Some Natural Floristic Areas in Boreal America Author(S): Hugh M

Home , Aquilegia brevistyla, Arnica lessingii, Continental Ranges, Ribes glandulosum, Salix commutata

Some Natural Floristic Areas in Boreal America Author(s): Hugh M. Raup Reviewed work(s): Source: Ecological Monographs, Vol. 17, No. 2 (Apr., 1947), pp. 221-234 Published by: Ecological Society of America Stable URL: http://www.jstor.org/stable/1943266 . Accessed: 05/04/2012 11:13

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecological Monographs.

http://www.jstor.org SOME NATURAL FLORISTIC AREAS IN BOREAL AMERICA

HUGH M. RAUP Harvard Univer sity Cambridge, Massachusetts

IN A SYMPOSIUM ON

ORIGIN AND DEVELOPMENT OF NATURAL FLORISTIC AREAS WITH SPECIAL REFERENCE TO NORTH AMERICA TABLE OF CONTENTS

PAGE INTRODUCTION...... 223

GEOGRAPHIC AFFINITIES OF TIlE BRINTNELL LAKE FLORA: Preliminary Remarks, Including Notes on Local and Cosmopolitan Species ...... 223 Wide-ranging Forest Species ...... 225 Wide-ranging Arctic-alpine Plants ...... 226 Plants of the Timber Line Region ...... 228 Alaskan-cordilleran Species ...... 228 Cordilleran Species ...... 229 Discussion ...... 230

APPICATION OF II1ULTEN 'S HYPOTHESIS ...... 230

ORIGINS OF THE BRINTNELL LAKE FLORA ...... 233

GENERAL CONSIDERATIONS ...... 234

LITERATURE CITED ...... 234

[222 1 SOME NATURAL FLORISTIC AREAS IN BOREAL AMERICA

INTRODUCTION The actual number of these patterns depends of course upon the degree of refinement that enters into A comprehensive discussion of natural phytogeo- the selection of criteria. Needless to say, the scanty graphic areas in boreal America, even if I were able existing knowledge of geographic factors in boreal to produce it, would not be compressible into the America makes it impossible to carry the refinement time allotted for my contribution to this symposium. very far. Comparison of the maps soon brings out, I have, therefore, chosen to outline a single problem however, two phytogeographic transition zones which which has led me into a study of some boreal plant appear to overshadow any others that appear. These ranges. These ranges, when they are superposed and are the arctic timber line, and the coastal mountains sorted into patterns, supply evidence of natural flo- of southeastern Alaska, British Columbia, and west- ristic areas, and afford some implications of the his- ern Washington. Some indication of the significance torv and manner of development of these areas. of these "boundaries" may be had from the following I have had occasion, during the past few years, to figures. Of the 271 species considered in the maps study- the vegetation of the southwestern part of the 199 approach or cross the arctic timber line, and 164 District of Mackenzie, in northwestern Canada. In- of them are either stopped by it or cross it to a lim- cluded in this area are the eastern slopes of the ited distance. Nearly all 283 of the Brintnell Lake Mackenzie Mountains, a range extending in a great plants extend to the western coastal mountains or are northwestward from the Liard River into central forests, but about 196 of them avoid all or part of Yukon (Fig. 1). The highest altitudes attained in these habitats. In view of the obvious phytogeo- these mountains are as yet unknown, but peaks 7,000- graphic importance of the coast range and the arctic 9,000 feet above sea-level are not uncommon. Tim- timber line, I have drawn up a