1953-1954, p. 336-341).
40. A few Umayyad Syrian glass coin weights are known. Glass weights are known with the names
of the Fatimid caliphs who ruled in lfriqiya and Sicily only, and it is possible that some of those with
the names of later Fatimids are also from these countries.
54 MICHAEL L. BATES

stamps for measuring containers, and all these have to be studied together since they
have the same kind of inscription and were made in the same workshops; but only the
coin weights are directly relevant to the monetary system. I
A splendid recent catalogue of these objects "1 has elucidated many of their mysteries,
but the question of the use of the coin weights is left open. Morton seems reluctant to
admit that they were used for coins at all. I think he is unduly cautious. We have on
the one hand a group of objects that are explicitly labelled as weights for coins, including
three different denominations of dinars; dirhams and half dirhams based on four
standards explicitly identified by the inscriptions on the weights; and fals of a number
of different standards also explicitly identified. On the other hand, we know what sort
of money was circulating in Egypt in the era of the coin weights.
Gold coinage over the years included Byzantine solidi, Damascus Umayyad dinars,
probably Umayyad Magribi dinars from North Africa and Spain, and 'Abbasid dinars
from Iraq. Each of these coinages had its own weight standard and each, like all
pre-modern hand-made coinages, varied somewhat around that standard, with more
coins being below the standard than above it. Silver consisted of imported dirhams
issued to a variety of weight standards (the 7/10 mitqal standard, or 2.975 grams, is a
myth), and much more irregular in weight than the gold. Copper was minted in Egypt,
but varies enormously in weight with no weight standard apparent. The coinage,
therefore, would have to be weighed, and there is contemporary evidence that it was
usually weighed "’. Specifically from the early period, there is a papyrus dated 57 H.,
that is 676-677, before the introduction of the Islamic dinarand of the Arabic glass
weights, that refers to 118 1/6 nomismas (solidi) with a value of 109 19/24 nomismas, that
is, there were 118 1/6 by count, but with a weight and thereforea value of only 109 19/24“.
In 90-91 H. the governor, Qurra b. Sarik, ordered one of his tax officials to accept
dinars and fractions from the arrival of the letter only at the treasury weight (wazn bayt
al-mal), indicating that coins had to be weighed, whether the text is interpreted as meaning
that only coins conforming to the standard were acceptable or as meaning that whatever
coins were accepted had to be evaluated by weighing them against the oflicial standard "°.
Given that the glass weights were used to weigh coins in some fashion, and postponing
the question of how they were used, it follows that the standards of the glass coin weights

41. A.H. Morton, A Catalogue of Early Islamic Glass Stamps in the British Museum, London, 1985.
42. Adolf Grohmann, Ein/iihrung, p. 186-187, cites numerous examples from the papyri, but drawn
from diverse periods, not only that of the Umayyad and 'Abbasid coin weights.
43. PERF 573, quoted in Grohmann, Einfiihrung, p. 183-184.
44. Cair. B.E. 342 = Adolf Grohmann, APEL III, no. 149. Incidentally, this text might be an
indication of the date of the introduction of the Islamic dinar standard in Egypt. It is noteworthy in
this respect that one of Qurra’s known weights is for solidi, at 4.48 grams (cited above), while the rest
are for dinars of about 4.23 grams. Against this dating of the change in weight standard are the dinar
weights of al-Amir 'Abd Allah, plausibly identified as 'Abd Allah b. 'Abd al-Malik b. Marwan, which
weigh 4.14, 4.218, and 4.225 grams (examples from Morton’s catalogue and the ANS collection),
referring evidently to the weight standard of the Islamic dinar.
 COINS AND MONEY IN THEARABIC PAPYRI 55

