William E. Parry (1821: 273), during his 1819–20 voyages in the Canadian Arctic, appears to have recorded the first botanical observation along the east coast of Baffin Island near Scott Inlet, but he apparently made no collections (see Suppl. material 1). The first plant collections made in the flora area were those of James Taylor, an Aberdeen-based surgeon who travelled aboard whaling ships during the period 1856–61 (Taylor 1863; Simmons 1913). He made collections on both the Canadian and Greenlandic sides of Baffin Bay, visiting at least five localities on Baffin Island from Cumberland Sound north to Cape Adair. Two sites are located within the flora area: Cape Adair (71°29'44"N, 71°33'55"W), now officially named Talluruti Tulliit, and Scott’s Bay/Scott’s Inlet (71°6'30"N, 71°6'30"W, official name Scott Inlet). Scott Inlet encompasses a large area, 25 × 20 km, at the mouth of Clark and Gibbs fiords where they meet. We located three specimens at MEL from Scott’s Inlet that can be attributed to Taylor, although they do not bear Taylor’s name (see Suppl. material 1) (herbarium acronyms according to Thiers 2025 [continuously updated]).

James Goldie (J.G.) McMillan, a geologist with the Geological Survey of Canada, collected a single specimen at Clyde River (housed at CAN) on 6 September 1909, while on the return voyage of the DGS Arctic on Captain J.E. Bernier’s 1908–1910 expedition through the Northwest Passage, which had overwintered at Melville Island (Bernier et al. 1910).

Several botanists visited Clyde River in the 1920s and 1930s, where a Hudson Bay trading post had been established in 1924. Malte Oskar Malte, a botanist with the Biology Division, Geological Survey of Canada, made collections at Clyde River on 18 August 1927, while accompanying the Canadian Arctic Expedition of 1927 aboard the SS Beothic, led by Captain George P. Mackenzie (Collins 1929). The 20 specimens seen by us are housed at CAN, MT, and QFA. Nicholas V. Polunin visited Clyde River twice, on 9 September 1934 and 15 September 1936—both times aboard the SS Nascopie (Polunin 1940; Cockburn 1983). The 43 specimens examined here are housed at CAN, F, GH, MIN, and US. Rev. Father (Père) Arthème Dutilly also traveled to Clyde River twice—like Polunin, both times aboard the SS Nascopie (Smith and Lackenbauer 2014; Eastern Arctic Patrol 1941). Dutilly collected at Clyde on 15 September 1936 (one specimen at QFA was collected on 9 September) and on 21 September 1941; the 60 collections examined by us are housed at CAN, CM, DAO, MIN, MT, QFA, S, and US. Norman Bethune Sanson, then curator of the Banff Park Museum, joined the 1938 voyage of the SS Nascopie (Hallworth and Jackson 1985), which called at Clyde River on 5 and 6 September (Shortt and Peters 1942). Two of his specimens (at TRT) have been examined, but there likely are others in undigitized collections (e.g., citations in Ball 1950). John George Oughton, a zoologist with the Royal Ontario Museum, visited Clyde River on 7 September 1939, aboard the SS Nascopie on the Eastern Arctic Patrol (Oughton 1939; Chitty 1940), and collected four specimens that we found at TRT.

Three non-botanists collected in the flora area during the 1930s and 1940s. Rupert W. Bartlett accompanied the Greenland Expedition of 1938 with Captain Robert A. Bartlett, collecting plants along the ship’s route (Smallwood 1981). Their only Canadian landing was at Hoffman Cove, “a small sheltered cove inside” Cape Raper (Nutt 1938) on the north side of Isabella Bay, on 16 August 1938. We located 33 of these collections deposited at CAN, MT, and US. Rutherford Platt, photographer and nature writer, collected at Cape Hewitt on 19 August 1947 (or 1948), while part of the MacMillan Arctic Expedition (Platt 1948, 1956) (see Suppl. material 1 for details regarding the year collected). We located 18 specimens, all at NY. D.B. Coombs, working as a surveyor for the Geodetic Survey of Canada, made collections of vascular plants and mosses at Coutts Inlet in August 1948 (Anonymous 1948). Some of these collections include the locality “Manning Fjord,” which is not a current or historic place name in Canada (Canadian Geographic Names Database, https://geonames.nrcan.gc.ca/search-place-names/search, accessed 20 January 2025). Daniels (1956) contains the only published reference to a “Manning Fjord”: “Manning Fjord on the N.E. coast of Baffin Island, lying some 75 miles E.S.E. of Erik Harbour, penetrates approximately 40 miles inland in a south-westerly direction.” Imprecise coordinates given on these specimens suggest this refers to the upper part of Coutts Inlet, which is located about 75 km SSE of Kangiqługaapik [Erik Harbour]. We found 50 of his collections at DAO, MO, and US.

