No people in any age or country had so great practice and experience in the construction of works for irrigation...
Tennent, Ceylon (1859)
It is possible that in no other part of the world are there to be found within the same space, the remains of so many works for irrigation, which are at the same time, of such great antiquity and of such vast magnitude as in Ceylon...
In, Egypt, Syria, Persia, and in India, there are remnants of far greater works, and in these countries, works of far greater antiquity, as well as magnitude, but probably no other country can exhibit works so numerous and at the same time so ancient and extensive, within the same limited area, as this Island...
Bailey, Report on Irrigation in Uva (1859)
Thus did two awestruck British officials of the nineteenth century view the most distinctive achievement of the people of the Anuradhapura kingdom—their masterly organization and maintenance of an irrigation network spread over the dry zone, a network that was remarkably attuned to coping with the geological and geographical peculiarities of its location: ‘Problems of intermittent streams, gross yearly variations, undulating relief, high evaporation some 80 from the Equator, poor groundwater resources, indifferent soils and marked seasonal concentration of rainfall with its risk of disastrous floods....’1 The dry zone afforded excellent conditions for the cultivation of rice: the high constant temperatures and received solar radiation, as well as the comparatively gentle relief of the region in contrast to the more rugged terrain of the wet zone of the south-west quadrant. But as against this, the rainfall was largely restricted to the period between September and January, less reliable and less ‘effective’ than in the wet zone. The topography of the dry zone with its gently undulating plains, the succession of small shallow stream valleys and low interfluves made irrigation more difficult than in a single great river basin or on a really flat plain. Besides, ‘the irrigation problem’ is much more formidable in an area with alternate wet and dry periods and a vanishing water table than in one with perennial streams and wells and a more even rainfall pattern.
The earliest projects were no doubt directed more at conserving than at diverting water on any large scale. But by the first century ac, large-scale irrigation works were being built. The reign of Vasabha (67–111) is regarded as a period of prolific activity and he is credited with the construction of twelve reservoirs and canals.2 Most of these that can be identified are located in the Anuradhapura area. The construction of tanks, canals and channels which this involved exhibited an amazing knowledge of trigonometry and the design of the tanks a thorough grasp of hydraulic principles.
In 1979, there was a demonstration of how skilled the ancient engineers had been. Canadian engineers, making an aerial survey searching for the most appropriate spot to build a dam for the Madura Oya Project in the Polonnaruva region in the eastern part of the dry zone, identified such a spot only to find after preliminary investigations at the ground level that it had been the very spot chosen by Sinhalese kings centuries earlier. When the forests were being cleared they found, not far below the surface, a well-preserved sluice and part of a dam. The Canadians moved to a site some short distance away, leaving the old site undisturbed for archaeologists to study.
