Measuring seining strategies and fishing success in the Philippines

Russell, Susan D

Maritime anthropologists often have argued that different fishing strategies or skipper skills partly account for variability in fishing success within a fleet, but statistical support for such strategies has been difficult to acquire. This article analyzes a mixed species, tropical seine fishery in south central Luzon, Philippines where boat size is similar and electronic fish-finding gear, mechanized hauling, and formal navigational training of skippers are absent. We review the qualitative and quantitative evidence for different fishing strategies in this fleet and then examine the degree to which these strategies account for differential fishing success. We suggest that one way to detect the existence of different seining strategies is to measure the locations fished, number of trips and type of species caught per time period. In fisheries where one is concerned to determine the degree to which fishing strategy accounts for variability in catch, we suggest that defining fishing success as mean catch per trip is more useful than conventional definitions of total catch over a season. The advantage of measuring fishing success as mean catch per trip rather than total catch is that it controls for effort, thereby allowing one to discern more clearly the variables that explain fishing strategy. Key words: maritime anthropology, fishing; Philippines, Luzon

The maritime anthropological literature has long held opposed positions about the causes of variation in fishing success within a fleet, and the "skipper effect" debates arise out of the desire to explain the reason why some boats catch far more fish than others. The "skipper effect" refers to the contribution of a skipper's skill to fishing success irrespective of boat size and effort (Durrenberger 1993). One group of scholars argues that a skipper's skill is a significant factor in explaining the variance in catch among boats (e.g., Acheson 1977, 1988; Hilborn and Ledbetter 1985; Jepson et al. 1987; Thorlindsson 1988; White 1992; Bjarnason and Thorlindsson 1993). Another group claims that the variance in catch is largely a function of quantifiable, impersonal variables such as differences in fishing effort and the size of a boat (e.g., Pilsson and Durrenberger 1982; Durrenberger and Palsson 1983; Durrenberger 1993). The latter authors argue that there may be a strong skipper effect in some fisheries and a negligible effect in others, while a social construction that attributes fishing success to a skipper may vary independently from empirical evidence (Palsson and Durrenberger 1984). They complain, however, that "no one has presented any definitive statistical demonstration that skippers do make a difference aside from differences of effort and boats"(Pilsson and Durrenberger 1990:138).

There are many reasons why the statistical measurement of a "skipper effect" has created a charged debate (e.g., Gatewood 1984a; McNabb 1985; White 1992). First, the variables that a captain must take into account in deciding where to fish, when to fish and how to fish can be listed, but the process by which a captain actually makes such decisions has been characterized as "reasonable" rather than rational owing to the multiplicity of variables to consider (Gatewood 1983). Second, the kind of data that lend themselves to statistical analyses is often limited in scope, duration, and largely confined to industrialized settings where the landings of fishing catches are regularly recorded. Hence, to date, almost all of the discussion of a skipper effect has focused on more technologically advanced fisheries in Iceland and North America rather than in the householdoperated fleets in developing countries (Russell and Alexander 1996). Third, White (1989) suggests that fisheries that target mobile species may experience a "fleet effect," wherein the tendency of boats to congregate in an area once another boat has found fish may level catches throughout the fleet, thus disguising the effects of differential skill in finding fish.

Few studies have examined comparatively the different skills and strategies that adhere to different kinds of fishing. Acheson (1988:101) notes that trap placement is an important skill in the lobster fishing industry, while good eyesight on the part of a skipper is significant in many peasant seining fleets that lack fish-finding gear (Kottak 1983). Palsson and Durrenberger (1990) note that a skipper effect may be more likely in places where captains are not subject to rigorous training prior to becoming qualified to operate a boat, while Bjarnason and Thorlindsson (1993:377) argue that a skipper's skill may be easier to detect in fleets composed of small boats rather than larger boats.

Most studies of fishing strategy have directly or indirectly focused on the issue of whether searching for fish independently is more successful than searching for fish with a larger group of boats. In contrast to Barth's (1966) earlier observation, studies by Orbach (1977), Gatewood (1984b), McGoodwin (1989), and White (1989) all suggest that independent fishing is generally less successful or desirable compared to relying and acting on information about where other boats are catching fish. These suggestions, like Barth's (1966) observation, are based on qualitative ethnography rather than a representative sample of where and when skippers actually fish. The few quantitative studies available have been conducted with industrialized, capitalized fishing fleets only in Maine (Acheson 1977), Iceland (Durrenberger and Pllsson 1983, 1986; Thorlindsson 1988; Bjarnason and Thorlindsson 1993) and British Columbia (Hilborn 1985; Hilborn and Ledbetter 1985). These studies provide important insights into the nature of fishing strategies, but complicate comparative analyses by dealing with cases where 1) variability in boat size or the grid system used for locations are so great that they may disguise the existence of intrafleet tactics (Durrenberger and Palsson 1986; Palsson 1988), 2) quotas artificially restrict variation in effort (Gatewood 1984b; Hilborn 1985), or 3) informal territorial rules or established "hook-offs" constrict the range of tactics available in terms of where a skipper decides to fish (e.g., Acheson 1988; Gatewood 1983).

In an earlier paper (Russell and Alexander 1996), we analyzed the folk model of the skipper effect for a Tagalog seining fleet in Batangas, Philippines and compared it with the quantitative evidence of variables affecting variation in total catch among boats. We showed that the folk model identifies a skipper's skill as a significant determinant of variation in catch, while statistically most of the variance in catch size within the fleet is accounted for by the number of trips and the number of crew per boat. In this article, we provide quantitative evidence of four distinct fishing strategies within this fleet, each comprised of a number of different tactics regarding when to fish, where to fish, and what species to pursue. Each strategy demonstrates a significant difference in total catch for the fishing season, mean catch per trip, and average catch size. In contrast to the fishing fleets of industrialized nations, where advances in electronic fish finding gear, the large and variable size of boats, and sophisticated navigational equipment have partly diluted the comparative advantage of traditional fishing skills, the Tagalog fleet represents a case where technological differences are insignificant. Hence, we have a controlled case from which to examine how fishing skill and strategy affect fishing success without the intervention of fishing technology variables.

We argue three main points in this article. First, too much of the literature has focused on the question of whether a skipper's skill is relevant or not, present or absent, and too little research has focused on the specific strategies pursued by skippers within a fleet that may effect greater fishing success. Second, we argue that the definition of fishing success -- usually defined as total catch per boat throughout a specific fishing season - may not yield the best measure for understanding fishing strategies within a fleet. In many household-based fisheries a better measure of fishing success would be mean catch per trip - a definition which mirrors the combined subsistence and commercial nature of such fisheries, where each successful fishing trip results in food for the crew and owner, funds to pay current other household expenses, as well as the income necessary for undertaking future fishing trips. Also, while most studies rely on total catch for a season as a measure of fishing success, we argue that mean catch per trip is a better measure of fishing strategy and skipper skill since it controls for "effort", or number of trips. By controlling for effort, we are then able to identify the degree to which specific "skill" or "strategy" variables account for differential fishing success. Finally, we suggest that the study of fishing strategies should begin with the assumption that strategies arise from a combination of tactics, inclusive of effort (frequency or duration of fishing trips per period of a fishing season), choice of area to search for fish, and selection of species to be pursued. The size of a boat and type of technology employed are factors that may influence strategy in some fisheries, but they also may mask differential skill among skippers or make identification of separate strategies more difficult to detect.