are a better indication of the weight standards for gold, silver, and copper coins than
anything we can obtain from the coins themselves. The overwhelming majority of the
weights for dinars, and for thirds and halves of dinars when multiplied appropriately,
fall in the range 4.22-4.23 grams in the Umayyad period and 4.22-4.26 grams in the
'Abbasid period. The standard seems essentially unchanged throughout the eighth
century. Such weights could be used not only to test individual coins for conformity
to the standard, but also as a measurement of value in themselves: by weighing any
batch of acceptable gold coins, Byzantine, Damascene, Magribi or Iraqi, against standard
weights, one could evaluate those coins in Egyptian dinars °°. It is not certain, for
example, that the Byzantine solidus of about 4.50 grams was demonetized in Egypt
when the reformed Islamic dinar of 4.23 grams was introduced. Provided that the gold
fineness of the Byzantine coins was regarded as equal to that of the new dinars, a payment
could be made by weighing any number of either coinage (or the two mixed) against
standard weights totalling the desired amount. Because the Egyptian standard, say
4.23 grams under the Umayyads, was slightly lower than the Damascus standard of
4.25 grams, the Damascus dinar was overvalued in Egypt by about 1/2 %, surely a
factor that helped Egypt attract and retain gold coinage without a mint of its own.
For dirhams, four different weight standards are attested by the weights. The most
common, usually designated merely << dirham » on the weights, is identified on one issue
as <<dirham of two-thirds» which is exactly what it was, a dirham of two-thirds the
weight of the mitqal, or 2.83 grams. This was the earliest dirham standard in Egypt,
until the end of the Umayyad caliphate.
The next dirham defined by the glass weights is the <<dirham of 13 harriiba» or
<<dirham of 13 qirat» which is otherwise unattested. It had two successive weight
standards. One group with the designation << family of Muhammad» (/fl Muhammad),
by consensus assigned to the first years of the 'Abbasid caliphate, has a weight standard
of 2.62 grams, while a second group from the administration of three different governors
of 135-143 H. has a weight standard of 2.50 grams "°. P

45. This statement, however, ovcrsimplifies the procedure for cases in which the foreign coins were
not regarded as equivalent in value to Egyptian currency. In the latter case, the foreign coins were
weighed to determine their amount in Egyptian dinars. This amount in foreign coins would then be
evaluated at the current rate of exchange between the Egyptian and the foreign currency. When an
Egyptian document states that a specified foreign dinar has a specific value in Egyptian dinars, it does
not mean that a single foreign coin has that value but rather that an Egyptian dinar’s weight of that
currency has the value stated.
46. Morton, Catalogue, p. 18. These weights, like all small groups of weights with the same stamp,
are very precise; of the five /I1 lvluhammad weights in the ANS, four weigh 2.62 grams and one 2.64 grams;
eleven of the twelve weights of the second group weigh 2.50 or 2.51 grams, with one, perhaps invisibly
damaged, of 2.48. An undefined <<dirham» of the Al Muhammad is known, published twice with
divergent weights of 2.88 and 2.92 (the same specimen), and two undefined << dirham » weights of one
of the three governors are known, one weighed at 2.90 grams. This fragmentary evidence may suggest
another weight standard. Otherwise the relatively common stamped glass of these officials includes
no dirham weights, only dirhams of 13 harruba.
56 r MICHAEL L. BATES

The dirham of two-thirds was then revived; it is not a coincidence that the only glass
weights for this standard that are explicitly labelled << dirham of two-thirds » are those
that immediately follow the dirhams of 13 barriiba ‘*7. The last of the weights Morton
considers to be part of the << official series» are for this weight standard, but they are
followed by a group of <<uno1ficia1 weights» for the <<dir/ram kayl», the dirham of
7/10ths of a migqdl. When this became established in turn as the ordinary dirham
weight, the qualifying adjective was dropped. Surviving specimens weigh 2.92 to 2.95
grams, slightly below the 2.96 gram level that would be expected as 7/ 10ths of the Egyptian
migqiil of 4.23 grams. The precise date of the introduction of the dirham kayl is not
certain, but the weights with this designation are all ulterior, afterthe end of the main
Umayyad/'Abb2'1sid weight series in the l70’s "8 and after the introduction of gold and
silver minting. Documents using the term <<di/‘ham kayl» should be assigned to the
late second century or the the first half of the third "9.
In sum, during the century of the Umayyad and 'Abbasid Egyptian glass weights, the
standard of the country’s silver money was changed four times. The dirham standard
of the Umayyad period was 2/3 migqiil, which was changed to a 13 [zarrflba standard
based on a barrziba of 0.2015 grams at the time of the 'Abbasid conquest, changed again
to a 13 barrfiba standard based on a lrarrfiba of 0.1923 grams (very nearly the normal
figure used for the lzarriiba) from 136 to 143, thenreverted to the 2.83 gram standard
until sometime after 170 when the << weight of seven» standard of 7/10 migqiil was
introduced.
These changes in weight standard were not accompanied by any changes in the weight
of silver coinage. In fact, during most of this period, Egypt had no silver coinage of its
own. The surviving specimens of the Egyptian dirhams after 170 H. are too few to
enable any estimate of their weight standard. No corresponding changes can be
detected in the easternand Magribi Umayyad and 'Abbasid dirhams which would have
made up the bulk of Egypt’s silver money. In any event, no change in the coinage was
necessary. The government could devalue or revalue the silver coinage simply by changing
the standard of the glass weights used to measure value in silver.
This evidence shows how careful one must be in interpreting exchange rates in the
papyri or indeed in any documents. During the course of only a century, the weight
standard of the dirham was changed four times. During the Fatimid, Ayyubid and
Mamluk periods the alloy of the silver coinage was often changed, and the weight

