Yellow-Billed Loon Subsistence Harvest in Alaska: Challenges in Harvest Assessment of a Conservation Concern Species

The yellow-billed loon Gavia adamsii breeds in restricted areas in Arctic Alaska, western Canada, and eastern Russia, has small global population (16,000–32,000 breeding birds) and is sensitive to changes in habitat and adult mortality (Figure 1; North 1994; Earnst 2004). These factors make the species vulnerable to threats and motivated conservation groups to request listing of the species as threatened or endangered under the U.S. Endangered Species Act (ESA; ESA 1973, as amended). During the ESA process, subsistence harvest in Alaska was identified as a potential threat, although harvest of yellow-billed loons is not legally authorized in the United States (U.S. Fish and Wildlife Service [USFWS 2014]).

Subsistence users tend to value sustainability of the biological resources they depend upon (Berkes and Turner 2006). However, conservation actions by government agencies can disrupt traditional practices and governance of subsistence communities (Stanek et al. 2006; Braund and Huntington 2011). Listing of the yellow-billed loon under the ESA could lead to restrictive subsistence harvest regulations and increased law enforcement efforts across the species range in Alaska. Other loon species could be closed to harvest because loons are difficult to tell apart, or the subsistence bird harvest as a whole could be unauthorized. Fisheries regulations could also be affected because loons entangle in fishing nets (Bentzen and Robards 2014).

However, the harvest estimate that prompted identification of subsistence harvest as a potential threat to yellow-billed loons was unusually high (1,077 birds in 2007 in the Bering Strait region of Alaska) and inconsistent with data on the species' abundance and distribution, suggesting potential issues with harvest estimates and species identification in harvest surveys (Naves 2010, 2014; USFWS 2014). Basically, because yellow-billed loons naturally occur in small numbers (3,000–4,000 loons breeding in Alaska and 8,000–20,000 in Canada), it was questionable that these loons could occur anywhere in large enough numbers to be harvested by the thousands (Earnst et al. 2005; USFWS 2014).

Because of concerns over yellow-billed loon conservation and effects of an ESA listing on subsistence harvest management, the USFWS and the Alaska Migratory Bird Co-Management Council expanded efforts to better understand harvest and inform the ESA process. The council implements collaboration among indigenous (Alaska Native), federal, and state entities in recognition of subsistence uses and conservation needs of migratory birds (AMBCC 2017). In this context, our objectives were to clarify 1) loon species harvested, 2) numbers of loons harvested, and 3) loon harvest methods and uses by local subsistence communities. To meet these objectives, we conducted harvest surveys and shore-based bird counts and gathered ethnographic information. Our study exemplifies the importance of using multidisciplinary approaches and engaging subsistence users to improve harvest assessment and inform wildlife conservation policy. Detailed information from this study needed to address specific questions raised during the ESA process was reported in Reference S2 (Supplemental Material).

The Bering Strait region of Alaska includes subsistence communities on the mainland and on the St. Lawrence and Diomede islands in the Bering Sea. Data on loon ecology, distribution, and harvest on the mainland area indicated relatively low loon numbers potentially available for harvest as well as low harvest levels (North 1994; CSIS 2017). Yellow-billed loons breeding in Alaska and Canada migrate westward over the Beaufort Sea and southward across the Bering Sea, nearing St. Lawrence Island, to winter in Asia (Schmutz 2009; Schmutz and Rizzolo 2012; USFWS 2014). Unknown migratory concentrations near St. Lawrence Island could make large loon numbers accessible for harvest. Because the human population (and thus the harvest potential) on Diomede Island was much smaller than that on St. Lawrence Island, we focused efforts to clarify loon harvest and distribution on the latter.

The human population on St. Lawrence Island is represented by the two indigenous (St. Lawrence Island Yupik) communities of Gambell and Savoonga (Figures 2 and 3). These communities primarily engage in subsistence activities and harvesting follows seasonal cycles of availability of animals and plants. Marine mammals are their main subsistence resources. Birds and eggs are only 4% of edible pounds harvested, but they add diversity to the diet and their harvesting and sharing are culturally important (Reference S4, Supplemental Material). Bird hunting occurs together with other subsistence pursuits near the communities and at remote sites inland, freshwater bodies, seashores, and at sea. Loons are not a main subsistence resource and have been harvested in relatively small numbers (References S1 and S4, Supplemental Material; Naves 2014). However, a relatively small harvest could represent a significant proportion of yellow-billed loon populations, because these populations are naturally small.