Skill, Strategy, and Success

Acheson (1981:290) argues that "in many fishing societies the kinds of skills necessary for success are very much the same." Pre-eminent among the skills of experienced fishers are ways to find the fish being pursued. He identifies four different kinds of "search" skills: 1) navigational accuracy; 2) good knowledge of the ocean, including its depths, currents, reefs, and bottom types in order to avoid damaging or losing fishing gear; 3) a detailed knowledge of the species of fish being pursued; and 4) the ability to know what other fishers are doing (Acheson 1981:291).

Studies of specific fishing strategies and tactics, as opposed to more general reflections of the kinds of skills needed to be a successful skipper, are mostly ethnographic in the maritime anthropological literature. Many of the discussions of fishing strategies have focused on purse seining, which some have argued exhibits the greatest degree of reciprocal interdependence, teamwork and skill within a crew (e.g., Wadel 1972; Norr and Norr 1978; Gatewood 1985).' One of the earliest (and most contentious) points was made by Barth (1966:10), who argued that the chances of finding herring are no doubt greater if a vessel strikes out on its own rather than staying with the bulk of the fleet. However, some skippers lack the confidence to pursue independent search strategies for fear of inviting criticism from their crews, especially if they fail to find fish. Hence, an "unadaptive" pattern wherein most boats congregate together in small areas of the sea tends to be the norm for the fleet (cf. Heath 1976).

Since that time, several researchers have commented on various patterns of intra-fleet movement and the degree of independence skippers exhibit when looking for fish. Orbach's (1977:77-78) study of the Californian seining fleet identified "hunters," or skippers known to rely on their own experience and hunches for finding fish, and "chasers," or those who rely primarily on radio information from other skippers to make their decisions. Chasers also tend to be skippers with small boats who lack gas money or confidence in their own expertise at finding fish. Orbach emphasizes that hunters and chasers are ideal types and that some skippers use both strategies in different mixes, depending on the situation. He also identifies three different stalking patterns used by skippers, each of which varies according to the general kind of fish being pursued (Orbach 1977:90).

Whether an independent fishing strategy will likely result in a greater catch obviously depends on the type of marine species being pursued. Acheson (1981:286) observes that the sharing of information is most commonly found among skippers who pursue mobile, migratory species so as to track them once they are spotted. In these circumstances generally, it is difficult to understand how an independent strategy would translate into greater fishing success. In contrast, skippers who exploit sedentary species such as clams or lobster often maintain secrecy or spread misinformation about their fishing spots in order to maintain preferential access to a desirable location over a long period of time. Here, an independent fish strategy makes obvious sense. White's (1989) study of Alabama shrimp trawlers suggests that information sharing is most common among inshore fleets, which are characterized by multiplex personal ties between boat owners and crew, in contrast to the more restricted or irrelevant relationships between boat personnel that operate in offshore fleets.

With the exception of Barth (1966), few if any studies of fishing strategy among skippers pursuing mobile species indicate that innovativeness or independence is a goal of most skippers. While not attempting to measure the actual dispersion of a fleet directly, Orbach (1977:109-110) notes that the ideal for most California seining skippers is to "fish an area," meaning a part of the sea where there are congregations of schools of fish (and hence other boats). "Scratching," or finding isolated schools of fish, typically leads to smaller catches but exemplifies an independent search strategy -- usually for porpoise (Orbach 1977:109-110). White (1989) describes a range of different types of strategies in trawl fishing, but argues that most prefer to work in fleets. He also suggests that independent skippers may have larger catches in newly located spots, but that the additional search time required to locate shrimp cancels out any benefits they may gain (White 1989:34).

Studies of fishing fleets in developing countries also note an absence of innovative and independent decision making on the part of skippers as to where to fish. McGoodwin's (1989:149-151) study of Mexican shark hunters observes that the fleet coordinates their boat searching patterns into a random "starburst" pattern when migratory schools of sharks cannot be located. Once one of the boats experiences a good catch, the fleet exhibits a "concentration" pattern, wherein they all focus their effort in one area. He also notes, however, that a few independent skippers remain aloof from the rest of the fleet's coordinated hunting strategy. Barnes (1997:292) found that Indonesian sea mammal fishers also search as a fleet so as to capitalize on 1) the tendencies of whales to gather around an injured member of the pod; and 2) the chances of whales escaping a single boat's harpooning efforts.

Gatewood and Mace (1990:336-337) used computer simulation to determine the optimum threshold strategy for setting on schools of herring. Some Nova Scotian seiners espouse a "hard work" strategy, meaning they like to set their nets on relatively smaller schools. Others prefer a "patience strategy," meaning they like to wait for a few large schools. While threshold strategy makes a substantial difference in terms of a skipper's seasonal catch, many pursue suboptimal strategies owing to other intervening factors. One type of intervening factor that has been suggested in a study of salmon seiners in British Columbia is area specialization. Hilborn and Ledbetter (1985:55-56) speculate that area specialization may well account for why seining catches are consistently higher in one area rather than another.

The most pertinent statistical analyses of fishing search strategies and the independence of skippers are those of Durrenberger and Palsson (1986) and Palsson (1988). Following their earlier analyses of Icelandic skippers, which showed that statistically the size of the catch was mostly a function of boat size and number of fishing trips (Palsson and Durrenberger 1982; Durrenberger and Pglsson 1983), they demonstrate that skippers' fishing tactics are not significantly different within the fleet. Successful skippers visit relatively fewer locations than less successful ones, follow conservative fishing strategies in terms of area diversity, and are no more independent in terms of their likelihood of fishing away from the fleet than less successful skippers (Palsson 1988:21-22).

Coastal Seining in San Andres

The coastal community of San Andres, Batangas, is home to roughly 4,000 people, mostly members of the Tagalog ethnolinguistic group of south central Luzon. Different kinds of fishing characterize this community, but the most important is pukot, or baby purse seining. In this community, there are roughly 50 canoe owners, 62 owners of baby purse seiners, 1 owner of a lift net outfit, and around 500-600 crew members who work on the seining boats. Some canoe owners join the crews of seining boats during the peak season, and otherwise fish on their own during the off season.

Prior to World War II, seine fishing was conducted from non-motorized boats of 15-20 meters in length with two outriggers. Spoehr (1980:28) suggests that the round haul seine was probably a Tagalog innovation that spread into the Visayan islands of the central Philippines by the 1920s or 1930s, and early forms of seine fishing in Batangas were of this type. Crews were composed of 15-25 men and boats were powered by large sails. Large crews were required in order to haul the seine net and to row the oars attached to each outrigger. The seine nets were of two types: large ones (500 m. long by 30 m. wide) for tuna fishing, and smaller nets (350 m. long by 50 m. wide) with closer mesh for other kinds of schooling fish. By the 1960s, most boat owners adopted purse seine nets and 225 horsepower inboard diesel engines but retained their double outriggers in the converted vessels. Otherwise, the only significant changes were a reduction in the number of crew needed to put to sea and an expanded share for the owner of the boat and fishing gear (Russell 1994). The fleet expanded to its present size in the late 1980s, following several years of excellent prices for fish.

