pp. 259-61 ; for a reference to a Nanda king of Magadha who con structed a canal in the Tosali division near the city of Kalinga, see D.C. Sircar,
Select Inscriptions,
Vol. I, 2nd e d . , Calcutta,Kalinga. ( 9 2 If we accept that the Mauryan penetration could have
A reservoir surrounded by an embankment dating to this
( 9 3
period was exposed in the vicinity of Dharanikota. At the
well have been used for irrigation purposes. Similarly useful
skills can be inferred from the excavations at Chandavaram which
revealed three monastic complexes on man-made terraces with
(94 retaining walls and flights of steps.
Contemporary inscriptions record the excavation of
reservoirs. An inscription from Adoni
taluk
in Kurnool districtrecords the construction of a reservoir by a householder
(gahapati)
(95of the village of Vepuraka. Similarly, we have the construction
of a reservoir near a Buddhist establishment by a lay-worshipper
(96
Bodhisiri along with her relatives. A Nagarjunakonda inscrip
tion, besides the construction of a temple by various individuals (97 with royal titles, also records the construction of two tanks.
We also have an inscription from Alluru in Nandigama
taluk
ofKrishna district which records the dedication of an already (98 existing reservoir to the local Buddhist community.
From Nagarjunakonda we get an excellent example of the use
of gravity flow. Excavations have revealed a canal: 1 km in length, (99
9.14 m in width and 1.82 m in depth. It has been dated to the
third and fourth centuries A.D., when the Iksvaku dynasty ruled over
the lower Krishna basin with their capital at VijayapurT, in the
neighbourhood of the canal. A detailed study of the canal reveals
that the gradient of the land, the slopes of the hills, a long strip
9h Ibid., 1976-77, p. 9.
95 V.S. Sukthankar, E . I . , Vol. XIV, pp. 153-55* 96 Based on an unpublished Nagarjunakonda epigraph. 97 D.C. Sircar, E.I., Vol. XXXIV, pp. 127-20U. 98 K. Gopalachari, E.H.A.C., Appendix A, p. 9^*
99 I.A.A.R., 195^-55, PP. 22-23, pi. XLV - A,B; 1958-59, p. 8, pi. XII, A; H. Sarkar and B.N. Misra,
Nagarjunakonda,
New Delhi, 1972, p.2h.
of depression and the local pattern of streams and rivulets had been
carefully studied before its construction. its sloping em
bankments and floor were reinforced with a hard lime gravel mixed
with kankar and stones to withstand the rush of water. A rubble
wall about 1.4 m wide was constructed on the southern side, so as
to prevent any land-slips from the adjoining hill-slopes. The
canal itself led the water out of a long strip of depression in the
eastern corner of the valley where the conjunction of hills is
angular. The canal was fed by rainwater from the hills. Several
streamlets fell into the canal. The manoeuvring of water is also
evident in the bunds around the canal.
Although the purpose of this canal is not clear,
considering the large number of people who resided in the valley
during the third and fourth centuries A.D., we may assume that a
large canal served different purposes. Even though the immediate
area around the canal itself had no religious or secular remains,
we may assume that during the dry summer months, the stored water
would have been indispensable for people living in the centre of
the valley. The habitational areas were situated within the four
kilometre span separating the river Krishna from the canal. The
rather meagre occurrence of any buildings in the canal area definitely
underlines its importance in irrigation, as does the enormous planning
and labour expended on waterworks at least two kilometres away from
the heart of the urban centre. It was a useful strategy to
combat the long dry season in this valley which had a relatively low
rainfall.
36
100 I.K. Sarma, ’Ancient Canals', pp. 309-11.
37
There may have been other waterworks although on a smaller
scale. An example is from the monastic area east of the citadel in
the same valley. Here an isolated brick drain or channel was
excavated. it ran for about 14.5m from east to west, taking off
from a brick platform (
3.2
x1
.6
m ) which was covered with a thick rubble wall. It served as a bund against the water coming from thenearby hills during the rains. The structure was probably designed
not only to divert water from the important buildings in the low-lying
area but also to collect the water in a storage tank for the dry
summer months. Thus, natural channels may have been exploited by
the construction of artificial courses leading off from them.
There is sufficient archaeological evidence underlining
the use of wells which are found both in the inland and coastal
(103 . „ ^ . , . „ „
districts. Material from excavations, however, is mainly from
habitational sites as these are the areas explored and excavated.
There were not only ring wells but also wells lined with bricks.
At Peddabankur, archaeologists have discovered remains which
apparently were meant for two wooden poles carrying a pulley over a
well to draw water. Numerous oblong shaped wells have been
102 I.A.A.R., 1957-58, p. 9.
103 Ring-wells from Kadali and rock-cut w e l l s , one of them with an inscription, from Kodavali in east Godavari district, have been discovered. I.A.A.R. s 1962-63, p. 66; Liiders List3 No. 13^+1; Dhulikatta and Peddabankur from Karimnagar district have yielded numerous wells. I.A.A.R., 1976-77, p.U; 1968-69. p. 2; 1970-71,
103 I.A.A.R., 1972-73, p. 2. p , 2 ‘
discovered at Nagarjunakonda. They were located not only in
and around the citadel but even along the river bank. The latter
may have acquired greater importance in the hot season when the
river bed was dry. In fact, while the role of wells would have been
of a supplementary nature in the coastal districts where plenty of
water was available, they could have been crucial and in some places
the only source in the inland districts.
