Structure and function of wildfire and mountain pine beetle forest boundaries

Landscape boundaries, in particular those created by natural disturbances, are fundamental structures in landscape functioning. At the stand scale, forest boundaries show a wide range of characteristics varying from sharp to diffuse, wavy to straight, enclosed areas (patches) to open ones; all of these have different effects of forest vegetation. The objectives of this study were 1) to characterize the structure of forest boundaries (width, steepness, heterogeneity) and 2) to investigate forest vegetation functional response in the boundary zones created by wildfires and mountain pine beetle (MPB) outbreaks. We studied 11 mountain pine beetle and 7 wildfire boundary zones (4–7 y since disturbance) using field‐based two‐dimensional grids in southern central British Columbia (Canada). Boundary zone delineation was determined through spatially constrained clustering algorithms at three sampling unit resolutions (5×5 m, 10×10 m, and 20×20 m). Then, to characterize boundary width and shape, we developed two new boundary indexes (maximized and minimized boundary zones). The identified boundary widths ranged from 0 to 127 m with a mean width of 51 m at the 20×20 m resolution for both fire and MPB. Although the widths were comparable between disturbance types, fires generally had steeper boundaries (more pronounced) than MPB, largely due to higher peak tree mortality within the disturbances (89 vs 43%). Most of the forest vegetation response variables (understory vegetation diversity, plot‐level species richness, evenness, and multivariate community measures) in the boundary zones tended not to be intermediate in value between the intact forest and disturbance area, in spite of intermediate tree mortality. Tree mortality heterogeneity in the boundary zones was often highest in fires and was equal to the internal heterogeneity in MPB disturbances. Using historical natural disturbance patterns as a conservation strategy, this study proposes that forest management should create boundary complexity (width and shape), thereby creating landscape heterogeneity similar to landscapes influenced primarily by natural disturbances.     

Analyzing landscape diversity in time: The use of Rènyi's generalized entropy function

This paper analyses the use of Rènyi's generalized entropy function to identify landscape temporal transformation. In order to demonstrate its application we analyze recent historic landscape diversity changes focusing on the evolution of boundaries between patches in the town of Isernia (central Italy).Firstly, three 1:25,000 land cover/vegetation raster maps (1954, 1981 and 1992) were selected. Next, changes in landscape diversity were measured analyzing the variation in the extension of boundary types between adjacent land cover categories (patches) from 1954 to 1992. To do this, Rènyi's diversity curves for each year were built and compared.The major advantage in applying the Rènyi generalized parametric diversity function to analyze landscape changes in time is that diversity profiles do not display a single index. In fact, a family of indexes is shown, many of which are currently applied and widely used in landscape ecology.     

What is the role of local landscape structure in the vegetation composition of field boundaries?

Question: How distinct is the flora of field boundaries? How does the structure of field boundaries determine the composition of vegetation? Location: Estonia, six 4 km × 4 km agricultural areas. Methods: We studied the vegetation of fields and field boundaries using 2 m × 2 m sample plots. We estimated the frequency of species in both habitat types, applied an MRPP test to analyse the vegetation composition of field boundaries with various combinations of landscape features (ditches, roads, tree and bush layers) illustrating this by DCA ordination, and used indicator species analysis to determine the characteristic species of each boundary type. Results: Ca. 45% of the flora of field boundaries comprised species found on agricultural land. Most typical species in fields — agrotolerants — were also the most common in field boundaries. The vegetation of road verges and grassy boundaries consisted mainly of disturbance‐tolerant species. Woody boundaries were characterised by shade‐tolerant and nitrophilous species. Ditch banks included species typical of moist habitats and semi‐natural grasslands. Few threatened or protected species were observed. Conclusion: The vegetation composition of field boundaries varied due to the complex effects of landscape structure around and in these boundaries. Plant species in agricultural landscapes can be classified into two broad emergent groups on the basis of their different responses to agricultural disturbances — agrotolerant species and nature‐value species. Agrotolerant species are promoted by agriculture, nature‐value species include rare weeds and habitat specialists. We suggest that high‐nature‐value species should prevail in monitoring the effects of land‐use intensification on biodiversity rather than total species richness.     