These characteristics of the local fishing economy have changed little today. The primary difference is that crew members are more difficult to recruit locally owing to other job opportunities in urban areas nearby. The core crew of many boats today is constituted around relatives of the boat owner, with an average of 13-14 crew per boat during the peak fishing season. Slightly over half of the 59 seining boats active during 1991-92 in San Andres were captained either by the owner or his son.2

For most fishing households in this community, fishing is the primary form of income and a very important source of protein. All crew members receive a small share of fish for consumption whenever there is a catch, with the owner retaining the right to sell the remaining catch to local fish buyers. Boat owners decide on their own whether to pay their crew in cash shares or in fish for their labor on the boat, but roughly 75 percent of all fish caught are sold though the market.

Seine fishing in this area can be characterized as petty commodity fishing - a form of household-based enterprise ownership operating in a market economy (Russell and Poopetch 1990; Russell and Alexander 1996). Almost all boats are owned and operated by households, with both related and unrelated crew hired to enhance household labor resources. The ability to pool household labor and income is one factor which enables petty commodity producers to compete against more heavily capitalized firms, since they can adjust their costs rapidly to cope with fluctuations in prices, wages, and output (Smith 1986).3 While most boat owners are in debt to financiers who also market the bulk of their fish, boats and gear are locally owned and only two boat owners live outside the community. The ownership of a boat and a house distinguish boat owners from most crew members, but otherwise the joint experience of making a living from the sea tends to mute the overt expression of different class identities locally (Russell 1997).

Deciding When, Where, and What Species to Pursue

In contrast to commercial trawler skippers who have navigational training and often are addressed elsewhere in the Philippines by the title of maestro (teacher) or kapitan (captain), seining skippers in this community are more informally acknowledged as katiwala ("manager" or "trusted one"). The role of skipper has evolved over the years from a previous 'foreman' or coordinator role in a largely subsistence form of production (e.g., Palsson and Durrenberger 1983) to one where he now receives a larger share of the boat's earnings in return for making the important decisions at sea in regard to catching and marketing the fish. Tagalog seining boats are not equipped with mechanized gear, fish-finding radar, safety equipment, or navigational aids of any sort. Nor are skippers subject to any formal training other than a form of "hands on" experience. Since the government classifies the boats operated by San Andres skippers as "municipal" (inshore) rather than "commercial" (offshore), they are not required to undertake formal navigational training. Both inshore seiners and offshore fishers, however, operate in a larger commercial economy. When seine boat owners get too old to captain their own boat, they generally prefer to pass this role on to one of their own sons in order to keep the money within the family. This practice does not necessarily mean that skippers are hired on principles other than skill, but simply that sons have greater opportunities if they are so inclined to learn the skills necessary for being successful skippers.

In Iceland, where public explanations of fishing success give prominence to the skipper's role, skippers have a great deal of authority relative to boat owners and are said to deny the owner of a boat any say in how the boat and gear are used (Bjarnason and Thorlindsson 1993:374). This type of "unquestioned" skipper authority is true for boat owners who are also skippers in San Andres, but not true for the sons of boat owners. Boat owners vary in terms of the degree to which they involve themselves in the decisions of their sons as to when the boat should go out fishing. Most boat owners are former skippers and many feel more knowledgeable about fishing than their sons or hired skippers. Also, not all boat owners are wealthy. While they all tend to have nicer houses than their crew members, many also have huge debts. If only a few boats are having success at catching fish at the beginning of the fishing season, some boat owners are reluctant to finance fishing trips until the chances of success improve.

In this sense, commercialized household fishing fleets may be quite different from industrialized fleets where the number of trips made during a fishing season is viewed as a simple outcome of differences in a skipper's motivation to work hard (Durrenberger and Pilsson 1983) or a skipper's skill at handling the boat in rough weather (Bjarnason and Thorlindsson 1993). The number of fishing trips taken by a Batangas seining boat partly reflects a boat owner's confidence in his own or his skipper's skill, and partly reflects his economic position and wealth. While almost all boat owners are able to borrow money from fish merchants for fuel, some are so deeply in debt that they are reluctant to borrow further money for fuel until more boats are catching fish.

This "delaying" strategy is most apparent at the beginning and end of the fishing season, since boat owners who refuse to fund fishing trips until late in the season risk alienating skippers and crew. Skippers in turn find it difficult to attract crew members if their boat owners are perceived as lacking confidence in their abilities or too poor to support a full-time fishing operation. Few hired skippers will stay with a boat owner under these conditions. Finally, boat owners who are in debt to fish merchants cannot wait too long before initiating a fullscale season of fishing, since fish merchants will become angry if the delay continues once other boats are catching fish. Aside from the cases where boat owners may delay the inception of the season, boat owners generally allow their skippers to decide where to fish or what species to pursue. In addition, skippers and boat owners in small neighborhood and kinship-based groups often exchange ideas about what species other boats are catching and where they are catching them. Information about which boats have caught fish and where they have caught them on the previous day is well known to everyone in this small coastal town.

Seining in this area is an opportunistic, multi-species venture, but November to June is the primary season for tuna and mackerel. Besides deciding where to fish, a key decision skippers face revolves around what size school of fish or species of fish to set the net on once they are spotted.4 Schools of fish may be considered too small or moving so erratically that skippers will ignore them. Some species also are considered more difficult to catch than others, with frigate tuna considered to be one of the easier species to catch and skipjack one of the more difficult. A skipper who has not caught anything during a trip often will cast the net on even a very small school so as to acquire enough fish to give the crew food for their day's efforts. A large sized school of fish, regardless of the species, generally will spark an all out effort if the skipper thinks the school is moving in a predictable fashion and that he can position the boat effectively.

During the season for pukot seining, fishing boats ply the waters of Batangas Bay and nearby coastal areas looking for schools of fish. The skipper generally stands on a raised platform in the front of the boat, scanning the surface of the sea. Boats typically set out around 4 or 4:30 a.m. for the fishing grounds around the mouth of Batangas Bay, as just after dawn is considered the premier time for fish to come to the surface. Skippers guide the boat and set the desired speed through hand signals to crew members. Once a school is spotted, the boat tracks the school at a faster speed in an attempt to parallel the movements of the fish. If the skipper determines that the currents and the movements of the school are favorable for a cast, he gives a signal for the boat to pick up speed. He also is in charge of signaling when the crew should drop the net into the water. It is not unusual for seining skippers to recognize another boat that has spotted a school of fish owing to the greater speed suddenly exhibited. If the boat is not too far away, some skippers will move in to try and make a cast on a school that slips through the net of the first boat. These occasions sometimes lead to arguments between skippers, but generally trying to snag a school of fish that has escaped the cast made by the original spotter is expected and acceptable behavior among skippers (Russell 1996).

Batangas fishermen recognize different levels of skill and experience among skippers, as is true for seiners elsewhere in the Philippines (Veloro 1994). Skippers are said to need several different kinds of skill (kasanayan), and each skipper has his own "style" or "technique" (diskarte) for fishing. In order to elicit a rank ordering of skills, 48 active skippers were asked to rank the skills solicited in open-ended interviews from most to least important. The most important skills of a successful skipper are ranked as: 1) knowing where to find schools of fish; 2) knowing how to read the currents and movements of the waves, 3) knowing where and when to set the net; and 4) knowing the geography of the sea bottom.

