The Nestedness of Forest Dependence: A Conceptual ...nofc.cfs.nrcan.gc.ca/bookstore_pdfs/18724.pdfThe Nestedness of Forest Dependence: A Conceptual Framework and Empirical Exploration

The Nestedness of Forest Dependence: A Conceptual Framework and Empirical Exploration

THOMAS M. BECKLEY

University of Alberta and Canadian Forest Service Edmonton, Alberta, Canada

Previous research on forest dependence has been limited by (1) an incomplete treat­ment of the topic (an overemphasis of timber dependence) and (2) careless specifica­tion of inferences based on data employed. This article presents a conceptual frame­workfor understanding the nested nature of forest dependence, and demonstrates how forest dependence changes when examining progressively smaller units of analysis. North Americans derive a wide variety of goods and services from their forests. A combination of human uses, including timber use, tourism/recreation use, subsistence, other nontimber use, and ecological use, comprises forest dependence at any given unit of analysis. A vailable data from the northeastern United States are used to show how forest dependence differs as one shifts between nested, spatial units of analysis.

Keywords community, forest dependence, New England, timber dependence, units of analysis

Human dependence upon forests, whether at the individual, household, community, or re­gional level, is a multifaceted phenomenon. This is due to the fact that forests provide a diverse stream of benefits to humans. Forests, or their component parts, provide timber and nontimber commodities, recreational experiences, and sustenance to active forest users. Passive human users of forests-individuals who attach cultural value to forests­also derive both economic and noneconomic benefit from the existence of forests.

This article examines the complexity of human dependence on forests and describes the geographically nested nature of forest-dependence at different spatial scales. Nested­ness refers to the fact that forest dependence takes different forms depending upon the unit of analysis chosen. This may seem obvious, yet few scholars who treat forest depen­dence have articulated the important differences between forest dependence at various levels of analysis. Indeed, these differences are often assumed away. The most common example of this involves the routine use of county-level data to draw inferences regarding forest dependence at the community level.

This article contains three sections. The first treats the concept of nestedness in greater detail and illustrates both spatial and temporal dimensions of this phenomenon. The second section presents a typology of forest dependence that outlines the broad spec­trum of humans uses of forests. Accompanying the typology is a conceptual framework for situating studies of forest dependence. Much of the previous literature is imprecise as

Received 8 August 1996; accepted 11 December 1996. This research was initiated while the author was a PhD candidate at the University of Wiscon­

sin-Madison. Institutional support has been provided by the Canadian Forest Service and Depart­ment of Rural Economy at the University of Alberta. The author thanks Dr. Lloyd Irland, Dr. Bill White, Adam Wellstead, Penny Dugay, the Rumford Historical Society, and the anonymous re­viewers for their useful comments and insights.

Address correspondence to Tom Beckley, Northern Forestry Centre, 5320122 Street, Edmon­ton, Alberta, Canada, T6H 3S5.

101 Society and Natural Resources, II: \0 1-120, 1998

Copyright © 1998 Taylor & Francis 0894-1920/98 $12.00 + .00

102 T. M. Beckley

to what dimensions of forest dependence are addressed. The common error in this case is to make claims about forest dependence when the data employed only allow for infer­ences regarding timber dependence. In the third section, empirical data from the north­eastem United States are used to demonstrate the mUltiple dimensions of forest depen­dence at different levels of analysis.

The Nestedness of Forest Dependence

Most people are familiar with Russian and Ukrainian "stacking dolls" or matroshka. At first glance, a matroshka appears to be one doll. However, when pulled apart, the doll re­veals a smaller but similar doll inside. Within that doll, yet another is contained, and within that one, another. Each doll is related to the other. The smaller ones are a part of the larger ones, yet each has unique qualities as well. The phenomenon of forest depen­dence is similar to a matroshka (Figure 1). An individual's use of, and dependence upon, the forest is different but related to the dependence on forests by other members of that person's household. That individual's dependence on the surrounding forest is nested in his or her household. That household's dependence is different but related to the depen­dence upon the forest of the community in which that household is situated. That commu­nity's dependence is different but related to the dependence of the county within which that community is nested. County profiles of forest dependence differ from but are related to state or provincial profiles of forest dependence, and the same relationship holds for states and provinces in relation to suprastate or provincial regions. Each level of analysis is nested in another. Due to the differences in forest dependence at various levels of analysis, macro studies of forest dependence at the regional level may be of little use in understanding the dynamics of forest dependence in a particular community or county. Similarly, one must be careful in generalizing about forest dependence at the state, provincial, or regional level, based on localized data.

This description of nestedness refers to spatial relations with respect to units of analy­sis. Others have described the nested nature of communities of interest within communi­ties of place. Within regions (or counties, or towns, etc.) there exist occupational commu­nities, of loggers, for example (Carroll 1995), individuals with shared class locations related to forest dependence (Brown 1995), and individuals or groups with shared cultural attitudes or value orientations toward forests (Steel et al. 1994). Machlis and Force (1988) suggest that these "communities of interest" are also nested in larger contexts of forest de­pendence. Conceptually, nestedness is more clear in a geographical context. Each of these spatial units of analysis is contained within another, larger geographical scale, just as the individual dolls in a matroshka are enveloped by a larger doll. The remainder of this article treats the spatial or geographic interpretation of nestedness, while recognizing that com­munities of interest are also nested within forest-dependent places.

The nested units of analysis just described lend themselves to different indicators and different methods of measurement. The unfortunate reality with studies of forest depen-

. dence is that data availability and cost are often determining factors in the nature and the scope of research that is done. This is, in part, what leads to misspecification of infer­ences. Scholars often wish to address community stability or other community issues, but find county data to be the least expensive and most readily available. Rather than visiting communities to conduct interviews, collect labor and employment data at this smaller unit of analysis, and make an accurate assessment of individual communities, many re­searchers make a huge cognitive leap and simply assume that county-level data may be used to draw inferences to the diverse communities contained within them.

Nestedness of Forest Dependence 103

� Individual

..

tt� Household

..

�i_"; �1If1l Community

...

� County

..

D State

...

Region

Country

World

Figure 1. Nested units of analysis in the study of forest dependence.

