VERHANDELINGEN

VAN HET KONINKLIJK INSTITUUT

VOOR TAAL -, LAND - EN VOLKENKUNDE

218

SMALLHOLDERS AND STOCKBREEDERS

History of foodcrop and livestock

farming in Southeast Asia

Edited by

PETER BOOMGAARD and DAVID HENLEY

KITLV Press
Leiden
2004
 JAN WISSEMAN CHRISTIE

The agricultural economies of

early Java and Bali

Introduction: the prehistory of rice in Java and Bali

The antiquity of rice cultivation in Southeast Asia has been the subject of debate
for several decades (see also Hill in this volume), the breadth of opinion on the
subject stemming largely from the dearth, until recent years, of adequate data.
However, archaeobotanical data now suggest that cultivated rice (Oryza sativa,
Linn.) - which appears to have been domesticated about 8,000 years ago in the
region of southern China that borders on northern Southeast Asia and north-
east India - had spread to mainland Southeast Asia before the mid third mil-
lennium BC (White 1995:63 note 2; Higham 1996:309), as populations adapted
to Holocene environmental changes (Kealhofer 2002:183-84). Archaeological
evidence further points to the introduction of rice to island Southeast Asia dur-
ing the early to mid second millennium BC at the latest (Glover and Higham
1996; Higham and Lu 1998). Dates for domesticated rice in Sarawak have been
pushed back to at least 2000 BC, and possibly much earlier (Doherty, Beavitt
and Kurui 2000:150), and in Sulawesi to about 2000 BC (Glover 1979).
Dates for cultivated rice in we~tern Indonesia are more difficult to ob-
tain, due to geological conditions and to local burial and farming practices.
However, domesticated rice was clearly present on these islands by the first
century AD (Bellwood 1997:117), and there seems no reason to doubt that
it was planted in this region as early as it was in Borneo, that is, by about
2000 Be. Rice appears to have been introduced to Java and Bali, as it was
elsewhere in the Indonesian and Philippine archipelagoes, by Austronesian
speakers. The key terms associated with rice and rice farming in Old Javanese
and Old Balinese - pari/padi (rice plant), wras/bras (husked rice), sawah/huma
(irrigated rice field) - all derive from Proto-Austronesian (Wurm and Wilson
1975:171). Rice, along with the rearing of pigs, dogs and chickens, appears to
have formed the basis for farming in Java and Bali from the late Neolithic pe-
riod onwards. Such alternative staples as foxtail millet (Setaria italica, Beauv.),
Job's tears (Coix lachryma-jobi, Linn.), taro (Colocasia antiquorum,Schott), yams
(Dioscorea alata, D. pentaphylla, Linn.) and sago (Metroxylon sagus, Rottb.) (see
48 Jan Wisseman Christie

Roy Ellen in this volume), which played more important roles in other parts
of the archipelago (but see David Henley in this volume), appear to have
played a secondary role on these two islands. The dominant alternative sta-
ples now grown on Java or Bali - either as dry-season crops or as main crops
in marginal fields brought into cultivation under population pressure - are
maize (Zea mays, Linn.) and cassava (Manihot utilissima, Pohl.), both New
World imports. This suggests that rice had begun to dominate the landscapes
of the two islands at such an early date that no other indigenous grains, tu-
bers or tree starches were developed as effective maincrop alternatives.

The economies of early historic Java and Bali

Some of the most productive agricultural land in the world is found on the
islands of Java and Bali. The volcanically enriched soils and moderate mon-
soonal climate of parts of central and east Java and southern Bali provided the
ecological context for the early development of population densities on a scale
much closer to those of the heartlands of mainland Southeast Asian states than
to those of other islands in the archipelago. Javanese and Balinese inscriptions
of the ninth to fifteenth centuries contain the names of hundreds of communi-
ties, some of which, judging from the length of the lists of inhabitants and com-
munity officials, were very large. In some areas these communities were also
closely packed: one early tenth-century Javanese benefice charter (Stutterheim
1940) lists witnesses drawn from over sixty neighbouring communities.
Palynological evidence from the region around Borobudur in central Java
(Nossin and Voute 1986:858) indicates that by the ninth century the landscape
around the monument supported the mixture of grain fields, house gardens,
palm groves and forest that the contents of the inscriptions lead one to expect.
Suggestions, based upon an interpretation of the plants illustrated in selected
relief panels on Borobudur, that foxtail millet rather than rice was the impor-
tant grain in Java at the time (Bernet Kempers 1976:242-3; Steinmann 1934b:583-
4), are not substantiated by the texts or by the majority of illustrations of grain
(Christie 1992:11). The use of rice chaff as temper in bricks made during the
eighth and ninth centuries in central Java (Teguh Asmar and Bronson 1973:35)
tends to confirm the textual evidence that the grain planted was rice.
Compared to their island neighbours, however, Java and Bali are poorly
endowed with toolmaking stone and metal ores. The wealth of these two
islands, in the pre-modern era, was thus based upon agriculture, upon the
labour of the large populations that agriculture supported, and upon an ac-
tive and extensive trading network. Since toolmaking stone and metal ores
were in short supply on Java and Bali, the inhabitants of the two islands were
forced, as early as the Neolithic period, to look outwards to the other islands
 The agricultural economies of early Java and Bali 49

in the archipelago in order to meet some of their most basic technological

needs. By about 500 BC, Javanese and Balinese ports were trading regular-
ly with mainland Southeast Asia, south China, and the east coast of India
- exporting both produce of the two islands and spices and sandalwood of
eastern Indonesia. By the first century AD, these ports and their archipelago
trade network were drawn into the Old World sea trade system that linked
Han China to the Roman Mediterranean (see Glover 1990). Cloves from the
Moluccas reached Rome, presumably via Java, in the first century AD (Miller
1969: 47-51), and Romano-Indian rouletted ware pottery, made during the
first three centuries AD in eastern India, has been found on the coasts of Java
(Walker and Santoso 1977) and Bali (I Wayan Ardika and Bellwood 1991).
By the ninth century AD, when the first surviving inscriptions were pro-
duced in local languages, Java and Bali had been trading overseas for well
over a thousand years. Not only were the islands' cosmopolitan connections
reflected in borrowed aspects of the cultures, religions and languages of their
ports and courts, but these outward-looking, mixed economies stretched in-
land from the ports to the fertile rice-growing areas of the interior. Outward
trade connections had a major impact, over time, on the range of crops and
animals available to Javanese and Balinese farmers. They also encouraged the
commodification of crops and labour, with the result that, by the ninth centu-
ry AD, both had monetary value. The surviving early Javanese and Balinese
records are overwhelmingly financial in nature, and references to productive
land, crops, livestock, and foodstuffs appear - in a number of different con-
texts - in many of these inscriptions.