Surface pollen-vegetation relationships on the Atlantic seaboard: South Uist, Scotland

Studies of modern pollen rain from remote islands have raised a number of interesting issues concerning the spatial precision of present‐day pollen spectra in relation to their parent plant community types. This paper examines the relationships and degree of correlation between a sequence of contemporary vegetation types, environments and their associated surface pollen spectra from a transect across the island of South Uist in the Outer Hebrides of Scotland. Paired data on both contemporary vegetation and associated surface pollen assemblages have been collected and analysed using methods of numerical classification and ordination. In general terms, the modern pollen rain on South Uist reflects the major changes in vegetation pattern and the major community types fairly closely. The major boundary between the alkaline machair sand dune communities and the various acidic upland vegetation types is particularly clear. However, both variability in the vegetation and the effects of the strong prevailing westerly and south‐westerly winds tend to blur the boundaries of the various communities within each of these larger categories. On average 86.6% of the palynomorphs come from in‐community quadrat sources, while only 1.3% are from off‐island sources. The limited present‐day tree distribution on the transect is discussed in the context of the more widespread distribution of arboreal pollen. Overall, there is a strong numerical correspondence between vegetation, pollen and environmental variables. The various problems inherent in examining surface pollen spectra are reviewed.     

Mapping a forest mosaic – A comparison of vegetation and bird distributions using geographic boundary analysis

Many areas of ecological inquiry require the ability to detect and characterize change in ecological variables across both space and time. The purpose of this study was to investigate ways in which geographic boundary analysis techniques could be used to characterize the pattern of change over space in plant distributions in a forested wetland mosaic. With vegetation maps created using spatially constrained clustering and difference boundary delineation, we examined similarities between the identified boundaries in plant distributions and the occurrence of six species of songbirds. We found that vegetation boundaries were significantly cohesive, suggesting one or more crisp vegetation transition zones exist in the study site. Smaller, less cohesive boundary areas also provided important information about patterns of treefall gaps and dense patches of understory within the study area. Boundaries for songbird abundance were not cohesive, and bird and vegetation difference boundaries did not show significant overlap. However, bird boundaries did overlap significantly with vegetation cluster boundaries. Vegetation clusters delineated using constrained clustering techniques have the potential to be very useful for stratifying bird abundance data collected in different sections of the study site, which could be used to improve the efficiency of monitoring efforts for rare bird species.     

Soil and vegetation differences across ecological boundaries in an arid South African ecosystem

Boundaries are the most reactive nodes in landscapes and may be hypersensitive to global change in climate and land use. Investigations on how soils govern vegetation boundaries are scant, particularly in arid and semiarid ecosystems. The Tankwa Karoo National Park (TKNP) is a unique arid biodiversity hotspot with an unrivalled aridity gradient from < 100 mm MAP to about 700 mm in < 10 km. We investigated the abruptness of four soil‐vegetation boundaries separating eight communities. Two 50 m transects were established across four boundaries for 24 descending point transects, in which the cover‐abundance of each plant encounter at 1 m intervals was recorded. In addition, three soil samples were collected from the top 5 cm in each of the four boundaries and twelve patches. Soil and vegetation parameters altogether indicated three boundary syndromes that were context dependent: (a) a sharp boundary, (b) a gradual boundary or (c) no boundary exists. Soil respiration recorded here, and perhaps other ecosystem processes, was mediated by the soil‐vegetation boundaries. These nodes should be the focus of ecological studies since they reveal much more than the constituent patches themselves.     

Vegetation patterns,microtopography, and soils on a Chihuahuan desert playa

Abstract. Spatial patterns of vegetation on an ephemeral lake bed (playa) in the northern Chihuahuan desert were closely correlated with elevation. Distributions of species along a 265 m transect from the center to the edge of the playa showed abrupt boundaries at 110 m and 220 m. These boundaries seem related to the maximum elevation covered with standing water following complete flooding, and to an edaphic boundary between heavy clay soils at the center of the playa and adjacent soils. The complete turnover in species composition across this transect, with an elevation change of only 85 cm, indicates that environmental gradients resulting from flooding and soil development are steep. Vegetative patterns within the lowest part of the playa were patchy, with patch diameters ranging from 2 to 5 m. These patches appear to be produced by differences in the duration and frequency of flooding between small knolls and depressions resulting from gilgai microtopography.     