The tanks had broad bases which could withstand heavy pressures and at suitable points in the embankment there were outlets for the discharge of water. A British engineer working in Sri Lanka in the late nineteenth century and early twentieth century pointed out that the Sinhalese were the ‘first inventors of the valve pit’ (bisokotuva)3 counterpart of the sluice which regulates the flow of water from a modern reservoir or tank. He went on to claim that the engineers of the third century BC or earlier who invented it had done their work with a sophistication and mastery that enabled their successors of later centuries merely to copy the original device with only minor adaptations or changes, if any.4 Sri Lanka owes more to the unknown inventors of this epoch-making device than to all but a handful of kings whose virtues are extolled in the Mah•vaṁsa and Cūlavaṁsa. Without the technological breakthrough which the bisokotuva signified, irrigation works on the scale required to sustain the civilization of ancient Sri Lanka—the construction of artificial lakes of outsize dimensions like the Minneriya Tank attributed to Mah•sena (AD 274–301) and the Kal•väva built by Dh•tusena (ad 455–73) where vast expanses of water were held back by massive dams—would have been all but impossible. During the early Anuradhapura period, the agricultural surpluses made available by the multitude of irrigation tanks scattered in rich profusion over much of Sri Lanka’s dry zone, along with the profits from external trade, made possible, in substantial measure, the enormous investment which the architectural and sculptural splendours of Anuradhapura required. That same agricultural surplus and the profits of external trade provided the capital for investment in irrigation works, dams, tanks and canals. Indeed, the early Anuradhapura period was the most creative and dynamic era in the history of irrigation in ancient Sri Lanka.5
A variety of seemingly intractable technical and physical problems were confronted and overcome, and the skills acquired and experience gained in this period were a rich lode mined by future generations. In the first century ad, the main problem had been that the water resources of the Kal• and Malvatuoyas, dry-zone rivers which dwindled to a mere trickle of water for much of the year if they did not dry up altogether over long stretches, were unequal to the demand for an abundant and dependable supply of water set off by population growth in a vigorous civilization. This could only be ensured by the diversion of water from rivers like the Mahaveli and others closer to the wet zone south of Anuradhapura which carried a perennial supply of water. The most notable of the irrigation projects of this early period was the Älahära Canal which took the waters of the Ambanganga, a tributary of the Mahaveli to the Anuradhapura region. This canal, first mentioned in the Mah•vaṁsa’s account of the reign of Vasabha,6 stretched about 48 km from a weir across the Ambanganga. Its length was testimony to the maturity and competence of the irrigation engineers of ancient Sri Lanka. During the reign of Mah•sena, the Älahära Canal became the main source of water supply for the Minneriya Tank which he built, and which was by far the largest tank up to that time. Mah•sena is credited with the construction of sixteen tanks and canals, four of which are in the Anuradhapura area and one in the Puttalam district.7
Three notable trends in the development of irrigation facilities during Mah•sena’s reign were: a resolute endeavour to harness the waters of the Mahaveli and the Ambanganga, the most important project being the massive Minneriya Tank; the improvement of facilities for water conservation in the north-western part of the island; and the attempt to develop the south-western part of the dry zone on the periphery of the wet zone. Together they accelerated agricultural development in the vicinity of Anuradhapura and opened up new areas for cultivation in the east and south-west. All the major irrigation projects initiated by him were achieved by a prodigious investment of labour resources on an unprecedented scale and they reflect, too, a notable advance in irrigation technology on the island.
The reign of Dh•tusena (455–73) matched, if it did not surpass, the achievements of Mah•sena and Vasabha in the extension of the island’s irrigation network. He is said to have added to the irrigation works in the Mahaveli region by building a dam across that river. But the main focus of attention in irrigation activity during his reign seems to have been the development of water resources in the western part of the dry zone. By far the most impressive achievement by this period is the construction of the Kal•väva, which tapped the Kal• Oya and helped to supplement the supply of water to Anuradhapura and the area around the city.8
The Kal•väva was an outstanding technological feat on its own. It had an embankment 5.2 km long and rose to a height of about 12.2 m, with a bund constructed of blocks of dressed granite mortised together to enable a very close fitting. Even more amasing was the 86.4 km Jayaganga linking it to the capital city to augment the waters, in its environs, of tanks such as the Tissa, the Nagara and the Mah•d•ragatta. Apart from irrigating an area of about 466 sq. km, the gradient in the first 27.4 km of its length was a mere 15.24 cm in 1.6 km (6 inches to a mile). R.L. Brohier, the historian of Sri Lanka’s irrigation systems, points out:
The Jayaganga, indeed an ingenious memorial of ancient irrigation which was undoubtedly designed to serve as a combined irrigation and water supply channel, was not entirely dependent on its feeder reservoir the Kalaweva for the water it carried. The length of bund between Kalaweva and Anuradhapura intercepted all the drainage from the high ground to the east which otherwise would have run to waste. Thus the Jayaganga adapted itself to a wide field of irrigation by feeding little village tanks in each subsidiary valley which lay below its bund. Not infrequently it fed a chain of village tanks8 down these valleys—the tank lower down receiving the overflow from the tank higher up on each chain.9
There was also the Yodaväva in the Mannar district, attributed to Dh•tusena, formed by building an embankment about 11.3 km long. Fed by a 27.4 km canal from the Malvatu Oya, the Yodaväva covered a vast area. It was a shallow reservoir and its efficiency in water storage must have been severely affected by the heat and aridity of the region, but the topography of the area made it impossible to construct a tank with a greater depth without resorting to techniques of lift irrigation. Carefully laid-out canals flowing at a low gradient distributed the water from the Yodaväva to a multitude of village tanks around it. Together with the P•nankulam, this tank was a vitally important asset in an area—Mannar—which records some of the lowest rainfall on the island.10
By the end of the fifth century, two major irrigation complexes had been developed, one based on the Mahaveli and its tributaries and the other on the Malvatu Oya and Kal• Oya. The skill with which these were elaborated further in subsequent centuries into a web that linked the Mahaveli and dry zone rivers, tributaries of the Mahaveli, and irrigation canals from coast to coast was demonstrated by Brohier in 1937.11 The two cities of Anuradhapura and Polonnaruva, the vital centres of political power and cultural activity which contained the most impressive monuments of Sinhalese civilizations were the cores of this network. Anuradhapura was much the larger of the two, and necessarily so, for during the first ten centuries AD it was, with brief interludes, the capital of the island.
There was the third core of Sinhalese civilization apart from the Anuradhapura and Polonnaruva areas, located in the dry zone of the south-east in Rohana where the climate was more severe and the rainfall much less reliable. Rohana was settled by the ancient Sinhalese nearly as early as the Anuradhapura region itself and was just as dependent on irrigation as the latter, with the difference that instead of the large tanks which dominated the landscape of the R•jarata, it had a distinctive irrigation pattern, a multitude of small and medium-sized projects, most if not all of which were the product of local initiatives. Apart from a few small tanks in the reign of Dh•tusena, and the building of a weir across the Valave River in the ninth century, there is hardly any recorded evidence of kings of the R•jarata devoting as much attention to the development of irrigation in the south-east of the island as they did to the main centres of civilization in the northern plain.
In the R•jarata, the Mahaveli complex provided the more dependable and abundant source of water, and while contributing substantially to the prosperity of the Anuradhapura region its prime function was to nurture the well-being of Polonnaruva and the outlying zone in its vicinity. The irrigable land area around Polonnaruva was further extended by Moggall•na II by the construction of the Padaviya Tank which utilized the waters of the M• Oya.
As we have seen in the previous chapter, proximity to the Mahaveli, the longest river in Sri Lanka, increased the economic potential of this region. Mah•sena had built the famous Minneriya Tank there and between the fourth and ninth centuries, a number of smaller tanks in the region helped sustain a considerable local population whose agriculture produced a substantial surplus for the state to extract. With the growth of commercial relations with China and south-east Asia, in which the port of Gokanna (modern Trincomalee) would have played a prominent part, the economic importance of the region was further enhanced. Thus the adoption of Polonnaruva as the capital of the Sinhalese kingdom by four kings of the period between the seventh and tenth centuries, and the final abandonment of Anuradhapura in its favour, were determined as much by considerations of economic benefit as by strategic and military factors.12
A thriving civilization dependent on irrigation for its sustenance has an insatiable demand for water and the search for a dependable and permanent supply of water is a never-ending one. Breakdowns caused either by some structural fault or by the depletion of water supply in periods of drought were inevitable, with the result that not every unit or link in this chain of interconnected tanks and channels was working at peak efficiency (or for that matter working at all) at any given phase of the island’s history in these centuries. Quite apart from essential repairs and maintenance, renewal was vitally important, as too were extensions of the irrigation network. Major initiatives in irritation activity called for a tremendous burst of energy, and these were not very frequent. Most rulers were content with keeping the tanks and channels which formed the country’s stock of irrigation works functioning at a reasonable level of efficiency.