Community, land and farming in early Java and Bali

Javanese and Balinese communities of the late first and early second millen-
nia AD were neither primitive nor isolated. Many were too large and complex
in structure - often comprising more than one settlement cluster - to be called
mere 'villages' . There is no evidence that even the smallest of these communi-
ties farmed their fields communally, although members of the communities
apparently had access, either continually or in rotation, to certain types of
community common land. Early Javanese and Balinese communities were
internally stratified, and the larger communities contained a mixture of core,
landowning families who were represented on the village council; profes-
sional traders and artisans; residents of relatively high status in the state or
religious hierarchies who were not represented on the village council, but
who were allowed to purchase land in the village; and others of lower status
who were - either voluntarily or through some form of bondage - dependent
upon farming households in the village (Christie 1994:32-4).
50 Jan Wisseman Christie

Early Javanese and Balinese communities had a great deal of control

over their immediate environment. The oldest terms for these communities
- wanua in Old Javanese, and banua in Old Balinese - predate states on the is-
lands (Christie 1994:27-28), and so, apparently, did their rights over land. The
territories belonging to communities were described in such terms as 'their
forests, dry fields and rivers, in the valleys and on the hills' or 'their valleys
and mountains, their wet rice fields [... J their orchards' (Christie 1994:29). In
Java, the community territories appear to have included far more land than
that under cultivation at anyone time, providing scope for settlement fis-
sion and the establishment of pioneering daughter-settlements (babadan). In
Bali, regulations found in such tenth-century inscriptions as Srokodan (Goris
1954:63, II.3-4), relating to the clearing of community land by outsiders for
swidden farming, suggest a similar scope of land tenure.
Community rights to hind, recorded in ninth- and tenth-century inscrip-
tions of both islands, were probably based upon older swidden-farming ter-
ritories. Hanks (1972:166) and others have noted the modern density of be-
tween 50 and 90 persons per square mile that can be supported by swidden
agriculture in the archipelago, although the more fertile parts of Java and Bali
could have supported conSiderably more. It seems likely that, in early Java,
community territories expanded with population growth until they met the
expanding boundaries of neighbouring communities. Subsequent popula-
tion growth might theoretically have forced either the setting up of pioneer-
ing daughter-settlements at a distance, or changes in agricultural practices.
The fact that many ninth- and tenth-century central Javanese communities,
even at the heart of the state, retained forest reserves suggests, however, that
changes in agricultural practices occurred well before available resources had
been exhausted. The driving factors in the intensification of land-use were
probably not so much the negative ones of overcrowding, as suggested by
Boserup (1965), but the positive ones of increased labour efficiency (Hunt
2000:274), enhanced fertility through harnessing the mineral-rich run-off
from volcanoes (Christie 1992:11), and changes in the structure of intracom-
munalland tenure favouring individuals and families (Christie 1994:29-30).
State encouragement undoubtedly also played a role (Christie 1992:12).
By the ninth century AD, when Old-Javanese and Old-Balinese language
inscriptions were first written on permanent materials, irrigated rice had be-
come the economically dominant crop on the two islands. Irrigated rice fields
(sawah in Old Javanese / huma in Old Balinese) were the focus of most of the
early charters of both islands. Other types of productive land was, however,
mentioned, including dry or hill rice fields (gaga, gagan); dry or swidden fields
(tegal/parlak); fallowed swidden land (lmah suket); orchards or perennial gar-
dens (kbuan/ngmal, mmal); house Iand with gardens (pomahan); sirih gardens
(paserehan); taro fields (patalesan); cotton fields (pakapasan); meadow or grass
 The agricultural economies of early Java and Bali 51

land (dukut/padang); treeless, uncultivated land (harahara); forest (alas); marsh

(reni!k); river banks/margins (tpitpi); rivers (luah, kali, bangawan); lakes, ponds
(ranu/danu); salterns (lmah asinan or pawuyahan). In Old Javanese language
literature the forest-edge garden, or managed forest-garden with bamboos,
trees and shade-tolerant perennials and annuals (talun), was also mentioned.
In fact, the range of natural and man-made 'ecozones' recognized as produc-
tive in early Java and Bali appears to have included all but the large-scale
plantations introduced under colonial rule.
Different types of productive land within the boundaries of early Javanese
and Balinese community ter,ritories fell into different classes of tenure.
Irrigated rice fields, orchards and houseland were owned and inherited by
individuals and families, In Java, other types of land in the community's ter-
ritory were apparently owned by the community as a corporation; in Bali, in-
heritance regulations suggest that in some cases even grassland and dry rice
fields may have been owned by individuals. Communally held land came
under the direct or indirect authority of the community council, and was sub-
ject to regulation by community officials, This land included salary fields at-
tached to community offices; swidden land, presumably used by households
in rotation; and forest, fallowed swidden land, grassland, marshes and bodies
of water that were apparently open to communal grazing, fishing, hunting
and collecting, under the oversight of specified community officials.
Community officials in Java also held authority over their community's
irrigation arrangements. The situation in Bali was complicated by the difficul-
ties of dealing with water in a more dissected landscape, a situation which led
to the development of subak, or irrigation societies, that cross-cut community
boundaries (Christie 1992:16), In addition, Balinese inscriptions highlight state
regulations relating to the protection of valuable and productive trees, Rulers
of the small states of Bali appear to have had more direct interest in internal
community affairs than did their contemporaries in the larger states of Java.
Some insights into the manner in which the land was farmed are pro-
vided by illustrations in the relief panels on temples. Most of these illustra-
tions focus on the farming of rice - with the possible exception of one or two
scenes of forest-edge fields depicted on the ninth-century central Javanese
temple of Borobudur, in which the grain illustrated may represent foxtail mil-
let (Steinmann 1934b:plate Ib). Other reliefs on Borobudur depict plowing of
paddy fields with humped zebu cattle, and the harvesting of rice using the
ani-ani harvesting knife (Bernet Kempers 1976:plates 156, 158). A later east
Javanese relief from Trawulan, which dates to the fourteenth century, illus-
trates the transplantation of rice seedlings (Bernet Kempers 1976:plate 160).
This last provides support for the observation of the early fifteenth-century
Chinese visitor Ma Huan that rice was being double-cropped in the Brantas
delta region of east Java (Ma Huan 1970:91-2),
52 Jan Wisseman Christie

The scene, in a panel on Borobudur, of rats consuming rice in a field while

two men and their dog sit inattentatively in a temporary field shelter (Bernet
Kempers 1976:plate 162), illustrates one of the perennial hazards of farming.
Both Javanese and Balinese inscriptions include neglect of fields amongst
fineable offences - and such Balinese inscriptions as Dawan, dating to 1053 (I
Wayan Ardika and Ni Luh Sutjiati Berartha 1998:136, Il.a.6-II.b.1) - provide
considerable detail concerning the fines levied on neglect resulting in damage
by rats, floods and collapse of rice-field bunds, fire, and so forth. Balinese in-
scriptions also provide many details concerning the mechanics and the regu-
lation of irrigation.