47. E.g., Morton, Catalogue, p. 101, no. 244, of the governor Yazid b. Hatim (144-152). This is the
only issue with << dirham of two-thirds»; later stamps for his standard revert to << dirham ».
48. Morton, Catalogue, p. 18-21. Weights of the dirham kayl standard continued to be made and
used long after the term <<kayl» ceased to appear on the weights.
49. Examples include P. Michaelides 30; PER Inv. Ar. Pap. 8651, cited Grohmann, Einfiihrung,
p. 213, n. 6; PER Inv. Ar. Pap. 772, cited Grohmann, APEL II‘, 1936, p. 103; PERF 691, quoted and
assigned to the 2nd-3rd centuries by Grohmann, From the World of Arabic Papyri, Cairo, 1952,
p. 147-148. .
 COINS AND MONEY IN THE ARABIC PAPYRI 57

standard sometimes changed also. A given exchange rate has quite different implications
depending on the alloy and fineness of the dinars and dirhams that are involved,
making it necessary, in interpreting these rates, to date the documents as closely as
possible and determine what dinars and dirhams were in use, their alloy, and what
weight standards were used to evaluate them. Conversely, any change in alloy or fineness
of either coin resulted in a new normal figure for the exchange between dinars and dirhams;
to the extent that we can learn the history of the coins and determine typical exchange
rates for each period, exchange rates quoted in documents can serve as an indicator of
their approximate date. At this time, however, the work of identifying the successive
gold and silver issues of Egypt and their fineness and weight standards has scarcely begun.
Successive changes in the weight standard of the copper currency of the eighth century
can now also be identified, thanks to Morton’s observation that the glass weights for
copper coins, usually identified as << weight for fals of so-and-so many qird; » or
<< barrfiba » (the two terms seem synonymous), can be arranged in chronological
sequence according to the number of qirrit per fals. These fals weights are known
from the timeof Qurra, 90-96, until the late l60’s; that is, the glass weights for copper
end at the same time as the thick dumpy fabric Egyptian copper coinage ends, just
before the introduction of gold and silver minting in Egypt. It may seem odd to have
weights for measuring payments in copper, which normally seems to have circulated at a
token face value even in pre-modern times, but during the era in which glassfals weights
were made copper coins were the only native minting of Egypt and must have played
a larger role in everyday economic life with a higher purchasing power than we would
normally attribute to copper. It is probably not coincidental that copper disappeared
in Egypt within a half-century after the introduction of silver coinage there. Prices of
everyday commodities in Egypt are quoted in dirhams from the third century until the
seventh. Before the introduction of silver coinage, copper would have been used in
larger quantity to pay larger sums. The absence of special weights for fals coins in the
era of the thin light copper coinage from 186 to 227 and in the Tiilfinid period suggests
that these coins circulated at a fixed value, regardless of their weight, though for large
sums weighing was always easier than counting and could be carried out with ordinary
commodity weights. t
As before mentioned, the copper of the seventh and eighth century, until 170, is
extremely irregular in weight, to the point that we would not be able to determine its
weight standard from the coins. Payments in copper were thus necessarily measured
by weight; and so the government was able to change the nominal value of the copper
coinage merely by changing the standard weights used to measure its value. In addition
to the changes in the number of qird; per fals, there were also changes in the weight
of the qird; itself. Someof the fals weights are based on the heavier qird; used also for
the first series of dirhams of thirteen barrfiba, while most are based on what is considered.
the normal qird; of 0.195 grams. As Morton says, << the metrological picture is still far
from tidy» and we are still not in a position to draw up a chronological tabulation of
the different denominations of fals weights and compare them to contemporary coins,
58 MICHAEL L. BATES

