The population of the rufa subspecies of the Red Knot (Calidris canutus), which breeds in the central Canadian Arctic and mainly winters in Tierra del Fuego, has declined dramatically over the past 20 yr. Previously estimated at 100,000-150,000, the population now numbers 18,000-33,000 (18,000 if just the Tierra del Fuego birds are C. c. rufa, more if the Red Knots of uncertain subspecific status that winter in northern Brazil (7,500) or Florida (7,500) are also C. c. rufa). Counts show that the main Tierra del Fuego wintering population dropped from 67,546 in 1985 to 51,255 in 2000, 29,271 in 2002, 31,568 in 2004, but only 17,653 in 2005 and 17,211 in 2006. Demographic studies covering 1994-2002 showed that the population decline over that period was related to a drop in annual adult survival from 85% during 1994-1998 to 56% during 1999-2001. Population models showed that if adult survival remained low, C. c. rufa would go extinct within about 10 yr. After 2002, the population held up in 2003-2004, but plunged again by nearly 50% in 2005 increasing the likelihood of extinction within the next decade. Despite intensive studies, the reasons for the population decline and reduced adult survival are imperfectly known. During northward migration, most C. c. rufa stopover in Delaware Bay where they feed mainly on the eggs of horseshoe crabs (Limulus polyphemus) and lay down fat and protein reserves both to fuel the 3,000 km flight to the arctic breeding grounds and ensure their survival after they arrive at a time when food availability is often low. The crucial importance of Delaware Bay is demonstrated by studies that show that Red Knots with lower mass in Delaware Bay have lower survival than heavier birds and that from 1998-2002 the proportion of birds there at the end of May weighing more than the estimated departure mass of 180 g declined by >60%. This might be the result of the progressive failure of the food supply in Delaware Bay and/or a trend for birds to arrive there later and/or in poorer condition. In years when Red Knots experience reduced food availability and arrive late, the result may be an exacerbation of the effects of each of these deleterious factors. The main identified threat to the C. c. rufa population is the reduced availability of horseshoe crabs eggs in Delaware Bay arising from elevated harvest of adult crabs for bait in the conch and eel fishing industries. Since 1990 the crab population has declined substantially. Although significant uncertainty regarding the extent of the decline of the horseshoe crab population remains, there is general agreement that horseshoe crab stocks have declined to a level where increased management of the fishery is necessary and appropriate. The decline in crabs has led to a decrease in the density of eggs available to shorebirds. Because of the crab's delayed maturity, demographic models indicate that even if further exploitation of crabs ceases immediately, it will be some years before the horseshoe crab population recovers to its former level. Although clear evidence, as in 2003 and 2005, shows that the reduced availability of eggs is already having an impact in some years on the Red Knots ability to gain mass in Delaware Bay, it is likely that other threats to C. c. rufa exist and that these are the cause of some birds arriving in the bay late and/or in poor condition. It is not known what these are, but they could be related to Bahia Lomas, the main wintering site in Tierra del Fuego (because the largest reduction in recent years has occurred there and because northward migration from Bahia Lomas along the Atlantic coast of Argentina has taken place 1-2 wk later since year 2000). If it is proved that something leads Red Knots to arrive late in Delaware Bay and/or in poor condition, this does not diminish the importance of the Delaware Bay food resource. If anything, it is increased because it is of critical importance in enabling the birds to recover quickly and reach the breeding grounds on time and in good reproductive condition. Actions being taken to improve feeding conditions for Red Knots and other shorebirds in Delaware Bay include beach closures to prevent disturbance and exclosures to reduce competition from gulls. However, although these measures help, they are no substitute for a recovered horseshoe crab population. Actions to conserve horseshoe crabs have included reduced harvest quotas, more efficient use of crabs as bait, closure of the harvest in certain seasons and places and the designation of a sanctuary off the mouth of Delaware Bay. The latest information indicates that the crab population may have stabilized, but there is no evidence of recovery. Another Red Knot subspecies, C. c. roselaari, breeds in Alaska