Foodcrops and diet

Zhao Ru-gua's early thirteenth-century description of Java, a composit of the

reports of merchants and travelers of that time and of earlier periods, paints
a compelling portrait of the agricultural landscape of east Java:
It is a broad and level country; well suited to agriculture. It produces rice, hemp, mil-
let, beans, but no wheat. Ploughing is done with buffaloes. [... J They make salt by
boiling sea water. The country also abounds in fish, turtles, fowl, ducks, goats, and
they kill horses and buffaloes for food. The fruits comprise big gourds, cocoanuts,
bananas, sugar-cane and taro [ ...J. [...JTheir wine is derived from the cocoanut and
from the inner part of the hia-nau-tan tree [... J or else it is made by fermenting (the
fruits) of the kuang-Iang ([ ...J sago palm) and of the areca palm [...J. There is a vast
store of pepper in this foreign country [ .. .J. (Hirth and Rockhill 1966:77-8.)

This and other foreign accounts of Java complement descriptions, in local

written sources of the ninth century onwards, of the crops raised by the
Javanese and Balinese, not only for their own consumption, but also for sale.
Such information is provided both in descriptions of food consumed at feasts
and in lists of comestibles traded in the markets.
Descriptions and illustrations of food offerings made to temples and of
feasts served on ceremonial occasions provide some inSight into the food-
stuffs valued by the Javanese and Balinese of the late first and early second
millennia. The ninth-century feast scene illustrated in a panel on the central
temple of Loro Jonggrang at Prambanan in central Java, for instance, appears
to have included such foods as cakes, fish, crabs, sate and possibly sausage, as
well as sprouted onions (Bernet Kempers 1976:plate 177). In Java, contents of
charters issued early in the tenth century provide the most detail concerning
ceremonial feasts. The Taji charter of 901 (Sarkar 1971-72, II:4-14), for example,
includes in its list of feast foods, six buffaloes and a hundred chickens, along
with large quantities of dried salted meat and fish, rice and palm wine. King
 The agricultural economies of early Java and Bali 53

Balitung's charter of 902 (Van Naerssen, Pigeaud and Voorhoeve 1977:58-61)

contains one of the more detailed lists of foods consumed at a ceremony:

(Iv.a.5) The food that was enjoyed [included] tahu and wagalan-fish, haryyas, boiled
vegetables (kuluban), side dishes, minced meat, and so forth. There was nothing lac-
king. Likewise for the (IVb.l) grilled foods, the dried sea perch (tjang kakap), pom-
fret (karjiwas), Spanish mackerel (tangiri), hnus, prawns (hurang), bilunglung, all sorts
[of food]. All was complete. (2) [As for] drink, [there were] rum (sfddhu) and other
spirits (mastawa), kifica, fermented sugar cane (kilang), and palm wine (twak) [ ... ].

If the reading of the term tahu in the above passage is correct, then it appears
that by the tenth century the Javanese were both growing soy beans (Glycine
;lispida, Maxim.) and making bean curd. The Javanese term for soy beans
(ka4le) does not occur in published inscriptions, but does appear in the kidung
literature of the fifteenth century and later. Soy beans have a long history of
being grown as a dry-field succession crop to rice. Burkill suggests that the
bean arrived in Java at an early date from India, since the type grown in Java
resembles the Indian rather than the northeast Asian variety, and the name
appears to be derived from Tamil (Burkill 1966:1098-9).
Balinese inscriptions provide fewer lengthy descriptions of individual
feasts, but do refer to similar festive preferences for steamed rice, chicken,
alted fish, palm wine and rum. They also refer to offerings not only of ordi-
nary rice, but - much less frequently - of glutinous rice (ketan), rice-cakes (jaja
lukan), sugar syrup (juruh), mung beans (atak), sesame oil (lnga wijyan), millet
or rice (jawa) and Job's tears (jahli) (see I Wayan Ardika and Ni Luh Sutjiati
Berartha 1998:414, lILa), pork, duck and taro (kaladi). The feasts described in
early Javanese and Balinese inscriptions resemble, to a degree, early twenti-
eth-century selamatan-feasts, setting aside changes in diet precipitated by the
acceptance in Java of Islam and by the import of New World crops.
Balinese inscriptions contain particularly detailed lists of protected trees
and perennials (tumuwuh salinarangan) - that is, trees that had either religious
or economic value. Of trees protected because of their economic rather than
religious value, about half were important sources of food, spices (basa) or
timulants. These included coconut (nyu, tirisan), betel (sirih), areca (pucang),
ptung and hampyal bamboos, aren (hano), durian, mendi, jirek, cubeb pepper
(kumukus), tamarind (kamalagi, lunak camalagi), cardamom (kapulaga), and
candlenut (kamiri).
Judging from the recovered contents of a ship that was wrecked in pas-
age from south Sumatra to Java, by the middle of the tenth century demand
for candlenuts, in particular, must have outstripped supplies in Java. This
Indonesian-built ship, which went down in the Java Sea south of Bangka, in
a shipping lane now used between Palembang and the north Javanese ports,
54 Jan Wisseman Christie

has been dated to the mid tenth century on the basis of its contents and radio-
carbon dating of surviving timbers. The very mixed cargo appears to have
originated from the entrepot of Srivijaya, which was probably at that time still
located in Palembang. A large portion of the cargo comprised small Chinese
stoneware pots. Rather than waste space by shipping them empty, the own-
er of the cargo filled many of them with candlenuts, presumably sourced
in southern Sumatra (Flecker 2001a:223-5). The fact that candlenut remains
were also found in the Indonesian ship that went down in the same area, on
the same route, in the thirteenth century (Mathers and Flecker 1997:93-4),
suggests that the Javanese demand for these nuts remained above local pro-
duction levels over long periods of time. The export opportunity created by
Javanese demand may explain the inclusion of candlenuts amongst the most
prominent of the crops over which successive rulers exercised' compulsory-
purchase' tax rights in Bali (see below).
Foodcrops (tuwutuwuhan, tanemtaneman) mentioned in Javanese inscrip-
tions included rice (pari, wras), mung beans (atak, kacang), fruits (wwah), sugar
cane (tevu, gula), cucumber (?tmu), onions (bawang), sesame (lnga), and such
spices, flavourings and medicinals as ginger (pipikan, jahe), turmeric (kunir,
kufiit, jne), and galingale (laja). Balinese inscriptions provide more detail con-
cerning these crops, and add a number of others, under the rubric of 'proper'
crops (yogya tanem). General classes of crops mentioned in Balinese charters
included swidden crops (wgilwgil), fruits (phala), vegetables (gangan), and
rootcrops (mulaphala). Grains listed included not only paddy rice (pari, bras)
and hill rice (gagan, panggaga bras), but as noted above, more rarely, two types
of sticky or glutinous rice - red (laktan bang) and black (ifijin, laktan hireng),
and less frequently still, possibly millet, and Job's tears. Mung beans (hartak,
atak) are mentioned frequently; bananas (pisang) and taro (tales, kaladi) less so.
Both red onions (bawang bang) and garlic (bawang putih) are mentioned, along
with such flavourings as coriander (katumbar) and fennel (adas).
One indication of the importance of the early overseas trade connections
of Java and Bali is the number of crops listed in inscriptions that originated
elsewhere. These transplanted crops - aside from the grains, which were
probably brought in during the first millennium BC or earlier - probably ar-
rived in Java and Bali sometime in the last millennium BC or the first millen-
nium AD, at much the same time as cotton and Indian indigo (Indigojera tinc-
toria, Linn.). Amongst these transplants were coriander (Coriandrum sativum,
Linn.), fennel (Foeniculum vulgara, Mill.), tamarind (Tamarindus indica, Linn.),
and sesame (Sesamum orientale, Linn.), all of which appear to have originated
in southwest Asia or the Mediterranean region and must have reached the
islands via India. Mung beans (Phaseolus aureus, Roxb.) may have originated
in India, although they are now grown across the drier parts of the Southeast
Asian tropics. Garlic (Allium sativum, Linn.) may have come in from China.
 The agricultural economies of early Java and Bali 55