Figure 1 indicates the range of coastal areas fished by seining skippers in San Andres during the 1991-92 season. Fishers distinguish areas of the sea according to the nearest town on shore. Despite the choices open to skippers as to where to look for fish- a skill they identify as the one most critical to success - most trips made by all boats in the fleet are within the home port of Batangas Bay. Fishers say that their home bay has long been a popular area for tuna and mackerel, since it is deeper than the neighboring bay of Balayan. These days, however, most say it is difficult to know when and where the tuna are likely to appear. Some skippers readily acknowledge that if they hear that other boats are catching fish in another location from where they are fishing or planning to fish, they often feel enticed to head in that direction. Similarly, they say that if there are fish in their home bay, they will not be looking for them elsewhere. Skippers do not articulate specific strategies for searching for fish. They either say they are going to fish "here (there) only" (dito (diyan) lang), meaning just in their home bay, or "go far away" (dumayo). Some skippers say they prefer to fish only inside Batangas Bay rather than look for fish outside the area. This strategy is common among skippers who believe that searching for fish elsewhere is risky, since the movements of fish are unpredictable. They also complain about the cost of what they will spend on overnight trips and extra fuel. These skippers will search the area around the mouth of the bay early in the mornings, and then keep watch for fish as they gradually return to shore via a criss-crossing search pattern - all the while watching to see if other boats are catching anything. If several days go by with no one catching any fish, many of these skippers simply pull up the boat for awhile to change their luck or wait until they hear other skippers have made a catch.

Skippers say that fishing far away requires one to have "strength inside" (malakas ng loob), referring to the greater physical risk associated with fishing in distant waters. They identify the most dangerous waters as the Verde Island Passage and Calaca owing to the swift and cross-cutting currents. Also, they say the riskiest areas for fishing gear are Anilao and Mindoro owing to the presence of vast, submerged areas of coral reef. A skipper who casts in water that is too shallow or close to the reefs may tear his fishing net.

Tagalog skippers deny that an independent search strategy, meaning one that emphasizes fishing away from the fleet, is a tactic that leads to greater catches. Most skippers express amazement when told that this might be a pattern valued in other seining fleets. They say it makes good sense to go where one knows that other boats are catching fish. In the absence of such information, they say they just look for them on their own. In short, the articulated view of fishing strategy is one which tends to emphasize fleet consolidation when fish have been located, and which eschews independence as a goal in itself. Skippers also scoff at the notion that those boats that go out fishing everyday necessarily will catch more fish than other boats. They point out that if the skipper does not have a good technique (diskarte), is experiencing bad luck (malas), or if there really are no fish to be caught (talaga wala ng isda), then making a lot of trips is just wasting fuel.

Fishing Tactics and Strategies

Even though Tagalog skippers do not articulate a welldefined set of strategies for seeking fish, one can discern whether such strategies exist and determine which ones are the most successful through statistical analysis. In order to identify fishing strategies within this fleet, 59 of the 62 seining boats in San Andres were monitored daily over the course of the fishing season from November 1, 1991 to June 13, 1992. For this analysis we eliminated 13 of these boats from the final sample because either they did not fish at all in the peak months of January to May or because they practiced a type of fishing other than pukot. The final sample of 46 boats fished in each month of the peak months. The reason for eliminating these boats from the final sample, as argued by Bjarnason and Thorlindsson (1993:381-82), is that otherwise one confuses differential participation throughout the season with differential intensity of participation during an equal time span.

Information on the amount of fish and specific species caught each day, the prices received, the location where the boat fished, the number of trips made, the type of gear used, and to whom the fish were sold was collected. This information was supplied by the boat owner or his wife, or from hired skippers. Local fish merchants cooperated in this study by providing information on prices and actual weights of different bins of species caught. In cases where whole fish were the unit sold, they provided the average weights of the size of fish sold by each boat owner.

In order to determine the degree to which this fleet exhibits different fishing strategies and the degree to which those strategies are significant in explaining the variance in catch, we first identified the quantifiable variables that might represent relevant tactics used by skippers. We assume that a "strategy" represents a collection of different tactics. Tactics may be added to or dropped from a strategy, and the weight or proportion of different tactics within a strategy may change over time. Following Hilborn (1985:5), we relied on a relatively simple set of questions based on the idea that where to fish, when to fish, and what species to pursue are the key issues that skippers face. These variables parallel the comments of skippers reviewed above.

A further assumption we make is that aggregate seasonal data for a fleet may disguise certain fishing tactics. The problem with looking at the aggregate average behavior of a fleet in terms of understanding skipper strategies is that skippers are not "aiming" for a fleet "average." Instead, skipper strategies are part of a continuum and must be looked at that way; they are not static but dynamic through time. Hence, we give greater attention to "periodization" in our data compared to previous studies. In particular, we rely on the number of trips, the number of different locations fished, and the number of different species caught by a skipper per two week period in order to insert a time dimension into the data. The 1991-1992 fishing season observed in this study spans a total of 17 two-week periods.

Table 1 summarizes the descriptive statistics for the variables used in the analysis of fishing strategies, tactics, and skill. The issue of when to go fishing is represented by two variables. "Periods" refers to the number of two week intervals fished by a skipper during the season; hence, the higher the score on this variable the longer the duration of time fished by each boat. This variable indicates the degree of commitment to an occupation and perseverance insofar as a skipper searches over a longer or shorter period of time. To some degree, it may also reveal differential skill, insofar as skippers who score high on this variable may be more capable or confident of venturing out during rough seas that are caused by strong winds which occur from November to January and sometimes in the early part of the rainy season (June). "Trips per Period" refers to the mean number of trips per two week interval, and represents fishing effort, motivation and hard work. As noted above, a high score on this variable may represent both a skipper's capability of handling a boat in rough seas and his confidence that he may be able to catch fish when others fear failure. The issue of where to fish is represented by the variable "Bays per Period," which refers to the mean number of different fishing locations fished per time interval. A high score on this variable indicates that a skipper tries a diversity of fishing locations. The issue of what species to pursue is represented by the variable "Species per Period," which represents the number of primary and secondary species caught per time interval. Because the distribution of this variable was not normal, we transformed it by taking the log of each value. A high score on this variable indicates confidence and ability to catch different kinds of species and perhaps a willingness to go after smaller schools of fish. It also may indicate a greater knowledge of different species' behavior on the part of skippers.

Since no Tagalog skipper recognizes the sensibility of a strategy of intentionally fishing in a location "where no one has gone the day before," e.g., the measure of independence/ innovation used by Durrenberger and Palsson (1986), we developed a measure of independence/innovation defined as the total number of trips to bays with no reported catch the day before. We calculated this variable for the two peak fishing periods in January when most boats went fishing. This definition of independence/innovation conforms well to local fishers' view that it is smart to fish where others are known to be catching fish. The variable "Independence," then, refers to the number of trips to a location where no one had caught fish the day before, during the two peak fishing periods December 29. 1991 to January 25, 1992. A high score on this variable would indicate that a skipper is inclined to search for fish in places where no one else has caught fish the day before.

We then classified the sample of 46 boats into a limited number of strategy groups based on our analysis of several tactics variables. We employed two multivariate statistical techniques, cluster analysis and discriminant analysis, to assign the skippers to one of several subgroups representing distinct fishing strategies.5 The variables that proved most useful in creating the strategy groups were periods, bays per period, trips per period, and species per period. Cluster analysis is a procedure that classifies entities into an unknown number of groups based on a measure of similarity or mathematical distance among the attributes recorded for each entity.' It is frequently used as an objective technique in developing a numerical taxonomy or typology when the investigator needs to compare simultaneously a large number of attributes on a large number of individuals (Shennan 1988:192). Cluster analysis results in a dendogram that illustrates the articulations among groups of observations. In this case, the entities to be placed in groups are skippers, and the attributes are periods, bays, trips, and species per period. The cluster analysis indicated a solution of four strategy groups.7

We then used discriminant analysis to confirm whether the four strategy groups defined by the cluster analysis were indeed distinct (Baxter 1994).' Discriminant analysis is a descriptive technique that expresses the separation among a presumed known number of groups as a linear combination of variables (Afifi and Clark 1984). Strategy clusters, their mean scores on each variable, and the standardized coefficients of the linear discriminant function are shown in Table 2. The standardized coefficients indicate the relative contribution of each variable to the linear equation. All four are important for defining the separation among groups, especially "periods" and "species per period." Only one boat (less than 2% of the sample) was misclassified by this procedure.

