Strict access rules hut no conservation: the case of type-II societies

Back to contents - Previous file - Next file

General considerations
Let us now shift our attention away from resource-abundant societies to consider situations in which resources have become scarce in the sense that they are now the object of potential competition among users. This is taken to mean that the number of users is sufficiently high relative to the resource flows for any additional entry to cause the average product to fall or the average cost of extraction to rise. In other words, the resource is now subtractible in the sense that each individual user is potentially capable of subtracting from the welfare of other users. For the sake of illustration, consider the case of maritime fishing. When the resource is abundant—implying that the average product or cost is constant—there are numerous beaches from which fishing operations can be launched, fish can be caught in sufficient quantities near the shore, and anyone who so desires can therefore start to fish without hurting the interests of the existing fishermen. When the resource has ceased to be abundant, all the most productive beaches have already been occupied and additional fishermen have either to settle on less productive or more risky beaches (for example, those where the sea bottom is not as smooth or desirable for beachseine operations, or where the launching and landing operations are fraught with danger), or to get access to a high-quality beach through acceptance by the local community. In the former case, the new entrants have to bear the whole burden of the decreasing returns to fishing (or of the increased risk) while in the latter whether it is so will depend upon how they are integrated into the local community and, particularly, upon the nature of the access rights granted them. The burden of decreasing returns is shared among both new entrants and existing fishermen if no discrimination is practiced vis-ŗ-vis the former which is typically the case when they come from the natural growth of the resident population.

Now, it must be borne in mind that the decline in the average product (the increase in the average cost) resulting from the pressure of a growing number of resource harvesters (assuming constant technology) may be due (1) only to the enhanced competition for the resource that this growing pressure entails, or (2) both to this enhanced competition and to the erosion of the environmental capital stock. The first situation points to a purely distributive problem caused by the very scarcity of the resource: since the resource is increasingly difficult or costly to harvest, there arises the question as to how the cost of extracting it is to be shared among the various claimants. Such a situation therefore requires well-defined sharing rules which are obviously tantamount to access rules (it being understood that exclusion is the obverse of access). The second situation involves both a distribution problem and a genuine resource management problem. (In the socio-anthropological literature, unfortunately, the expression resource management is generally used to denote the ways and methods applied to solve both types of problems indiscriminately, thereby creating a lot of unnecessary confusion.) Consequently, it calls not only for sharing rules but also for conservation mechanisms lest the pressure on the resource, other things being equal, should be stronger in the future than in the present.

If we were to follow the logic of the romantic view, we would argue that societies confronting situations of the first aforementioned type tend to adopt well-defined sharing or access rules while those confronting situations of the second type tend to devise both sharing rules and conservation mechanisms. Our own position is different, however, because we do not believe that societies optimally and instantaneously adjust their rules or institutions to the evolving challenges which they meet. In the new institutional economics (NIE) - a new and rather eclectic branch of economics which deals explicitly with issues of institutional change—when such change is slow or hard to come (societies do not evolve the optimal rules which they need as soon or as effectively as they would need), it is necessarily because of some sort of collective action problem which prevents the new rule or institution from being supplied, thus frustrating the individual actors' demand for institutional change. In other words, people are assumed to be wise, lucid, and rational and the problem arises only from the fact that individually rational actors do not necessarily produce collectively rational outcomes as is epitomized in the PD story (bear in mind that any rule or institution is a public good and that its production is therefore vulnerable to the freerider problem).

Our thesis is at variance with that of the NIE inasmuch as we believe with McNicoll that, contrary to the assumption commonly made in the social sciences (including, to an increasing extent, in sociology), people are actually 'error prone, partially informed, culturally blinkered to varying degrees, and somewhat arbitrarily mixing short- and long-run considerations in their decisions' (McNicoll, 1990: 148-9; see also Lane, 1991: 52 for evidence from cognitive psychology). To put it in another way, societies may not optimally adjust to the evolving challenges which they face, not only because interactions among their members give rise to collective action problems, but also because the nature of the problem at stake is not correctly perceived or is simply misunderstood by the actors concerned. In the context of the present discussion, our hypothesis can be formulated as follows: people in traditional societies are naturally inclined to think that, when their income derived from the exploitation of CPRs happens to fall, this must be entirely caused by the encroachments of new producers (especially so, of course, if the latter are strangers endowed with superior techniques) whose activities have the effect of diminishing the share of the current resource flows available to the existing operators. The idea that decreasing incomes can also follow from a reduced stock of the resource as a result of overexploitation and mismanagement by all the producers involved (that is, as a result of their own harvesting behaviour as well as that of new entrants) does not readily come to people's mind.