104 T. M. Beckley

Quite contrary to the statement by Overdevest and Green (1995, 115) that "For most purposes, the county is an appropriate political, economic and social unit to analyze com­munity issues," I suggest that, given the nested nature of forest dependence, communities are the most appropriate units for analyses of community issues, and counties are the most appropriate units for analyses of county issues. Harp and Pauley (1993) demonstrate the danger of using county-level data for analysis of communities. I In their study of two counties in Idaho, economic profiles at the county level and input-output models at the community level showed very divergent results. Aggregate data at the county level show two counties with diverse economic bases incorporating tourism, forestry, mining, agri­culture, and government services. However, input-output models showed that a majority of the communities within those counties were highly dependent upon only one of these sectors. Thus, nearly all of the communities are vulnerable to commodity price or natural resource policy changes that affect the given commodity or service they depend upon. County-level analysis, taken alone, masks this community-level vulnerability and is therefore misleading from a policy perspective if community stability is a policy objec­tive. The county-level analysis suggests that little community development or economic diversification need be pursued, when in fact the individual communities are at risk due to their individual single-sector dependencies.

It is possible to provide detailed portraits of forest dependence that utilize different types of data at more than one level of analysis. Brown (1995) does this through de­tailed interviews with working people in a forest-dependent region in Oregon. She de­scribes the regional context and history of forest dependence with secondary and his­torical sources. She combines this analysis with face-to-face interviews to illustrate the unique dependencies of a particular subset of the region's population ("working peo­ple") at the individual and household level. The body of work by Kusel (199Ia), Kusel and Fortmann (1991), and Kusel (199Ib) uses multiple levels of analysis to treat forest dependence in northern California. However, they are careful to draw conclusions about counties from county-level data, and conclusions about communities from com­munity-level data.

A Typology of Forest Dependence

The fact that forests provide diverse benefits to humans has long been recognized by both academics and forest managers. The doctrines of multiple use, integrated resource man­agement, and most recently ecosystem management all incorporate the principle of man­aging forests for a broad range of benefits to a diverse population with many interests. However, studies of forest dependence rarely attempt to treat multiple dependencies. This is due to a number of factors. In some respects, the types of forest dependence outlined in Figure 2 are the purview of different academic disciplines. Timber dependence has, of course, been of interest to many disciplines: sociology, economics, policy sciences, and, from a more utilitarian perspective, forest management. Tourism has been traditionally associated with recreation studies, although the social sciences are taking more interest in the topic as tourism is hailed as a potential savior for declining timber-dependent commu­

nities. Subsistence forest dependence has traditionally been the bailiwick of anthropologi­

cal work (Brody 1982; Tobias and Kay 1994). There is little work at all done on non-tim­ber forest products, although some interesting work is emerging in this area (Hinrichs

1993; Cizek 1993; Schlosser et al. 1995). Some dimensions of forest dependence have been ignored because little data exist on

them. A common assumption is that these activities contribute little to society and the

Extraction Harvesting LogICbip Transport

Nestedness of Forest Dependence

Production Furniture, Specialty Wood Products Lumber. Veneer, Fiber-Board

Pulp andPa r

Forestn' Services Forest Resource Management

Regeneration/Silviculture Restoration

TouriSmlRecreation Wildlife Viewing

Flora Viewing (fall foliage) HikIDglCamping

ResortlDestination Non-Timber Products CottagelSecond Home Extracts, Cones,'..., M�uS�lSlIlIlSJ---�7::;-:��:::::::::::::-----_--..£Eco-�Tourism

Subsistence

Bio-physical Air. Soil and Water Quality Carbon Sequestering

Gathering Food Fuel Craft Material Building Material

Ecological Psycho-cultural Existence and Bequest Values Historical and Spiritual Values

Figure 2. A typology of human uses of forests.

105

economy. This is true for nontimber products and, at least in the United States, subsis­tence forest dependence. However, research has shown that at very small units of analysis (e.g., household or individual), such activities can contribute substantially to overall well­being. Maple sugaring provides an outlet for labor in a traditionally slow time in the New England economy, and also provides revenue that for some is critical for economic sur­vival (Hinrichs 1993). Similarly, hunting (both legally sanctioned hunting and poaching) have been acknowledged to be important for survival of individuals and households near the margins of poverty in remote, rural regions.

Determining the value attributable to forests from activities such as hunting, skiing, wildlife viewing, and other forms of tourism has proved a most intractable problem. Many of these activities involve nonmarket values. Even where market exchanges exist for some of these activities, it is difficult to determine the portion of expenditures on these activities that is attributable to the existence of forests. For example, some portion of a hotel or restaurant's eamings may be due to the fact that these establishments are lo­cated in, or adjacent to, a forested environment. But how much?

Reporting tourism data in the context of forest dependence also assumes that one can accurately determine the portion of hotel, restaurant, and other service industry earn­ings that actually come from tourists, as opposed to locals, business travellers,2 or oth­ers. There is agreement that forests do contribute to tourism. They may do so indi­rectly-downhill skiing facilities are usually in forests, but forests are not a necessary component for that activity. Forests may also contribute to tourism directly. In New England, a poor year for fall foliage can mean millions of dollars in lost revenues to businesses that depend on the tourist trade (Marshall Wiebe, Director of Public Informa­tion, Maine Department of Conservation, personal communication, 31 July, 1995). So while forests are recognized as significant contributors to tourism, it is difficult to quan­tify their importance.

106 T. M. Beckley

When categories of forest use or forest dependence overlap in Figure 2, one is po­tentially confronted with a dilemma regarding how to account for the value of that activ­ity. Recreational hunters often consume their quarry, so this activity has a subsistence as well as a recreational dimension. Many nonindustrial forest commodities are harvested for household use, or commercial use, or both. Forestry services (forest management, planning, consulting, and nonextractive contractors) have generally supported and been associated with industrial uses of forests (timber in Figure 2) and tourism/recreation. Some studies of forest dependence that use employment to measure dependence have in­cluded forestry service jobs (Overdevest and Green 1995); others have not (Lee and Cubbage 1993). In western regions of the United States, where the U.S. Department of Agriculture (USDA)-Forest Service sometimes has a large complement of individuals in small communities, this type of employment might account for a significant portion of forest dependence in that place. However, U.S. Forest Service employees often provide services to both industrial and recreational clients, so it is difficult to determine how much forest service employment should be attributed to timber dependence versus tourism dependence.

Finally, humans depend upon forests indirectly for things such as air and water qual­ity, biodiversity, carbon sequestering, and other "ecological services." People also demonstrate psychocultural needs for forests. Attitudes and policy preferences for preser­vation of forest land reflect such needs. However, it is difficult, if not impossible, for for­est users to express what percent of the nonmarket value of their experience is attribut­able to some specific activity and what percent is attributable to some more abstract psychological benefit or more general ecological service that forests provide.