Harvest surveys were conducted in 2011 and 2012 as part of the Harvest Assessment Program of the Alaska Migratory Bird Co-Management Council (AMBCC 2017). Surveys and ethnographic research followed ethical principles for social science research (Arctic Research Consortium of the United States 1999). Consent to conduct interviews was obtained from tribal councils and households. In-person interviews were conducted by one avian biologist or anthropologist and one local resident hired as a research assistant. To minimize recall bias in loon harvest reporting, interviews occurred in September–October to collect spring (2 April–30 June) and summer (1 July–31 August) data and November–January to collect fall data (1 September–31 October), when most loons were taken (Naves 2014).

The sampling goal was to survey all households in each community depending on their voluntary participation. Residents of noncontacted households were known to be out of town or could not be found after three reasonable attempts. Household participation rates (households agreeing to participate divided by contacted households) and sampling rates (surveyed households divided by total households) were high; thus, survey results were unlikely to be affected by potential bias in the sampled population (Table 1). Respondents were instructed to report harvest by all hunters in the household including birds and eggs given away. To avoid double-counting, respondents were instructed not to report birds or eggs received from other households and report only their portion (share) of harvests done by parties including more than one harvester.

The survey covered all bird harvest to portray the relative importance of loons as subsistence resources (results for all species were reported in Reference S2, Supplemental Material). To clarify loon species identification issues in harvest data, we addressed all loon species rather than focusing only on yellow-billed loons. Materials used for bird species identification included a) two-page identification guide with color drawings of dozens of species (appendices F and G in Reference S2, Supplemental Material), b) four-page loon identification guide (appendix H in Reference S2, Supplemental Material), c) species identification book (Dunn and Alderfer 2011), and d) compilation of St. Lawrence Island Yupik bird names (appendix K in Reference S2, Supplemental Material).

For data analysis, households were stratified a priori as harvester (harvested birds or eggs in any one of the three prior years) or nonharvester to better account for households that could not be contacted or that declined to participate. Harvest reported by surveyed households was extrapolated to nonsurveyed households by multiplying the arithmetic mean household harvest (number of birds and eggs per species) in each stratum (harvester and nonharvester) by the total number of households (Naves 2012). Harvest estimates did not account for crippling (birds struck but not retrieved). Confidence intervals were calculated for harvest estimates following conventional methods for stratified random sampling (Naves 2012). Island-level harvest estimates represented both communities.

In 2011, we used the standard harvest survey materials also used in previous surveys (Naves 2012; AMBCC 2017). However, we increased sampling effort (number of surveyed households) and ensured that survey staff was highly experienced in both harvest data collection and loon species identification. These steps minimized data collection errors and ensured that harvest estimates would not be excessively inflated by extrapolation of harvest to nonsurveyed households. Benefiting from ethnographic information also being collected (below), we learned that most loons were taken in nonbreeding plumage, although a loon in nonbreeding plumage was not depicted in survey materials used up to 2011.

In 2012, still relying on experienced field staff and increased sampling effort, we revised survey materials to facilitate identification of harvested loons (appendices F and G in Reference S2, Supplemental Material). We added a loon in nonbreeding plumage. Because loons in nonbreeding plumage are particularly difficult to tell apart, we combined species and age categories (nonbreeding adults and young). To minimize confusion involving uncommon species, loons were presented in a row from the most to least abundant from left to right, as people usually read. The common loon Gavia immer and the yellow-billed loon were presented side by side for easy comparison. Drawings depicted size differences among species. Because of confusion with loon names (especially with the English name “common loon”), names were omitted and drawings were numbered for reference. As in previous years, the Pacific and Arctic G. arctica loons were combined in one category because these species are particularly difficult to tell apart.