Depth of edge influence of the agricultural-forest landscape boundary,Southwestern China

The landscape boundary is an important component of a landscape and often plays an indispensable role in regulating ecological flows. The primary objective of this study was to estimate how far the edge effects on agricultural-forest landscape boundaries can penetrate into the forest and agricultural field. This will serve as a basis for understanding the interaction between forest and agricultural fields in the mountainous area of southwestern China and provide scientific basis for the practice of the policy of “returning agricultural field to forest.” Based on field investigations, three types of boundaries with six sampling transects were selected. We investigated the soil moisture, soil nutrients and vegetation diversity along the transect gradient and explored the depth of edge influence (DEI) with moving split-window techniques for analyzing the data. DEI for soil moisture changed with the seasons, ranging from 6 m in the pepper field to 2 m in the forest during the dry season, and from 12 m in the pepper field to 2 m in the forest after heavy rain. DEI based on soil organic matter ranged from 1.5 to 10 m in the forest field, while it was not detected based on other soil nutrient factors. DEI based on vegetation diversity varied from 4 to 26 m and from 10 to 31 m in the forest and agricultural fields, respectively. These results provide the scientific basis for the policy “returning agricultural field to forest”. Based on these field observations, reducing human disturbance and revegetating with natural shrubs and meadows could be more effective for vegetation conservation in terms of soil moisture and soil nutrient content in the arid valley of the Minjiang River.     

Can stable carbon isotopes (δ13C) in soil carbon be used to describe the dynamics of Eucalyptus savanna–rainforest boundaries in the Australian monsoon tropics?

The history of isolated patches of monsoon rainforest within large tracts of Eucalyptus savanna is poorly understood because of the scarcity of reliable palaeoecological records in the Australian monsoon tropics. Elsewhere in the world, the ratio of the stable isotopes ¹³C to ¹²C (δ¹³C) in soil organic matter has shed light on the dynamics of rainforest–savanna boundaries because tropical grasses with the C4 photosynthetic pathway have a distinct δ¹³C signature (–17 to –9‰) compared with that of woody plants with the C3 photosynthetic pathway (–32 to –22‰). In order to determine the magnitude of the variation in δ¹³C, unreplicated soil profiles were sampled beneath different vegetation types on three boundaries between Eucalyptus savanna and rainforest that were both growing on Tertiary age laterite parent material. Replicated (n = 3) soil profiles, which were also derived from Tertiary age laterite, were sampled from beneath: (i) dense stands of African grasses within a frequently burnt Eucalyptus savanna; and within the same long unburnt Eucalyptus savanna, (ii) patches of African and natives grasses and (iii) clumps of Acacia trees. The strongly negative δ¹³C values of soil organic matter derived from the frequently burnt and long unburnt grassy understoreys in the Eucalyptus savannas showed that a considerable amount of the soil carbon was derived from C3 (woody) species despite the presence of a ground layer dominated by C4 grasses. However, a feature of these data was the considerable variability among the three ‘replicate’ profiles. The surface soil samples from beneath three clumps of Acacia trees in the unburnt Eucalyptus savanna had much less variable δ¹³C values and were similar to two of the three monsoon rainforests sampled. The pattern of δ¹³C values from unreplicated soil profiles from different vegetation types across three rainforest boundaries was also very variable and not always obviously related the known disturbance history of the extant vegetation. Given the considerable variability within and between vegetation types with contrasting disturbance histories, it is concluded that the use of carbon stable isotopes to advance understanding of the dynamics of rainforest and Eucalyptus savanna boundaries will require further development, such as determination of the ¹⁴C age and δ¹³C values of different soil carbon fractions.     

Contrasting mechanisms of resilience at mesic and semi-arid boundaries of fynbos,a mega-diverse heathland of South Africa