By the end of the seventh century, there was a vast array of irrigation works spread over a substantial part of the dry zone of the country. The monumental scale of the large tanks is compelling evidence of a thriving economy and of a state structure which had the resources from an agricultural surplus and the profits from trade to invest in these projects as well as on religious and public buildings designed on a lavish scale.
It would be appropriate at this point to quote C.W. Nicholas’s assessment of this achievement.
This achievement, accomplished by the seventh century, is a very notable and distinguished one because many of these projects, if put in hand today [1960] would still be regarded as major undertakings. The chief engineers responsible for planning designing and construction were undoubtedly men of the highest technical ability as well as men of vision. We have no knowledge of what gaugings they made, what formulae they employed in their calculations and what instruments they used. All that information which must have been imparted orally, is irretrievably lost. We know, from surveys carried out in recent times for the restoration of ancient schemes, that the instruments they used were capable of the same accuracy as modern instruments. In some sections of the ancient canals the fall was only six inches to a mile, in most sections one foot to a mile.13
There were fewer new irrigation projects initiated in the eighth, ninth and tenth centuries than in the past. Many explanations have been made for this apparent lull, including the argument that it was all due to a sharp drop in the profits of external trade.14 This contention is too facile. By the end of the seventh century, the system was already nearly a thousand years old and there would have been a multiplicity of factors at play for this lull, of which an apparent fall of revenue from external trade would have been just one. Explaining that the reasons for this decline in activity are not easily discernible, Nicholas goes on to add:
It may be that what had been already accomplished was, with occasional extensions, sufficient: perhaps these earlier works did not begin to be utilised to the full extent of their productivity till the ninth century. Then apparently, if we may so infer from the references to famine and scarcity in the inscriptions of Kassapa V and Mahinda IV... food crises could occur if the irrigation system failed to function to the full through natural causes or neglect, and the time for constructing great, new schemes was approaching: but these undertakings would have been delayed by the invasions, disturbances and conquest during the second half of the tenth century.15
By itself the irrigation network of ancient Sri Lanka was a tribute to the ingenuity of her engineers and craftsmen and the organizational skills of her rulers. Nowhere else in south Asia does one find such a multiplicity of irrigation works as in the dry zone of Sri Lanka. The scale of comparison is not with the Indian subcontinent, but with the major hydraulic civilizations of the ancient world, the fertile crescent of west Asia and China itself. Despite its diminutive size, Sri Lanka belongs to this super league in regard to irrigation technology and creative achievement in irrigation works, for nowhere else in ‘the pre-modern world was there such a dense concentration of irrigation facilities at such a high technical level’.16
Ancient Sri Lanka was the example par excellence of a hydraulic civilization, but it does not figure at all in Wittfogel’s massive work17 on the theme. This was just as well, for Sri Lanka’s hydraulic experience, dispassionately reviewed, would have provided a refutation of some of the vital component elements of his theoretical framework.18
No part of Sri Lanka’s dry zone conformed to Wittfogel’s model of ‘full aridity’, and while it could be argued that in the two core areas of Sinhalese civilization in the R•jarata—the Anuradhapura and Polonnaruva regions—transfer of water from a distant locality was an essential prerequisite for the development of agriculture, the other part of the Wittfogelian theorem that ‘government-led hydraulic enterprise is identical with the creation of agricultural life’ has less validity for Sri Lanka. Irrigation projects were among the most important public works undertaken by the state in ancient and medieval Sri Lanka,19 but despite their crucial importance for economic development they were not ‘identical with the creation of agricultural life’ in any part of the dry zone. Certainly, the role of ‘government-led hydraulic enterprise’ in sustaining agricultural development in Rohana was comparatively minor, and even in substantial parts of the dry zone of the R•jarata outside the two focal points of civilization there, regional initiatives were just as important as state enterprise in sustaining the system. Indeed, in all parts of the dry zone, while major irrigation schemes were largely matters of state enterprise, such local initiatives—communal, institutional (especially monastic) and even individual—were responsible for the construction of a multitude of smaller reservoirs and village tanks which conserved water from the seasonal rains for agricultural development in their locality and which existed concurrently with and often independent of the main irrigation complexes.20
Nor did the state retain ownership of all the major irrigation works constructed under its direction. Dh•tusena ceded half the income of the Kal•väva to his brother. The long Älahära Canal was granted to a monastery not long after its construction. Monasteries, indeed, often had the resources to maintain irrigation works in their charge or control in good repair. Immunity grants of the Anuradhapura period record the transfer to the monasteries of the control of sections of the population together with the right to exact taxes and corvée labour from them; apart from these fiscal rights, administrative and judicial powers traditionally enjoyed by the king were also delegated to them by such grants. Similar immunities came to be enjoyed by the kullna gentry who claimed proprietary rights over some irrigation works and land.