When these conditions obtain, societies therefore tend to adopt strict access rules while neglecting to evolve conservation devices: they thus correspond to type-II societies in our basic typology. Note that another possible situation is that in which a stock is depleted owing to the more intensive harvesting efforts made by a constant number of existing producers (either because the latter use more effective harvesting techniques than before, or because they have increased their working time). In these circumstances, and following the same logic as that described above, people arc unlikely to attribute to their own behaviour any decline in the number of resource units which they are able to currently appropriate or use from their natural resource system. If this is the case, they do not respond to the worsening of their incomes by trying to halt the process of resource degradation so as to maintain their long-term incomeearning capacity. (Bear in mind that resources have been assumed to be scarce—people cannot easily move to other areas or start harvesting new resource stocks—and that people's livelihood crucially depends on what they are able to harvest from their natural environment.) In terms of our typology, what we encounter here is a variant of type-I or type-II society depending upon whether rules of access are relatively flexible or relatively strict, that is, depending upon whether there is a threat from potential users who do not belong to the local community.

We are thus faced with a set of three distinct situations which need to be elucidated. To recall, there is first the situation in which the sheer pressure of a growing number of users creates a problem of distribution of the resource flow currently available from the natural system concerned. To this situation, people react by developing stricter rules of access. Second, there is the situation where the above distributive problem is compounded by a problem of stock depletion/degradation, yet people react as though there were no management problem and the only issue were that involved in sharing the current flow from a given resource stock. Third, we have the situation in which stock depletion/degradation takes place with a constant number of users who exploit their natural environment in a rather intensive way. In such a case, people remain passive and there is continuous erosion of the environmental capital stock.

Regulating access to CPRs

How can we account for the above reactions and, in particular, how are we to explain the difficulty of traditional societies' members in admitting the possibility of resource destruction as a result of man-made efforts or human intervention or, at least, how are we to explain their unwillingness or inability to take active steps to remedy such kinds of situation?

The first thing to note is that, in a number of significant cases, the fact that new entrants directly compete with existing producers for the use of a resource is patent and it is just to be expected that the people concerned respond to the newly created problem as essentially one of pure distribution. Two striking illustrations of this phenomenon of straightforward competition are those of village woods on the one hand, and of fishing spots on the other hand, that is, situations in which the resource is well localized and either directly visible (like in the first example) or within almost tangible reach (like in the latter example). Now, when a CPR is manifestly strained by increasing population pressure, the spontaneous reaction of the local community is (a) to tighten the rules of access so as to make it more and more difficult for outsiders to use the CPR; and (b) to design more precise rules for the sharing of access and the avoidance of conflicts among the growing number of insiders. The latter process may involve de facto privatization of use rights or some sort of rotation.

Let us consider the problem of access to choice fishing spots in some detail. This is an instructive illustration since a recent study based on a sample of thirty fisheries concludes that fishing groups are usually quite effective in solving 'assignment problems' of the aforementioned kind while they systematically fail to overcome externalities that arise from competition for scarce fish in the open sea (Schlager, 1990, 1994).3 Depending on the resource available in a particular spot and on the level of labour productivity, a smaller or greater number of fishermen will be able to operate from this spot without impairing each other's catching capacity. Moreover, the fishing technology used may involve labour indivisibilities. This implies that the minimum number of fishermen required to exploit the targeted stock may sometimes be rather high (say, up to twenty-five to thirty-five people), as evidenced by the beachseining technique in which a large bag-shaped net (called a beachseine) with coir-wings of extensive length is slowly dragged from the shore after having been put out at sea with the help of a boat.