which may also have fluctuated in average weight as Morton suggests 5°. Here again,
however, prices given in fals in this period by the papyri need to be interpreted in the
light of changing weight standards for the copper coinage.
These olficial glass weights are not as common as we might expect, especially in
comparison to the very common glass coin weights of the Fatimid period, and were
probably not the only means of measuring payments by weight. First of all, there are
found in Egypt many bronze weights which are seldom inscribed and thus difiicult to
assign to a particular period, but which, it seems to me on the basis of my experience
at Fustat, come from every archeological level and probably every period. Enough
inscribed weights are known to say that these were used in the Roman, Byzantine, and
Islamic periods through to the nineteenth century, and indeed, of course, still today,
though not for coins. The chronology of the types of bronze weights is quite uncertain.
Nearly all these weights have gram weights corresponding to multiples of the dirham
or the dinars‘. Secondly, the fact that no multiple glass coin weights have survived
from the period of the unit and fractional glass weights does not mean they did not
exist. I would have said that no large multiple glass coin weightsever were made, and
that no glass weights were made in the Ottoman period, but in 1984 in the Islamic Museum,
with Judith Kolbas, It saw three glass weights explicitly labelled for 200 dirhams with
the names of the Mamluk sultan Qa’itbay and the Ottoman sultan Selim. Large
multiple weights were probably rare at any period and carefully controlled; in other
words, none were lost for us to find. The manufactureof such weights was technically
possible, as shown by the surviving pieces of glass weights for one and two ra_tl.
On the other hand, the large glass weights for wuqiyya or ounces and ratl or pounds
do not seem to have been usable for coins. The Egyptian ounce, which also varied in
weightiduring the eighth century, was always divided into twelve dirhams, but these
weight dirhams do not seem to correspond in any simple way to the weight standards
for dirhams. In one case only a correspondence may exist: the << 'Abbasid >> wuqiyya 5’,
if divided by twelve, yields a dirham of 2.6388 grams, perhaps not too far from the dirham
of thirteen barriiba of the early 'Abbasid period of 2.62 grams. i Unfortunately for this
hypothesis, the two weight units are not contemporary. 5
I would suggest, given the scarcity of the Umayyad and ‘Abbasid glass coin weights
and the care taken in their manufacture and inscription, that they were not used in

50. See Morton’s general discussion, Catalogue, p. 21-25. Another discovery of Morton’s, that the
vessels that were stamped with similar inscriptions to those used on the coin weights were used to
measure out one fals’ worth of a number of ordinary foodstufls and commodities, is very probably
related to the fals weights. By changing the standard weights by which fals payments were measured,
the government could easily change the official prices of commodities to adjust to market forces or for
expediency.
51. Paul Balog, << Islamic Bronze Weights from Egypt », Journal of the Economic and Social History
of the Orient, 13, 1970, p. 233-256; idem, << Pesi di bronzo islamici del XIII secolo», Quaderni ticinesi di
numismatica e anticliita classiche, 1973, p. 179-193. A
52. Morton, Catalogue, p. 29-30.
 COINS AND MONEY IN THE ARABIC PAPYRI 59