tentative classification ber line in the southern part of the range is at about of range patterns based in large part upon the plants' 4,500 feet, but is somewhat lower at the north. The behavior with relation to them. A third geographic range as a whole is separated from the Rockies, Cas- boundary is not so well defined in all cases, but is siars, Coast Ranges and Richardson Mountains by sufficiently clear to necessitate recognition. This is forested valleys and plateaus. marked by the eastern ranges of the Cordillera. The flora of the Mackenzie Mountains is still very About 25 percent of the Brintnell Lake flora does not incompletely known. Prior to our own collections extend east of the Rocky Mountains, or has limited only a handful of plants had ever been gathered extensions into the northern great plain of the con- there. In 1939 we spent two months at Glacier or tinent. Brintnell Lake, in-the central part of the range, and About 4.2 percent of the flora at Brintnell Lake in 1944 the route of the Canol pipeline road was (12 spp. and vars.) is composed of plants recently examined by Mr. A. E. Porsild who made extensive described as new, recently described from the lovyer collections (Porsild 1945). Also in 1944 Dr. V. C. Mackenzie district, or known in North America from Wynn-Edwards made some notable collections in this single area (Raup 1947). Tentative ranges can the eastern fringes of the mountains, and also along be drawn for a few of these, but so little has been the eastern part of the pipeline road. The only discovered about them that I have preferred to keep place, however, in which anything like a comprehen- them in a separate category, and I have not included sive collection has been made on the eastern side of maps of them. They are as follows: the mountains is at Brintnell Lake. The vascular flora of this area proved to be comparatively small- Picea glauca var. Porsildii Raup, Poa Brintnellii Raup, 283 species and varieties-with a certain amount of Carex Soperi Raup, Salix Barrattiana var. marcescens Raup, Salix Bebbiana var. depilis Raup, Salix glauca rather poorly defined endemism. var. perstipula Raup, Lychnis brachycalyx Raup, Saxi- fraga sibirica L., Rosa acicularis var. oucurbiforrmis GEOGRAPHIC AFFINITIES OF THE Raup, Arnica alpina (L.) Olin, Arnica Snyderi Raup, BRINTNELL LAKE FLORA Antennariwasp. PRELIMINARYREMARKS, INCLUDINGNOTES ON At the other extreme in range extension are three LOCALAND COSMOPOLITANSPECIES cosmopolitan species: Cystopteris fragilis, Equisetum In order to study the geographic affinities of the arvense, and Carex aquatilis s. 1. These are among Brintnell Lake flora I have mapped the ranges, in the most wide-ranging plants in boreal America, and Alaska, Canada and the northern States, of 271 of the first two are cosmopolitan throughout much of the 283 species and varieties collected there.' the world. In Canada and Alaska they thrive in When the maps are superposed it becomes possible nearly all parts of the arctic, alpine, and forest re- to designate a number of rather well-defined patterns. gions, without regard to boundaries that are limiting 1 The base map used for these ranges is by J. Paul Goode, to most other plants. Of the three, Equisetum arvense and is copyrighted and published by the University of Chicago Press. seems to have the most varied stations. 224 IIUGIIM. RAUP Ecological Monographs

~~~~~6~~~~~~~~?"~~~~~~~~~~~~~~~ ~o .17 o