Let us assume that (a) there are m suitable fishing-spots available in a given community's territory (in a beachseining community, for example, this means that there are m separate nearshore areas of sufficient size where the sea bottom is sandy and free from rocks and other obstructions so that the net can be dragged smoothly); (b) on each day, a particular fishing spot can be worked by h shifts of fishermen, assuming that each shift works to its maximum capacity (to stick to the above example, the laying of a beachseine and its subsequent operation takes about two or three hours and, as the net must be dried after use, it can be operated only once in a day, thus allowing two or three other nets to be operated from the same beach site), and (c) n fishermen make up a shift. Then, a situation of resource abundance obtains if the number of fishermen willing to operate from the fishing territory of that community (N) is smaller than or equal to m x h x n. If, on the contrary, N has become greater than m x h x n, a potential for conflicts arises because accommodation of the (N - m x h x n) excess fishermen would necessarily cause a reduction in the catches of the fishermen who were working before fishing space became scarce. In these circumstances, local fishermen immediately understand the threat to their incomes posed by the excessive number of operators in the area. As a result, they typically respond to the new situation by restraining and eventually stopping access of outsiders to their fishing grounds much in the same way as was observed in agriculture when pressure on land became to be felt by local residents. Thus, while in Africa up until recently strangers could easily obtain rights to clear virgin forest lands against the payment of a share of their crop to the customary custodian of the (unoccupied) land, as the land frontier became gradually exhausted, indigenous ideology began to reassert with vigour 'the inalienable rights of the native custodians of the land, and the inalienable rights of individual usufruct' (Robertson, 1987: 77).

Note that, as theoretically argued in Chapter 8 (Sect. 2), it is important, in order that exclusion can actually occur, that the identity of the winners and losers be well known before exclusion takes place. This is amply verified on the ground since 'outsiders' tend to be excluded while 'insiders" continue to have their customary rights of access protected.

The enforcement of increasingly exclusionary arrangements in the form of barriers to entry into the local fishery may not, however, be sufficient to prevent the multiplication of conflicts if population pressure also arises from the natural growth of the resident community. This is why many fishing communities have taken steps to regulate access to the available fishing spots even among community members. One method adopted towards this end consists of assigning long-term exclusive use rights to particular groups or families. It often favours groups or families which were the first to settle down in the area and therefore occupy the top of the social hierarchy. Thus, for example, on Ulul atoll in the Pacific, 'islanders can fish freely anywhere in the island's sea territory, apart from one particularly wide and fish-rich tract of reef flat near a large reef passage, which is reserved for the exclusive use of the chief's clan' (Ruddle, 1988: 80). In the more egalitarian Indian societies of the Lower Klamath area in California, a social group was just a loose aggregation of individuals and nuclear families clustered together in small villages along the rivers and coast. Among the Yurok, the Hupa, and the Karok, 'individual men claimed exclusive use of fishing sites, improved game yards, or gathering sites for themselves and their families. Good fishing-spots were the property of individuals or partners and were transferable. Others could not fish at or immediately below a private spot, although a few good pools or rapids were open to common access' (McEvoy, 1986: 34). There were none the less other exceptions to the privatization rule. For one thing, 'resources gradually lost their private character away from the center of the area and back from the rivers and coast, where land was of less economic value'. For another thing, important exceptions to the same rule 'occurred when people of one village or several together cooperated in building seagoing boats or salmon weirs' (ibid.).

In San Miguel Bay, Philippines, even in times of general abundance of fishery resources, 'near-shore areas were allocated to specific families for exclusive use for fixed gear, due to the limited availability of such sites' (Cruz, 1986: 123). In Bahia, Brazil, long-term private claims are assigned over brackish water spawning-grounds, reefs, and net fishing spots defined by the lunar-tide cycle (Cordell and McKean, 1986: 92). In Japan, in the small-scale fixed-net fishery individual operators are assigned fishing spots based on their traditional usage patterns, a practice vindicated on the grounds that these nets must be individually tailored to the topography of the sea bottom in each spot fished. In some areas, moreover, octopus holes within a joint rights area are owned and inherited as personal property (Ruddle, 1987: 46).