Biome boundaries are expected to be sensitive to changes in climate and disturbance, because it is here that ecological communities are at environmental, ecological or disturbance limits. Using palaeoecology to study ecosystem dynamics at biome boundaries provides opportunities for understanding ecosystem resilience or sensitivity at ecologically meaningful timescales, and under varying climatic and disturbance conditions.The fynbos biome is a megadiverse Mediterranean type shrubland, found only in South Africa, that is threatened by climate change, land-use change and invasion by alien species. We used palaeoecological records from the semi-arid and mesic boundaries of the fynbos biome to test hypotheses regarding ecosystem resilience over timescales of centuries to millennia. We hypothesised that fynbos would expand at its mesic boundary at the expense of afrotemperate forest under drier and / or more fire prone conditions. In contrast, we hypothesised that at the semi-arid boundary, fynbos would expand at the expense of succulent karoo under wetter and cooler and / or more fire-prone conditions. Contrary to our expectations, the fossil pollen record at both biome boundaries showed remarkable stability at centennial - millennial timescales. To explain our results, we generated new hypotheses exploring possible mechanisms that might confer resilience.At the mesic (temperate) boundary, we suggest that decreased seasonality of rainfall during drier phases favoured fire and fynbos persistence, while in wetter periods, increased seasonality of rainfall resulted in enhanced summer drought stress, inhibiting forest expansion. At this boundary, internal reorganisation from grassy to proteoid fynbos states conferred resilience through resistance. At the succulent karoo boundary, we suggest that increased aridity was offset by less seasonality of rainfall, which enhanced biomass and allowed fire to persist, favouring persistence of fynbos. At this boundary, fynbos sensu stricto retreated during arid phases but recovered during climate amelioration, consistent with resilience through recovery. In both cases, this mega-diverse, disturbance-adapted flora provided a range of traits that enabled fynbos to persist despite environmental perturbation. Our findings agree with general observations that for ecosystems in regions of ample resource availability (i.e. at the mesic boundary), biotic interactions and disturbance tend to become more important in ecosystem dynamics, whereas in regions of scarce resources (in this case water scarcity at the semi-arid boundary) abiotic stress is more important. Our findings contribute to debates over the mechanisms that confer resistance and resilience to environmental change. Understanding and conserving the processes and mechanisms underpinning its resilience will be critical to effective conservation planning.     

Vegetation controls vary across space and spatial scale in a historic grassland‐forest biome boundary

Ecological boundaries are critical landscape regions of transition between adjacent ecological systems. While environmental controls of boundaries may operate in a scale‐dependent manner, multiple‐scale comparisons of vegetation–environment relationships have been characterized for few boundary systems. We used approximately 250 000 point records on the occurrence of woody versus grassland vegetation in conjunction with climatic, topographical, and soils data to evaluate scale effects and spatial heterogeneity in a 650‐km section of the historic prairie–forest biome boundary of Minnesota, USA. We chose this as a model system because of the availability of historical vegetation data, a considerable spatial extent, a sharp ecological transition, and the ability to avoid confounding from more recent anthropogenic land use change. We developed modeling techniques using hierarchical variance partitioning in a spatially‐structured format that allowed us to simultaneously evaluate vegetation–environment relationships across two‐dimensional space (i.e. the prairie‐forest boundary) and across spatial scales (i.e. varying extents). Soils variables displayed the least spatial autocorrelation at shortest lag distances and tended to be the least important predictors of woody vegetation at all spatial extents. Topographical variables displayed greater spatial heterogeneity in regions dominated by forest compared with prairie and were more important at fine‐intermediate spatial scales, highlighting their likely control on fire regimes. An integrated climatic variable (precipitation minus potential evapotranspiration) displayed a trend of increasing spatial variance across the study region and was unambiguously the strongest biome boundary control, although its joint influence with fire was difficult to characterize. Spatially heterogeneous vegetation–environment relationships were observed at all scales, especially at finer scales. Our results suggest that the importance of environmental controls changes smoothly rather than discretely across scales and demonstrate the need to account for spatial non‐stationarity and scale to predict and understand vegetation distribution across ecological boundaries.     

Understory vegetation indicators of anthropogenic disturbance in longleaf pine forests at Fort Benning,Georgia, USA

Environmental indicators for longleaf pine (Pinus palustris) ecosystems need to include some measure of understory vegetation because of its responsiveness to disturbance and management practices. To examine the characteristics of understory species that distinguish between disturbances induced by military traffic, we randomly established transects in four training intensity categories (reference, light, moderate, and heavy) and in an area that had been remediated following intense disturbance at Fort Benning, GA. A total of 134 plant species occurred in these transects with the highest diversity (95 species) in light training areas and the lowest (16 species) in heavily disturbed plots. Forty-seven species were observed in only one of the five disturbance categories. The variability in understory vegetation cover among disturbance types was trimodal ranging from less than 5% cover for heavily disturbed areas to 67% cover for reference, light, and remediated areas. High variability in species diversity and lack of difference in understory cover led us to consider life-form and plant families as indicators of military disturbance. Life-form successfully distinguished between plots based on military disturbances. Species that are Phanerophytes (trees and shrubs) were the most frequent life-form encountered in sites that experienced light infantry training. Therophytes (annuals) were the least common life-form in reference and light training areas. Chamaephytes (plants with their buds slightly above ground) were the least frequent life-form in moderate and remediation sites. Heavy training sites supported no Chamaephytes or Hemicryptophytes (plants with dormant buds at ground level). The heavy, moderate, remediated, and reference sites were all dominated by Cryptophytes (plants with underground buds) possibly because of their ability to withstand both military disturbance and ground fires (the natural disturbance of longleaf pine forests). Analysis of soils collected from each transect revealed that depth of the A layer of soil was significantly higher in reference and light training areas which may explain the life-form distributions. In addition, the diversity of plant families and, in particular, the presence of grasses and composites were indicative of training and remediation history. These results are supported by prior analysis of life-form distribution subsequent to other disturbances and demonstrate the ability of life-form and plant families to distinguish between military disturbances in longleaf pine forests.     