Tax, trade and value

As noted above, although a range of types of productive land appears in

various inscriptions, the agricultural base of early Javanese and Balinese
states was clearly rice cultivation. Irrigated rice fields produced the largest
and most predictable yields, a!1d were the backbone of the tax-paying agri-
cultural economy. Surpluses produced by wet-rice farming were, by the late
eighth and ninth centuries, substantial enough on both islands to underpin
relatively large non-farming segments of the populations of several organ-
ized states. In central Java, these surpluses were also large enough to under-
write a massive campaign of stone-temple construction. Wet-rice agriculture
was encouraged by early rulers, presumably because it tended to anchor tax-
paying farmers in place, making their taxable harvests more visible. It also
. roduced a more densely populated and secure landscape (Christie 1992:12).
Javanese rulers, in particular, used tax incentives to encourage farmers in key
areas to convert from dry- to wet-rice cultivation: brush land, or fallowed
5widden land, was mentioned in ninth- and tenth-century Javanese inscrip-
:ions - see, for instance Jurungan (Machi Suhadi and Soekarto 1986:75, Lb.2),
_1ulak (Boechari 1986:28, La.1-2), and Taragal (Machi Suhadi and Soekarto
~986:99, La.4-5) - in the context of land purchases and tax grants made for the
? urpose of creating wet rice fields.
It appears, from the surviving records, that by the ninth century most
members of Javanese and Balinese states - even those living in farming com-
munities located some distance from the ports or from the palace - operated
L'1 a milieu in which both commodities and services were understood to be
3.\ ailable in exchange for payment in a standard, state-sanctioned currency
system, based upon coinage struck locally in gold and in silver alloy - a sys-
:em to which, by the late tenth or early eleventh century, imported Chinese
-opper cash (later copied locally) was added in order to meet the burgeoning
demand for small change (Christie 1996). -
Although various acts of reciprocity and prestation played key roles in
;:eremonial contexts, members of early Javanese and Balinese communities
calculated the value of their crops, their labour, their land and possessions
:n terms of the state's currency. Land was a saleable commodity, and it was
3.pparently possible for outsiders to purchase community land of all descrip-
:ions, either from the community council, or - in the case of irrigated rice
- elds, orchards and houseland - from individuals and families. In Java, by
e tenth century, land could be bought and sold on a freehold basis (dwal
:!JI1s), pawned (dwal saJ:lIja) against future redemption (tbus), and leased (dwal
aruk) (Christie 1994:31-2).
56 Jan Wisseman Christie

Foodstuffs and livestock in early Javanese and Balinese periodic markets

The Javanese periodic market (pkan) of the late first and early second millen-
nia AD circulated amongst groups of neighbouring villages on a schedule set
by the indigenous five-day market week that is still used in rural Java. Lists
of Javanese community officials regularly included the mapkan, or market of-
ficial. In Bali, the periodic market (pasar) circulated on a three-day schedule,
but appears in other respects to have been similar to its Javanese counterpart,
although the market officials (ser pasar) in the smaller states of Bali appear
to have been connected with the state administration rather than that of the
community.
Foodcrops and livestock regularly appear as commodities in Javanese and
Balinese inscriptions. In Java, shifts in the focus of the secondary contents of
benefice charters early in the tenth century brought taxable market activities to
the fore. Javanese inscriptions, particularly those issued between the early tenth
and the mid eleventh centuries, often contain lengthy lists of standard classes
of traders and trade goods to be found in local periodic markets, many of them
connected with the sale of comestibles. Balinese inscriptions of the same period
also mention goods traded in markets, but in a more cursory manner.
An early illustration of a rural Javanese market is found on the hidden
base of the central Javanese monument of Borobudur, in a panel that was
probably finished in the middle of the ninth century (Bernet Kempers 1976:
plate 169). The vendors illustrated are - aside from a man carrying salt fish
suspended from a shoulder-pole - predominantly women, who appear to
be selling fruit, cooked food, and other items in smallish quantities. By the
tenth century, taxable activities had been divided into two categories: one
encompassing the semi-professional, interrnittant, and very localized trade of
farming households within villages where periodic markets were held; and
the other including a variety of professional traders who circulated with the
market, and who made their living primarily through commerce.
Semi-professional activities in early Javanese markets included a fairly
standardized list of manufacturing, hunting and collecting activities carried
out by members of most farming households in most villages. The goods
produced and sold by these local, part-time traders normally relied upon raw
materials available locally, and most must have been consumed locally. This
list of goods changed very little over the centuries, and the manufactured
goods amongst them seem to have been relatively resistant to changes in
fashion. Aside from the snaring of wild animals, birds and fresh-water fish
and eels for the market, the main taxable part-time activity connected with
food-production was that of sugar making (manggula), either from sugar cane
or from aren palm.
Sale of farmers' produce - including rice, vegetables, fruit, betel and con-
 The agricultural economies of early Java and Bali 57

diments - is alluded to in other contexts in Old-Javanese language inscrip-

tions. That at least some of the crops sold at market were normally taxed is
suggested by the list of privileges granted in the Hantang charter of 1135
(Brandes 1913:154-8), which speaks of farmers being allowed to carry certain
quantities and types of produce to market free of tax:

(B.lO) [They] are also to be allowed to carry to market [free of tax] those crops
(tuwutuwuhan) that are [normally] tied in bundles [and] (11) carried in shoulder-
pole loads (pikul) [or] handled in wantayan-units, but not those [normally] carried
in kadut-sacks [or] karafijang-baskets [ ... ].