To help interpret harvest data and inform on the importance of loons as food and cultural resources, we recorded information on ethnography (harvest practices, methods, and uses); ethnotaxonomy (species naming and categorization); and local and traditional knowledge (loon occurrence, grouping behavior) in harvest surveys, key respondent interviews, and other opportunities such as community meetings. Based on scientific taxonomy, five loon species occur in Alaska in both breeding and nonbreeding plumages (Dunn and Alderfer 2011). Although scientific taxonomy is based on genetic relationships, ethnotaxonomies categorize animals based on appearance, vocalization, behavior, habitat, and uses by people (Nakashima 1991; Souza and Begossi 2007; Tidemann and Gosler 2010). In 2012, four semidirected key respondent interviews with local, knowledgeable people focused on loon ethnography and ethnotaxonomy. Interviews were guided by predefined questions and respondents were encouraged to share related information reflecting their worldview (Huntington 1998; Bernard 2011). Here we highlighted the main ethnographic findings; a complete compilation of notes was presented in Reference S2 (Supplemental Material).

We conducted shore-based bird point counts to determine the local species composition of loons and other birds as an indicator of harvest availability. Bird count data were used to interpret and adjust harvest estimates. Because most loon harvest occurred in fall, counts were conducted from mid-September to mid-October (Naves 2014). This period covers most of the postbreeding migration of all loon species (North 1994; Schmutz 2009; Zeller et al. 2011; Lehman 2016). Count stations were located on the seashore close to Gambell (n = 4) and Savoonga (n = 7). Some stations were traditional hunting blinds. Logistical difficulties precluded access to remote locations. A stationary observer facing seaward recorded birds in flight and on water using binoculars and a spotting scope (Richardson and Johnson 1981; Byers and Dickson 2001; Day et al. 2003, 2004). Counts were distributed along daylight hours to account for circadian movements of birds. Sampling distance depended on weather and sea conditions. Counts were not made with strong wind, fog, rain, and snow. An index of relative species abundance was calculated by dividing numbers of birds counted by hours of observation.

It is difficult to differentiate between Pacific and Arctic loons, thus counts of these species may be subject to observer effect. Nevertheless, our results on the relative abundance of these species were in agreement with other count data for Alaska (Stehn et al. 2013; Platte and Stehn 2015; Lehman 2016). Species that migrate mostly at night, high altitude, or occur far from shore may be underrepresented in counts. It was logistically impossible to deploy field efforts to correct for double counting of individual birds, detectability rates, and other sources of bias (Dawson 1981). Relationships between the index of relative species abundance and the actual local species abundance or population sizes are unknown. Despite limitations common to bird counts, our data represented the species composition most accessible to subsistence users: birds flying at low altitude, relatively close to shore, in daylight, and fair weather and sea conditions similar to those safely travelled by hunters.

Yellow-billed loons usually occur as single birds, pairs, or groups of few birds. Little is known about their migratory behavior and whether migratory aggregations occur (North 1994). Because such aggregations could increase harvest vulnerability, we recorded loon group size (number of individuals) in bird counts and harvest surveys. We recorded loon plumage (breeding, nonbreeding) in bird counts to serve as a basis of comparison to harvest composition obtained in harvest surveys. Loons in early molt had mostly breeding appearance and were categorized as breeding plumage.

Loons were commonly harvested in nonbreeding plumage, when species identification is most difficult. Thus, it was challenging to determine harvest species composition in the absence of a species identification system based on biological sampling (Carney 1992). For loons harvested in nonbreeding plumage, we estimated species composition a posteriori based on bird count data (Table 2). This approach was valid because we found no evidence of harvest selectivity for individual loon species. For 2011, when survey materials did not include a loon in nonbreeding plumage, we calculated the number of loons in nonbreeding plumage in summer and fall based on their proportion in the 2012 harvest (64%). This step was needed to account for harvest preference for young birds (regardless of species). Then, we calculated harvest species composition based on the average proportion of species for loons in nonbreeding plumage in bird counts (Pacific and Arctic loons: 97%, yellow-billed loons: 2%, and red-throated loons Gavia stellata: 1%).

The overall household participation rate was 90% for both spring–summer and fall data collection, and the sampling rate was 90% for spring–summer and 82% for fall (Table 1). The estimated annual bird harvest (all species combined) was 10,663 birds in 2011 and 7,754 birds in 2012; loons were 1 and 2% of these totals (Table 2). Eggs of loons were not reported as harvested. Overall, 52% of the bird harvest occurred in spring, 15% in summer, and 33% in fall, whereas 86% of loons were harvested in fall. In fall 2011, 66% of loons were reported as common loons. In fall 2012, using revised survey materials, 3% of loons were reported as common loons and 64% were reported as unidentified species in nonbreeding plumage (Table 2).