and is presumed to include those Red Knots that winter on the Pacific coast of the United States and Mexico. Two other Red Knot wintering populations are of uncertain subspecific status -one in the southeastern U.S. (mainly Florida) of about 7,500 and one on the north coast of Brazil also of about 7,500. These populations have not been the subject of regular systematic surveys, but it is not thought that either has suffered the same catastrophic decline as the C. c. rufa that winter in Tierra del Fuego. Substantial proportions of both pass through Delaware Bay during northward migration, but banding shows that these are distinct populations without interchange with the Tierra del Fuego birds. Moreover, genetic studies show that no exchange of genes has occurred between the southeastern U.S. and the Tierra del Fuego birds for at least 1,200 yr. Some progress has been made toward understanding why the Tierra del Fuego population has suffered a major decline, but the northern wintering birds have apparently remained more stable. It appears that physiological constraints mean that the southern birds, which mostly make a long, non-stop flight to Delaware Bay from at least northern Brazil, are more reliant on soft, easily-digested horseshoe crab eggs in Delaware Bay than the northern winterers, many of which feed on blue mussel (Mytilus edulis) spat or surf clams (Donax variablis) on the Atlantic coast of New Jersey. Evidence from Patagonia suggests that, for a reason that remains obscure, northward migration of Tierra del Fuego birds has become 1-2 wk later since the year 2000 and this has probably led to more Red Knots arriving late in Delaware Bay. Late arriving birds have been shown to have the ability to make up lost time by increasing their mass at a higher rate than usual provided they have sufficient food resources. However, late-arriving Red Knots failed to do this in 2003 and 2005 when egg availability was low. Although C. c. rufa Red Knots are spread thinly across a large area of the Canadian Arctic during the breeding season, for the rest of the year they occur mainly in large flocks at a limited number of key coastal wintering and staging sites. This review describes each of these sites and the threats the birds face ranging from oil pollution to disturbance and reclamation for development. Overall the goal of conservation activities throughout the flyway should be to increase the C. c. rufa population to at least the number of 25 yr ago - 100,000-150,000 by 2015. Given the uncertain genetic relationships between the three main wintering populations we suggest the following population increases: (1) Tierra del Fuego wintering population to 70,000-80,000 birds, (2) Brazilian wintering population to 20,000-25,000, (3) Florida wintering population to 20,000-25,000, and (4) other sites to 15,000-20,000. The means whereby such population increases might be achieved include: (1) recovery and maintenance of Delaware Bay horseshoe crab egg densities to levels sufficient to sustain stopover populations of all shorebirds including 100,000 Red Knots, (2) control impact of disturbance at all stopovers and wintering areas, particularly in high-importance, high-disturbance areas like Delaware Bay and the west coast of Florida, (3) by 2008, develop a system for the yearly determination of population demographic status based on counts, capture data, and resightings of banded individuals, (4) by 2008, determine the genetic and breeding status of the three main wintering populations (Tierra del Fuego, Maranhão, and Florida), (5) by 2008, identify all important breeding locations in Canada and recommend protection needs and designations for the most important sites, (6) by 2009, complete site assessments and management plans for all important wintering areas and stopovers in the flyway, (7) by 2009, delineate and propose protection measures for key habitats within the main wintering areas of Maranhão, Tierra del Fuego, and Florida, and develop management plans to guide protection, (8) by 2009, determine key southbound and northbound stopovers that account for at least 80% of stopover areas supporting at least 100 Red Knots, and develop coast-wide surveillance of birds as they migrate, and (9) by 2011, create a hemisphere-wide system of protected areas for each significant wintering, stopover, and breeding area. Also crucial to C. c. rufa's recovery is adequate funding to support the conservation actions and research needed. Despite the fact that much of the research, survey, monitoring, and conservation work has been carried out by volunteers and has been supported financially by state, federal government and non-government agencies, present funding levels are inadequate to sustain the work required.