During the 1950 Baffin Island Expedition led by P.D. Baird (Baird et al. 1950; Baird 1951, 1952), botanist Pierre Dansereau and ornithologist Vero Copner Wynne-Edwards both made extensive collections from various locations on east-central Baffin Island (Fig. 4A, B). Dansereau collected extensively from numerous sites within walking distance of the expedition’s base camp (“B2”) at the head of Clyde Inlet (Kangiqtugaapik) (Fig. 4A), as well as at Clyde River, the head of Arviqtujuq Kangiqtua [Eglinton Fiord] (Fig. 4B), Umiujaq [Agnes Monument], and Gee Lake (Baird et al. 1950). We located 594 of Dansereau’s collections, made between 30 May and 30 August 1950, at DAO, MIN, and MT (the primary set). Wynne-Edwards also collected primarily at the head of Clyde Inlet, as well as once at the head of Arviqtujuq Kangiqtua, from 1 June to 24 August 1950. There are 300 of his collections at CAN (primary set), with a few at E and LSU. Hans Röthlisberger, a mountaineer with the expedition, made scattered collections at Arviqtujuq Kangiqtua and Swiss Bay (where the expedition had two mountaineering camps) from 12 July to 11 August 1950, 10 of which were located at CAN and ZT (Baird 1952).

John E.H. Martin, an entomologist with the Department of Agriculture, Ottawa (now Agriculture and Agri-Food Canada), collected plants in Clyde River, at the head of Clyde Inlet, and on unspecified islands in Clyde Inlet, from 29 June to 18 August 1958, during the department’s insect survey fieldwork that year (Lotz 1963: 2). We located 63 of Martin’s collections at DAO, MT, QFA, RBCM, TRT, and UTC.

Neal G. Smith, an ornithologist conducting his PhD research on Larus gulls on Baffin Island in 1961, collected vascular plant specimens from the Home Bay area between 30 June and 6 August. We located 63 of these, all at CAN. While there is some controversy as to where Smith spent most of that summer’s field season, work by Snell (1991) indicated that Smith was almost certainly based at the head of Rocknoser Fiord (Kangigfik Fiord) (Fig. 4C) for the majority of the season.

Between 1965 and 1967, the Geographical Branch, Department of Energy, Mines and Resources, Canada, established a long-term base camp at the head of Inugsuin Fiord that hosted several botanists over three years of operations (Løken et al. 1966) (Fig. 4D). In 1965, Robert Hainault, based out of the Fowler Herbarium at Queen’s University, traveled to Inugsuin Fiord, collecting extensively around the Geographical Branch base camp, at a field camp near the mouth of Inugsuin Fiord, and in a valley inland at the head of McBeth Fiord, from 6 June to 28 August (Hainault 1966). We found 332 of his collections at ACAD, CAN, DAO, H, MT, QFA, and QK. From 21–22 July 1967, J.A. Parmelee and J.R. Seaborn collected near the Cape Hooper DEW Line site (FOX-4), and from 25 July to 2 August, they collected within walking distance of the Inugsuin Fiord base camp. The main set of their collections is at DAO (160 collections), with at least one collection (Parmelee & Seaborn 3933) distributed more widely to ALTA, BC, BISH, F, L, LD, MO, S, TRH, TROM, TRT, UAM, UTC, and W as a part of a DAO exsiccate set.