Loons, as most other birds, were harvested with shotguns. Loon harvest occurred mainly by boat (n = 222 reported loons, 2011–2012) together with fall seal hunting. Loons were also harvested from onshore blinds (n = 11), while hiking inland (n = 1), traveling along the beach (n = 1), and entangled in fishing net (n = 2). Entanglement in fishing nets was uncommon and loons were released alive or used for food. Harvest opportunities in fall were limited because storms, sea ice, and travel cost often precluded boat trips. When asked about preferences in bird harvest, people indicated preference for young birds (including loons) because they are tender and fatter than adults, but they did not indicate preference for harvesting some species over others (e.g., “We catch anything, as much as we need to eat. But we do not catch much of the parents because they are tough.”).

Local and traditional knowledge concurred that loons usually occurred as single birds and pairs and it added that loon groups were more common in fall than in other seasons (e.g., “In spring I see mostly loon couples, flying. In fall time I see groups of up to five, swimming.”). Loons were harvested as single birds in 17 of 33 harvest events. Occasionally, larger loon numbers occurred in feeding aggregations formed around forage fish (e.g., “This fall there were 60 to 100 young loons close to shore feeding on a concentration of bait fish. There were also shearwaters, murres—the sky was black—seals, and whales all together eating fish.”).

Currently, with few exceptions, loons and other birds are only used for food (e.g., “I use birds of all kinds just for food. Once in a while I get birds for dog food, no more sled dogs, just pets.”). Birds were eaten fresh or preserved frozen or air-dried for later consumption. Loons were roasted, boiled, or added to soup. Current uses of loons for crafts, cultural, and ceremonial purposes were not identified on St. Lawrence Island.

The St. Lawrence Island loon ethnotaxonomy differed from scientific taxonomy (Table 3). Most commonly, loons were identified by local people based on size and overall plumage, resulting in four main categories: large or small, in breeding or nonbreeding plumage [e.g., “I got the small ones, not the big ones. September-October we have loons out here, mostly young ones (pointing to nonbreeding plumage on identification guide), no adults.”]. The name nangqwalek was likely used for the two large loons, the common and yellow-billed loons (e.g., “I wonder which of these two species, yellow- and black-billed, are the exact nangqwalek, we call them nangqwalek, the largest of the loons. Probably both of them are really nangqwalek.”). Species-specific indigenous names of small loons were uncommonly used. Also, loon names were inconsistently used. It was unlikely that loons in nonbreeding plumage were identified and named at the species level.

The sampling effort was 168.3 h in 2011 (September 21–October 11) and 161.4 h in 2012 (September 17–October 11). Count duration averaged 1.7 h (SD = 0.8, range = 0.3–5.0 h). Loons were <0.1% of all birds counted each year (Table 4). Loons occurred mostly as single birds (66%), pairs (19%), and groups of three to nine birds (14%; 2011–2012 combined). Only Pacific loons were recorded in groups of 10+ birds (Figure 4). Pacific loons were 96% of all loons identified in 2011 and 94% in 2012. Yellow-billed loons were 4% of loons identified in 2011 and 5% in 2012. The common loon was not observed. In 2011–2012, 24% of all loons were in nonbreeding plumage and a larger proportion of Pacific (25%) than yellow-billed (9%) loons was in nonbreeding plumage (Table 5). In 2012, the smaller proportion of nonbreeding loons in bird counts (24%) compared to fall harvest (64%; Table 2) supported harvest preference for young birds. Yellow-billed loon harvest estimates adjusted based on bird counts to account for unidentified loons in nonbreeding plumage were five birds each year in 2011 and 2012 (Tables 2 and 5).

Local availability of loon species affects their harvest vulnerability. Yellow-billed loon migration routes and timing are poorly known (North 1994; Schmutz 2009; USFWS 2014). However, bird counts and local and traditional knowledge supported that, on St. Lawrence Island, loons typically occur as single birds or pairs and sporadically as groups of dozens of birds (Day et al. 2003; Lehman 2016; this study). In New Zealand, harvest of sooty shearwater Puffinus griseus chicks by Maori people was positively related to chick abundance, but harvest effort was the main predictor of harvest (Mckechnie et al. 2007). We found no evidence of large harvest effort focused on loons. Loons were commonly harvested, but harvest was opportunistic and in relative small numbers.