Most research on forest dependence includes some assessment (quantitative or qualitative) of the degree of dependence. For some studies, measuring the degree of de­pendence is the purpose of the research (Lee and Cubbage 1993; Fletcher et al. 1991); for others it is merely a means toward some other end (Force et al. 1993; Overdevest and Green 1995). However, degrees of dependence vary between the nested scales or units of analysis, and they vary depending upon which type of forest use is being con­sidered. When we combine the concept of nested scales of forest dependence (Figure l ) with the typology of forest dependence (Figure 2), we arrive at a multidimensional ma­trix of forest dependence. Figure 3 graphically' presents a conceptual framework for sit­uating studies of forest dependence. Most studies represent a point within this three-di­mensional matrix. For example, Bliss et al. (1992) show that timber-dependence in the state of Alabama is high and they discuss the consequences of that fact for well-being. Force et al. (1993) demonstrate that timber-dependence in a community is high. Some work assesses multiple types of dependence at one level of analysis-Beckley and Sprenger (1995) examine subsistence and industrial forest-dependence at the commu­nity level. Other work assesses a single type of dependence at multiple levels of analy­sis (Marchak 1983).

I do not mean to suggest that exemplary social science research on forest dependence must address all types of forest dependence at all units of analysis. This would be a daunting prospect for any researcher, even one with an unlimited budget and an army of research assistants. I do believe that researchers should be much more careful in drawing inferences from their data, and in specifying where their study lies on the three-dimen­sional matrix in Figure 3. The figure graphically illustrates the errors of omission and/or misspecification made when researchers discuss regional dependence with data from a single state, or when researchers discuss forest dependence with employment or income data from only industrial, timber-dependent firms.

Nestedness of Forest Dependence

Unit of Ana1�sis Nation

Region

State

County

Community

Household

Individual High

'" .§

- Q) C; '"' Co) Co) Q) Q) ::3 c:: Co) Co) .6'0 Go) .0 .�

- "0 £ e C<:I

£ 0

lU � '"

� .c;; '0 o c:: til Co) .0 � Go)

� ... ::3

Go)] S E til '" e Go) � '" .;:1 � Go) & ·c ,

::3 § 0 � Z

Figure 3. Forest dependence matrix.

Methods for Examining Forest Dependence

107

Degree of Dependence

In a perfect world, researchers could access data of any sort at any level of aggregation. In the real world, only a fraction of the desired data is usually available and there are cer­tainly budgetary and temporal constraints on the generation of new data. Even if all data were available, its meaning and utility vary according to the scale or unit of analysis. For example, it would be difficult to meaningful integrate and analyze the life histories of all 134,000 individuals in northern New England and New York that work the industrial for­est sector, if such a qualitative database of life histories existed. Similarly, statistical, quan­titative research tools are of little use for small units of analysis, such as the household. Certain levels of analysis lend themselves to certain types of data, and vice versa. That is not to suggest that one cannot integrate different types of data in one analysis. Analyses that combine qualitative and quantitative approaches are often the most illuminating.

The following data on forest dependence in the northeastern United States (and sev­eral lower order spatial scales within that region) come from varied sources. Regional, state, and county-level dependence are documented with secondary data from sources such as the U.S. Bureau of the Census, U.S. Fish and Wildlife Service, Maine State Cen­sus of Manufacturers, state employment statistics, and other reports that utilize secondary source material. The data used to document timber dependence in one community in the region are a combination of secondary (employment, tax, population) and primary (inter­view and archival) data.

Not all forms of dependence are documented at each level of analysis. An exhaustive account of all types of forest dependence at all levels of analysis is beyond the scope of this work. The purpose of this section is to describe the diversity of forest dependence within one region, one state, one county, and one community in the United States and to

108 T. M. Beckley

point to data sources and methods for measuring that dependence. While future research will likely continue to focus on specific types of forest use at only one or two scales of

analysis, this section demonstrates empirically the nestedness of forest dependence. This

empirical portrait of forest dependence in one region is used to highlight the importance of precise specificity of data and precise use of terminology. This section also illustrates aspects of forest dependence for which there is a paucity of data and therefore few acade­

mic analyses. These areas will hopefully be the subject of further research.

The Study Areas

The study area in question is a four-state region that contains a 26-million-acre uninter­rupted expanse of spruce/Conifer and mixed hardwood forest. The states are New York, Vermont, New Hampshire, and Maine. Maine is examined in greater detail. Within Maine, the Western Mountain Region and the Rumford Labor Market Area are featured, and a

profile of RumfordlMexico, a community within these jurisdictions, is presented. Qualita­tive data from interviews with residents of RumfordlMexico illustrate perceptions and in­terpretations of forest dependence at the household and individual levels. Differences and similarities between forest dependence at these levels of analysis are highlighted.

Forest Dependence in the Northeastern United States

In the aftermath of the Northern Forest Lands Study (Harper et al. 1990), researchers and policymakers have come to treat contiguous forested land of northern New York and the three northern New England states as a distinct geographical, cultural, and ecological re­gion. Previous work has treated a much larger area of the northeastern United States based on the USDA Forest Service region. Other work has focused on the forests of New England (including Massachusetts, Connecticut, and Rhode Island, but excluding New

York) (Irland 1982). This work refers to the forested region within the four already men­tioned states, or more specifically to the contiguous forest land within those states that comprised the Northern Forest Land Study Area (NFLSA) (Harper et al. 1990).

The Northern Forest has a unique land tenure structure associated with its depen­dence on forest resources. Eighty-four percent of the land in the NFLSA is privately

owned. Thirty-eight percent of all land within the NFLSA is owned by industrial owners, primarily large, vertically integrated forest products firms. Sixty percent of the private land is actively managed for timber and roughly half of the region is held in large blocks

of 5,000 acres or more. Various forest products firms exist in the region, but pulp and paper producers dominate in terms of ownership of land, employment, and value of ex­ports (Harper et al. 1990). At smaller units of analysis (community and county) much of

the data on pulp and paper manufacturers are suppressed for reasons of confidentiality.

However, data may be reported for larger aggregations such as states or regions, and in such cases the dominance of pulp and paper as the primary industrial forest user in the

Northeast is clear. In total, 86,050 persons were employed in wood-based manufacturing

in 1989. When all employment in the paper industry and lumbering and furniture-making employment are added, industrial forest-based employment is raised to 134,100 (North­

east Forest Alliance [NEFA] 1994a). This vast expanse of forest is home to 1 million residents and thus has a relatively

low population density. However, 70 million other North Americans live within a day's drive, and many of these neighbors of the Northern Forest visit the region for recreational

pursuits (Harper et al. 1990). While forest-based manufacturing generates more wealth,

Nestedness of Forest Dependence 109

the recreation and tourism sectors employ more people, roughly 140,600 in 1989 (NEFA 1994a). Authors of the NEFA study calculate tourism expenditures attributable to forests to be $7.2 billion, compared with $14.7 billion in wood-based manufacturing shipments ( l994a). Specific recreational activities that are significant include skiing, snowmobiling, fall foliage tourism, family camping, hunting, and fishing.