~~~~~~~i~~~~~~~~~~~~~il)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

N2C ~~~~~~~~~~~~~~~~ C~~~~~

-- C

The remaining' 268 species and varieties can be timber line, or in both the arctic lowland and in the arran-ed in five general categories: (1) species whose cordilleran alpine areas; (3) an intermediate group principal areas are in the region of the Canadian of wide-ranging subarctic species which cannot be coniferous forest, arid whose ranges are wi(le in this placed with either of the above, but which appear to r001( 11, (xtend(fing for the most part all the way across have their greatest concentration on or near the arctic the continnrt; (2) arctic or artic-alpine wi(le-ranging timber line; (4) a group of Alaskan-cordilleran spe- 5I)((i(5 with ranges eoliccntrated( north of the arctic cies wdhich have onlyv limited extensions east of the April, 1947 SOME NATURAL FLORISTIC AREAS IN BOREAL AMERICA 22 5 Rocky and Mackenzie Mountains, and most of which The other 11 species either avoid the coastal forests are alretic-alpine in general affinities; (5) a cordil- altogether, or enter them only in southern Alaska or leran group, most of which are alpine but not arctic, Washington. Thus they are more strictly continental and whose ranges are concentrated in the Rocky in the f ar west (Fig. 2B): Mountains and coast ranges. Further subdivisions Colamnagrostis neglecta, Car-ex vaginata, Habenaria ob- of these five categories can be made by noting first tusata, Salix planifolia, Belula glandu losa, Ranunculus the actual relation of the wider ranges to the arctic Gmnelinivar. Purshii, Ranunculus lapponicus, Draba lan- tiniber line, and second the relation of the ranges to ceolata, Potentilla fruticosa, Hedysarumn alpinurn var. the western coastal mountains or the Pacific slope arnericanum, VacciniumnVilis-Idaea var. minuts. forests. The following discussion will follow the above outline. The remainder of the widespread forest species are more In order to comapare the range maps more easily I rigorously limited by the timber line, though have drawn lines connecting marginal stations in such some of them occur far north on the Labrador coasts, a wvav as to outline the limits to which the various southern Baffin Island, or in Greenland. There are 80 of species are now known to extend. For common, much them, divisible into four groups on the basis of their geographic relation to the Pacific collected plants this can be done without difficulty, slope forests. The of but in cases of rare or poorly collected species the first these groups (Fig. 3A) contains 30 outlines of geographic limits must be regarded as species that appear in the coastal forests both in tentative. The most difficult cases are those with southern Alaska and in British Columbia and western stations separated by hundreds of miles. Here it Washington. Several of them, perhaps half a dozen, is often impossible to decide whether to consider the do not appear to reach the western coast of Alaska. ranges as continuous or disjunct. Of the ranges Dryopteris Linnaeana, Equisetum fluviatile, Equiseturn mapped in this study I have designated only nine as palustre, Lycopodium annotinum s. 1., Selaginella sela- being discontinuous, although it is possible that a ginoides, Juniperus conmmunis var. mtontana, Agrostis few others might be included. In defense of my scabra, Poa nemnoralis, Carex leptalea, Carex canescens, tendency to reduce the number of disjunct ranges as Carex Garberi s. 1., Juncus balticus s. 1., Smilacina stel- much as possible I can only say that one of the re- lata, Habenaria hyperborea, Salix Bebbiana, Arenaria lateriflora, Stellaria calycantha s. sults of the collecting and exploration in arctic and 1., Actaea rubra, Ribes lacustre, Ribes triste, Shepherdia canadensis, Cornus subarctic America during the past twenty years has canadensis, Cornus stolonifera, Moneses uniflora, Pyrola been to fill in and make continuous many ranges that asarifolia var. incarnata, Pyrola minor, Heracleum lana- were formerly thought to have separate eastern and tunt, Arctostaphylos Uva-ursi (inel. var. coactilis), Lin- western components. naea borealis s. 1., Viburnum edule.

WIDE-RANGING FOREST SPECIES The second group, of 13 species (Fig. 3B), occurs The most extensive plant ranges in Canada and in some of the coastal forests of southeastern Alaska, Alaska, with the exception of the three cosmopoli- but becomes more continental in British Columbia and tans already mentioned, are exhibited by a small Washington. A few of them, probably not more than group of about 19 species. These plants do not stop four, fail to reach the western Alaskan Coast. at the arctic timber line, but reach far out into the Equiseluni sylraticum. s. 1., Lycopodiurn clavaturn var. tundr a, particularly in the eastern arctic. Most of mnonostachyon, Lycopodiutrn cornplaniatuni s. 1., Carex them are spread throughout the timbered parts of gynocrates, Eriophorum brachyantheruvn, Cypripedi um Alaska. E ight of them (Fig. 2A) have the widest passerhinulm,Populuts Tacanah-acca, Populas tremulo ides, ranges because they also enter the forests of the Pa- Gcoca mdon lividum, Parnassia pal ust) is var. neogaea, cific slolpe frome southern Alaska to western Wash- Ribes hudsouianurn, Dryas Drunowonudii,Rosa acicularis s. 1. ing ton: Eq ui.setumn.scirpoic(sIs, Poa pratenxlsis s. 1., Corallorrhiza A third group of 14 species (Fig. 3C) is still more trifida, Alnus crispa s. 1., Rubvs Chainaeniorus, Epilobium continental in range, reaching the coast timber only angl stifoliIumu S. 1., Pyrola seciunda, Ledulu grocnlandi- in southwesterni British Columbia and western Wash- CUM)1. ington. They avoid the AN-etslope forests of southern

Nfr N

K 'A K /

A / B' /

K MX - 226226 HUGH A. RAUP Ecological Monographs ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Vol.17, No. 2

FIG. 3. Wide-ranging forest species whose northern limits are. at or near the arctic timber line.