Although the wells in the habitational sites, mainly for
domestic purposes, were lined, those used for irrigation probably
were not, especially in the coastal districts where the heavy monsoon
rains could have damaged the walls. Investing in permanent
structures could have been uneconomical. This pattern endures even
to this day. In fact, excepting where water pumps are used, the
mode of utilization of water appears to have changed little.
We can, based on the existing practices in some of the
inland areas of Andhra Pradesh, put forward the possibility of an
infield - outfield system. The immediate vicinity of the wells
where the land could have been best watered and best cared for, was
probably used for the cultivation of paddy. Millets and pulses,
however, were grown in the areas where water could be transferred.
In the drier months, with a shortage of water and with less input of
labour required, the crops were probably closer to the wells. Pro-
105 H. Sarkar and B.N. Misra,
op. oit.3
p.2h
.106 One of the wells at Peddabankur was H3 'courses' in depth,
I.A.A.R.,
1971-72, p. 2.duction was dependent not only on the amount of water available
but also the availability of labour for ploughing and the drawing
of water. Cattle which figure in donative records could have
been quite crucial. The rise and fall of water tables of the wells
would have led to the endemic instability of such irrigated tracts.
These irrigation systems were small and required only a
village level of organization. The excavation of wells and
reservoirs were probably concerns of the local élite who controlled
the agricultural land. As we have noticed earlier, pious laity and
householders could engage in such activities. Some major works would
have required on a large scale not only planning but also mustering
of labour. Although there is no evidence of state intervention,
we may assume from analogies elsewhere in early India, that corvee
labour was employed to build canals and channels. The availability
of labour would not have been a problem in summer months when the
demands on land were minimal and the water in the rivers and
streams was low or non-existent. It is not surprising that the
navigational channel and a reservoir were near the regional
capital of the Satavahanas while the canal was adjacent to that of
the Iksvakus.
The essential purpose of irrigation was to facilitate not
only the regular supply of food but also a greater quantity of it
to a non-food-producing population located in the urban centres and
scattered all over the countryside. No new major crop was introduced
during this period. Rice and
ragi
were the staples. But from thenature of the climatic variations and the distribution of early 39
Uo
historical settlements, we can surmise that paddy cultivation must
have been in vogue, a development from the contacts with North
India. This would have been a major change from the earlier dry
hill variety of rice, even to this day grown in the semi-arid
inland districts. If we assume that the proportional yield of
rice and millets, such as
jowar3 bajra
andr a g i
which are the mostproductive, have remained constant, then the former, whose yield is
twice that of the latter, would have acquired greater
significance. Moreover, in the 'new lands' with a greater amount
of precipitation and with alluvial tracts and with better facilities
for irrigation, more than one crop could be harvested annually.
These factors would have increased production considerably, making
the proliferation of agrarian settlements and monastic establishments
possible
/ 108
An important aspect of irrigation was that it facilitated
the cultivation of different varieties of paddy. In Andhra Pradesh
and the neighbouring regions, three such strains appear to have been
grown:
sali3 vrzhi
andsyamaka.
v/e have direct evidence of theknowledge of
s3li
in our area. This association is preserved in theancient name of an early Buddhist site, Salihundam, which was
Salipetaka
of the second-third centuries A.D. The name itself has been
translated as an emporium of rice. This reminds us of a similar name
107 V.S. Gananathan,
Economic Geography of India,
New Delhi, 19&7, P* 2U.108 Map III.
109 D. Das,
Economic History of the Deccan
, Delhi,1969
, p. 95. 110 R. Subrahmanuam,Salihundam3 a Buddhist Site in Andhra Pradesh3
kl
Dhanyakataka
, which can also be translated as an 'emporium' ofrice.^^^ A few centuries later, Hsllan-tsang observed that
An-to-lo
(Andhra) andTe-na-ka-che-ka
country (Dhanyakataka
orthe modern Amaravati-Dharanikota area) had fertile soil and that the
it was probably an important market place for rice. We may assume
that these places produced not only abundant rice but also different
varieties of it.
There is some evidence for the cultivation of other crops.
Inscriptional references to guilds of betel-leaf growers or sellers
(114
and of confectioners indicate the commercial importance of
betel-vines and sugar-cane as local crops. Part of the sugar
exported to the Mediterranean
^115
probably came from Andhra. From a later text we know that sugar was grown in parts of Andhra Pradeshand Maharashtra, and that a device called
gudayantrika
was used toextract the juice from the cane and this was then boiled to produce
molasses. 11^
There is evidence for other crops from the adjacent Maharashtra
for this period. Although at present such evidence is lacking in
Andhra Pradesh, based on the interaction between the two regions, we
may assume that those crops were also grown in our region. Of the
different types of oilseeds, mustard
(rajika)
, and castor(eranda)
^11^111 Taranatha translates the name as a 'rice heap', see A. Schiefner,
Taranathas Geschichte des Buddhismus3
St. Petersburg,1869
,p. lU