During the same period, several researchers associated with the Geographical Branch also made collections on east-central Baffin Island (Ives 2016). From 17 June to 18 August 1966, Jane T. Philpot, Penny A. Crompton, and June Ryder collected from several locations accessed via the Inugsuin Fiord base camp (Ives 2016), including Ekalugad Fiord, Kangok Fiord, Inugsuin Fiord, and Remote Lake on Remote Peninsula (Fig. 4E). We located 82 of these collections at COLO. In 1967, June Ryder collected at Ekalugad Fiord between 17 June and 23 July and Nudlung Fiord on 30 July; we located 24 collections at COLO and UBC. From 13–18 July 1966, Mike Church, studying glacial outwash plains for his PhD research at the University of British Columbia and the Geological Survey of Canada (Church 1972; Ives 2016), collected at the head of Ekalugad Fiord (Fig. 4F); we located eight of his specimens at COLO. He also collected plants in 1967 (dates unknown) at the head of Ekalugad Fiord, three of which are at UBC.

Other collectors were also active in the Home Bay area of east-central Baffin Island in 1967. From 7 July to 8 August 1967, botanist Patrick Webber collected at Ekalugad, Rocknoser, and Kangirlugag fiords; 108 collections are vouchered at CAN, COLO, QFA, and QK. Between 7 and 17 July, J. Richardson and P. Webber made collections from the Fox Charlie (FOX-C) DEW Line station on the southern shore of Ekalugad Fiord, 72 of which are also at CAN, COLO, QFA, and QK. We also located five collections at COLO and UBC made by R. Stock from the head of Ekalugad Fiord on 22 August 1967.

In 1983, Phil Sadler made one collection of Eriophorum scheuchzeri while working with a runway maintenance team at the Cape Hooper DEW Line site (FOX-4) (Sadler, personal communication, 2024). This specimen is at ASC.

Over the course of three years, Bruce C. Forbes collected specimens in Clyde River during the periods 19 June–14 August 1988, 27 June–5 July 1989, and 22 June–12 July 1990, while studying the ecology of archaeological sites and tundra disturbance (Forbes 1992, 1993, 1996). We located 116 of his specimens at DAO.

David Oswald of Royal Roads Military College (now Royal Roads University) collected vascular plants in the vicinity of the FOX-C DEW Line station; we located 31 of these, all deposited at QK. Oswald was affiliated with the college’s Environmental Sciences Group (as indicated on the specimen labels), and these plants likely were collected during this group’s study of the environmental impact of the DEW Line sites (Dushenko et al. 1996).

During the Tundra Northwest 1999 expedition (Eriksen et al. 2006), three researchers made collections at Cape Hooper on 30–31 August 1999: botanist Reidar Elven (80 collections found at ALA, CAN, and O), zoologist Johan Hammar (three collections found at CAN and O), and plant ecologist Ulf Molau (one collection found at O).

Several Canadian Museum of Nature(CMN) scientists visited the study area and collected in the early 2000s. In 2008, botanist Julian Starr collected at Swiss Bay on 11 August 2008, during an expedition with Students on Ice; to date only two of these specimens have been processed and are available at CAN. From 8–10 August 2017, Oluwayemisi Dare, on the Canada C3 expedition, made collections at Tingijattut on the western arm of Kangiqtualuk Uqquqti [Sam Ford Fiord], at Ravenscraig Harbour on the eastern shore of Arviqtujuq Kangiqtua [Eglinton Fiord], and in Clyde River; these 49 specimens are at CAN. On 16 August 2018, Roger Bull made two vascular plant collections (housed at CAN) from Coutts Inlet during an expedition with Students on Ice.

Between 22 July and 22 August 2021, we, accompanied by Clyde River wildlife monitors Jaypiti Inutiq and Leeno Apak, collected vascular plant specimens from 165 sites at 23 main localities across Agguttinni Territorial Park, northwest of Clyde River. Details regarding the fieldwork within the park are summarized in Gillespie et al. (2024). We collected vascular plants at an additional seven sites in Clyde River on 10 and 22 August, within the community and near the airport. In total, we made 850 vascular plant collections from Agguttinni TP and Clyde River; the primary set of these collections is vouchered at CAN, with duplicates at Nunavut Parks and Special Places (“npsp”), ALTA, ALA, CHARS, CORD, MIN, MO, MT, NFLD, NY, QFA, UBC, US, and WIN.