everyday market-place transactions. Rather, their use was probably limited, by law or
by practice, only to certain circumstances. They might have been standard weights
distributed as a basis for the manufacture of privately made weights, or to check such
weights. Or they might have been used only by tax officials, or muhtasib, or by the
bureaucracy in general, or only among the Arabs, or among city dwellers,’ or for the
army. The population of Egypt got along without officially made coin weights during
several long periods, and it is entirely possible that, during the era of the Umayyad and
'Abbasid coin weights, most people continued to use privately made weights in private
transactions, weights that might or might not have been regulated by the government,
or might have been regulated badly. One wonders why, in so many papyri, it was
necessary to specify that the payment was in << cash of the treasury and by its weight »
(naqd bayt al-mal wa-waznuh) 53, if there was not at least one alternative standard. Such
alternative standards may be suggested by the use of the adjective sfiqi << market » 5“,
or qa’im << usual or current » 55 or payments made << at the weight of the merchants » 5°.
Judging by its use in context, the term gawaz, which I will discuss later, can also mean
<< current money, ~what people use » as opposed to official or standard currency.
The phenomena I have described or alluded to may seem strange in today’s economy,
but they arise because in the pre-modern era--that is, before the 19th century, everywhere
in the world-—it was very difiicult for governments to guarantee anything about their
currency, and most governments seldom tried to. The picture shown by the papyri as
well as by later documents such as the Geniza texts and by the historical sources is of
a monetary economy which was largely unregulated and laissez~faire. Essentially the
stamp of the mint on a coin guaranteed nothing more than its alloy, or merely that it
was acceptable in payment of taxes. The value of the coin in relation to coins of other
metals was not fixed, giving rise to the varying exchange rates between dinars, dirhams,
and fulus which are familiar to any student of the papyri or the Geniza documents.
Fixed values between coins of different metals can be effective over the long term only
if the government or some other agency stands ready to exchange either of the coins for
the other at the set rate; but when the value of coins is closely tied to their metal content,
no government can do this without infinite reserves of the two metals. This was a
problem even in England until the early nineteenth century, and in the United States

53. Grohmann, Einfiihrung, p. 187 n. 1 cites many examples. Dated examples in my files range from
241 (P. Cair. B.E. 166 = Grohmann, APEL III, no. 182) to 312 (P. Cair. B.E. 238 recto = Grohmann,
ibid. II, no. 86). The phrase dahil bayt al-ma! wa-waznuh seems to have the same meaning (P. Cair.
B.E. 219 = Grohmann, ibid. III, no. 196 of 262 H.; PERF 835, cited Grohmann, ibid. II, p. 97, of
272 H.). The common use of these phrases in this period may reflect a special situation.
54. Louvre SN 183 == Jean David-Weill, << Papyrus arabes du Louvre II », Journal of the Economic
and Social History of the Orient, 14, 1971, p. 12-15, dated iawwtil 156: << four suqi dinars.»
55. Dated examples include P. Louvre 6963 A = Jean David-Weill, << Papyrus arabes du Louvre»,
Journal of the Economic and Social History of the Orient, 8, 1965, p. 289-291, dated gunzddci I 202 and
P. Cair. B.E. 87 = Grohmann, APEL II, no. 89 (both with dinar qa'im).
56. Grohmann, Einfiihrung, p. 186 n. 7.
60 MICHAEL L. BATES