Our last example is taken from a much more sophisticated fishery and illustrates the fact that regulatory efforts can emerge in new areas from traditional practices brought about by strongly knit immigrant ethnic groups. Thus, during the last quarter of the previous century, in the Bay Area shrimp fishery (California), we are told that, without any formal authority to do so, Chinese shrimp companies (after they had won control over this fishery by virtue of their comparatively low production costs) recognized each other's proprietorship over certain fishing areas and each company thus possessed two or more grounds that it worked at different times of the year. This regulatory system was devised with a view to protecting the substantial investment each company made in fishing assets. So strong was the system that it still functioned as late as the 1930s, long after most of the industry had been legislated away at Sacramento (McEvoy, 1986: 96-7).

As a result of the granting of long-term exclusive rights, some groups or families may be excluded from access to the available (scarce) fishing sites and be forced to outmigrate and, perhaps, to settle in less productive fishing grounds. For the sake of illustration, let us consider the recent evolution of the situation obtaining in the Niger delta (Mali). According to Bozo fishermen's customs, group control over the water channels and group rights over particular spots on the weirs built across these channels were easily transferable to families with which original settlers united by marriage: they could thus be given over by a father (or a brother) to a marrying daughter (or sister) and the latter's descendants would be entitled to inherit them; or they could be transferred as a betrothal gift to seal the alliance with the family into which the daughter was to marry. However, when competition arose during the last decades around the available fishing sites, it became usual for the descendants of a native woman who are geographically or genealogically distant to waive their traditional use rights (Fay, 19906: 2245).

It is worth noting that secrecy often operates as an effective mechanism that assigns choice fishing spots, thereby minimizing competition among fishermen (White, 1977;

Forman, 1980: 20-2; Carrier, 1987: 147; Ruddle, 1987: 53-4; Acheson, 1989: 362-3). Secrecy is possible simply because knowledge and skill for locating dense concentrations of fish or choice fishing sites are difficult to acquire. Thus, for example, Ruddle remarks that in Japan 'by far the commonest way in which temporary private tenure to communal sea space arises is through the exclusive use of fishing spots as a result of closely guarded personal knowledge' (Ruddle, 1987: 53-4).

Another method of regulating access to scarce fishing sites inside a local community—a method which is probably more widespread than the de facto privatization of these sites—is to lay down or recognize use rights by rotation so as to promote equitable access to resources and to minimize interpersonal conflicts between fishing units. Under such circumstances, all the members of the community are expected to share in the costs of its natural growth in terms of loss of individual welfare. Rotating access to scarce fishing spots naturally implies that only temporary use rights are assigned to individual fishing units. The period of time for which these rights are granted may vary significantly yet in most recorded cases it does not exceed one day and can even be as short as a few hours. Note that, when the period is relatively long, it is presumably because, rather than a scarcity of fishing locations relative to the number of fishing units, there prevails a scarcity of productive locations or a strong disparity in the productivity of different available locations. In these conditions, rotation can be slower because the objective pursued is to equalize expected incomes over the medium or the long term rather than to share a limited number of spots among the fishermen so as to enable all of them to earn a minimum income in the short term.

On the other hand, it is a common practice in many fishing communities to let the firstcomer's priority principle decide who will be assigned a given fishing spot on a particular day. In Yaeyama and off the main island of Okinawa (Japan), for example, 'the priority right to a fishing spot by a firstcomer is absolutely sacrosanct . . . it is ironclad in customary law' (Ruddle, 1987: 56). It is thus important to emphasize that acknowledgement of the rights of a firstcomer to a fishing location may involve well-established informal tactics or codes of conduct. This can be vividly illustrated with the system of sea tenure discovered by Cordell and McKean in southern Bahia in the Brazilian north-east:

When a captain wishes to fish in a particular spot outside the system of lunar-tide property rights, he announces his intention—including what tide level or series he will use in casting nets—several days in advance at a local bar where fishermen like to congregate. All that is required is for another fisherman to be present as a witness. To ensure the claim, the captain must follow his proclamation by going to the chosen spot the day before fishing to leave a canoe anchored with paddles sticking up in the air. This forewarns competitors that the casting space has been taken. Fishing captains go to considerable lengths to support each other in this routine, which is part of the sea tenure politics that shore up the entire fishing system. (Cordell and McKean, 1986: 96)

There is another widespread system of allocating fishing spots for temporary use and it seems to be all the more predominant as locations are more productive and therefore more intensely sought after. This system consists of predetermining a com plete sequence of net-laying rights for regulating access to fish resources during a whole fishing season or even during only a single day's fishing operations. In many cases, the mechanism used for determining the sequence of rights is a simple lottery, a mechanism which is still in force nowadays in some advanced countries' fisheries, as in some mussel fisheries of southern France. In other cases, however, the order of (rotating) access to a particular fishing spot is set by a local customary authority: thus, among the Somono fishermen of West Africa, it was the responsibility of the so-called river's master (the Ba Tigi in local parlance) to decide the fishing turns (which lasted twenty-four hours each) by taking into account the intensity of demand, the abundance of fish, and the principle of equality between lineages (Jeay, 1984: 5).

The literature provides solid evidence of the ability of traditional communities to devise sometimes ingenious systems for allotting fishing sites to their members on a rotating basis. One such system has been observed by Taylor in the small estuarine salmon fishery of Teelin Bay (on the Glen River) in Ireland where it was still in operation in the early 1970s. Taylor's description is as follows:

Seven eighteen-foot open boats sat motionless around the perimeter of Teelin Bay, half a mile across at its widest. Each boat was manned by four or five men sitting quite still and staring in various directions over the water. A seine net lay in the stern of each boat, from which a single line led to another crew member standing on shore. After a while a salmon leaped just downriver from one of the boats, the crew of which immediately began to row furiously upriver and around the leaping fish, paying out the net as they went. The shore man held the line taut, holding one end of the net stationary as the crew rowed and brought the other end of the net around to complete the circle. All boat hands then pulled the bag thus formed, and its contents, aboard. Having stowed their salmon in one end of the boat and neatly piled their net for the next dol (as each casting of the net is called in Gaelic), the crew looked up and around toward the other boats scattered about the estuary. One of these boats was in fact rowing toward the crew that had just taken a dol and yet another vessel was heading for the position vacated by the second boat. Within a few minutes, about half of the boats had thus exchanged positions, and all with no words other than a polite greeting. The same system operated upriver in the illegal channel, where each boat, after taking a dol. would simply row upstream to the end of a virtual queue. (Taylor, 1987: 296-7)

The sea tenure system discovered by Alexander (1980, 1982) and by Amarasinghe (1988, 1989) in the beachseine fishing communities of southern Sri Lanka clearly involves an initial assignment of fishing spots through a lottery and then rotating these among the fishermen in a predetermined manner so that all of them get equal chances not only to fish but also to fish in the best spots (Hannesson, 1988: 9).

As we have already mentioned, beachseining is a fishing technique that requires a rather large water space (since it is intended for catching a whole school of fish) located close to the shore and with a smooth sea bottom. Furthermore, the laying of a beachseine and its subsequent hauling takes only a few hours so that, on a particular day, only a maximum of four nets can be operated on each suitable location. It must also be borne in mind that the probability distribution of incomes from beachseining is significantly affected by the timing of fishing operations, within a daytime as well as across seasons. For one thing, expected incomes are always higher when the net is laid in the (early) morning troth because big catches tend to occur early in the day and because the earlier fish is brought to the market the higher its price. For another thing, there are considerable interseasonal variations in fish catches and, in particular, a sizeable portion of yearly incomes from beachseining are obtained during a flush season which lasts only one month and the exact occurrence of which can never be safely predicted.