From 4–16 August 2022, botanists Martha Raynolds and Helga Bültmann made 44 collections of vascular plants at a camp near Generator Lake at the southern edge of the Barnes Ice Cap. These specimens are at CAN.

Despite a long history of exploration, collecting, and documenting the vascular flora of east-central Baffin Island, only a small subset of the collection data has been published. Only Taylor (1863), Polunin (1940), and Hainault (1966) provided lists or accounts of species in the flora area. Of these authors, only Polunin (1940) provided vouchers for some species; however, most voucher information was skeletal, giving only a collector’s name, general locality, and sometimes the year, and he rarely cited specific collections. Subsequent authors of regional floras that include this area simply provided dot maps of species distributions, and no specimen details were given (Porsild 1957; Porsild and Cody 1980; Aiken et al. 2007). Substantial effort is needed to match dots with voucher specimens, and the size of the dots makes it difficult—sometimes impossible—to align a dot with a specific voucher collection. It is also unclear whether all of these dots were based on voucher collections; some may have been based on the above literature sources. Saarela et al. (2020b) also noted in their flora of Victoria Island that at least some of Porsild’s dots were based on unvouchered observations.

Complementing our fieldwork in Agguttinni Territorial Park and Clyde River described above, we searched for existing specimens from the study area online and examined Arctic specimens in various herbaria. We manually searched collections at the National Herbarium of Canada at the Canadian Museum of Nature(CAN), the National Collection of Vascular Plants at Agriculture and Agri-Food Canada(DAO), the Marie-Victorin Herbarium at the Université de Montréal(MT), the Bell Museum Herbarium at the University of Minnesota (MIN), the Green Plant Herbarium at the Royal Ontario Museum(TRT), the Fowler Herbarium at Queen’s University(QK), and the monocot collections at the Finnish Museum of Natural History (H). We also manually searched a complete set of images of University of Oslo (O) Canadian Arctic specimens received from O in 2008. The recently completed full digitization of Arctic vascular plants at CAN greatly assisted our search of CAN specimens. Our study also benefited from the recent processing at CAN of several backlog collections, as described in Gillespie et al. (2024).

Through the Global Biodiversity Information Facility (GBIF), we conducted various queries to locate additional specimens from the study area in other herbaria. These included two bounding-box searches for georeferenced plant specimens from the area (GBIF.org [30 Nov 2021] GBIF Occurrence Download https://doi.org/10.15468/dl.5rx9mr; GBIF.org [18 Mar 2024] GBIF Occurrence Download https://doi.org/10.15468/dl.ppa2at), two searches of non-georeferenced plant specimens using locality text (GBIF.org [2 Dec 2021] GBIF Occurrence Download https://doi.org/10.15468/dl.bkyhur; GBIF.org [29 Jan 2023a] GBIF Occurrence Download https://doi.org/10.15468/dl.j888bp), targeted searches for plants collected by Hans Röthlisberger (GBIF.org [29 Jan 2023b] GBIF Occurrence Download https://doi.org/10.15468/dl.k6ccd5) and V.C. Wynne-Edwards (GBIF.org [29 Jan 2023c] GBIF Occurrence Download https://doi.org/10.15468/dl.av2hwm), all non-georeferenced plant specimens from Canada collected in July 1950 (GBIF.org [29 Jan 2023d] GBIF Occurrence Download https://doi.org/10.15468/dl.92czcp), and all Canadian specimens at the University of Colorado Herbarium (GBIF.org [30 Jan 2023] GBIF Occurrence Download https://doi.org/10.15468/dl.49avyj). All records from the latter two searches were manually examined to locate digital specimen records from east-central Baffin Island.