The Panumbangan inscription (Brandes 1913:159-63) of five years later con-

tains similar mentions of tax relief on crops sold by farmers at market. No
ceilings were placed upon the semi-professional commercial activities that
could be carried out by farming households within communities whose tax-
es had been transferred by a ruler for the support of religious foundations.
However, by the tenth century, this sector of commercial activity had become
so economically important that the taxes were no longer transferred in toto.
Traders and artisans in the 'commercial' (masamwyawahara) category were
full-time professionals who moved from market to market in overlapping
circuits during the market week. These professionals derived all or most of
their income from their trade, and the taxes they paid were clearly of some
importance to the palace. This class of market trader included producer-ven-
dors (pakarmma), such as toddy-tappers (amahat) and butchers (ajagal), as well
as providers of transport (who hired out pack animals, horse-drawn carts and
ox-drawn wagons), and those who sold the produce of others. The levels of tax-
free trade carried on by these producer-vendors was, at times, relatively high.
Livestock were soM in the market by professional livestock farmers . These
livestock were apparently expected to be traded across community bounda-
ries, and in some cases even further afield. In Bali the Sukawana All inscrip-
tion of 1054 (I Wayan Ardika and Ni Luh Sutjiati Berartha 1998:175, I1l.b.3.2)
specifically related the livestock tax exemption to those who bought livestock
in other communities. The Dawan inscription of the year before (I Wayan
Ardika and Ni Luh Sutjiati Berartha 1998:136, I1.a.4) mentions 30 water buf-
falo that had been bought by the Balinese ruler from the neighbouring island
of Lombok (Curun). The tax-free ceiling placed on the number of professional
livestock vendors was normally three or four dealers in each type of animal
per benefice community. The tax-free ceiling on the number of animals han-
dled by each herdsman appearing in these markets was also standardized:
the mid eleventh-century Kambang Putih charter (Brandes 1913:252-4), for
example, contains the following list:
58 Jan Wisseman Christie

(7) herders of water buffalo (kbo), [2]0 head per person; herders of cattle (sapi), 40
head per person; herders of sheep / goats (wt;lus), 80 head per person; those who
keep (8) pigs (cu[ung), 100 pigs per person; those who raise ducks (a1Jt;lah), one
wantayan per person [... ].

Early Balinese tax lists include even longer lists of domestic and wild animals,
as in the Sukawana AI (Goris 1954:53-4) inscription of 882:
(ILa.I) [... ] buffalo (karambo), cattle (sampi), besara (banteng?), (2) sheep / goats (kam-
bing), scrub-land pigs (culung sukift), wild pigs (buru babit elephants (?paficayan),
kijang-deer (daker), quail (puruh) [and] dogs (asu) [ .. .].

The Bwahan (Goris 1954:83-6) inscription of 994 adds domestic pigs, chickens
(hayam), turtle doves (manuk kitiran), doves (putir), wild pigeons (wuruwuru)
and forest chickens (hayam alas) to this list.
Illustrations on Borobudur (Steinmann 1934a:plate 4b) indicate that Indian
zebu cattle (Bas indicus) were known in the islands, and it is possible that the
besar listed in Balinese insciptions referred to the native banteng cattle (Bas
sundaicus), which are mentioned in later Old-Javanese language literature.
The Borobudur illustrations also suggest that the w4us was wooly enough to
be classed as a sheep.
Pigs, which appear by the late ninth century in Balinese inscriptions, and
which were undoubtedly kept by most early Javanese households as well,
do not appear in surviving Javanese lists of professional livestock farmers
until the eleventh century (see Brandes 1913:253, line 8). Chickens, which are
mentioned in ceremonial contexts in Javanese inscriptions - as offerings, sac-
rifices and feast foods - were not listed amongst the animals raised profes-
sionally, although they appear in some Balinese tax lists. Horses (kuda/ajaran)
are mentioned more rarely: horse breeding (tangkalik kuda) is mentioned in a
few Balinese inscriptions of the tenth and eleventh century, most prominently
in that of Batur Pura Abang A (Goris 1954:88-94).
It seems clear from the above that by the ninth century AD the Javanese
and Balinese had added enormously to the core Austronesian triad of indig-
enous livestock: pigs, chickens and dogs. Sheep or goats, which originated
in western Asia, had apparently been carried as far east as Timor before the
first millennium BC (Glover 1986:196). The indigeno'us water buffalo and the
red/black banteng cattle may have been domesticated at about the same time.
The fifth-century AD inscriptions of west Java and of east Borneo mention
gifts of large numbers of cattle, one of the Bornean inscriptions describing
them as 'tawny' (Chhabra 1965:90). Indian humped zebu cattle and horses
were certainly present by the time that the reliefs of Borobudur were carved
in the ninth century AD. The brown runner-type duck found in Java and Bali
may also have been brought in from India at an early date.
 The agricultural economies of early Java and Bali 59

Also classified as professionals were the highly mobile, small-scale traders

(adagang), transporters (mangunjal) and shoulder-pole carriers (apikul), who
sold in the market goods that had been produced by others. Numbers of the
pedlars based in a benefice community who were allowed to trade free of tax
were normally limited to three or four per type of commodity carried, and
each was limited in the tax-exempt volume of trade to five bantal in weight.
Some of the goods carried by these vendors were clearly not local, or even
regional in origin. Comestibles carried included not only regionally-available
sesame cake (pacjat lnga), sugar, rice, mung beans, fruit, sesame seeds, tama-
rind, black pepper (mirica) and betel, but also salt (wuyah/garam) and salt fish
(pj a) from the coasts. Balinese lists of taxable foodstuffs are similar, focusing
particularly upon rice, mung beans, sesame cake (lnga watu), oil (lngis/minak),
dried fish (be sudang), onions, and long pepper (cabya). Although such foreign
pices as Chinese star anise were apparently imported by the mid ninth cen-
tury - as evidenced by the ship of Arab build that went down off Belitung, en
route from south China to Java, carrying not only Chinese ceramics but also
star anise (Flecker 2001b:349) - they do not appear to have spread beyond the
courts and ports into the market circuits until later.
The trading networks in which these pedlars were involved were, by the
ninth century, reasonably well-articulated, and fairly far-reaching, with trade
even in basic foodstuffs crossing community boundaries. The early tenth-
century Panggumulan charter (Sarkar 1971-72, II:24-41), for instance, lists
amongst those who attended the ceremonial feast held at Pang gum ulan to
mark the benefice grant:
(IILb.2) [... J four persons from Tunggalangan, who happened to be passing by at the
time on the main road, on their way to sell rife in the market at Sir:t<;lingan [ ...J.