At the regional level, substantial data exist on recreational hunting and fishing. Again, how to properly divide this activity between subsistence and recreational uses of forests is problematic. The 1991 National Survey of Fishing, Hunting, and Wildlife-Asso­ciated Recreation produced by the U.S. Fish and Wildlife Service (U.S. Department of the Interior 1993) contains some of the best data on these activities. Unfortunately, data are not aggregated in the same manner as in the Northern Forest Lands Study or the New England Forest Alliance study. Maine, New Hampshire, and Vermont are included as part of New England, while New York is included in the Middle Atlantic States. How­ever, state data are reported separately and can be aggregated to suit researchers' needs. These data include quantitative measures, such as expenditures, as well as data with qual­itative implications, such as participation rates in wildlife-associated recreation. Partici­pation-rates suggest that the Northern Forest provides substantial benefit to residents of the region. Seventy-five percent of Maine, New Hampshire, and Vermont residents par­ticipate in hunting, fishing, or wildlife viewing, compared to a national average of 57%. When New York is included, the average drops to 48%, but that includes millions of New York City residents that have few opportunities to engage in wildlife-related activities. Northern New York participation rates would likely be more similar to the northern New England states.

Nonindustrial forest products in the region include maple sugaring (which produced $21.6 million in sales in 1989), Christmas trees (3 million trees grown by over 3,000 pro­ducers for a value of $46 million), wreaths ($1 l .5 million), mushrooms (1.5 million pounds produced $1.3 million in sales), and other horticultural products such as cedar oil (NEFA 1994a). These numbers should be viewed cautiously. Much of the trade in nontim­ber forest products exists in the informal economy. Therefore, these official statistics may underestimate the value of product or gross amounts of these goods that are harvested.

Forest Dependence in Maine

The state of Maine is nested in the Northern Forest, and is the most timber-dependent state in the region. The pulp and paper industry is and has been a dominant economic force in the state for nearly a century. Natural factors-abundant water and timber re­sources-initially drew the forest products industry to the region. The lack of quality agricultural land, a harsh climate, an underdeveloped infrastructure, and remoteness to markets are factors that have historically discouraged more widespread settlement and economic diversification in the northern sections of the state. The northwestern portion of the state is virtually uninhabited, as are significant blocks of east central Maine.

Maine is the most forested state in the United States. Eighty-nine percent of the land in Maine is in timberland (Birch 1986). Virtually all of Maine's land that is covered in forest is classified as commercial timberland (able to produce a commercially viable crop of 20 cubic feet per year) (Nevel et al. 1985). Ninety-six percent of Maine's timberlands are in private hands. The proportion of timberland owned by forest products manufactur­ers is higher in Maine than in any other state in the region (Nevel et al. 1985). The 15 largest private landowners hold 46% of the land in the state. Seventy-three percent of the entire state of Maine is owned by less than 1 % of landowners (Lapping and Clemenson

110 T. M. Beckley

1983). According to Irland (1982), 49% of Maine land is held by pulp and paper compa­nies and an additional 20% is held by "nonindustrial" owners (family ownerships, undi­vided ownerships, and land management companies) that are primarily in the business of raising trees that supply pulp and saw mills in the United States and Canada. U.S. Forest Service landownership statistics (Birch 1986) confIrm Irland's (1982) estimates.

The forest products fIrms that dominate the land tenure structure of the state are also primary economic players. The forest products sector employs thousands of workers and makes up signifIcant portions of the state's gross product. Paper has been ranked fIrst among the state manufacturing industries every year since 1955 in wages paid and value of product. It has ranked fIrst in employment for manufacturing since 1985. Combined with lumber and wood products manufacturing, the Census of Maine Manufacturers data overwhelmingly confIrm the dominance of forest products in the state's manufacturing economy. Over a third of all manufacturing wages are currently paid by forest products manufacturers. Over 40% of the value of Maine's manufactured products is accounted for and over a quarter of manufacturing employment is provided by forest products fIrms (Maine Department of Labor 1992). Maine's forest-based manufacturers paid out over $660 million in wages and salaries to 24,600 workers in those industries (NEFA 1994b).

Private forest land in Maine, particularly the large industrial holdings in the north, is widely accessible for recreational opportunities ranging from snowmobiling and cross­country skiing to hiking, hunting, fIshing, and camping. It is diffIcult to measure the ex­tent of these activities given their dispersion across a wide area. However, estimates have been made of the economic impact of forest-based tourism and recreation. According to a 1992 study by Davidson-Peterson Associates, $1.7 billion of expenditures in forest-re­lated tourism were made in Maine in 1990 (Davidson-Peterson Associates 1992). These values only account for actual expenditures, and not nonmarket values associated with recreational activities. Work by Boyle et al. (1990) asserts that over $100 million of addi­tional or "surplus" value exists for the activities of hunting and fIshing alone. That is, per­sons partaking in these activities would be willing to pay an additional $100 million in order to continue these activities. This is a broader measure of the value of the forest, but only accounts for two additional activities. The addition of surplus values for noncon­sumptive wildlife viewers, hikers, campers, skiers, and other users would boost the over­all value of Maine's forests considerably.

NEFA (1994b) calculated Maine's forest-based recreation employment to be 25,740 jobs. These workers received a total payroll of $223 million in 1989. While there are 1,140 more jobs in recreation than in wood products, wood products manufacturing gen­erates three times more payroll dollars than forest-based recreation.

Forest Dependence in Maine's Western Mountains

As we explore more localized nested levels of forest dependence, specifIc forms of de­pendence become dominant in certain regions and sites. Such is the case when we focus on the subs tate region of northwestern Maine, and the specific community of RumfordlMexico. This substate region is highly timber dependent. Unfortunately, census data are not very illuminating with respect to forest dependence at the county level for Oxford County, Maine. Data on forest manufacturers are suppressed due to the domi­nance of one large pulp and paper mill in the county. Similarly, recent national land own­ership data do not contain a large enough sample size to say anything with confIdence re­garding forest land ownership in the county (Thomas W. Birch, personal communication, 20 May 1996). Some data are available for a supracounty region and for a labor market

Nestedness of Forest Dependence 111

area at the subcounty level. While there are pockets of tourism/recreation dependence in the western mountains of Maine, in communities like Rangeley and Newry, this sector does not have nearly the importance it does in areas of southern Vermont, in the Adiron­dacks of New York, or in the White Mountains of New Hampshire. Rangeley is and has been a popular four-seasons recreation destination for nearly a century, and Newry is home to one of the fastest growing ski areas in the Northeast. Despite the proximity of these recreation/tourism forest-dependent communities to timber-dependent communities such as RumfordlMexico and Jay, Maine, data from detailed interviews at the community level and nonparticipant observation (the author lived in the area for several months) re­vealed that very little interaction occurs between these places, either culturally or eco­nomically.