Interpretation of harvest survey data requires ancillary information and participation of the involved communities, especially if data are questionable. Loon availability and harvest success depend on ecological and socio-economic factors leading to large annual variation in harvest numbers. Several years of data are needed to depict harvest level (Bromley 1996). Bird harvest data for St. Lawrence Island were available for 11 years (1993–2012; compiled in Naves 2014; this study). Despite large annual variation in harvest numbers for loons and other birds, there was no evidence to support Bering Strait regional 2007 estimates, which reflected unusually high harvest estimates only for Savoonga. Harvesting, processing, and sharing wild foods are fundamental to subsistence communities and unusual harvests are memorable. In community meetings for harvest data review, local residents were adamant that the 2007 data were incorrect. The current understanding is that Savoonga 2007 data may have been affected by data collection errors (Naves 2014). In subsistence harvest assessment, communication and logistics are complex because of diverse cultural settings and remoteness of the studied areas. Federal, state, and indigenous partners involved with the Harvest Assessment Program of the Alaska Migratory Bird Co-Management Council have strived to refine harvest surveys (AMBCC 2017). Despite challenges, harvest assessment is needed for resource conservation and protection of sustainable harvest opportunities.

Yellow-billed loon harvest is likely much lower than previously estimated, but it is difficult to be accurately enumerated. Based on all harvest and population data available for Alaska, Canada, and Russia, the USFWS concluded that “likely few yellow-billed loons are harvested annually in Alaska and that tens to few hundreds are harvested across the species distribution range” (USFWS 2014:48). Also, because of harvest preference for young birds, loons were commonly harvested in fall, in nonbreeding plumage. Based on molt timing, most Pacific and yellow-billed loons in nonbreeding plumage in summer and fall likely were hatch-year birds (at least half of the total harvest). Adult Pacific loons retain breeding plumage into October–November (Russell 2002). Adult yellow-billed loons likely molt later than Pacific loons, considering the longer chick-rearing period for yellow-billed loons. Immature yellow-billed loons and some nonbreeding adults remain year-round at wintering grounds (North 1994). Hatch-year birds have higher mortality rates than adults and lower reproductive value for populations (Martin 1995; Mauk et al. 2004; Mitro et al. 2008). Predominance of hatch-year birds in harvest lessens effects of harvest on loon populations.

Listing of the yellow-billed loon under the ESA was unwarranted in part because our study did not find evidence of high harvest (USFWS 2014). By using a multidisciplinary approach and engaging subsistence users, our study was key to inform conservation policy (Balmford and Cowling 2005; Newing and St. John 2013). Yellow-billed loon counts conducted in the Arctic Coastal Plain of Alaska showed a stable or increasing population and were also critical in the ESA listing decision (Stehn et al. 2013). The population trend reflects population-level responses to all potential threats, including harvest. Yellow-billed loons remain closed to harvest, but this law is not always known or observed in remote subsistence communities. Also, species identification issues make compliance difficult. Continued communication is needed to raise awareness of and engage subsistence users in conservation efforts.

Based on the relative loon species abundance derived from bird counts, most loon harvest was likely Pacific loons. Harvest surveys alone were unable to depict species-specific harvest composition because a large proportion of loons were harvested in nonbreeding plumages, heightening species identification issues. Information on ethnography and ethnotaxonomy did not support harvest preference for individual loon species. Harvest selectivity may indirectly occur if spatial and temporal distributions differ among loon species affecting their harvest vulnerability. However, this process was unlikely to significantly affect harvest composition because, based on bird counts, Pacific loons outnumbered other loon species.