Strategy 1 includes the boats that fished the most periods during the season and who made lots of trips per period. These skippers fish a diversity of bays (although not the highest), and also tend to go after more than one kind of species. These boat skippers and crew are the hardest workers within the fleet. Strategy 2 boats fish a fewer number of periods during the season compared to Strategy 1 boats, and also make fewer trips per period. They tend to fish a lower diversity of locations, but like to set their nets on different kinds of species. This strategy group has the largest number of boats in the fleet. Strategy 3 represents a fairly conservative and specialized group of skippers. They fish fewer periods, tending not to go out fishing unless it is really the peak season and other boats are catching fish. They have a very high number of trips per period and fish the highest diversity of locations per period of any strategy group. They catch only frigate tuna (tulingan), which is the type fish most frequently found in the coastal waters at this time of year.

Strategy 4 represents boats that fish relatively few periods compared to other boats, and also do not have very many trips per two week period. They fish the lowest diversity of locations, preferring to remain in the home bay of Batangas. They also tend to catch only frigate tuna, when they catch anything.

Table 3 illustrates the comparable performance of each strategy group. First, in terms of the percent of total catch of the fleet caught by each strategy group, Strategy I clearly produces the greatest overall productivity (47.4 percent). Yet Strategy 3 boats are also very successful and make a significantly larger percentage of trips outside Batangas Bay. Together, boats following these two strategies are not only significantly smaller than boats used by the rest of the fleet, but they caught 61.7 percent of the total fleet catch. These figures suggest that successful fishers may indeed pursue more than one strategy, either by following an opportunistic one in terms of the kinds of species sought (Strategy 1) or by intensively seeking the most common type of fish in this region (Strategy 3).

Second, Strategy 4 is also a specialized group of skippers insofar as they catch only frigate tuna, but different from Strategy 3 in that they made less than half the number of trips of boats in Strategy 3 and fished primarily in Batangas Bay. In short, Strategy 4 may represent less of a "strategy" as opposed to a supplementary form of income earning or the only strategy available to skippers with limited trip financing.

Third, the average number of crew clearly changes with each strategy group and effectively follows their rankings in terms of catch performance measures. This pattern substantiates our earlier argument (Russell and Alexander 1996) that the number of crew may well serve as a tracer variable of a skipper's skill (or, in this case, strategy) in this region. This pattern has been noted elsewhere as well, as crew members tend to seek employment with the most skilled captains owing to the practice of being paid a share of the catch (Acheson 1981).

While the number of crew per boat is strongly correlated with total catch, mean catch per trip, and the skill rank of a skipper, it is also moderately correlated with the number of different species caught and the number of trips per period (Table 4) [cf. Russell and Alexander 1996:Table 2]. We view this association as a function of the fact that 1) more highly skilled skippers are likely to go after and be successful at catching a larger range of species than less highly skilled skippers; and/or 2) boats with a larger number of crew are more capable of going after schools of fish that boats with a smaller number of crew would avoid. The larger number of crew will make an unsuccessful cast and haul less exhausting for everyone concerned, and hence may increase a skipper's motivation to cast the net on schools of a size or type of species that others may avoid or view as not worth the effort.

Table 4 illustrates the relationships among selected variables through correlation analysis. Fishing strategies reflect differences in skipper motivation and available financing, and wealth is a significant factor in explaining the number of trips that a boat makes in this particular fleet (cf. Russell and Alexander 1996:Table 3). To some degree, we can hypothesize that if wealth determines fishing effort, then one would expect wealth to be correlated with trips, periods fished, and bays per period. If skipper motivation is the key issue determining number of trips, periods, and fishing location, then one would expect that wealth is not correlated with these variables. Our data show that wealth is weakly correlated with bays per period, but moderately correlated with trips, trips per period and periods. Hence, wealth (e.g., the ability to afford fishing trips) strongly affects the actual number of trips and length of the season, and has somewhat less influence on where a boat fishes. This covariation mirrors the points made by Tagalog skippers, insofar as they view differential wealth as a factor which may limit the number of trips, especially at the beginning and end of the season, but is less influential in determining where a boat will fish.

While fishing a diversity of bays is clearly related to fishing success in this fleet, our measure of innovation, "Independence," was not significantly related to strategy type, nor to total catch or average catch per trip. Independence is not even related to the number of different locations fished by a skipper. It is negatively and weakly correlated with a boat owner's debts to fish merchants, however (Spearman's r=-0375; p

If a boat owner's debts to fish merchants are significant in constricting the periods fished throughout the season, the number of trips per period, or the locations fished then one would expect a negative covariation between debt and periods. In our data, these variables were not significantly correlated. This suggests that skippers who are heavily in debt fish more intensely during periods when fish are most likely to be caught, but do not otherwise fish in the periods at the beginning and end of the season or during lulls in the catching power of the fleet. Debt is significantly correlated with Wealth Rank of a boat owner (Spearman's r=-0.462; p

The implications of our correlation analysis of factors influencing choice of fishing strategy are ultimately confounded by the relationship between wealth and debt. Wealthier households are more capable of financing fishing trips and yet fish merchants claim they prefer to loan the largest sums of money to boat owners who enjoy a reputation for being hardworking skippers (or for employing a hard-working skipper). The fact that a high level of fishing effort throughout a season, coupled with the tendency to fish a diversity of bays, characterize the most successful strategy in the fleet suggests that wealth is a more important determinant of fishing strategy than is the amount of debt to fish merchants. High debts to fish merchants appear to act primarily as a conservative deterrent to searching for fish at the beginning and end of the season when the risk of not catching anything is greatest.

Fishing Strategy vs. Fishing Success

In order to determine how fishing strategies are related to fishing success, we elaborated a correlation and regression analysis of two measures of fishing success, total catch and mean catch per trip. In particular, we wished to identify which of the several tactics that define fishing strategies in this fleet actually make a difference in fishing success. In an earlier paper, we analyzed the quantitative variables that explained fishing success, defined as total catch (kgs) over the course of the season (Russell and Alexander 1996:447). A multiple regression model was used to show that 0.78/0.77 (r2/adj. rt) of total catch is explained by the number of trips a boat makes during the season and the number of crew per boat.

In our data, boat size is not a significant variable in explaining total catch, nor does it correlate with other variables. Variables such as the cost of boat and gear, even when depreciated, or size of engine (which is virtually identical throughout the fleet), also fail to correlate with total catch. We also performed a regression analysis for "efficiency,'' defined as mean catch per trip. This model indicates that the number of crew per boat accounts for 0.35/0.34 (rz/adj. r2) of the variance in efficiency. Given that boat size is not correlated with total catch, we suggested that a boat's mean catch per trip and the number of crew both serve as reliable, indirect measures of a skipper's skill in this fleet (Russell and Alexander 1996). We also noted that in order for "efficiency," or the mean catch per trip, to be considered a legitimate measure of a skipper's skill, it is necessary to show that skippers of different levels of "efficiency" also demonstrate different fishing tactics.