The land ownership profile for northwestern Maine (Oxford and Franklin Counties) is reflective of the state as a whole. Birch (1986) organized the data into slightly different categories than state and regional studies, but rough comparisons with these data are still possible. Industrial ownership predominates in this substate region, as it does in the state as a whole. Forty percent of all land in northwestern Maine is under industry control. Ninety-four percent of the land in the two-county area is privately owned, and manage­ment for timber is the dominant use of that private land.

Timber Dependence in the Rumford Labor Market Area

The Rumford labor market area (LMA) consists of 34 townships in the northern portion of Oxford County, only half of which are inhabited. Rumford and Mexico are the largest towns in the LMA, and together constitute half its population. Using state labor market em­ployment data and company records of employment for salaried individuals at the Rumford mill, a forest dependence index was calculated for this area. This index measures only tim­ber dependence (extraction and manufacturing employment). Employment at local ski areas and other forest-related tourism facilities (fishing, hiking, camping) is not included, al­though again, interview respondents emphasized that persons employed in tourism and recreation jobs lived in the communities where the majority of those jobs exist, which are outside the Rumford LMA. Few ski school instructors or outfitters chose to live under the ubiquitous plume of the pulp mill and commute to their jobs 30-40 miles away.

As measured by employment, timber dependence has recently declined in the LMA, but over a third of all jobs in the area remain directly tied to wood extraction and process­ing. Less than 10 years ago, forest products accounted for over 40% of all jobs. Marchak's (1983) assertion that paper-mill employment is more stable than logging is demonstrated by these data (see Table 1). Forty-four percent of lumber and wood products jobs have been lost in the LMA in less than a decade. Vail (1993) and Lansky (1992) document the process of mechanization in logging that has led to this decline.

Timber Dependence in Rumford/Mexico, Maine

In the greater Rumford area, forest dependence is not only concentrated in a single sector (timber), it is concentrated in a single large firm. This has been the caSe for nearly a cen­tury. One way to demonstrate RumfordlMexico's timber dependence is by looking at the relationship between mill employment and community population over time. Employ­ment levels in the local pulp and paper industry have varied considerably in the hundred years that pulp and paper mills have been in operation. As employment levels have risen and fallen, so too has the population in the community. Figure 4 demonstrates the dra-

Table 1 Timber dependence index for the Rumford Labor Market Area

Lumber and wood Other manufacturing Mill salaried Total manufacturing Timber dependence Year products employment" employmentb employment employment indexc

1984 1140 1680 362 7630 .417 1985 1040 1580 329 7110 .415 1986 910 1350 347 7080 .368 1987 890 1500 343 7500 .364

...... 1988 750 1560 387 7550 .357 ;::; 1989 670 1630 406 8100 .334

1990 640 1660 422 8060 .338

Note. The LMA includes 34 minor civil divisions with a total population of 2 1,099, and an average labor force of 7,537 from 1984 to 1990. Sources: Maine Department of Labor, Bureau of Employment Security ( 1984-1990), William Peterson (Human Relations Director, Boise Cascade, Rumford, ME, personal communication, 22 July 1992), and Robert Woodruff (Director of Woodlands Division, Boise Cascade, Rumford, ME, personal communication, 24 June 1992).

a In 1992, Boise Cascade's Woodlands Division (Rumford Office) reported that they had direct contracts with 34 firms that employed 302 individuals. In addition, the company bought wood from an undetermined number of independent contractors.

bFrom 1984 to 1990, between 1,241 and 1,358 hourly workers were employed in the Rumford mill. The average for the period was 1,291. This represents 82% of all manufacturing employment in the area. The remainder of manufacturing employment was in smaller sawmill facilities and specialty wood products.

C The Timber dependence index represents the sum of wood extraction and wood processing jobs divided by the total employment base for the region. The index represents all jobs that directly involve cutting, hauling, and processing wood. Managerial jobs in those industries are included only for the Rumford mill, so the index is a conservative estimate of industrial forest dependence.

Nestedness of Forest Dependence 113

16000

14000 •

12000 c

I 10000

'0 8000 ..

.B 6000 E 4000 ::J z

2000 Mill Employment 0

0 i

0 0 � � 0 0 0 0 � 0 0 � 0

:2 ,.... :8 C;; � C') � � S; � to to Ol Ol Ol - - -

Year

Figure 4. Population and mill employment in RumfordlMexico, ME, 1850-1990. Sources: Popula­tion data are from the Maine Register, and the U.S. Department of Commerce, Bureau of the Cen­

sus. Mill employment data are from the Oxford Log, Board of Trade Journal, and from the files of the Boise Cascade Human Relations Department.

matic increase in population with the arrival of the mills in the late 19th century, and the subsequent decline in community population following significant layoffs in the 1960s.

At one time there were four mills operating in the community. Company documents claim that 91 % of the town of Rumford's payroll was accounted for by forest products manufacturers in 1903 (Leane 1958). The first peak in wood products employment oc­curred in 1920, when four mills related to the pulp and paper industry were operating in­dependently. Two corporate mergers occurred in the 1920s, and one of the merged com­panies failed during the Depression, leaving only one major paper processing facility in the community. As employment levels declined in the 1930s, population growth slowed. In the post-World War II economic boom, orders for paper increased, leading to in­creased employment in the remaining mill. Job numbers at the Oxford Paper Company peaked at 3,250 in the 1950s, but were stable at around 3,000 employees from the late 1940s to the late 1960s. The combined populations of Rumford and Mexico also reached a plateau of roughly 15,000 persons for this 20-year period.

The pulp and paper industry fell on hard times in the 1960s. The Oxford Paper Com­pany was sold and its new owners embarked on an aggressive rationalization strategy. Employment levels in the mill dropped 30% in the 9-year tenure of the new owners (from 2,850 workers in 1%7 to 2,000 workers in 1976), and population decline in the commu­nity followed suit. In 1976, the mill was sold again, this time to Boise Cascade. Employ­ment levels in the mill stabilized under Boise Cascade's ownership, but the continued dramatic reduction in the workforce in timber extraction has led to a continued decline in community population. As employment opportunities in the wood products sector have declined, other forest-related sectors, such as tourism and nontimber forest products, have not made significant gains.