Alaska, and none of them extends to the western arctic archipelago, 26 of them are commonly found coasts of Alaska. Four of these plants seem to have far south of the arctic limit of trees (Fig. 4). In disjunct ranges: Arctostaphylos Uva-ursi var. adeno- so doing they achieve the most extensive ranges of tricha, Betula papyrifera var. commutata, Arabis all our arctic-alpine plants and, as did the forest spe- Holboellii, A ntenuwria subviscosa. I have included cies that crossed the timber line, they approach a them here because a hypothetical joining of their cosmopolitan distribution in Canada and Alaska. parts would result in ranges similar to the others in The most extensive ranges are those of 7 species Fig. 3C. (Fig. 4A) which occur not only on the Rocky and Carex aurea, Betula papyrifera var. commutata, Arenaria Mackenzie Mountains but also on the coastal ranges dawsonensis, Arabis Drummondii s. 1., Arabis Holboellii, all the way from the Alaska Peninsula to Washing- Geum macrophyllumn var. perincisum, Potentilla norve- ton. They extend southward in the central plains at gica, Pyrola virens, Aretostaphylos Uva-ursi var. adeno- least as far as Lake Athabaska. tricha, var. kamtschaticus, Erigerorn Erigeron angulosus Equisetum variegatum, Lycopodium Selago, Festuca elatus, Antennaria subviscosa, Se'necio indecorus, Senecio brachyphylla, Poa alpina s. 1., Carex scirpoidea, Salix pauperculus. arctica s. 1., Empetrurn nigrum. The fourth, and almost strictlv continental, group Another group, of 8 species, is predominantly con- contains 23 species (Fig. 3D). With one or two pos- tinental south of the mountains in southeastern sible exceptions these plants avoid the Pacific slope Alaska (Fig. 4B). They occur on the coast and forests entirely, and all but five of them fail to reach Alaska ranges of southern Yukon and Alaska, as the west coasts of Alaska. A few are even more re- well as on the mountains around the Lynn Canal, but stricted and do not extend west of the northern in British Columbia and Alberta they are largely Rocky Mountains. confined to the Rockies and Selkirks. Picea glauca s. 1., Picea mariana, Agropyron trach'yeau- Poea glauca, Trisetum spicatum s. 1., Carex capillaris s. 1., lum var. novae-angliae, Agropyron trachyeautlm var. Juncus castaneus, Polygonum tri- unilaterale, Calamagrostis canadensis var. robusta, Carex viviparum, Saxifraga cuspidata, Dryas integrifolia, Astragalus alpinus. concinna, Carex diandra, Carex media, Goodyera repens S. 1., Li.stera borealis, Salix myrtillifolia, Alnus tenui- A third group of these extremely wide-ranging folia, Betula occidentalis, Aquilegia brevistyla, Rorippa arctic-alpine plants contains 11 species (Fig. 4C) islandica var. microoarpa, Ribes glandulosum, Ribes and is almost completely continental in the west. oxyacanthoides, Anzelanchier florida (incl. A. humilis), Fragaria glauca, Rubus acaulis, Rubus strigosus, Mer- Southeast of the Prince William Sound region of tensia paniculata, Valeriana septentrionalis. Alaska these plants are confined to the more northerly ranges of the coastal mountains, and to the Rockies 'WIDE-RANGINGARCTIC-ALPINE PlANTS and Selkirks. Onlv two of them come out to the There are 66 wide-ranging arctic-alpine plants in coast in northwestern Washington. the Brintnell Lake flora. As among the forest spe- Dryopteris fragrans s. str., Calantagrostis purpurascens, cies, one can first divide them into two groups ac- Arenaria verna var. pubescens, Ranunculus hyperboreus. cording to their actual relation to the arctic timber Cardamine pratensis s. 1., hi)aba cinerea, Chrysosplenium line. In spite of the fact that the principal ranges tetrandrurn, Saxifraga aizoides, Potentilla sivea s. 1., of nearly all of theni extend throughout much of the Pyrola grandiflora s. 1., Taraxacum lacerum. April, 1947 SOME NATURAL FLORISTIC AREAS IN BOREAL AMERICA 227

FIG. 4 Wide-ranging arctic and arctic-alpine species wshich extend far south of the arctic timber line.