We maintain that mean catch per trip is a reasonable, if indirect measure of a skipper's skill. It expresses "success per unit of effort" or the amount of catch while holding effort (trips) constant.'o Variability in mean catch per trip, therefore, describes the dimension of fishing success due to the skipper's skill, separate from effort or differences in gear or technology. We elaborated our existing multiple regression models for efficiency in order to investigate how variability in mean catch per trip is related to different fishing tactics (cf. Durrenberger and PAlsson 1986). Subsequently, we added the tactics variables to the multiple regression model for "Total Catch" to determine whether they would explain any additional variability in the standard measure of fishing success.

Table 4 indicates that the number of locations fished per two week period (Bays per Period), the number of two week periods fished throughout the season (Periods), and the number of different species caught (Species per Period) are all moderately and significantly correlated with Total Catch and Mean Catch Per Trip. The average catch size when there is a catch is also correlated with Bays per period, Trips, and Periods. These correlations together suggest that fishing effort (represented by a high number of trips throughout the entire fishing season); going after different species of fish; and trying a diversity of locations throughout the season on a regular basis are important variables that partially explain larger catches and the sizes of catches relative to the number of trips.

Two separate multiple regression analyses were carried out on the sample of 46 boats for the dependent variables "Total Catch" and "Mean Catch per Trip." The most parsimonious models that account for the greatest proportion of variability in the dependent variable, measured by r2 and adjusted r2, are presented in Table 5 and highlighted in boldface.11 The "best" regression model and the significance of each independent variable was assessed by examining the individual t-tests for each parameter, Ho: fi=0, the standardized regression coefficients (Afifi and Clark 1984:149), and the F-tests for the uniqueness index of each independent variable (Hatcher and Stepanski 1994: 407-415).'2 These statistics are presented in Table 6. Fifty-four percent of the variability in Efficiency (mean catch per trip) is accounted for by the number of crew, bays per period, and species per period (Table 5). The number of periods fished did not significantly improve the model (Table 6).

When the tactics variables were added to the regression model for fishing success, as measured by total seasonal catch, we found that 78% of the variability in catch was accounted for by the number of trips, the number of crew, and the number of bays fished per period. The number of species caught per period did not add significant information to the regression model (Tables 5 and 6). It is interesting to note that when we added our tactics variables into our original multiple regression model for total catch, their explanatory value is minimal. Only the bay in which the skipper fished is significant in the most parsimonious model, and then the overall amount of variance explained by the model is only marginally better than our original model based on trips and number of crew.

However, when the tactics variables are inserted into our multiple regression model of "efficiency," or mean catch per trip, both the bay in which the skipper fished and the number of species caught are significant in explaining the variation in mean catch per trip. This strongly suggests that fishing strategies and tactics do make a difference in fishing success, when one controls for differential effort, and that strategy affects fishing efficiency (defined as mean catch per trip) in this fleet (Table 7). Specifically the number of different bays fished per period and the number of species caught per period represent tactics that significantly influence fishing success. Therefore, skippers whose overall fishing strategies emphasize these skills are more productive and successful. Also, as we showed in an earlier paper (Russell and Alexander 1996), skippers of different levels of efficiency differ significantly among each other in terms of the specific locations where they like to fish.

Conclusion

Our primary purpose in this article is to illustrate a methodology for detecting the existence of different intra-fleet strategies that influence fishing success. We have shown the utility of disaggregating fleets according to changes in tactics that are measurable, including when skippers fish, where they fish, and what species they pursue. Since a skipper's tactics may change throughout a fishing season, we suggest further delimiting the analysis into two week periods to obtain a more refined measure of strategies. Finally, for the Tagalog fleet that we examine, we argue that mean catch per trip is a better definition of fishing success since it mirrors our informants' concerns with the regularity of income from fishing and because it controls for effort, or number of trips. By controlling for effort, we are able to detect more clearly the degree to which strategy accounts for differential fishing success.13

Fisheries management policies are a subject of considerable debate in terms of how best to regulate the harvest of marine resources. Regulatory options include privatization or restrictions of access to a resource; the establishment of aggregate or boat quotas on allowable catch size or total catch; the setting of seasonal limits to participation; or creating barriers to entry into the fishery by raising the cost of fishing through taxation or transferable quotas. Determining which policies are the most "fair" and reasonable in terms of reducing inefficiencies and overharvesting are hampered by a lack of reliable data and disagreement over what data are significant (Smith 1990). Fisheries management models have been accused of conflating the behavior of household economies with that of capitalist firms (Durrenberger 1997), of failing to understand and "manage" fishermen (Hilborn 1985), of ignoring small fishing enterprises while basing policy on larger fishing enterprises (Russell and Poopetch 1989), of favoring larger fishing enterprises at the expense of smaller fishing enterprises (Apostle and Barrett 1992), and of misunderstanding the degree to which stock fluctuations are predictable (Wilson and Kleban 1992; Acheson and Wilson 1996). Furthermore, economists disagree among themselves as to whether smaller, inshore fishing fleets are more efficient or less efficient than vertically integrated corporate firms (e.g., Doeringer and Terkla 1995).

Determining the degree to which variance in fishing success is a function of variance in fishing strategies is critical for forming appropriate policies for preserving a biologically sustainable level of fishing effort (Hilborn 1985; Durrenberger and Pdlsson 1986). In fisheries management, economists and biologists assume that the amount of fish caught is largely a result of fishing effort. Hence, most policies are designed so as to reduce effort in order to restore fish stocks to either economically or biologically sustainable levels (Acheson 1981:300-301). To the degree that a skipper's skill and/or strategy are what account for variance in fishing success, however, policies intended to reduce fishing effort may fail (Hilborn 1985; Durrenberger 1993).

Owing to the changing nature of the management policies enacted by state regimes and the resources and technology used by fishers, it is not surprising that few fishers are able to articulate detailed strategies in interviews. The large number of variables that skippers consider when they formulate specific strategies inhibits more precise exegesis (Gatewood 1983) and contributes to the persistent mystique or "ideology" of the skipper effect (Palsson and Durrenberger 1990). These difficulties underscore the need to shift the debate away from discussion over whether a skipper's contributions are best understood as individualized skill (Thorlindsson 1988) or as the collective outcome of a larger field of crew and social relations (Pilsson 1994). The study of fishing strategy does not require us to subscribe to the notion of an autonomous skipper, isolated from the influences of his crew, technology, environment and the larger array of political economic influences that surround fishing today. What it does require are more well-grounded, comparative, and empirical studies of what fishers actually do.

We argue that a less contentious and more productive way to analyze factors that contribute to differential fishing success is for scholars and policymakers alike to devote increased attention to variations in fishing strategies within and between fleets over time. Skill itself is a multifaceted phenomenon that is resistant to statistical analysis; strategies, however, are multifaceted phenomenon that can be measured and which may well indicate significant variability in fishing success within a fleet. We expect that variables such as the level of technology in use, the range and type of species sought, and whether one is dealing with commercial household fishers or artisanal fishers will produce different findings regarding the degree to which fishing strategies vary within fleets and account for differential fishing success. Certainly, in many contemporary industrialized contexts, a skipper's actions are influenced by the role of large fish processors, corporate owners, or state-enforced quotas on a boat's catch. While these influences may control or collapse differences in fishing effort within a fleet (e.g., Durrenberger 1997), some strategic choices may still be under the control of individual skippers.