Contemporary employment figures illustrate the high degree of timber dependence in RumfordlMexico and the high degree of dependence on the single forest products facility in the community. A comparison of the number of forest products jobs in the Rumford labor market area with jobs provided by the Boise Cascade's Rumford mill reveals this concentrated industrial structure. In 1992, 302 loggers were directly subcontracted by the mill. These subcontractors worked independently, but all their output was purchased by

114 T. M. Beckley

the mill or marketed by the Boise Cascade's woodlands division. Much of the subcon­tractors' harvest was of mill-owned timber. Additional contractors also supply pulpwood

to Rumford's mill, but not on an exclusive basis. Hourly employment at the mill for 1990 was 1,248. Salaried employees comprised an additional 422 employees (William Peterson, Human Relations Director, Boise Cascade, Rumford, ME, personal communication, 22 July 1992). The total number of jobs directly tied to the Rumford mill is 1,972. This ac­

counts for 72% of all manufacturing forest products jobs in the area, and nearly a quarter of total employment in the entire LMA.

While timber dependence as measured by employment has declined in recent years, other measures of timber dependence have increased, at least in Rumford proper. The contribution that wood-based manufacturers make to Rumford's tax base has increased both in actual dollars and as a percent of overall revenue sources for the town. This study treats Rumford and Mexico as one community for the purposes of this analysis, but the town of Rumford is more timber dependent due to this issue of municipal taxes.3 As illus­trated in Figure 5, the pulp and paper mill (and its associated biomass recovery boiler) ac­count for nearly three-quarters of the assessed valuation of the town of Rumford. The high level of municipal services Rumford residents have enjoyed over the century is di­rectly related to the tax contributions of the mill. The entire mill complex is located in Rumford, including wood yards, pulping operations, paper machines, energy production facilities, and offices for managerial staff. The only facilities that the Boise Cascade oper­ates in Mexico (and thus the only facility it pays taxes on there) are a sludge dump and a small branch of its woodlands department.

100

- 60 c CD 50 u

"-CD Q. 40

30

20

10

0 0 0 0 0 0 0 0 0 0

0> 0 T"" C\I ('I) q- LO (0 1"-. ex> 0> 0> 0> 0> 0> 0> 0> 0> T"" T"" T"" T"" T"" T"" T"" T"" T"" T"" T""

Year

Figure 5. Forest products firms' contribution to the assessed property value of Rumford, ME, 1890-1990. Source: Town of Rumford Tax Assessors Office.

Nestedness of Forest Dependence 115

The significance of the mill's contribution is not lost on area residents. In Rumford,

the tax rate in 1991 was 19 mils (Town of Rumford, ME 1991). Rumford's town budget

for 1992 was $11 million, $8 million of which were contributed by the mill, according to

town officials and salaried mill employees. Mexico, which incidentally receives the bulk of the airborne and water-borne effluent from the mill, relies on its homeowners and

small businesses for its tax base. In 1991, Mexico had a tax rate of 36 mils (Town of Mexico, ME 1991). Mexico has attempted a variety of strategies to "level the playing

field" with respect to local services and taxes. These strategies have ranged from trying to

interest Rumford in a merger to raising (exponentially) the assessed value of the Boise Cascade sludge dump located in Mexico. Mexico has been trying to interest Rumford in a municipal merger since the tum of the century, but to date, neither the merger nor at­

tempted mill tax hikes have been successful. Qualitative data are also useful for demonstrating timber dependence in Rumford!

Mexico. The community conceives of itself as an industrial community, despite its re­mote setting and natural surroundings. Personal and familial identities are closely con­nected to occupational associations with mill work or woods work. Local residents are "addicted" to the wood-based manufacturing in ways that Freudenburg has outlined

(1992). High wages, tradition, lack of other exploitable resources, and entrenched institu­tional structures have limited the community's development options. Most resident can­not envision their community without a paper mill in it.

The bulk of anecdotal evidence to support the claim that the community is heavily dependent on the industrial forest sector comes from contemporary documents and inter­views. However, some historical sources document the community's concern over the de­gree of single-sector dependence. As early as 1907, an editorial from a local newspaper expressed hope for a more diversified local economic base in the future.

The general hope is that some manufacturing enterprise not dependent upon the paper or pulp trade will establish here. This is not because the paper mills are less desirable than other industries, but because a variety of manufactur­ing is considered a benefit, on the same principle that a variety of opinions

among men is beneficial. (Anonymous 1907)

The passage is telling, not only in that it equates democracy with a diversified economy,

but because in the ensuing 89 years, the author's vision has not been realized. In-depth, face-to-face interviews reveal a great deal concerning the depth and nature of

timber dependence in the community. Many individuals stated their perceptions of the com­

munity's dependence directly. The following observation from a senior manager at the mill expressed the historical importance of the mill to the very existence of the community:

This place [the mill] came and actually brought the community with it. It had

more than an ann's length, low profile, corporate citizen's role to play here.

It was the town. It brought people here. It built the community .... It was the employer.

Another mill executive echoed these sentiments and discussed the fact that Rumford was es­sentially a planned, industrial community, built around the manufacture of pulp and paper.

It wasn't just, "put the mill there and ... be productive". It was, "put the

mill there and put the town around it, and have everything focussed upon the

mill" . . . a sort of a planned community.

116 T. M. Beckley

Mill executives were not the only interview respondents to emphasize the importance of

the mill. A local journalist expressed the community's dependence upon the mill-and

some of the ensuing consequences of that dependence.

We have been dependent upon a paper mill for the last seventy five years or so .... Sure as hell they are the life-blood of the community. We would

blow away if it wasn't here, so we need it desperately. They pay well. And

basically the whole town respects and supports them, I think. On top of that,

they cater to the town, and the town responds by giving them just about what­ever they want.. . . The politics of the mill . . . spills over into the commu­nity. We have no zoning here that amounts to a hill of beans. There are no en­vironmental laws here that amount to a hill of beans. They [town government] have never fought the company.

A local banker, born and raised in the community, had the following comments about RumfordlMexico's reliance on its major manufacturing facility and the prospects for the community's future.

Rumford and Mexico and all these surrounding towns are geared to that mill. . . . There is nothing else here but that paper mill. But again too with Boise

Cascade, the mill modernizing, and new technologies, they are having a smaller and smaller and smaller work-force. It is the only game in town. And if they employ fewer and fewer people every year then we are going to have a smaller population. We are not a bedroom community for [regional trade cen­ter]. I mean this is the end of the road, Rumford, Maine.