From Alexander's account (1980: 97-102; 1982; ch. 7), we can infer the following set of social rules that determine the local sea tenure system as a non-market organization ensuring equal access to productive resources:

  1. Membership in the village community, whether hereditary or acquired in a lifetime, involves a right of access to the community-controlled sea area (local fishermen belong to the same kinship group).
  2. Behaviour of the rentier-capitalist type is not allowed, which implies that ownership of a net carries the obligation to work it when required.
  3. There is no labour market and, since a beachseine normally requires eight fishermen to operate, joint ownership of nets is the rule and the usual ownership share in a net is 1/8. In fact, each net is divided into eight sections or shares but 'once the net is in operation individuals have no particular rights to the sections they have contributed' (Alexander, 1982: 142).
  4. Access to the sea is preconditioned upon the ownership of a net share. Nevertheless, this does not lead to exclusion of a fraction of the community in so far as all members, rich or poor, have access to capital. Indeed, a common way for a poor fisherman to acquire a share in a net is to receive it on credit (payment being postponed until after a good catch) from other shareholders in need of an additional partner (ibid.: 143).
  5. There is no market for net shares which can be transferred only through inheritance. The importance of such share transfers lies in the fact that the heir thus acquires the right to fish at the local beach: a net share therefore plays the role of an entry ticket into the local fishery (ibid.: 203).

Notice, incidentally, that when families grow in size, they may choose between either maintaining joint family ownership of their fishing assets—in which case all the siblings get fractions of one-eighth shares in old and newly constructed nets or giving over one-eighth shares in the existing nets to some sons and one-eighth shares in newly constructed nets to others. While the former method has the advantage of spreading production risk over several nets (Alexander, 1982: 143), the latter enables the heirs to be more independent of each other for the management of fishing assets.

Now, given the comparatively high productivity of early-day fishing operations, there arises an intense competition for taking the first turn on each available spot as soon as the total number of nets (k) exceeds the number of spots (m). Designing rules to regulate access to the sea within daytime is then necessary to avoid serious conflict over the use of the suitable water space. In addition, rules governing rights of access across days will be required when the number of existing nets exceeds the maximum number of hauls per day, that is, the number of fishing spots (m) times the maximum number of daily hauls on each spot (h). To sum up:

regulation within days must occur if k > m; regulation within and across days
must occur if k > hm = 4m.

The sea tenure system in south Sri Lanka's beachseine fishery is based upon the principle of equal access: every net should have equal chances not only in terms of access to the water but also in terms of access to catch and income opportunities. And when new individuals enter into the fishery and new nets are being constructed beyond the point where k = 4m, access to the available water space for existing nets will be proportionately reduced. Equal access is guaranteed through a turnover system that determines turns in a sequence of net-hauling rights. Thus, in Gahavšlla—the village where Alexander conducted his in-depth study—the fishing area is divided into two stations: the harbour side (from which most big catches come) and the rock side. The net cycle begins on the harbour side and, after a net has had the dawn turn on that side, it is entitled to the dawn turn on the rock side on the next day. Subsequently, it may be used on the rock side each day once the net immediately following it in the sequence has been used. The sequence of net use over a period of five days is shown in Table 10.1.

As can be seen from the table, twelve nets have been used at least once in a period of five days. Furthermore, nets 5-8 have been worked a maximum number of five times during this period; nets 4 and 9 have been used four times; nets 3 and 10 three times; nets 2 and 11 twice; and, finally, nets 1 and 12 only once. If the total number of nets is twelve, net I will reappear in the harbour site on the sixth day (where it will have the last turn) and in the rock site on the tenth day (where it will have the first turn). As for net 5, it will have a new dawn turn in the harbour site on the thirteenth day: a complete net cycle lasts twelve days. Over the full net cycle, each net will have been operated eight times (the total number of possible turns per day in all the existing fishing stations), that is, for two out of three days on an average: the intensity of use of each net is thus equal to hm/k. It is worth noting that, for each net, the period of use is strictly continuous: in the above example, once its first turn has come, a net will be worked during eight days in succession and, thereafter, it will be left idle for four days.