Comparative understanding of the role of a skipper's skill or strategy in determining fishing success has been difficult because, while ethnography suggests that skippers pursue different tactics, aggregate statistical analysis collapses tactical variables into gross seasonal measures which tend to obscure the detection of intra-fleet strategies. By looking at aggregate fleet behavior to explain total catch, one gathers a broad general view of the impact of fishing effort, boat size and technology on fishing success. However, it is only by looking at mean catch per trip - a definition of fishing success more compatible with petty commodity fishers' concerns to balance effort, cost, and success - that we are able to detect strategies that vary significantly among Tagalog purse seiners. Mean catch per trip (or other unit of effort) also has long been used as an index of stock abundance and a way of comparing the catching power of boats (Hilborn and Ledbetter 1985:51). Ethno-archaeologists have used similar measures of productivity per unit of time or effort to determine the degree to which skill accounts for differential hunting success (e.g., Kent 1996). The virtue of such a way of measuring success in fishing or hunting is that it controls for effort, thus allowing one to focus on variables that directly or indirectly reflect skill and/or strategy.

NOTES

'Davenport (1960), Forman (1967), Cordell (1974), and Acheson (1977,1988) also have discussed strategies in relation to various kinds of fishing, although not seining.

ZIn our sample of 59 boats, which represent 95 percent of all seine boats in the community, 24 were captained by the boat owner, 10 by sons of the owner, 13 by other close relatives (cousin, nephew, or brother), and only 12 by individuals unrelated to the boat owner. 3Also see Apostle and Barrett (1992) for a recent discussion of the differences between artisanal and commercial fishers. "The primary species caught include Aulis thazard (frigate tuna), Rastrelliger kanagurta (Indian mackerel), Rastrelliger brachysoma (short mackerel), Katsuwonus pelamis (skipjack tuna), Mene maculata (moonfish), and various species of round and big-eyed scad. 5 See, e.g., Afifi and Clark (1984); Aldenderfer and Blashfield (1984); Baxter (1994); Sherman (1988); The Sas Institute (1996). 6See, e.g., Aldenderfer and Blashfield (1984); Manly (1986). 'We used Wards minimum variance method of the SAS Cluster procedure which yielded the clearest distinctions among subgroups in

the data (The SAS Institute 1996). All variables were normally distributed, although `species per period' was transformed by taking the log of each value in order to eliminate outliers. The distance measure is the ANOVA (or error) sum of squares added over all variables. The number of strategy groups was determined by examining the graph of the coefficient of fusion (the numeric value of distance at which clusters are joined) [Aldenderfer and Blashfield 1984] which revealed a four group solution.

sThe SAS discrim procedure was used to conduct Fisher's linear discriminant analysis (The SAS Institute 1996). All variables were normally distributed. Species per period was transformed by taking the log of each value so that it would conform to the normality assumption. The homogeneity of within covariance matrices was tested and found to be not significant at the 0.1 level, therefore a pooled covariance matrix was used in the discriminant function. Scatterplots among pairs of variables also indicte good separation among strategy groups.

9We followed Barbara Grandin's (1988) method of community ranking for ranking skippers by skill and boat owners by wealth. Lower

scores indicate greater wealth than higher scores. See Russell and Alexander (1996) for a full discussion of the procedure.

'oP5lsson and Durrenberger (1982) also recognize the desirability for statistical models of holding effort constant, although their concern is to explain total catch throughout the season rather than mean catch per trip.

"All models presented in Table 5 meet the assumptions required for multiple regression. Scatterplots indicate that linear relationships between the independent and dependent variables are plausible. All variables are normally distributed and free of outliers. A thorough analysis of normality, homogeneity of variance, outliers, residuals, and measures of influence was performed on the models that accounted for the greatest amount of variability in the dependent variable. Trips

per period was not included in the regression model for efficiency, because to do so would violate the assumption of independence between the independent and dependent variables. 1[The individual t-tests for each parameter should indicate that the null hypothesis is rejected, because independent variables with coefficients of zero do not add information to the regression model. Standardized regression coefficients are the parameter estimates that would be obtained if the variables were standardized prior to analysis. They indicate the relative contribution of each variable to the regression plane. The uniqueness index is a measure of the percentage of variation in the dependent variable that is accounted for solely by a single independent variable. It can be tested for significance using the Fstatistic.

"Space limitations prevent us from offering a more complete explanation of the factors that canalize the choice of fishing strategy in this Tagalog fleet. We explore these issues in a future publication.

REFERENCES CITED

Acheson, James M.

1977 Technical Skills and Fishing Success in the Maine Lobster Industry. In Material Culture: Styles, Organization and Dynamics of Technology. Heather Lechtman and Robert Merrill, eds. Pp.lll-138. St. Paul, MN: West Publishing. 1981 Anthropology of Fishing. Annual Review of Anthropology 10:275-316.

1988 The Lobster Gangs of Maine. Hanover, NH: University Press of New England.

Acheson, James M. and James A. Wilson 1996 Order Out of Chaos: the Case for Parametric Fisheries Management. American Anthropologist 98(3):579-594. Afifi, A. A. and V. Clark

1984 Computer-aided Multivariate Analysis. London: Wadsworth, Inc.

Aldenderfer, Mark S. and Roger K. Blashfield 1984 Cluster Analysis. Quantitative Applications in the Social Sciences. London: Sage Publications. Apostle, Richard and Gene Barrett, eds. 1992 Emptying their Nets: Small Capital and Rural Industrialization in the Nova Scotia Fishing Industry. Toronto: University of Toronto Press. Barnes, R. H.

1997 Sea Hunters of Indonesia: Fishers and Weavers of Lamalera. Oxford: Clarendon Press. Barth, Fredrik

1966 Models of Social Organization. Occasional Paper No. 23. London: Royal Anthropological Institute of Great Britain and Ireland.

Baxter, M. J.

1994 Exploratory Multivariate Analysis in Archaeology. Edinburgh: Edinburgh University Press. Bjarnason, Thoroddur and Thorolfur Thorlindsson 1993 In Defense of a Folk Model: the "Skipper Effect" in the Icelandic Cod Fishery. American Anthropologist 95(2):371-394. Cordell, John

1974 The Lunar-tide Fishing Cycle in Northeastern Brazil. Ethnology 13:379-92. Davenport, William H.

1960 Jamaican Fishing: a Game Theory Analysis. Yale University Publications in Anthropology 59:3-11. Doeringer, Peter B. and David G. Terkla 1995 Troubled Waters: Economic Structure, Regulatory Reform, and Fisheries Trade. Toronto: University of Toronto Press. Durrenberger, E. Paul

1993 The Skipper Effect and Folk Models of the Skipper Effect

among Mississippi Shrimpers. Human Organization 52(2):194202.

1997 Fisheries Management Models: Assumptions and Realities or, Why Shrimpers in Mississippi are not Firms. Human Organization 56(2):158-166. Durrenberger, E. Paul and Gisli P5lsson 1983 Riddles of Herring and Rhetorics of Success. Journal of

Anthropological Research 39:323-35.

1986 Finding Fish: The Tactics of Icelandic Skippers. American Ethnologist 13(2):213-229. Forman, Shepard

1967 Cognition and the Catch: the Location of Fishing Spots in a Brazilian Coastal Village. Ethnology 6(4):417-26. Gatewood, John B.