One individual expressed the difficulty in establishing other successful enterprises in the region. He attributed that difficulty to the development path chosen over one hundred

years ago, and the community's steadfast support of timber over other forest uses.

Boise Cascade owns everything around. They own all the riparian rights, all the rivers, all the way to New Hampshire. They own all the potential building sites, dump sites, everything. What industry could come in? They talk about

it, but what industry could come in? You either work in Boise, or you have your own business, or you're working for one of the school systems ... or you leave town. The undertakers, they do well here. People still die, they [the mill] can't control that!

The same respondent suggested that many people's lives are completely wrapped up in the mill.

A lot of these people, work in the mill and they don't do anything else. It is

their whole life. They go home, eat, sleep and come back to the mill.

This last testimonial underscores the point that certain households and individuals nested

within this timber-dependent community are differentially affected by the operation of the mill. A major strike in 1987 created bitter divisions within the community. When

field work for this study was conducted in 1992, some families were still not on speaking terms. Those households and individuals whose economic livelihoods depend directly on the mill are the most dependent of all. They likely do not care that RurnfordlMexico is

33% timber dependent based on employment. Of concern to them is the fact that they are 100% dependent upon the mill.

Nestedness of Forest Dependence 117

Conclusion

At the regional level, the dominant form of forest dependence in the Northern Forest is

timber dependence. Measures that illustrate this dominance include landownership (acres held by industrial owners), contribution to the regional economy (income), and primary land management objectives (fiber management). Tourism and recreation are certainly important, and appear to be growing in importance from year to year. However, it is diffi­cult to measure the total contribution the forest makes to this sector. The link: between timber dependence and forests is much more direct, and partly explains why social scien­tists focus on this particular human use of forests. What is critical to acknowledge, given the theme of this article, is that for certain subregions, counties, and communities within the Northeast, tourism and recreation are the leading economic sectors. Because different forms of forest dependence are nested in larger units of analysis, one cannot assume that the dominant forest use at one level will be the dominant use in all the subunits of which that geographic scale is comprised. A county profile of forest dependence in the North­east may tell one very little about the social reality in communities within that county, particularly when forest use in individual communities is very focused on skiing, paper manufacturing, hunting and fishing, and so forth.

The dominant form of forest dependence in RumfordlMexico is timber dependence. Many households nested within that community still "live by the whistle," at the local pulp and paper mill. Persons in that community hunt, fish, snowmobile, and ski. Many harvest wild blueberries in late summer. While they use their forest for more than wood fiber, most view their forest and their trees as the raw material for industrial production. Partly as a re­sult of this view, the local society and culture in RumfordlMexico are remarkably different from those of Newry and Rangeley, two tourism-dependent communities that lie within 40 miles of RumfordlMexico in opposite directions. The very different character of these com­munities is masked if one limits oneself to secondary, county-level data. By employing multiple methods at multiple levels of analysis one can demonstrate the nestedness of forest dependence and show differences in forest dependence between the various nested units of analysis. Most researchers are more interested in dealing with one unit of analysis, and one type of forest dependence, at a time. This is perfectly acceptable, providing that the proper inferences are made. For example, certain aspects of RumfordlMexico' s timber dependence are similar to, and comparable to, other aging mill towns in the region-communities like Millinocket and Jay, Maine, Berlin, New Hampshire, and Ticonderoga and Ft. Edward, New York. In other respects, these places are unique. They have distinct histories, and are nested in slightly different county and state contexts of forest dependence. However, few generalizations about forest dependence at higher levels of analysis can be made using data from RumfordlMexico. If one made such generalizations, one would inevitably overstate the importance of the timber sector at the expense of other forest uses.

Elsewhere in North America, nontimber forms of forest dependence may be domi­

nant, even at a regional level. The Lower Liard Valley in the Northwest Territories of Canada and much of the northern boreal forest are best characterized as being subsistence dependent with respect to their forest resources. In the Lower Liard Valley, harvests of wild game and fuel wood are economically significant, while commercial harvests of tim­ber harvests are not (Beckley and Hirsch 1997). In other regions, timber dependence has been historically dominant, but that dominance is being challenged by nontimber users of the forest. This is nowhere more evident than in the Pacific Northwest, where sport and commercial fishers, Native Americans, mushroom pickers, and environmentalists are ex­erting significant pressure on decision makers to practice alternative management regimes.

118 T. M. Beckley

Scholars need to begin taking a broader view of forest dependence. We need better tools to measure forests' contributions to tourism and recreational experiences. We need to exam­ine in greater detail how individuals and households use nontimber forest-based activities as a part of household survival strategies. And since nontimber uses of forests are at the root of increased conflict over forest management, researchers are wisely beginning to examine changes in how society ascribes cultural meaning and value to pristine forested ecosystems.

This discussion of types of forest dependence and units of analysis is not merely of concern to academics. Policymakers also need to start taking a broader view of forest de­pendence. The political power of timber interests is facing a serious challenge by non­commodity users of forests. While it is clear that many are not happy with the status quo, policymakers are struggling to learn just what it is the public want from their forests, be they local or national, private or public. The framework presented in this article suggests that there are huge gaps in our knowledge regarding different forms of forest dependence at particular units of analysis. Future research should begin to address some of these areas. Such research could help forest decision makers and managers take a broader view; to recognize, at the same time, the diversity of human uses of forests. They must then de­termine how policy and practice may best provide benefits to all human uses and users, while not compromising the ecological integrity of the forest in question.

Notes

1. The use of county data to make community inferences is more problematic in some re­gions of the United States than others. In the Southeast and Mid-Atlantic states, counties are small

and researchers may be safe drawing community inferences from county data. Other regions, such as northern New England, the northern Great Plains, and virtually all of the West, are characterized

by large counties, many of which contain several communities with unique dependencies on their natural resource endowments.

2. A full and total accounting of forest dependence would attribute business travelers' ex­penditures at hotels and restaurants to industrial forest dependence and tourists' expenditures at the same establishments to tourism/recreational forest dependence. Again, the data do not exist to sup­

port such a fine-grained analysis. 3. Rumford and Mexico share fire and police services, a school system, Chamber of Com­

merce, and social service clubs, as well as a shared culture and industrial heritage that is unique among the rural communities in the region.

References

Anonymous. 1907. Rumford Citizen, 3 August 1907. Beckley, T. M., and B. Hirsch. 1997. Subsistence and non-industrial forest use in the Lower Liard

Valley, Northwest Territories, Canada. Information Report. NOR-X-352. Edmonton, Alberta: Natural Resources Canada-Canadian Forest Service.