TABLE 10. 1. Sequence of net use in the village of Gahavšlla

Day Harbour Rock
  Dawn Night Dawn Night
One 5 6 7 8 4 3 2 1
Two 6 7 8 9 5 4 3 2
Three 7 8 9 10 6 5 4 3
Four 8 9 10 11 7 6 5 4
Five 9 10 11 12 8 7 6 5

Note: Adapted from Alexander (1982: 145).

Now, it is important to notice that the interseasonal variability of catches does not create major problems in the above procedure as long as the total number of nets does not exceed by too large a margin the total number of possible turns (or hauls) per day (in all fishing stations): this should be evident from the foregoing example. However, if the total number of beachseines increases significantly due to population pressure, increasing market integration, and rising fish prices, or any other reason, as the period of net use is strictly continuous, the annual returns of any beachseine will be governed mainly by its position in the net cycle: most incomes will accrue to the nets which happen to have many turns during the flush period. This is exactly the situation which was observed in Southern Sri Lanka from 1940 onwards: in Alexander's area of study, for example, only twenty-five of the ninety-nine nets received turns during the flush period in 1970-1 (Alexander, 1980: 105-7; 1982: 147, 203-8).

The aforementioned problem must nevertheless be viewed in its right perspective. Indeed, it is over a single year that the customary means of regulating access to the sea tend to make the distribution of the catch more and more unequal as the number of nets increases: over a longer period, returns should be expected to even out. Thus, Alexander writes: 'a particular net's turn is not tied to any point in the year, and over a four-year period each net will be used in each month. As the fishermen are unable to predict the flush period with any great accuracy, each net has an equal chance of good catches each year' (Alexander, 1982: 147). Yet, this clarification should not be taken to mean that the multiplication of nets does not raise problems. Even assuming that the economic pie is sufficient to provide for the livelihood of all households, there remains the question as to how year-to-year fluctuations in income will be buffered by them. Clearly, risk-pooling and intertemporal redistribution mechanisms become increasingly necessary to compensate for the partial failure of the traditional method of risk management.

In the case of the Alanya fishery (Turkey) studied by Berkes (1986), the devising of a rotating system of access to the sea is of much more recent origin than that found in Sri Lanka. Here, indeed, it is only in the early 1970s that the local fishermen began experimenting, on a trialand-error basis, with a mechanism for regulating the use of the available fishing sites in an equitable manner. This institutional innovation was actually motivated by the need to reduce the incidence of conflicts that had become increasingly frequent and violent due to enhanced competition among fishermen for the better sites, as well as by the need to decrease the costs resulting from more difficult search for, and more intense fighting over, a fishing spot. The guiding principles of the system eventually adopted are the following. First, within the area normally used by Alanya fishermen, all suitable fishing locations are named and listed.

Note that the adjacent sites are defined in such a way that no externality is created: the nets set in one spot cannot block the fish that should be available at the adjacent sites. Second, a list of all eligible fishermen is laid down and in the beginning of the fishing season the eligible fishermen meet in the coffee house to draw lots and are thus assigned to a given fishing site except those who draw blanks because the number of candidates exceeds that of the available sites. Third, once the season has started, each fisherman makes a daily move eastward to the next site and, after the middle of the season has been reached, each fisherman moves west so as to give all participants equal opportunities at the stocks that migrate from east to west during the first half and from west to east during the second half of the season. We are also told that Alanya fishermen are so pleased with this system that some of them set up similar arrangements for other activities (for example, in the case of chartering boats to tourists during the summer).

In a lagoon fishery located near SŤte in southern France, mussel fishermen use a similar system. Here, however, despite the scarcity of fishing sites, rights of use are assigned for a full year. If rotation is thus much lower than in southern Sri Sanka or in Alanya, it is because mussels are cultivated in the sea and they have a long crop cycle making a fishing site akin to a land plot. As for the losers in the lottery—the lottery decides not only how the best sites are to be allotted but also who is going to be denied access to the lagoon—as in the Alanya fishery they have to fall back on other activities and wait till the next year's registration process to try their chance again (personal observation of Platteau).