It is interesting to note that vendors of rice were standard members of this

\·ery mobile category. There were probably several reasons for the movement
of rice around the countryside. There were clearly enough non-farming resi-
dents of most communities to provide a steady demand for rice in the market.
At the same time, rice surpluses in some regions appear to have been large
en ough to provide farmers with supplementary income. Javanese farmers of
:he early second millennium may have dealt with their problems of rice stor-
aae and cash flow in ways similar to their more recent counterparts - by
selling off a portion of their harvest immediately, and then buying rice in at a
later date, if necessary. In addition, since royal taxes were assessed on farm-
ing communities in currency, it is possible that some rice was sold in order to
;:neet tax obligations.
60 Jan Wisseman Christie

Bakul trade and the market network

The early Javanese market network appears to have consisted of a web of inter-
secting market circuits - many of which must have crossed internal administra-
tive boundaries. By the tenth century, in L\e large markets of the Brantas delta
in east Java, the lynchpin of this system was the abakul (Christie 1991:38-9). The
abakul was a market-based trader, similar to the modem bakul (Dewey 1962:76),
who bought produce from mobile traders and resold it either retail to custom-
ers or wholesale to other traders. The commodities handled by the abakul were
largely foodstuffs, spices and stimulants, and other agricultural crops with a
reasonable storage life and a sufficiently widespread demand for them to have
been traded between markets and regions. For instance, the Kubukubu char-
ter of 905 (Barrett Jones 1984:173) mentions abakul handling areca nuts, betel
leaves, onions and ginger; and'the early eleventh-century Madhawapura char-
ter (Van Naerssen 1941:58-9, iv.A.3) lists abakul handling salt fish. In the Cane
inscription of 1021 (Brandes 1913:120-5), the abakul was connected with the sale
of sugar, betel leaves, rice, safflower dye, mulberry and hirnbung-flower dye.
The Garaman charter of 1053 (Boechari n.d.) speaks of
those abakul who trade in areca nuts and betel leaves [... ] sesame and oil [ ... ] all the
produce of the marshes [ ...] tamarind [ ...] cotton [who] receive goods from other,
distant regions and sell them in the neighbourhood [ ...].

The Tuhanaru and Kusambyan charter of 1323 (Brandes 1913:198-204) refers

more generally to:
(7.a) those who may act as bakul traders (adagang bakulan)' handling all of the
contents of the dry rice fields (gaga), all the contents of the wet rice fields (sawah),
all the contents of the marshes, all the contents of the sea, all the contents of the
rwang [ ... ].

Tax and the export trade of early Java and Bali

Agricultural produce played a large role at all levels of trade, both in Java and
in Bali. A few Javanese inscriptions deal with the crops traded in the port areas,
and the taxing of trade conducted by the more highly capitalized merchants
and merchant groups trading between the ports and the interior. One of the
most important of the Javanese port inscriptions connected with port trade is
the eleventh-century Mananjung charter (Stutterheim 1929; Christie 1998:373-
74), which deals with regulations concerning crops stored for export:
(8.a.5) [They] must be mindful that they are obligated to undertake at all times
to protect [the quality of the produce]. They must not allow to become damp the
black pepper, [mung] beans, fennel, safflower dye, medicinal jamuju (Cuscuta,
 The agricultural economies of early Java and Bali 61

Linn.) seeds, coriander, wungkUi;lu-dye, and above ail, the rice [ ... J. (6) [ForJ black
pepper, one kulak volume-measure must henceforth weigh one kati [...J. (7) [...Jsalt
cakes (wuyah pasagi) [... J garlic (jasun) [... J.

ugar is mentioned in another passage in the same inscription. The exports

w arehoused in this port were, aside from salt, all agricultural produce - a
mixture of foodstuffs, spices, medicinals and dyestuffs. These exports can be
divided into two groups by probable destination. The rice - here identified
as the port's most important export - along with the beans, salt, sugar, garlic
and a portion of the other spices, were almost certainly bound for other ports
in Java's archipelago trade network. These would have been exchanged for
d oves and nutmeg from the Moluccas and sandalwood from Timor to be re-
exported to China and the Indian Ocean, as well as for metals and other raw
materials to feed Java's domestic industries.
The black pepper, wungkuqlu and safflower dyes, jamuju seeds, some of the
sugar, fennel and coriander were almost certainly bound for China. This sec-
ond list is of particular interest, because so few of the crops were indigenous
to Java. Black pepper (Piper nigrum, Linn.) and safflower-dye (Carthamus tinc-
.orius, Linn.) both originated in south India, which exported large quantities
of both to China during the first millennium AD. Fennel (Foeniculum vulgare,
~ill.) and coriander (Coriandrum sativum, Linn.) originated from even further
west, and probably also reached Java via India.
By the early second millennium, Chinese records indicate that Java had
displaced southern India as China's main supplier of black pepper, and Java
and Bali together had become the main suppliers of safflower to the Chinese
market (Christie 1998:352-53). International trade opportunities had clearly
precipitated some important shifts in the types of crops grown in port areas.
This may also explain Ma Huan's mention of double-cropping of rice in the
Brantas delta region during the early fifteenth century (Ma Huan 1970:91).
Since there was no apparent pressure on land, it may have been commercial,
export-related, pressures that made the planting of short-season rice attrac-
tive in port-catchment regions.
The difference in the way that produce was collected for overseas trade
in Javanese and Balinese states reflects the differing financial arrangements
made by rulers on the two islands. In Java, the abakul wholesale traders, so
important in the domestic marketing network, particularly in the major mar-
et centres on the lower Brantas, are not mentioned in inscriptions from the
orts themselves. The overseas trade of the ports in Java appears to have
been handled by highly capitalized merchants (both Indonesian and foreign)
who acted as tax farmers, making payments, both in cash and in kind, to the
court in return for the right to collect certain quantities of selected crops from
-armers in the port regions (Christie 1998:361). The abakul themselves were
62 Jan Wisseman Christie