Beckley, T. M., and A. Sprenger. 1995. Social, Political and Cultural Dimensions of Forest-Depen­dence: The Communities of the Lower Winnipeg Basin. RDI Report Series 1995-1 . Brandon, Manitoba: Rural Development Institute.

Birch, T. W. 1986. Forest-Land Owners of Maine, 1982. Research Bulletin NE-90. Broornhall, PA: Northeast Forest Experiment Station, U.S. Forest Service, USDA.

Bliss, J. c., C. Bailey, G. Howze, and L. Teeter. 1992. Timber dependency in the American South. Paper presented at the 8th World Congress for Rural Sociology, Penn State University, Uni­

versity Park, PA, 1 1-16 August. Boyle, K. J., S. D. Railing, M. Tesl, and M. L. Phillips. 1990. Study of Impacts of Game and Non­

Game Species on Maine's Economy. Staff paper 423. Orono, ME: Department of Agricultural and Resource Economics, University of Maine.

Nestedness of Forest Dependence 119

Brody, H. 1982. Maps and dreams: Indians and the British Columbia frontier. Vancouver, BC: Douglas and McIntyre.

Brown, B . A. 1995. In timber country: Wolting people's stories of environmental conflict and

urban flight. Philadelphia. PA: Temple University Press. Carroll, M. S. 1995. Community and the Northwestern logger: Continuities and changes in the era

of the spotted owl. Boulder, CO: Westview Press. Cizek, P. 1993. Guardians of Manomin: Aboriginal self-management of wild rice harvesting. Alter­

natives 19(3):29-32.

Davidson-Peterson Associates. 1992. The Economic Impact of Expenditures by Tourists on Maine: Calendar Year 1991. York, ME: Davidson-Peterson Associates.

Fletcher, S., W. White, W. Phillips, and L. Constantino. 1991. An Economic Analysis of Canadian Prairie Provinces' Forest Dependent Communities. Project Report No. 91-05. Edmonton, Al­berta: Department of Rural Economy, University of Alberta.

Force, J. E., G. E. Machlis, L. Zhang, and A. Kearney. 1993. The relationship between timber pro­duction, local historical events, and community social change: A quantitative case study. For­

est Sci. 39(4):722-742. Freudenburg, W. F. 1992. Addictive economies: Extractive industries and vulnerable localities in a

changing world economy. Rural Sociol. 57(3):305-332.

Harp, A. J., and T. Pauley. 1993. A Social Profile of Seven Communities in Custer and Lemhi Counties, Idaho. AER 93-13. Moscow, ID: Department of Agricultural Economics and Rural Sociology.

Harper, S. C., L. L. Falk, and E. W. Rankin. 1990. The Northern Forest Lands Study of New Eng­

land and New York. Rutland, VT: U.S. Forest Service. Hinrichs, C. 1993. Sideline or lifeline?: Social constructions of maple syrup enterprise in Vermont

and Quebec. Paper presented at the 56th Annual Meeting of the Rural Sociological Society, Orlando, FL, 7-10 August.

Irland, L. C. 1982. Wildlands and woodlots: The story of New England's forests. Hanover, NH: University Press of New England.

Kusel, J. 1991a. A new approach to the study of forest community well-being. Paper presented at the Annual Meeting of the Rural Sociological Society. Columbus, OH, 1 8 August.

Kusel, J. 1991b. Well-being in forest-dependent communities, vol. II. Berkeley, CA: Forest and Rangeland Resources Assessment Program and California Department of Forestry and Fire Protection.

Kusel, J., and L. Fortmann. 1991. Well-being in forest-dependent communities, vol. I. Berkeley, CA: Forest and Rangeland Resources AssesslUent Program and California Department of Forestry and Fire Protection.

Lansky, M. 1992. Beyond the Beauty Strip: Saving what's left of our forests. Gardiner, ME: Tilbery House.

Lapping, M. B., and H. A. Clemenson. 1983. The tenure factor in rural land management: A New England perspective. Landscape Planning 10:255-266.

Leane, J. 1 958. The Oxford Paper story: A history of the Oxford Paper Company, 1847-1958. Rumford, ME: Oxford Paper Company.

Lee, K. J., and F. W. Cubbage. 1993. Timber-dependency in Georgia. Paper presented at the 56th Annual Meeting of the Rural Sociological Society. Orlando, FL, 7-10 August.

Machlis, G. E., and J. E. Force. 1988. Community stability and timber-dependent communities. Rural Sociology 53(2):220-234.

Maine Department of Labor, Bureau of Employment Security. 1984-1990. Nonfarm wage and salary employment, Rumford LMA, by month and annual average. Augusta. ME.

Marchak, P. 1983. Green gold: The forest industry in British Colombia. Vancouver: University of British Columbia Press.

Nevel, R. L., P. R. Lammert, and R. H. Widmann. 1985. Maine Timber Industries: A Periodic As­sessment of Timber Output. Resource Bulletin NE-83. Broornhall, PA: Northeast Forest Ex­periment Station, U.S. Forest Service, USDA.

120 T. M. Beckley

Northeast Forest Alliance. 1994a. The Economic Importance of the Northeast Forest. Saranac Lake, NY: Northeast Forest Alliance, Inc.

Northeast Forest Alliance. 1994b. The Economic Importance of Maine's Forest. Saranac Lake, NY:

Northeast Forest Alliance, Inc.

Overdevest, C., and G. P. Green. 1995. Forest-dependence and community well-being: A seg­

mented market approach. Soc. Natural Resources 8: 1 1 1-13 1.

Schlosser, W. E., K. A. Blatner, E. G. Schuster, and M. S. Carroll. 1995. Potential for expansion of

the special forest products industry in the northern Rockies. Western 1. Appl. Forestry 10(4): 138-143.

Steel, B. S., P. List, and B. Shindler. 1994. Conflicting values about federal forests: A comparison

of national and Oregon publics. Soc. Natural Resources 7: 137-153. Tobias, T. N., and J. J. Kay. 1994. The bush harvest in Pinehouse, Saskatchewan, Canada. Arctic

47(3):207-22 1. Town of Mexico, ME. 1991. Annual Report. Town of Rumford, ME. 1991. Annual Report. U.S. Department of the Interior, Fish and Wildlife Service and U.S. Department of Commerce, Bu­

reau of the Census. 1993. 1991 National survey of fishing, hunting, and wildlife-associated recreation. Washington, DC: U.S. Government Printing Office.

Vail, D. J. 1 993. The internal conflict: Contract logging, chainsaws, and clear-cuts in Maine

forestry. In Who WiU Save the Forests: Knowledge, Power and Environmental Destruction,

eds. T. Banuri and F. Apffel Marglin, pp. 142-189. London: Zed Books.