It may be noted that, since there may exist substantial interseasonal variations in the abundance of fish or in the availability of fishing locations (due, for example, to tidal fluctuations), the system of fishing turns may be organized only during the low season when the chance of conflict over water space is great: fish is scarcer and consequently the better spots are more intensely sought after or the suitable fishing spots are fewer (see, e.g., Cordell and McKean, 1986: 94-5; Levieil, 1987: 115). Also worth mentioning is the fact that rotation-based spatial access codes may serve to regulate access to the water among different fishing communities. Thus, for example, in southern Bahia, competitive encounters may easily arise in the intervillage buffer zones if several boats arrive simultaneously, sometimes from different ports, running after a sizeable school of fish whereas the tide offers room and time for only one net cast. To cope with this kind of situation, 'net bosses follow a standard procedure of drawing lots to decide who will cast first. Once an order is established, a tide marker, usually a pole stuck in the bank, dictates a sequence of net-shooting rights. Not more than one tide-level change is allowed each boat. On this basis, captains decide whether to remain' (Cordell and McKean, 1986: 96-7). When conflicts break out, it is usually because a boat does not close its seine and draw in the catch in the specified time, thereby disturbing the queuing pattern. If the next boat in line decides to begin its operations regardless, the two nets can become fouled (ibid.: 97).

Another interesting illustration of rotating access to fish resources between neighbouring communities is found among the Indian river fishermen of California. It is easy to understand why salmon weirs built over the entire width of a river formed a potential cause of serious quarrel among bands inhabiting different parts of that river. It is precisely to avoid such conflicts that an understanding existed for the timely dismantling of weirs, or the opening of portions of them so that upstream tribes could take their shares of the run. None the less, given the extreme importance of the stake involved, this type of understanding is inherently unstable. Yet, in California, legitimization through myth apparently helped to make it sustainable over the long term (McEvoy, 1986: 37-8).

Notice finally that examples of rotation systems to regulate access to scarce resources abound in other sectors than fishing. Thus, in Uttarakhand (Uttar Pradesh, India), we learn that, traditionally, many villages had fuel reserves even on common land measured by the government, which the villagers cut over in regular rotation by common consent (Guha, 1985: 1940). In Nepal, systems of spatial control adopted by some communities in the middle hills of the country have ensured equal access to both nearby and more distant areas to all members of the community (Arnold and Campbell, 1986: 436). In Japan, rotational access was traditionally provided—possibly by lottery—when forest resources showed variability in spatial or temporal productivity. Thus, grass varied enormously in quality according to how thickly it grew in different spots and how much extraneous undesirable plant matter was included. Therefore, since the grass in the fodder commons was collected by all households together and then tied up into equal bundles, a lottery system was used to assign the bundles randomly to the different member households (McKean, 1986: 556-7).

It is also possible to refer to traditional grazing-control systems which aimed at ensuring fair access to the available pastures for different herders or groups of herders. For instance, the dina system, which was established in 1818 by a powerful Fulani leader assisted by a council of eighty territorially based chiefs, served the purpose of regulating the sequence in which various pastures in the Niger river delta (Mali) were grazed along with the grazing rights of different groups. Rules of access to grazing were defined and livestock movements were co-ordinated throughout the delta, in such a way that grazing pressure was fairly evenly distributed (Lawry, 1989a: 4). This being said, the sector for which rotation procedures are best documented is certainly that of irrigation or water distribution. Just to give one example, in Marpha village (Mustang District) in Nepal, an ancient local custom ensured that irrigation water was distributed in an equitable manner over the course of the year among all the villagers. It thus enjoined that the barley crop be watered from the top of the north fields downward; that is, the fields closest to the head received first water. Then, for buckwheat, the second crop cultivated in Marpha, the order was reversed so that the tail-end fields were watered first (Messerschmidt, 1986: 463; for other examples, see Ostrom, 1992; Tang, 1992). As noted above for fishing locations, schemes for rotating water may be enforced only periodically if there is wide variability in the level of the resource availability: thus, in the Erguita mountain tribe of Morocco, in a year with good rains, 'there are no water allotments to distribute, and no rotations to observe' (Mahdi, 1986: 181).