subject to such collections by tax farmers. Javanese rulers, although interested

in the luxuries and income that trade could provide, appear to have played
no direct role in the collection of produce for export. They collected the major-
ity of their agricultural taxes in cash, calculated on the basis of the size of the
irrigated rice and other fields farmed by communities (Christie 1994:29-30).
In Bali, where the states were smaller, and the domestic tax-income of
rulers proportionately lower, taxes were collected both in cash and in kind.
Overseas trade, and the income it produced, appears to have been of greater
direct interest to the courts. In addition to handing on some areas of produce-
collection to tax farmers, Balinese rulers themselves used the tax system to
collect exportable produce. The Balinese system of compulsory-purchase levy
(pamli, tumbasan) is mentioned frequently in inscriptions from the late ninth
century onwards. The commodities collected by the courts in this manner
included not only items for their own use, and possibly for local resale or
redistribution, but also for export. The tax exemptions listed in the Sembiran
AI inscription of 922 (Goris 1954:65-7) provide an indication of some of the
goods that were normally subject to compulsory purchase in the hinterland
of a major port in north Bali:
(II.a.2) [ ... ] The compulsory-purchase levy of the ruler (pamli sang ratu), [to be col-
lected] each month of Magha, to the value of 1 ma$aka, and on the Great Ninth,
(3) 1 ma$aka; the compulsory-purchase levy of the nayaka official [to be collected]
each month of Magha, to the value of I ma$aka, and on the Great Ninth, I ma$aka;
the compulsory-purchase levy on commodities (pamli bhm:lIja), 4 ma$aka; the talikur
nayaka levy, I ma$aka. The following prices are to be paid by edict (arghayanangna):
(4) for 2 tala-measures of cotton, for 2 gunja-measures of thread, for 20 gunja-
measures of mung beans, for 20 gunja-measures of sesame cake, [the price is to
be] I kupang per official (panghurwan) gunja-measure. There will be no compulsory
purchase on rice, oil, long pepper, candlenuts (tingkir).

The Bwahan A inscription of 994 (Goris 1954:83-6) provides a similar list:

(II.4) [T]here is to be no [compulsory] purchase (tumbasan) of oil, long pepper,
mung beans, wungkUi;lu-dye, thread, cloth, if it is not [normally] produced by the
community [ ... ]. (III.I) [T]he royal compulsory-purchase levy (pamli haji) of the
month of Magha [on] the Great Ninth [is to be]: onions, at a price of 3 saga for I
sukat and 1 catu-measure; safflower dye, at a price of 3 saga for I catu; (2) dried
dlag-fish, at a price of 3 saga for two fish; sheep / goats (wljus) at a price of 3 saga
multiplied (?Ipihen) [... ].

Other commodities subject to compulsory-purchase levy in the tenth and

eleventh centuries included several types of dried fish, cow hides and sheep
skins, spices (basa), and copra (kalapa kring).
It is possible that the Balinese compulsory-purchase levies and the Javanese
tax-farming practices provided the models upon which the Dutch VOC later
 The agricultural economies of early Java and Bali 63

based their system of produce collection - a system that was further refined
in the nineteenth century, and labelled the' cultivation system'.

Conclusions

Although the above discussion is largely empirical, several points can be

made concerning early foodcrops and livestock in Java and Bali. First, it is
clear that rice was, from the first millennium onwards - and probably for a
long time before - the major staple crop of the two islands. Although dry-field
or hill rice continued to be mentioned in inscriptions, the expansion of wet
rice agriculture was promoted by states on the two islands, partly as a means
of settling strategic parts of the landscape, partly to improve taxable yields
and to make tax-paying population more visible (Christie 1992:12).
Second, because both islands, although rich agriculturally, were poor in
toolmaking stone and metal ores, their inhabitants were forced, from the
Neolithic period onwards, to trade outwards for basic necessities - much in
the way that their contemporaries in the lowland plains and river valleys of
the mainland traded with the inhabitants of the surrounding mountains. The
trade networks of the two islands extended through the archipelago and to
the mainland by the second millennium Be. By the later first millennium Be,
trade with India had commenced; by the first centuries AD, Java and Bali
had been drawn into the Old World trade system that linked Rome to Han
China by sea. Many of the crops, and a number of the types of livestock, that
were mentioned in Javanese and Balinese inscriptions of the ninth century
onwards, were apparently brought in by this trade - principally from the
Indian Ocean, mostly via Indian ports. Some lof these transplanted crops later
became mainstays of Java's and Bali's trade with China.
Third, external trade shaped the way in which both foodstuffs and the
labour that produced them were commodified. This, in turn, shaped the do-
mestic market and tax systems of the two islands, just as irrigated-rice farm-
ing shaped their landscapes. Foodstuffs and other crops, grown not only for
subsistence but also for domestic trade and for export, formed the core of the
early economies of Java and Bali.
64 Jan Wisseman Christie

Bibliography

Ardika, I Wayan and Peter Bellwood

1991 'Sembiran; The beginnings of Indian contact with Bali', Antiquity
65:221-32.
Ardika, I Wayan and Ni Luh Sutjiati Berartha
1998 Perajin pada Masa Bali Kuna Abad IX - XI Masehi. Denpasar: Fakultas
Sastra Universitas Udayana.
Asmar, Teguh and Belmet Bronson
[1973] Laporan ekskavasi Ratu Baka. Jakarta: Lembaga Purbakala dan
Peninggalan Nasional.
Barrett Jones, Antoinette M.
1984 Early tenth century Java from the inscriptions; A study of economic, social
and administrative conditions in the first quarter of the century. Dordrecht:
Foris Publications. [KITLY, Verhandelingen 107.] [Originally PhD the-
sis London, 1976.]
Bellwood, Peter
1997 Prehistory of the Indo-Malaysian Archipelago. Revised edition. Honolulu:
University of Hawaii Press. [First edition 1985.]
Bernet Kempers, A.J.
1976 Ageless Borobodur; Buddhist mystery in stone, decay and restoration, Mendut
and Pawon, folklife in ancient Java . Wassenaar: Servire.
Boechari
1986 Prasasti koleksi Museum Nasional. Jilid I. Jakarta: Proyek Pengembangan
Museum Nasional.
n.d. Prasasti Garaman. [Unpublished manuscript.]
Boserup, Ester
1965 The conditions of agricultural growth; The economics of agrarian change
under population pressure. London: Allen and Unwin.
Brandes, J.L.A.
1913 Oud-Javaansche oorkonden; Nagelaten transcripties van J.L.A. Brandes.
Uitgegeven door N.J. Krom. 's-Gravenhage: Nijhoff, Batavia: Albrecht.
[Bataviaasch Genootschap, Verhandelingen 60.]
Burkill, LH.
1966 A dictionary of the economic products of the Malay Peninsula. Kuala
Lumpur: Ministry of Agriculture and Co-operatives. 2 vols. [Reprint
of 1935 original.]
Chhabra, B.Ch.
1965 Expansion of Indo-Aryan culture during Pallava rule (as evidenced by in-
scriptions) . Delhi: Munshi Ram Manohar Lal. [Originally PhD thesis
Leiden, 1934.]
Christie, Jan Wisseman
1991 'States without cities; Demographic trends in early Java', Indonesia
52:23-40.
1992 'Water from the ancestors; Irrigation in early Java and Bali', in: Jonathan
Rigg (ed.), The gift of water; Water management, cosmology and the state
in South East Asia, pp. 7-25. London: School of Oriental and African
Studies, University of London.
 The agricultural economies of early Java and Bali 65