1983 Deciding Where to Fish: the Skipper's Dilemma in Southeast Alaskan Salmon Seining. Coastal Zone Management Journal 10(4):347-367.

1984a Is the "Skipper Effect" Really a False Ideology? American Ethnologist 11(2):378-79.

1984b Cooperation, Competition, and Synergy: Informationsharing Groups Among Southeast Alaskan Salmon Seiners. American Ethnologist 11(2):350-370. 1985 Actions Speak Louder Than Words. In Directions in Cognitive Anthropology. Janet W.D. Dougherty, ed. Pp.199219. Urbana: University of Illinois Press. Gatewood, John B. and Pamela M. Mace

1990 Hard Work versus Patience: Simulation of Fishing Strategies in Nova Scotian Herring Purse Seining. Journal of Quantitative Anthropology 2:329-352. Grandin, Barbara E.

1988 Wealth Ranking in Smallholder Communities: A Field Manual. Rugby, England: Intermediate Technology Publications.

Hatcher, L. and E. J. Stepanski

1994 A Step-by-step Approach to Using the SAS System for Univariate and Multivariate Statistics. Cary, NC: SAS Institute Heath, Anthony F.

1976 Decision Making and Transactional Theory. In Transactional Meaning: Directions in the Anthropological Study of Exchange and Symbolic Behavior. Bruce Kapferer, ed. Pp. 263-80. Philadelphia: Institute for the Study of Human Issues. Hilborn, R.

1985 Fleet Dynamics and Individual Variation: Why Some People Catch More Fish Than Others. Canadian Journal of Fisheries and Aquatic Science 42:2-13. Hilborn, R. and M. Ledbetter

1985 Determinants of Catching Power in the British Columbia Salmon Purse Seine Fleet. Canadian Journal of Fisheries and Aquatic Science 42:51-56.

Jepson, M. E., J. S. Thomas, and M. C. Robbins 1987 The "Skipper Effect" among Gulf Coast Shrimp Fishermen. Sociology and Social Research 72(1):20-24. Kent, Susan

1996 Hunting Variability at a Recently Sedentary Kalahari Village. In Cultural Diversity Among Twentieth-century Foragers: An African Perspective. Susan Kent, ed. Cambridge: Cambridge University Press. Kottak, Conrad

1983 Assault on Paradise: Social Change in a Brazilian Village New York: Random House. Manly, Bryan

1986 Multivariate Statistical Methods, A Primer. London: Chapman and Hall. McGoodwin, James R.

1989 Do Randomizing Devices Aid Marine Hunters? Shark Fishermen in Pacific Mexico. Maritime Anthropological Studies 2(2):134-153.

McNabb, Stephen L.

1985 A Final Comment on Measurement of the "Skipper Effect." American Ethnologist 12(3):543-44. Norr, James L. and Katherine L. Norr 1978 Work Organization in Modern Fishing. Human Organization 37(2):163-171. Orbach, Michael K.

1977 Hunters, Seamen, and Entrepreneurs: the Tuna Seinermen of San Diego. Berkeley: University of California Press. Palsson, Gfsli

1988 Models for Fishing and Models of Success. Maritime Anthropological Studies 1(1):15-28. 1994 Enskilment at Sea. Man (n.s.) 29(4):901-927. Pdlsson, Gfsli and E.Paul Durrenberger

1982 To Dream of Fish: the Causes of Icelandic Skippers' Fishing

Success. Journal of Anthropological Research 38:227-42. 1983 Icelandic Foremen and Skippers: the Structure and Evolution of a Folk Model. American Ethnologist 10(3):511-528. 1984 Reply to Gatewood. American Ethnologist 11(2):378-80. 1990 Systems of Production and Social Discourse: the Skipper Effect Revisited. American Anthropologist 92(1): 130-41. Russell, Susan D.

1994 Institutionalizing Opportunism: Cheating on Baby Purse Seiners in Batangas Bay, Philippines. In Anthropology and Institutional Economics. James M. Acheson, ed. Pp. 87-108. Lanham, MD: Society for Economic Anthropology and University Press of America.

1996 Of Beggars and Thieves: Customary Sharing of the Catch and Informal Sanctions in an Open Access Fishery in the Philippines. Paper presented at the American Anthropological Association meetings, San Francisco.

1997 Class Identity, Leadership Style, and Political Culture in a Tagalog Coastal Community. Pilipinas: Journal of Philippine Studies 28 (in press).

Russell, Susan D. and Rani T. Alexander 1996 The Skipper Effect Debate: Views from a Philippine Fishery. Journal of Anthropological Research 52(4):433-460. Russell, Susan D. and Maritsa Poopetch 1990 Petty Commodity Fishermen in the Inner Gulf of Thailand. Human Organization 49(2):174-187.

Shennan, Stephen

1988 Quantifying Archaeology. Edinburgh: Edinburgh University Press.

Smith, Carol A.

1986 Reconstructing the Elements of Petty Commodity Production. Social Analysis 20:29-46. Smith, M. Estellie

1990 Chaos in Fisheries Management. Maritime Anthropological Studies 3(2):1-13. Spoehr, Alexander

1980 Protein From the Sea: Technological Change in Philippine Capture Fisheries. Ethnology Monograph No. 3. Pittsburgh: Department of Anthropology, University of Pittsburgh. Thorlindsson, Thorolfur

1988 The Skipper Effect in the Icelandic Herring Fishery. Human Organization 47(3):199-212. Veloro, Carmelita E.

1994 Suwerte and Diskarte: Notions of Fishing, Success and Social Relations in Two Palawan Villages. In Fishers of the Visayas: Visayas Maritime Anthropological Studies 1: 19911993. Iwao Ushijima and Cynthia N. Zayas, eds. Pp.133-163. Quezon City: University of the Philippines Press. Wadel, Cato

1972 Capitalization and Ownership: the Persistence of Fishermenownership in the Norwegian Herring Fishery. In North Atlantic Fishermen: Anthropological Essays on Modern Fishing. Raoul Andersen and Cato Wadel, eds. Pp. 104-19. St. John's: Memorial University of Newfoundland. White, David R. M.

1989 Knocking 'Em Dead: Alabama Shrimp Boats and the "Fleet Effect." In Marine Resource Utilization: A Conference on Social Science Issues. J. S. Thomas, L. Maril, and E. P. Durrenberger, eds. Pp.25-37. Mobile: University of South Alabama Publication Services.

1992 Belief in Differential Skill among U.S. Gulf of Mexico Shrimp Fishermen. American Anthropologist 94(2):450-52. Wilson, James A. and Peter Kleban

1992 Practical Implications of Chaos in Fisheries: Ecologically Adapted Management. Maritime Anthropological Studies 5(1):67-75.

Susan D. Russell is an Associate Professor in the Department of Anthropology, Northern Illinois University, DeKalb and Rani T Alexander is an Assistant Professor in the Department of Sociology and Anthropology, New Mexico State University, Las Cruces. Susan Russell's field research for this project was funded by the National Science Foundation (Grant Nos. 9009745 and 9107136) and the Fulbright-Hays Foundation during the summer of 1990 and June 1991 to June 1992. Their support is gratefully acknowledged, as is the sabbatical leave she received from Northern Illinois University.

Copyright Society of Applied Anthropology Summer 1998
Provided by ProQuest Information and Learning Company. All rights Reserved