Lists of officials which occur in inscriptions of the ninth and tenth centuries, when the irrigation network of Sri Lanka was most extensive and highly developed, have been cited as evidence of the existence of a hydraulic bureaucracy. Quite clearly the services of men with a high degree of technical skill were necessary for the construction of large and complex irrigation works, for their maintenance in good repair and for the regulation of irrigation water to fields. But this is not conclusive evidence of an irrigation bureaucracy on the Wittfogelian model, of a phalanx of technically competent officials who formed the key ingredient in an authoritarian political structure in which power was concentrated in the king and his bureaucracy. On the contrary, hydraulic society as it developed in Sri Lanka was not a centralized despotism, rigidly authoritarian and highly bureaucratic, but a structure and society in which power was shared between the monarch and other groups in society, including the gentry and monastic institutions.
The more important state-sponsored irrigation works boosted the island’s agricultural economy by enabling extension of the area under cultivation and habitation and facilitating more intensive exploitation of agricultural resources without upsetting the balance between land and population. Instead of a single annual crop, large-scale irrigation works ensured the production of two or three crops a year. The resulting agricultural surplus was adequate to maintain a large section of the population not engaged in food production and to sustain a vibrant and dynamic civilization. It provides an effective demolition of yet another of the key features of Wittfogelian theory—stasis as a characteristic of hydraulic civilization.
We need to end this brief survey of the hydraulic civilization of ancient Sri Lanka on a more sombre note. Irrigation civilizations by their very nature are critically vulnerable to natural disaster and foreign invaders. For such a society is like a complex machine with an extraordinarily delicate mechanism. It could function with amazing efficiency but could break down just as easily if maintenance were neglected or as the result of some seemingly manageable damage to the mechanism. With increasing complexity, inertia and negligence could be as insidiously detrimental to its smooth functioning as the more palpable threats from natural disaster or foreign invasion.
As we shall see in later chapters of this volume, the irrigation system was repaired, indeed reconstructed, and expanded by the Polonnaruva kings once Anuradhapura was abandoned after the Cōḷa invasions. Thereafter it lay damaged and neglected, and under a crushing cover of forests and jungles for six centuries, when some repairs were initiated under British rule in the 1860s and thereafter. The reconstruction was sustained with greater commitment and enthusiasm in the 1930s and 1940s. The first new major irrigation work since the days of the Polonnaruva kings came in the 1950s, in the early years of independence. Thirty years later came the dams and canals of the Mahaveli Scheme, whose magnitude and spatial reach matched, if not surpassed, those of the Sinhalese rulers who constructed similar works in the heyday of irrigation activity in the Anuradhapura kingdom. By this time Sri Lanka was no longer a hydraulic civilization, but the latter had left an enduring legacy for posterity, a legacy which had a powerful appeal for politicians and the people.