This digital search uncovered specimens at an additional 32 herbaria: the E.C. Smith Herbarium at Acadia University(ACAD), the Herbarium at the University of Alaska Museum of the North(ALA), the Herbarium at the University of Alberta(ALTA), the Deaver Herbarium at Northern Arizona University(ASC), the B.A. Bennett Herbarium in the Yukon (BABY), Institut Botànic de Barcelona(BC), Herbarium Pacificum at the Bishop Museum(BISH), Meise Botanic Garden(BR), Carnegie Museum of Natural History(CM), University of Colorado Museum of Natural History(COLO), Royal Botanic Garden Edinburgh (E), Field Museum of Natural History (F), the Gray Herbarium at Harvard University(GH), Naturalis Biodiversity Center (L), the Shirley C. Tucker Herbarium at Louisiana State University(LSU), the National Herbarium of Victoria at the Royal Botanic Gardens Melbourne(MEL), the University of Michigan(MICH), Missouri Botanical Garden(MO), the William and Lynda Steere Herbarium at the New York Botanical Garden(NY), O, the Chinese National Herbarium at the Institute of Botany, Chinese Academy of Sciences(PE), the Herbier Louis-Marie at the Université Laval(QFA), the Swedish Museum of Natural History (S), the Norwegian University of Science and Technology(TRH), The Arctic University Museum of Norway(TROM), the University of Arkansas at Monticello(UAM), the University of British Columbia Herbarium at the Beaty Biodiversity Museum(UBC), the United States National Herbarium at the Smithsonian Institution(US), the Intermountain Herbarium at Utah State University(UTC), the Royal British Columbia Museum (V), Naturhistorisches Museum Wien (W), and the Herbarium of the University of Zürich (Z).

Most specimens collected prior to 2012 either lack coordinates for their collection locations or have general, often imprecise, coordinates. Where possible, we georeferenced these collection sites using the point-radius method or assigned coordinates based on literature sources; the rationales for each site are provided in Suppl. material 2. We assigned uncertainties for all secondary georeferenced coordinates based on the precision of locality information and the terrain accessible for collecting.

The three largest historical collections required extensive georeferencing: Dansereau and Wynne-Edwards’s 1950 specimens from the head of Kangiqtugaapik [Clyde Inlet], Hainault’s 1965 collections, and Parmelee and Seaborn’s 1967 collections from the head of Inugsuin Fiord.

Dansereau’s 1950 specimens bear pre-printed labels with a general collection location circled (i.e., “Clyde” or “Head of Clyde Inlet”) and usually a short, typed location description, often with an unofficial place name. These locations were also used on Wynne-Edwards’s labels, which were typed in full. A hand-drawn map included in Dansereau’s field notes stored at MT indicates the location of many of these features. Comparing this map to ESRI satellite imagery in QGIS 3.28 (https://qgis.org/) allowed us to georeference most of these specimens (Suppl. material 3).

The historical collections from the head of Inugsuin Fiord either bear no coordinates (Hainault) or general coordinates for the fiord location (Parmelee and Seaborn), but most also include a locality description indicating a collection location with a bearing and distance (in either kilometers or miles) from the head of Inugsuin Fiord. These locality descriptions formed the basis for our georeferencing of these collections (see Suppl. material 2).

The first townsite of Clyde River developed around the Hudson Bay Post established in 1924 on the east side of Patricia Bay. The hamlet was relocated to the northwest shore of Patricia Bay at the mouth of Clyde River in 1970–73 (Forbes 1992, 1993). Pre-1970 Clyde River collection localities (often labeled simply “Clyde” on many historic collections) have been georeferenced to the historic townsite on the east side of Patricia Bay, whereas later collections are georeferenced to the current hamlet location. For vague historical collection localities, such as Clyde River (pre-1950), we applied a standard 5 km radius uncertainty based on the likely distance collectors may have traveled in a day by foot from a known base camp or specific site (minimum 10 km return).