Our study raised awareness about species identification issues in harvest surveys and some involved factors. First, local ethnotaxonomies commonly differ from scientific taxonomy (Tidemann and Gosler 2010). Second, in western and indigenous cultures, a small proportion of people develops advanced skills for bird species identification, especially for those species and plumages that are difficult to tell apart (Irving 1958). Third, vernacular species names can be confusing, especially in cross-cultural and bilingual settings. Previous common loon harvest estimates seemed excessively high compared to numbers known to occur on St. Lawrence Island (Lehman 2016; this study). Omission of loon names in the 2012 survey resulted in great reduction of common loon harvest reports. Common loons in previous harvest surveys likely referred to “the most common loon.” Confusion with the word “common” in English vernacular names applies to other species such as common eider Somateria molissima and common goldeneye Bucephala clangula. Following findings of this study, harvest surveys of the Harvest Assessment Program of the Alaska Migratory Bird Co-Management Council across Alaska have omitted loon names so that species identification is based on images only. Despite survey design improvements, loons in nonbreeding plumage remain unidentified and species identification of loons in breeding plumage likely remains unreliable.

There are many morphologically similar species across taxa. Species identification in shark fisheries is a challenge for harvest management worldwide and has prompted the development of genetic and morphological tools (Hernández et al. 2010). The composition of bird sport harvest in the United States and Canada is derived from wings and tails submitted by hunters (Carney 1992; Wilhelm et al. 2008). Biological sampling in subsistence bird harvest surveys has not been implemented and is challenging due to cultural acceptance issues and because birds are harvested and consumed at remote locations (Bromley 1996; Zeller et al. 2011). In this context, a better understanding of local ethnography and ethnotaxonomy can allow refining harvest surveys methods and design, leading to more accurate harvest data. Although some issues may remain unresolved, ethnography and ethnotaxonomy can clarify achievable levels of detail in species identification, and this information is particularly relevant when dealing with species of conservation concern.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article

Reference S1. Ahmasuk A, Trigg EW, Magdanz JS, Robbins B. 2008. A comprehensive subsistence use study of the Bering Strait region. North Pacific Research Board Project Final Report Project #643. Nome: Kawerak Inc. Subsistence Resources Division.

Found at DOI: http://dx.doi.org/10.3996/112016-JFWM-086.S1 (7611 KB PDF).

Reference S2. Naves LC, Zeller TK. 2013. Saint Lawrence Island subsistence harvest of birds and eggs, 2011–2012, addressing yellow-billed loon conservation concerns. Anchorage, Alaska: Alaska Migratory Bird Co-Management Council, Alaska Department of Fish and Game, Division of Subsistence. Technical Paper 384.

Found at DOI: http://dx.doi.org/10.3996/112016-JFWM-086.S2 (5337 KB PDF).

Reference S3. Romanenko O, Taylor DL, Kanishero V, Gologergen O, Schaeffer P. 1997. Biota of central Beringia with English, Russian, and Native names. Anchorage, Alaska: U.S. National Park Service, Beringian Heritage International Park Program.

Found at DOI: http://dx.doi.org/10.3996/112016-JFWM-086.S3 (3442 KB PDF).

Reference S4. Tahbone ST, Trigg EW. 2011. 2009 Comprehensive subsistence harvest survey, Savoonga, Alaska. Nome: Kawerak Inc. Subsistence Resources Division.

Found at DOI: http://dx.doi.org/10.3996/112016-JFWM-086.S4 (1046 KB PDF).

We thank the communities of Gambell and Savoonga for collaboration sharing traditional knowledge and reporting harvest. We thank J. Fall, R. Oates, D. Dewhurst, and N. Stellrecht for technical support. D. Marks, L. DeCicco, E. Labunski, P. Scully, A. Bankert, G. MacDonald, and D. Pavlik conducted bird counts. T.Z., L.N., J. Van Lanen, M. Marchioni, K. Gwynn, and E. Nageak conducted harvest surveys. B. Slwooko, S. Apassingok, D. Apassingok, D. Apangalook, J. Slwooko, M. Annogivuk, D. Iya, and T. Seppilu were local research assistants for harvest surveys. C. Noongwook, S. Mokiyuk, K. Iworrigan Sr., and M. Apatiki Sr. were key respondents. L. Naves, T. Zeller, and J. Van Lanen conducted key respondent interviews. R. Alowa, S. Alowa, C. Koonooka, D. Campbell, and H. Irrigoo provided logistic support. C. Koonooka translated interviews. The Associate Editor and an anonymous reviewer contributed improvements to manuscript drafts.

This study was funded by the USFWS (Alaska Region Fisheries and Ecological Services-Fairbanks Fish and Wildlife Field Office, Division of Migratory Bird Management, and cooperative agreements 70181-7-J653 COOP-07-118 and F12AC00653).

Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.