1994 'Wanua, thani, paraduwan: The "diSintegrating" village in early Java?',

in: Wolfgang Marschall (ed.), Texts from the islands; Oral and written
traditions of Indonesia and the Malay World, pp. 27-42. Bern: Institute of
Ethnology, University of Bern. [Ethnologica Bernensia 4.]
1996 'Money and its uses in the Javanese states of the ninth to fifteenth
centuries A.D.', Journal of the Economic and Social History of the Orient
39:243-86.
1998 'Javanese markets and the Asian sea trade boom of the tenth to thirt-
eenth centuries A.D.', Journal of the Economic and Social History of the
Orient 41:344-81.
Dewey, Alice G.
1962 Peasant marketing in Java. New York: Free Press of Glencoe.
Doherty, Chris, Paul Beavitt and Edmund Kumi
_000 'Recent observations of rice temper in pottery from Niah and other
sites in Sarawak', Bulletin of the Indo-Pacific Research Association, pp.
147-52.
Flecker, Michael
_OOla The archaeological excavation of the 10th century Intan shipwreck. [PhD
thesis, National University of Singapore.]
_OOlb 'A ninth-century AD Arab or Indian shipwreck in Indonesia; First evi-
dence for direct trade with China', World Archaeology 32:335-54.
Glover, Ian C.
1979 'Prehistoric plant remains from Southeast Asia, with special reference
to rice', in: M. Taddei (ed.), South Asian archaeology 1977; Vol. 1, pp. 7-
37. Naples: Istituto Universitario Orientale. [Series Minor Seminaria di
Studi Asiatici 6.]
1986 Archaeology in Eastern Timor, 1966-67. Canberra: Department of
Prehistory, Research School of Pacific Studies, Australian National
University.
1990 Early trade between India and Soutlteast Asia; A link in the development of
a world trading system. Second, revised edition. [Hull]: The University
of Hull, Centre for South-East Asian Studies. [Occasional Papers 16.]
[First edition 1989.]
Glover, Ian C. and Charles F.W. Higham
1996 'New evidence for early rice cultivation in South, Southeast and East
Asia', in: D.R. Harris (ed.), The origins and spread of agriculture and
pastoralism in Eurasia, pp. 413-41. London: University College London
Press.
Goris, Roelof
1954 Inscripties voor Anak Wungr;u. Band I[; Inleiding, transcripties, inscripties
in het Sanskrit]. [Bandung]: [Masa Bam]. [Prasasti Bali 1.]
Hanks, Lucien M.
[1 972] Rice and man; Agricultural ecology in Southeast Asia. Chicago: Aldine
Atherton.
Higham, Charles
1996 The Bronze Age of Southeast Asia. Cambridge: Cambridge University
Press.
66 Jan Wisseman Christie

Higham, Charles and Tracey L.-D. Lu

1998 'The origins and dispersal of rice cultivation', Antiquity 72:867-77.
Hirth, Friedrich and w.w. Rockhill
1966 Chau Ju-kua; His work on the Chinese and Arab trade in the twelfth and thirt-
eenth centuries, entitled Chu-fan-chi. Translated from the Chinese and
annotated by Friedrich Hirth and W.W. Rockhill. New York: Paragon.
[Reprint of 1912 original.]
Hunt, Robert C.
2000 'Labor productivity and agricultural development; Boserup revisited',
Human Ecology 28:251-77.
Kealhofer, Lisa
2002 'Changing perceptions of risk; The development of agro-ecosystems in
Southeast Asia', American Anthropologist 104:178-94.
MaHuan
1970 Ying-yai sheng-ian; 'T~e overall survey of the ocean's shores', (1433). Trans-
lated from the Chinese text, edited by Feng Ch'eng-Chiin; with intro-
duction, notes and appendices by J.v.C. Mills. Cambridge: Cambridge
University Press.
Mathers, William M. and Michael Flecker
1997 Archaeological report; Archaeological discovery of the Java Sea wreck.
Singapore: Pacific Sea Resources.
Miller, J. Innes
1969 The spice trade of the Roman empire, 29 B.C. to A.D. 641. Oxford: Clarendon
Press.
Naerssen, Frits Herman van
1941 Oudjavaansche oorkonden in Duitsche en Deensche verzamelingen. N.p.:
n.n. [PhD. thesis, University of Leiden.]
Naerssen, F.B. van, Th.C.Th. Pigeaud, and P. Voorhoeve
1977 Catalogue of Indonesian manuscripts. Part 2: Old Javanese charters,
Javanese, Malay and Lampung manuscripts, Mads Lange's Balinese let-
ters and official letters in Indonesian languages. Copenhagen: The Royal
Library. [Catalogue of Oriental Manuscripts, Xylographs etc. in Danish
Collections 4, part 2.]
Nossin, J. and C. Voute
1986 'Notes on the geomorphology of the Borobudur Plain (Central Java,
Indonesia) in an archaeological and historical context', in: M.CJ.
Damen, G. Sicco Smit, and H.Th. Verstappen (eds), Remote sensing for
resources development and environmental management; Proceedings of the
seventh international symposium [. ..J, ISPRS Commission VII, Enschede,
25-29 August 1986; Vol. 2, pp. 857-64. Rotterdam: Balkema.
Sarkar, Himansu Bhusan
1971-72 Corpus of the inscriptions of Java (Corpus inscriptionum Javanicarum), (up
to 928 A.D.). Calcutta: Mukhopadhyay. 2 vols.
Steinmann, A.
1934a 'De dieren op de basreliefs van de Boroboedoer', Tijdschrift voor Indische
Taal-, Land- en Volkenkunde (TBG) 74:101 -22.
 The agricultural economies of early Java and Bali 67

1934b 'De op de Boroboedoer afgebeelde plantenwereld', Tijdschrift voor

Indische Taal-, Land- en Volkenkunde (TBG) 74:581-612.
Stutterheim, W.F.
1929 'Transscriptie van een defecte oorkonde op bronzen platen uit het
Malangsche', Oudheidkundig Verslag 1928:105-8.
1940 'Oorkonde van Balitung uit 905 A.D. (Randoesari I)', Inscripties van
Nederlandsch-Indie 1:3-28.
Suhadi, Machi and M.M. Soekarto
1986 Laporan penelitian epigrafi Jawa Tengah. Jakarta: Proyek Penelitian
Purbakala Jakarta, Departemen Pendidikan dan Kebudayaan. [Berita
Penelitian Arkeologi 37.]
Walker, Michael J. and S. Santoso
1977 'Romano-Indian rouletted pottery in Indonesia', Asian Perspectives
20:228-35.
White, Joyce C.
1995 'Modeling the development of early rice agriculture; Ethnoecological
perspectives from Northeast Thailand', Asian Perspectives 34:37-68.
Wurm, S.A. and B. Wilson
1975 English finderlist of reconstructions in Austronesian languages (post-
Brandstetter). Canberra: Department of Linguistics, Research School
of Pacific Studies, Australian National University. [Pacific Linguistics
Series C, Books, 33.]