Several collection localities appeared to have inaccurate or very imprecise coordinates. The coordinates (68°15'N, 68°45'W) on Ryder’s Nudlung Valley collections indicate a barren upland area at ca. 500 m elevation about 15 km ENE of the upper reaches of the river valley extending from the head of Nudlung Fiord. In addition to being an unlikely collecting locality, this set of plants is not consistent with this elevation and habitat. We assume that the coordinates are incorrect. Coordinates of 68°15'N, 67°45'W would place the locality between the heads of the two arms of Nudlung Fiord, closest to the southeast arm. We tentatively georeferenced the locality here with a 10 km radius uncertainty that encompasses the entire valley of the southeast arm and the lower reaches of the larger valley at the head of the northwest arm. Coombs’s Coutts Inlet locality has primary coordinates that lie in Baffin Bay about 7 km from the nearest land, which is an unnamed island in the mouth of Coutts Inlet. We re-georeferenced this locality to the southeast tip of that island, with an uncertainty of 10 km encompassing the island, mouth, and adjacent land. Coordinates for Coombs’s Manning Fiord locality appear to be equally imprecise; lacking any additional information, we retained the primary coordinates (which lie in a barren mountainous area surrounded by glaciers, about 4 km from an arm of Coutts Inlet) and recognized that the uncertainty would be at least equally large. Primary coordinates for Bartlett’s “Hoffman Cove” locality (an unofficial name) place the locality on a large island in the mouth of Isabella Bay, whereas expedition notes indicate a location near Cape Raper on the north side of the mouth (Nutt 1938); we assigned coordinates based on the latter location.

Georeferencing of several vague localities is described below. Collections by Smith from “Home Bay” are considered here to be collected at or in the vicinity of the head of Rocknoser Fiord, based on where Snell (1991) argued Smith likely spent most of his time. This locality is likely correct for most, but perhaps not all, specimens, and it should be noted there remains considerable uncertainty as to exactly where these specimens were collected. Martin’s “islands in Clyde Inlet” collection locality most likely refers to a group of islands in the mouth and lowermost reaches of Kangiqtugaapik [Clyde Inlet], scattered over a distance of 36 km (rather than two islands near the head of Kangiqtugaapik). We georeferenced this locality on the largest island in the center of this area with an uncertainty of 18 km encompassing all the islands in the mouth of Clyde Inlet, plus many of the islands at the mouth of Inugsuin Fiord.

With few exceptions, we were not able to study the only known 19^th-century collection, that of Taylor. The majority of specimens are not yet digitized, and the location of his main collection is unknown.

All collection data were amalgamated into a spreadsheet and manually cleaned. Duplicate specimens of a collection were combined into a single field, and additional columns were added in order to maintain the herbaria where individual specimens are deposited.

We also searched iNaturalist (inaturalist.org) for all vascular plant photograph-based observation records within the flora area.

Vascular plant taxonomy and species identification were based on the relevant global taxonomic literature, including Elven et al. (2011), numerous treatments in the “Flora of North America north of Mexico” (Flora of North America Editorial Committee 1993+), “Flore nordique du Québec et du Labrador” (Payette 2013–2024), and recent taxonomic revisions and nomenclatural updates, including Banfi et al. (2022), Barberá et al. (2019), Elvebakk and Bjerke (2024), Gilman and Testo (2015), Jørgensen et al. (2006), Morin (2020), Schönswetter et al. (2001), Solstad (2009), and Wiegleb et al. (2017).

We reidentified or confirmed almost all collections cited, including most duplicates. A small subset of material was confirmed based on review of specimen images available online. Determinations of a few specimens not seen were accepted based on the authority of previous determiners, especially for taxa that are well known and not taxonomically problematic and/or the determiner is an expert in that particular genus or family. We also reidentified or confirmed records based on photographs on iNaturalist when accurate identifications were possible.

We reviewed relevant Canadian Arctic floristic literature (Polunin 1940; 1957, 1964; Porsild and Cody 1980; Aiken et al. 2007). We searched the literature for publications on specific localities or by specific collectors (e.g., Taylor 1863; Hainault 1966). We also consulted an unpublished database of specimen records that was developed and used in the production of maps for the “Flora of the Canadian Arctic Archipelago” (Aiken et al. 2007).

We determined the number of taxa recorded in the flora area in Taylor (1863), Polunin (1940), and Hainault (1966). We also counted the number of taxa that were mapped within the flora area in Porsild (1957), Porsild and Cody (1980), and Aiken et al. (2007). We then determined the number of taxa in these floras based on current accepted taxonomy.