Multivariate analysis of plant communities in the Narok district, Kenya: The influence of environmental factors and human disturbance

Species type in 42 samples was examined using ordination analysis to establish whether the vegetation pattern displays gradation due to environmental factors, and whether the pattern was being altered by human disturbance. The analysis displayed two patterns. First, the ordination on axis 1 demonstrated vegetation gradation from xeric communities on the plains to humid forests on the highland. This pattern is explained on the basis of moisture and nutrient variations. Vegetation vectors based on species presence-absence were correlated with average rainfall, carbon, nitrogen and zinc.The second was separation of disturbed and undisturbed forest and bushland site categories on axis 2 of the ordination space. This pattern suggests that disturbance response is interactive with moisture so that only in moist communities does a significant vegetation response to disturbance occur.     

岷江上游干旱河谷农林边界影响域的研究

对岷江上游农林边界的影响域进行研究,以提高该区管理农田和林地的水平．共调查3种类型农林边界10条样带,采用移动窗口法对植物多样性的数据进行分析,结果表明,当窗口宽度达到6～10时。SED曲线的变化趋向稳定,并且在曲线上有一或两个峰值出现．不同类型边界的影响域是不同的,但均在距边界50m内．各类型边界的影响域多在12～30m之间．6条林地样带只有M2和M6样带林地的影响域被确定,而4条农田样带的影响域均被确定．影响域的大小取决于边界两侧斑块类型和地形以及小气候等因子,但坡向对其影响不大;移动窗口法能有效地刻画边界动态,是一种分析边界简单而有力的工具．这些结果有利于进一步理解干旱河谷区农林间的相互作用．     

Vegetation establishment, succession and microsite frost disturbance on glacier forelands within patterned ground chronosequences

Aim The impact of microscale frost disturbance on vegetation colonization and successionary trends was examined within patterned ground features of Little Ice Age chronosequences. The goal was to investigate and compare vegetation response to micro‐site frost disturbance with that of previous studies done at a coarser landscape scale. Location The study sites occur on Little Ice Age glacier forelands within Jotunheimen, Norway (61°–62° N). The forelands of the glaciers Slettmarkbreen, Styggedalsbreen and Vestre Memurubreen have been well studied providing chronological controls for landscape studies. Sorted patterned ground features are found within the chronosequences, typically declining with frost intensity and disturbance with increasing terrain age. Methods Micro‐plots (8.3 × 8.3 cm) were placed at the inner borders and centres of patterned ground features. Species were identified and per cent species cover and per cent cover of life‐form category were noted. Nonparametric Kruskal–Wallis and Mann–Whitney U‐tests were used to test for differences between percent cover of life‐form categories within patterned ground features as well as to identify thresholds of successional change across the chronosequences. Results Significant relationships between life‐from groups and patterned ground positions of varying ages were deduced using nonparametric statistics. Findings were then used to discuss trends of succession within patterned ground features and across the chronosequences. Vegetation establishment occurs at the border positions of young (< 30 years) patterned ground features. With time and distance from the ice margin, vegetation encroaches inwards toward the disturbed centres. Succession within patterned ground exhibits several stages: (1) bryophytes/crusts and lichens, (2) grasses/sedges and (3) woody shrubs. The occurrence of forbs was sporadic and generally non‐significant. Main conclusions Frost disturbance in patterned ground appears to delay successional trends of vegetation communities when compared with previous studies on ‘stable’ terrain, producing micro‐site lag effects. These small patches of disturbed ground are therefore important regarding vegetation assemblages across the landscape and are unlikely to be detected at the landscape scale.     

