We know remarkably little about the current status and numbers of lynx in the contiguous United States. However, what little we know indicates that the subpopulations are not large. Until we better determine the current location and sizes of lynx subpopulations, it is unwise to assume that these populations can be reduced or further isolated without sharply increasing the risk of their individual and collective extinction. II, 14

Preliminary telemetry data suggest that lynx in Montana and Wyoming have large home ranges; this result supports the Koehler and Aubry (1994) contention that lynx from southern lynx populations have large spatial-use areas. XI, 1

Because my study period coincided with the increase phase of a hare cycle (C. J. Krebs, personal communication), assuming geographic synchrony, my results are consistent with suggestions that hare densities remain relatively low in southern British Columbia and the northwestern United States (Chitty 1950; Dolbeer and Clark 1975; Wolff 1980)... Annual home ranges of resident lynx in the southern Canadian Rockies were considerably larger than those reported for most other studies employing the same estimators, regardless of prey density or geographic locale (Koehler and Aubry 1994; Poole 1994; Slough and Mowat 1996). XII, 6-7

Southern lynx populations appear to exist at lower densities and to have lower reproductive rates than northern populations (Chapter 13). Habitat heterogeneity is also much higher than in the taiga (Chapter 3); therefore, southern populations likely occur partly in areas which act as "sinks "and are dependent on immigration from "sources." II, 9-10

For most of the islands in a metapopulation to be occupied at given time, the colonization rate must greatly exceed that of local extinction. If subpopulations on islands are extinction-prone, then a stable metapopulation requires a high rate of dispersal between the islands. II, 5

Arguably, the lynx in the northeastern United States are effectively isolated from the main Canadian populations by the St. Lawrence Seaway and from the Great Lake States by agricultural conversion and urban development. Similarly, boreal forests in Colorado and Utah are separated from the larger areas of boreal forest in Wyoming by at least 100 km. If these areas have become effectively isolated, then they represent special areas of concern and, if they still contain native populations, these populations are particularly endangered. II, 13

Because translocations and reintroductions are inherently risky even under optimal conditions, they should be used only when other management options have failed. Conservation objectives are more likely to be met by maintaining and/or enhancing habitat conditions for lynx to provide opportunities for them to recolonize former portions of their range. V, 12

We do not believe that a system of reserves maintained by policy or regulation, embedded in a increasingly fragmented and non-natural matrix, can provide for sustainable lynx populations. Inevitable large-scale catastrophes preclude this being a sustainable system from the standpoint of lynx conservation; catastrophes will remove reserves, and there is no mechanism for their replacement or restoration in a reserve/matrix approach. Allen and Hoekstra (1992) have argued that one role of management in sustainable ecosystems is to replace functions that are generally provided in the natural context, but have been lost due to human influences. For this reason, we argue for a system of active management in those land designations where it is permitted. Management of these lands must be cognisant of and limited by the dependency of various species on the condition of these areas given the likelihood of divergent management in adjacent private lands, as well as inevitable natural disturbances... The necessary regulatory framework must also be in place... XV, 6-7

Our challenge, from the perspective of maintaining lynx and their prey in the context of ecosystem management, is to design management strategies that result in dynamic, sustainable landscapes that approximate the composition of natural systems. XV, 10

Fragmentation of habitats occupied by lynx (including increased openings, higher road densities, exurban residential development, and wider use of snowmobiles and devices that compact snow in areas with deep, soft snow) is a plausible mechanism for the questionable conservation status of the lynx in the contiguous United States. IV,13

Because of the fragmented nature of landscapes and heterogeneous distribution of topographic, climatic, and vegetative conditions in western forests, lynx and snowshoe hare habitats there are more prone to a metapopulation structure, which has important implications for their population ecology (Chapter 2). This condition is exacerbated by the presumably greater human-caused fragmentation of lynx habitat in the south. XIV, 5

In a cyclic system, there is no formal mechanism to generate stands or elements older than rotation age... escape of older elements is not effectively mimicked by embedding fixed reserves in a landscape of cyclic management... in areas subjected to large wildland fires, or where inholdings are managed for short-rotation timber yield using even-aged management, timber harvest (including salvage) may be inappropriate for decades... we would argue for a planned rate of disturbance somewhat less than the historically derived estimate. It is easy to increase the disturbance rate if desirable, but much harder to recover old-growth forests removed through overly aggressive management... X, 11-12

... models suggest that large proportions of young forest were characteristic of the boreal region but that fair amounts of quite old forest were part of the mosaic as well. These natural landscapes diverged significantly from even-aged silvicultural systems: fewer roads, more coarse woody debris, older age structures, and complex juxtaposition of different-aged stands. III, 35