To facilitate floristic analyses, we divided the study area into four main areas. From south to north, these are Home Bay, the Clyde area, Agguttinni TP, and Coutts Inlet (Fig. 5). The Home Bay region includes all fiords draining into Home Bay and adjacent inland areas. The Clyde area includes Kangiqtugaapik [Clyde Inlet], Inugsuin Fiord, Isabella Fiord, and the hamlet of Clyde River. Agguttinni TP includes all areas within the outer boundary of the park, including Inuit Owned Land that is not part of the park (for more details, see Gillespie et al. 2024). The Coutts Inlet region encompasses the area surrounding the Coutts Inlet fiord complex. No collections from the large area between Agguttinni TP and Coutts Inlet were found. We further grouped collection sites into 38 main localities: 21 in Agguttinni TP, eight in Home Bay, seven in the Clyde area, and three in Coutts Inlet (Table 1, Fig. 5). The Agguttinni TP main localities are identical to those in Gillespie et al. (2024), except for Gee Lake and Generator Lake, which were treated there as sublocalities under Barnes Plateau S. Sites visited by us in Agguttinni TP in 2021 are numbered in the order visited, with the day of the month followed by a site number for that day, following Gillespie et al. (2024). All other collection sites are numbered by locality with an acronym and site number. Locality data for all sites are provided in Table 1.

To characterize floristic diversity patterns, we scored the presence of species/taxa in the four main areas and the 38 main localities. For the 10 most well-collected localities, we estimated the area (km²) and determined the number of days spent collecting (i.e., the number of days on which collections were made). The locality area was calculated using QGIS 3.28 to draw convex hulls around the sites at each locality, exporting into Google Earth, and using the polygon measuring tool to determine the area of each locality. For Kangiqtualuk Uqquqti, we excluded site 31.1, which was distant from the other sites and where we spent very little time; no species unique to this locality were found at site 31.1. For Inugsuin Mouth, a locality with a single collecting site, we calculated the area based on the estimated coordinate uncertainty, then subtracted the area occupied by large bodies of water such as fiords. For Clyde River we excluded two distant and/or very imprecise sites (CR-4 and CR-5, representing four collections) and calculated the area based on the estimated coordinate uncertainty of the site with the largest uncertainty (CR-1), minus the area occupied by large bodies of water; this area encompassed all other sites. Two species found at these sites, Silene sorensenis and Taraxacum ceratophorum, were not collected elsewhere in the Clyde River and were excluded from there in our diversity analyses.

We also scored the presence of taxa in the three ecoregions in the flora area. Based on the map in Ecological Stratification Working Group (1995), we included 13 localities in the Baffin Coastal Lowlands ecoregion (ordered north to south): Scott Inlet, Remote Peninsula, Refuge Harbour, Arviqtujuq Kangiqtua NW, Ravenscraig Harbour, Niaqurnaaluk–Qassialuit, Kuugaaluk, Tasialuk N, Clyde River, Cape Hewett, Inugsuin Mouth, Hewett Cache, and Hoffman Cove. We included Gibbs Esker, Marble Lake, Gee Lake, Generator Lake, and McBeth Valley in the Baffin Uplands ecoregion. The remaining 15 localities were included in the Baffin Mountains ecoregion, including two fiord heads, Kangiqtualuk Agguqti and Kangiqtualuk Uqquqti, that are not included in this ecoregion on the above map but should be based on their climate, low elevation, and mountainous surroundings.

We generated maps showing the locations of all collecting sites using QGIS 3.28. Based on our dataset, we calculated and plotted the number of collections by decade, the cumulative number of collections by year, and the cumulative number of taxa by year. We generated heat maps using QGIS 3.28 to show the density of collections and species richness at each locality.

We explored the relationship between how widespread a taxon is and how many times it was collected by plotting the number of localities where a taxon was found against the number of collections for that taxon. We also performed linear regression between these variables using Excel.

To characterize floristic affinities of the flora, we recorded the global distribution of each taxon following Elven et al. (2011). Distributions were modified based on new information and for simplicity. The pattern Amphi-Atlantic (W) is restricted here to occurrences in Svalbard and/or Iceland in Europe and cited as Amphi-Atlantic (Svalbard) or Amphi-Atlantic (Iceland). Distribution patterns were grouped into major distribution types for analyses.