until the middle of the century. The existence of fluctuating exchange rates in Islamic
Egypt is so familiar that I will not cite examples.
Beyond this, the stamp of the mint on a coin was not even a good guarantee of the
value of that coin in relation to other coins with the same stamp. There was no mechanism
by which faulty underweight coins could be exchanged for full weight coins without
penalty, except by returning the coin to the mint; but at the mint, the coin became
bullion, raw material, and a charge was levied to melt it, make the blank, and strike
it with a new impression. As bullion, the value of the coin at the mint was directly
dependent on its weight, so obviously an underweight coin could not be turned into a
new full weight coin (of course, no one went to the mint with a single coin in any case).
In addition, except for those who lived in cities with a mint, there were transportation
and security charges to consider, and, realistically, an ordinary person with one or a few
defective coins did not go to the mint himself, but sold the coins for what he could get
from a moneychanger, banker, or jeweller--and the profits of these and other middlemen
were also reflected in the price he could get for his coins. As a result, underweight coins
stayed in circulation. Naturally those who accepted coins in payment would protect
themselves by weighing the coins to insure full metal content, but in a situation where
a large part of the circulating money, perhaps even most of it, was below the legal
weight, it was not practical to demonetize, that is to reject, all underweight coins; it was
simpler to accept them at a reduced value basedon their actual weight. As a result,
since the monetary value of each coin was proportionate to its weight, the easiest way
to make payments was by weighing the coins in bulk, basing the value of the payment
on the total weight of the coins, not on their number. In this environment, the monetary
unit--dinar, dirham, or fals-was different from the actual coins even though a single
coin was also called dinar, dirham, or fals. In the papyri, this situationis reflected
sometimes for a single coin: <<I gave him a dinar worth 21 1/2 qirdt» (as opposed to
the normal 24 qircit) 57 but more commonly for coins in bulk. Grohmann cites several
examples of this 58, and many more could be cited from the Geniza documents 59 which,
however, are beyond the chronological limits of this paper. Weighing was so much
accepted as normal that << to weigh», wazana, became a synonym for << to pay», as in
a fourth century papyrus letter where the writer says <<1 will weigh him his dinars; do
not mix others with them » 6°. We should assume that all payments, unless there is
specific contrary evidence in the document, were made by weighing coins in bulk 9'.

57. P. Cair. B.E. 829, quoted by Grohmann, Ein/Tihrung, p. 188.

58. Ibid., p. 187-188. ,
59. S.D. Goitein, A Mediterranean Society: The Jewish Communities of the Arab World as Portrayed
in the Documents of the Cairo Geniza, I: Economic Foundations, Berkeley - Los Angeles, 1967, p. 230-231.
60. P. Cair. B.E. 321 verso = Grohmann, no. 302; the translation, however, is from Grohmann,
Einfuhrung, p. 186. Goitein, loc. cit., makes the same point.
61. Beginning in the twelfth century in Egypt (earlier in the East), gold coins had no fixed weight
standard, forcing transactions by weight. This continued until the 15th century, when the competition of
the conveniently uniform Venetian ducat forced the Mamluk mint to begin issuing dinars of fixed weight.
 COINS AND MONEY IN THE ARABIC PAPYRI 61

Clearly, however, some payments were made by count, when counting is specified
by the adjective << counted », 'adad, without any mention of weighing in the document.
This seems to have happened most often in the Umayyad period, judging by the citations
I have collected. One, dated Safar 91, is from the governor Qurra, demanding a payment
of 1042/3 dinars counted; another bilingual letter from him dated dii I-Qa'da 91
specifies in Greek arithmia nomismatia, << counted nomismas», although the Arabic
parallel says merely dananir min al-‘ayn, <<dinars of gold »62. Grohmann cites other
examples ‘*3. Of course, a demand from a governor may be a special situation, but
counting can be explained in two circumstances which might have to existtogether.
First, if the difference between the value of the coins and their bullion value is high,
variations in weight make less difference; there is less incentive to withdraw overweight
coins from circulation to melt and remint, so it becomes possible to accept coins by
count because their average weight is likely to be at the standard. This situation is
likely to have been obtained in Egypt before it had its own mint for gold, when the trouble
and expense of sending bullion or coins to Damascus or Baghdad for minting would
have made the-coins in the country too valuable to export. The second requisite for
circulation of coins by count is la high degree of uniformity of weight, which might well
have been true for the Umayyad dinars of the early eighth century which were uniformly
struck in the first place and had not been in circulation long enough to become irregular
through wear. Frequent references to weighing or acceptance of coins by weight do
not seem to appear until the end of the second century, when Egypt had its own dinar
mint which ironically would have had the effect of lowering the premium of coins over
bullion——that is, of making new coins cheaper to obtain--and therefore of increasing
culling, clipping, and the necessity to weigh payments. I believe that even in the Umayyad
period many gold payments were still weighed—-the existence of coin weights indicates
a need to weigh coins--and, probably, many Byzantine solidi were still in circulation,
with a weight substantially different from the new Muslim standard.
The probable overlap in the use of Byzantine solidi and Islamic dinars illustrates
another interesting difference between pre-modernand modern currencysystem: the
absence of a strict rule of legal tender. In the United States until the middle of the
nineteenth century we used many foreign coins, English, French, Spanish and Portuguese
among others-—sometimes to a greater extent than our own currency. In fact, the
comparatively primitive minting system of the early United States encouraged the import
and use of foreign coins. In early Islamic Egypt also, there is reason to believe that
coins other than the Islamic dinar and dirham were in circulation. Demonetization
occurred in pre-modern economies, when old or foreign coins were called in, but
attempts at demonetization were often unsuccessful because the terms of exchange for
new coins were economically unattractive to the possessors of the recalled money.