How simple grazing rules can lead to persistent boundaries in vegetation communities

Natural landscape boundaries between vegetation communities are dynamically influenced by the selective grazing of herbivores. Here we show how this may be an emergent property of very simple animal decisions, without the need for any sophisticated choice rules etc., using a model based on biased diffusion. Animal grazing intensity is coupled with plant competition, resulting in reaction-diffusion dynamics, from which stable boundaries spontaneously emerge. In the model, animals affect their resources by both consumption and trampling. It is assumed that forage consists of two heterogeneously distributed competing resource species, one that is preferred (grass) over the other (heather) by the animals. The solutions to the resulting system of differential equations for three cases a) optimal foraging, b) random walk foraging and c) taxis-diffusion are presented. Optimal and random foraging gave unrealistic results, but taxis-diffusion accorded well with field observations. Persistent boundaries between patches of near-monoculture vegetation were predicted, with these boundaries drifting in response to overall grazing pressure (grass advancing with increased grazing and vice versa). The reaction-taxis-diffusion model provides the first mathematical explanation for such vegetation mosaic dynamics and the parameters of the model are open to experimental testing.     

A remarkable case of mosaic parapatry in millipedes

The parapatric boundary between Tasmaniosoma compitale Mesibov, 2010 and Tasmaniosoma hickmanorum Mesibov, 2010 (Polydesmida: Dalodesmidae) in northwest Tasmania was mapped in preparation for field studies of parapatry and speciation. Both millipede species can be collected as adults throughout the year, are often abundant in eucalypt forest and tolerate major habitat disturbance. The parapatric boundary between the two species is ca 100 m wide in well-sampled sections and ca 230 km long. It runs from sea level to 600-700 m elevation, crosses most of the river catchments in northwest Tasmania and several major geological boundaries, and one portion of the boundary runs along a steep rainfall gradient. The location of the boundary is estimated here from scattered sample points using a method based on Delaunay triangulation.     

Multi-profile fine-resolution palynological and micro-charcoal analyses at Esklets,North York Moors,UK, with special reference to the Mesolithic-Neolithic transition

Multi-proxy palaeoecological data from two peat profiles at Esklets on the North York Moors upland provide a record of vegetation changes for much of the Holocene. Possible vegetation disturbance in the late Mesolithic and activity in the Neolithic and Bronze Age are recognised. In both profiles fine resolution analyses have been applied to the period leading up to the mid-Holocene Elm Decline which in this upland has been dated to ca. 4,800 bp (uncalibrated ¹⁴C years). Disturbance impacts at the Esklets Elm Decline are low scale, but phases of woodland disturbance, which include cereal (Hordeum)-type pollen, occur in both profiles ca. 5,200 bp, some centuries before the Elm Decline on the North York Moors, but similar to dates for this key palynological horizon in nearby lowland areas. A protocol is presented for the separation of Hordeum (cultivated species) and Glyceria (wild grass) pollen. The Esklets sites record disturbances during the late Mesolithic-Neolithic transition. These pre-Elm Decline disturbance phases represent either early penetration of neolithic cultivator-pastoralists into this upland or the activities of final mesolithic foragers. No neolithic archaeological sites occur nearby, but a ‘Terminal Mesolithic’ flint site dominated by microlith ‘rod’ forms occurs close to the palaeoecological sites. Such rod sites are dated in northern England to the centuries leading up to 5,000 bp and so are contemporary with the disturbance phases that included Hordeum-type pollen at Esklets. The cultural context of these disturbance phases and the role of ‘rod’ microlith sites during the Mesolithic-Neolithic transition require further focused research to clarify all issues relating to this important period.     

Elementary strategies of ethnic boundary making

This article offers a new taxonomy of how actors may change ethnic boundaries. I distinguish between five main strategies: to redraw a boundary by either expanding or limiting the domain of people included in one's own ethnic category; to modify existing boundaries by challenging the hierarchical ordering of ethnic categories, or by changing one's own position within a boundary system, or by emphasizing other, non-ethnic forms of belonging. The taxonomy claims to be exhaustive and accommodates a considerable number of historical and contemporary cases both from the developed and the developing world. It aims at overcoming the fragmentation of the literature along disciplinary and sub-disciplinary lines and prepares the ground for an agency-based comparative model of ethnic boundary making.