Snowshoe hares generally occur in areas of dense forest cover, including shrubs and "doghair" thickets of small trees (Chapters 6 and 7). These structures are common in naturally regenerating areas after fire, but do not result from standard, even-aged forestry practices (Daniel et al.1979). Thus, we believe that natural regeneration and stand development will likely benefit hares and, ultimately, lynx. XV, 14-15

Creative silviculture is required when management goals extend beyond wood production. This might translate to heavier reliance on natural regeneration after wildland fires. Where harvest mandates artificial regeneration in a short time frame, the planting might be clustered such that the time of early seral dominance will be spatially variable across a unit. If pre-commercial thinning is considered, it should be recognized that the usual objective of increasing the diameter growth of residual trees may be inappropriate for lynx. Leaving some doghair stands may be good for lynx; thinning others so heavily that additional conifer regeneration occurs in the unit may also be appropriate. Because fires leave large amounts of woody material standing and down, management should also leave substantial amounts of woody material in representative size classes, regardless of treatment. XV, 15

For the boreal forest, there are several general principles that must be addressed in landscape management plans if harvesting or fire disturbance is to be incorporated:

1. The habitat will be a product of the cumulative effect of all disturbances. The substitution of logging for fire, for example, is only meaningful if fire can be successfully removed from the environment, and even then, logging will not totally mimic fire as a disturbance process because of roading and coarse woody debris differences.

2. Typical models used for fire history studies suggest that fire selected a range of stand ages to burn. Harvesting only the oldest ages will decrease the average stand age of the landscape and will remove the complex boreal stand structure that may be critical for lynx denning. Young and old stands need to be part of any landscape disturbance plan.

3. Size and juxtaposition of stands is critical. Most fires are small. Most of the landscape, however, is affected by larger fire patches, with unburned areas inside the fire perimeter (stringers, islands, fire skips). III, 34

We caution against uncritical application of ecological understandings derived from the North to southern lynx and we emphasize the potential importance of late-successional forests as habitat for hares, red squirrels, and lynx in the southern part of the range of the lynx. XIV, 1

In mesic forests... we argue that hare densities should be bimodal with stand age: highest in early seral conditions, minimal in closed-canopy mature forests, and reaching moderate densities in extremely old gap-phase forests. We do not, however, expect this pattern to be common on dry sites where erratic regeneration and the lack of tall deciduous brush often lead to very open conditions for extensive periods after disturbance. In these stands, hare habitat will only be produced in later seres. XIV, 10

We question the generalization that the best way to provide habitat for lynx prey, including hares,is to create early-successional forest. On mesic sites, the highest hare densities are generally found in sapling-stage forests after disturbance, but squirrels are absent from these sites. We believe the data, though sparse, also suggest that old gap-phase forests provide a combination of high numbers of squirrels and moderate densities of hares on many sites. Additionally, site moisture appears to be an important correlate of shrub and sapling densities in the early stages following disturbance and, in dry areas of the West where post-disturbance regeneration is erratic, gap-phase old-growth forests may provide relatively higher densities of small-diameter woody stems compared to earlier stages of succession on the same sites. In areas where boreal forests are highly fragmented, older forests may be an important stabilizing element. Because small areas of boreal forest can be highly altered by single disturbance events, such as fire, hare dynamics associated with post-disturbance conditions will be transient. Older forests are temporally stable, will produce hares in lower but more reliable numbers, and provide squirrels as alternative prey. On dry sites, where young forests do not reliably produce hare habitat, older forests may provide a critical source of hare habitat. XIV, 14

In both northern and southern regions, lynx occur predominantly in habitats where snowshoe hares are abundant, especially early successional stands with high stem densities. However, in southern boreal forests, such habitats appear to be used primarily for hunting; all known den sites in southern regions were located in mature forest stands with large woody debris. XIII, 2

The structural components – mature forests and high woody debris – associated with the natal den in Wyoming were similar to those associated with dens in Washington. XI, 12

In general, intensive forest management and concomitant fragmentation of habitat via timber harvesting and road systems are more prevalent in southern boreal forests compared to the taiga. Thus, human-mediated increases in potential competitors and predators may present a more serious threat to lynx populations in southern boreal forests than in the taiga. XIV, 13

Recreational snowmobile use has expanded dramatically in the contiguous United States in the past 25 years, with hundreds of thousands of km of trails (>19,000 km of groomed trails in Maine alone) within the pre-settlement range of the lynx (Maine Snowmobile Association, World Wide Web site, Zesiger 1997). Various unpublished accounts describe snowmobile and snowshoe trails facilitating access by coyotes to areas used by hares and lynx. In the Yukon, coyotes use both snowshoe and snowmobile trails (O ’Donoghue, personal communication). This facilitation of travel, in general, could help explain possible lynx reductions in the West via human-facilitated competition from coyotes and other generalist predators. Better understanding of this postulated relationship is critical. IV, 12-13