62. P. Cair. B.E. 335 = Grohmann, APEL III, no. 160; PERF 593, in Grohmann, From the World
of Arabic Papyri, p. 130-132.
63. Einfiihrung, p. 187.
62 MICHAEL L. BATES

Was the solidus demonetized in Egypt (I do not speak of Syria here) when the new
dinars were introduced? One has to doubt that it was, because to do so would have
meant withdrawing the entire gold coin. stock of Egypt, transporting it to Damascus,
and rettlrning the newly minted coins to Egypt. Who would pay the costs of such an
operation? It seems unlikely that Byzantine solidi circulated in Egypt in the 3rd/9th
century, but it is not impossible that some survived from the seventh century or were
imported. At any rate, that is one possible interpretation of a papyrus of 238-242 H.
that lists various amounts of money in dinars, 121, 107, 341, and 405, then gives the
total in mitqal, 912, and the weight of the total, which is 974. The total of the different
amounts is also 974, showing that this figure, not 912, is the monetary value of the
whole 6". This, then, is an unusual but not unique example 65 where the value by weight
of an assembly of coins is greater than their count, showing that these particular coins
were on the average heavier than the standard against which they were measured by 6.80 %.
The ratio 974/912 corresponds to a coin of 4.52 grams, nearly the weight of the solidus,
as compared to the dinar of 4.23. The alternative interpretation, that the coins were
dinars of normal average weight (4.23 grams), implies that they“ were being evaluated
against a dinar standard of 3.96 grams, remarkably low and not attested by any other
evidence.
Legal tender meant, at most, that only specified coins would be accepted by the
government in payment of taxes. This monopoly, over time, wouldprovide an incentive
to remint the old coins, but one has to assume that the reminting process was slow.
It is possible, moreover, that new Byzantine coins were imported into Egypt even after
the introduction of the Islamic dinar. Another instance of failure todemonetize old
and defective coins was referred to previously. The Egyptian dinars of the early
fourth century, especially those issued with the name of al-Radi (322-329), were severely
debased. Mul_1ammad.al-Ibsid, when he ordered the-striking of dinars with his own
name, brought the fineness of the coinage back to its previous high standard. Nevertheless,
the Radi dinars, as they were called, continued in circulation until 363, when, according
to Ibn Muyassar, great economic harm was done to the populace by the refusal of the
new Fatimid government to accept them 66.
This latter episode leads to consideration of a point raised by the papyri that is difficult
to resolve but useful as an illustration of the benefits of combining thenumismatic and
papyrological evidence. The question is the meaning of the word magsiil which appears
frequently as an adjective for dinars in the fourth century 6’. The word may be either

64. PERF 761, p. 201, quoted Adolf Grohmann, From the World of Arabic Papyri, p. 136-137.
65. Another is << 10 dinars by count, weight 12 » (10 dinar“ 'adad al-wazn 12), PER Inv. Chart. Ar.
7303 quoted Grohmann, Einflihrung, p. 187.
66. Ibn Muyassar, p. 45; Maqrizi, Itti'a;, p. 172; Maqrizi, Hiya], II, p. 6; Maqrizi, Sudur, p. 35-36,
112-113.
67. I have collected more than fourteen uses of the term in the papyri, dating from 312 (P. Cair.
B.E. 238 recto = Grohmann, APEL II, no. 86: bi-'i§rin dinar" 'uyun"" mataqil magstila bi-naqd bayt
al-ma‘! wa-waznih) to 356 (P. Cair. B.E. Ta’ril) 2166d = Grohmann, ibid. II, no. 97: talatat dandnir wa
 p COINS AND MONEY IN THE ARABIC PAPYRI 63