The likelihood of highway crossings by resident lynx can be expected to vary among home ranges according to proximal habitat conditions, width of road allowance and traffic volume, and perhaps by the animal’s sex and reproductive status. Although my analysis did not account for these factors, all lynx crossed highways less than random expectation within their 95% UD home ranges, suggesting that highways influenced lynx movements. Although my analysis only considered the influence of highways within home ranges, they may also influence home range selection just as dominant natural features can (Koehler and Aubry 1994), and this would decrease the substantive influence apparent within home ranges. XII, 9

... bobcats in Wisconsin selected home ranges with lower densities of secondary roads and crossed paved highways less than expected, a function of vehicle traffic levels and juxtaposition of preferred habitat to roads (Lovallo and Anderson 1996). XII, 9

As in the taiga, we found little evidence that roads represented a significant disturbance or mortality factor for lynx. Roads into lynx habitat may, however, provide access to generalist competitors, such as coyotes and bobcats. XIII, 2

Roads into areas occupied by lynx may pose a threat to lynx from incidental harvest or poaching (Koehler and Brittell 1990), increased access during winter for competing carnivores, especially coyotes (Chapter 4), disturbance or mortality from vehicles, and loss of habitat. XIII, 19

Our analysis of the current and historic distribution of lynx populations in the United States indicates that large, contiguous areas of suitable habitat are necessary for population persistence (Chapter 8). XV, 8

Because lynx occupy large home ranges and occur at low densities (about one lynx/50 km 2 ; Chapter 13), the long-term viability of lynx populations cannot be achieved at the spatial scale of relatively small parcels of public land, or even larger units such as individual National Forests or National Parks. Consequently, we believe that lynx conservation in the contiguous United States can only succeed as part of an ecosystem management strategy that is designed to address the needs of a variety of potentially conflicting resource uses over long periods of time and broad spatial scales. XV, 2

The following calculation (included here for illustrative purposes only) provides a sense of the smallest scale at which conservation planning will need to be addressed. The estimated density of a local population of 25 resident lynx in north-central Washington studied with radiotelemetry from 1980 to 1987,was about one lynx per 50 km 2 (Chapter 13; Brittell et al. 1989, unpublished; Koehler 1990). Thus, conservation of this small group of animals, which is probably not viable as a geographically closed population (see Chapter 2), would require a planning area at least 1,250 km 2 in size (25 x 50 km 2 ). However, this estimate represents a minimum area requirement over a relatively short period of time; to provide sufficient habitat for the population size to fluctuate around a long-term mean of 25 animals, the actual conservation area would need to be considerably larger. XV, 9

We cannot assume that lynx populations in the contiguous United States will be maintained by dispersal of lynx from Canada, nor that connectivity with larger habitat areas in Canada will be maintained in perpetuity. Although cooperative conservation efforts with Canadian land management agencies should be explored in all areas of adjacent lynx habitat, we believe that lynx conservation efforts in the contiguous United States should be addressed at geographic scales that will provide for the persistence of resident populations of lynx, regardless of periodic augmentations that may occur from other areas. Clearly, ecoprovince-wide planning is necessary to provide the broad-scale information necessary for effective conservation of lynx. XV, 9

Although legal harvest is no longer a conservation concern, human-caused mortality is believed to be additive in the low-density lynx populations characteristic of southern boreal forests (Koehler 1990; Table 13.3). If so, illegal or incidental harvest could significantly reduce population numbers of lynx in southern regions. XIII, 19

Preliminary telemetry data suggest that lynx in Montana and Wyoming have large home ranges; this result supports the Koehler and Aubry (1994) contention that lynx from southern lynx populations have large spatial-use areas. XI, 1

The study area in Montana is located in the Clearwater River drainage, near the town of Seeley Lake. This area is about 1,800 km 2 , extending east to west from the Swan Range to the Mission Mountains, and north to south from Lindbergh Lake to Salmon Lake. Lynx harvests (1977-1994) and track surveys suggest this area may support the highest density of lynx in Montana (Brian Giddings, personal communication). XI, 2-3

In Wyoming, the female produced a litter of four kittens (two males, two females) on about 27 May 1998; all kittens were alive on 14 June 1998 when they were ear-tagged. However, based on snow tracking, the kittens were not with the female in November and presumably had died. In May 1999 the same female produced two additional kittens. XI, 10

Understanding the ecology of lynx in Wyoming is critical to conservation planning because these animals represent the southernmost known population. XI, 2

Until we understand the nature of geographic variation in lynx populations,it would seem prudent to assume the existence of important genetic and non-genetic differences among populations, especially those that are distant and/or relatively isolated... it is critical to consider both inherited and non-inherited variation when manipulating lynx populations or managing their habitat. We recommend that managers assume the existence of important genetic and non-genetic differences between distant and/or relatively isolated lynx populations until proven otherwise. V, 11-12