ma'siiI, which means << honeyed, sweet» or magsfil meaning << washed». The latter seems
more likely, though it is not easy to see what either meaning has to do with gold coins.
In my old notes I found a suggestion of my ownthat perhaps the coins were literally
washed, because, if the coins were weighed in payment, clean coins would weigh less
than dirty ones, but an examination of all the citations shows this hypothesis to be
ridiculous. I have not found any indication that the word magsiil, though it has the
obvious metaphorical meaning of << pure » in spirit or deeds, has ever been used to
indicate << pure » in the sense of unalloyed composition. As for ma'sz7l, << honeyed,
sweet», there seems no sensible way to apply this to coins.
The term appears several times in opposition to the term gawdz previously mentioned.
Some texts make it clear that these are two different kinds of dinar, the gawaz dinar being
the worse. For example, P. Giss. 108: << You set the workers’ payment at a gawaz dinar.
By God, if they hear that they will flee and not one of them will remain,so let them
have the magsul dinar as you gave them the first year for carrying the grain in Aswah,
a dinar magszil » 68. More specific is P. Cair. B.E. Ta’rih 1741/5: <<I had paidyou with
a purse weighing 13 1/2, containing 4 1/2 gawiiz, so the total magsiil muhtaqq is 12 >> 69.
One can easily calculate the value of the gawiiz dinar: if the purse contained 9 magsfil
dinars and 41/2 gawiiz dinars, and was worth 12 magszil dinarsin all, then the gawaz
dinars were worth 2/3 of a magsiil dinar.

This rings a bell, because Ibn Muyassar’s account of the events of 363 says that the
extortionate taxation of the Fatimid officials and their refusal to accept any but Fatimid
dinars led to a drop in the value of the Riidi dinar to 2/3. of a dinar. Thedecline, it is
true, was artificially produced and the implication isthat the normal value of the Radi
dinar was higher, but the ratio may indicate a floorvalue related to the bullion content
of the coin. Some of the dinars of the period 313-329 analyzed by Ehrenkreutz in fact
have as little as 67 ‘X, gold; othersihave more, but it is very likely that the value of the
Radi dinars dropped even below their average -bullion value under the pressure of the
need to change them into Fatimid dinars, which the banking system and the mint could
not do overnight. Certainly it cannot be only a coincidence that the appearance of the
paired terms magsiil, obviously meaning full value pure gold coins, and gawaz, an inferior
dinar worth as little as 2/3 of the full dinars, comes in precisely the period from the first
issuance of debased dinars in Egypt until the time when all these were forced out of
circulation by the new Fatimid monetary and taxation policies. Magsfll, then, whatever

nisf magsula yaqbadu fi kull Jahr sob‘ qarcirit magsul) and finally to the reign of the Fatimid al-Mu'izz
(359-365); PER Inv. Chart. Ar. 25611, quoted by Grohmann, Einfiihrung, p. 201: sab'at 'as'rat artal
magsula; but this latter text does not mention coins as such). Grohmann, Einfuhrung, p. 201-202, cites
many more references. \.
68. Adolf Grohmann, Die Arabischen Papyri aus der Giessener Universitdtbibliothek, Abhandlungen
der Giessener Hochschulgesellschaft, IV, Giessen, 1960, p. 58-59, no. 16, undated.
69. Grohmann, APEL, V, p. 101, no. 316; quoted Grohmann, Einfuhrung, p. 201 n. 5, with some
variants.
64 p MICHAEL L. BATES

the origin of theterm as a monetary term and its precise meaning, must refer to dinars
of full alloy, including those issued before the debasement began and those of the
Ilfsidids, in opposition to gawiiz, the debased dinars of 304-330, which apparently
remained in circulation in large quantity for more than three decades in competition
with the official issues of the Egyptian mint 7°. 4

1 70. I wish to thank my friends Professor Jere Bacharach and Dr. Lutz Ilisch who read this article
in draft and took the time to make numerous comments and suggestions that saved me from errors
and helped toclarify or improve many arguments. Naturally they bear no responsibility for the
remaining faults.