For the remaining 40 arctic-alpine plants the arctic, Another group of 10 species ranges the coastal timber line appears to be a more formidable barrier. mountains only in southern and southeastern Alaska They either stop at the timber line or come only a (Fig. 5B). In British Columbia and Alberta they short distance south of it. Like the preceding group reach their southern limits in the Rockies and Sel- they can be arranged in three divisions depending kirks, and are strictly continental. upon whether they extend to the mountains of the HierochloW alpina, Carex bipartita, Eriophorum Scheuch- Pacific coast. zeri, Luzula Wahlenbergii, Salix reticultat, Salix Bi-. First are 12 species of wide range on the coastal chardsonii, Oxytropis foliolosa, Cassiope tetragona, Pedi- mountains (Fig. 5A). One of these, Dryas octo- cularis oapitata, Pedicularis sudetica. petala, has a broken range. It is widespread in the Cordillera and reappears in northwestern Greenland. Finally, a group of 18 arctic-alpine plants, with one exception in western Washington, avoid the Agropyron Luzula latiglumne,Poa aretica, spicata, Oxyria coastal mountains entirely (Fig. 5C). digyna, Silene acaulis var. exscapa, Cardamine bellidi- folia, Saxifrtgaa oppositifolia, Saxifraga rivularis, Dryas Calamagrostis lapponica s. L, Carex nermbranacea, Lu- octopetala, Epilobium latifolium, Epilobium anagallidi- zula confusa, Arcnaria humifusa, Arenaria Rossii, Lych- folium, Erigeron unalaschkensis. nis apetala, Stellaria longipes var. Edwardsii, Ranuncu-

FIG. 5. WXide-ranlging aretice and aretic-alpl)iIe species whose southerni limits are at or near the aretic timibl)er line. 228 228 HUGH M. RAUP Ecological ~~~~~~~~~~HGHM APVol. Monogiaphs17, No.2 lats nivalis, Draba fladnizei.s~is var. heterotricha, Draba Carex deflexa, Carex nmicroglochin, Tofieldia palustris, glabella, Saxifraga nivalis, Potentilla emarginata, Oxy- Salix glaitca s. 1., Arenaria sajanensis, Sagina Linnaei, f ropis Maydelliana, Epilobiumr davitricum, Vaccinium Parnassia Kotzebu-ei, Hedysarum Mackenzii, Oxytropis uliginosum var. alpinumn, Pedicularis lanata, Erigeron hudsonica, Oxytropis ixodes, Epilobium lactiflorum. (riocephalus, Crepis nana. Cornus canadensis f. purpurascens, Arctostaphylos rubra, Gentiana propinqua, Pedicularis labradorica, Veronica Two features of the arctic-alpine plant ranges alpina var. unalascheensis, Pinguicula villosa, Solidago should be noted especially. Nearly all of them ex- multiradiata, Antennarta isolepis, Taraxacum cerato- tend to the western Alaskan coasts, in contrast to the phorum, Taraxacumn lapponic-um. forest plants among which 39 do not accomplish this. Second, there are a few of the arctic plants, probably ALASKAN-CORDILLERAN SPECIES not more than 8, which do not occur in the Rocky An arrangement of Alaskan and cordilleran com- Mountains at all, but find their southern limits in ponents of the Brintnell Lake flora on the basis of the Mackenzie Mountains and in southern Yukon or their relation to timber line proves to be imprac- Alaska. ticable. Of the 72 species counted in these categories, only ten can be regarded as forest plants, and most PLANTS OF THE TIMBER LINE REGION of these ten are likely to be found far above or be- There are 31 species which constitute something of yond the timber line. All the others are plants com- al enigma in that, although they have wide subarctic monly found in the alpine or western arctic tundra. ranges across most of the continent, they cannot However, a great many alpine plants of the Macken- properly be classed either with forest or arctic-alpine zie and northern Rocky Mountains are commonly categories. They extend into the tundra, but avoid found in the timbered areas at low levels. About most of the high arctic regions. At the same time half of the western alpine plants were found below they are common to abundant in the northern part of timber line at Brintnell Lake, and no doubt further the forest. They occupy a wide range of habitats, search would yield more of them. It is not possible, although none of them is characteristic of rich woods, therefore, safely to classify them in relation to tim- and about two-thirds of them are perhaps most ber line. abundant in muskegs or wet tundra. Most of them I have arranged the 72 species in two main groups, appear in the northern Rockies, but two or three do the first of which ranges throughout central and net come south of the 60th parallel. All but two or western Alaska, many of them north to the arctic three reach the western Alaskan coasts. coast. Most of them come into the mountains far- A glance at Fig. 6 will show that most of these ther south, so that I have called them the Alaskan- "timber line" species are continental in the more cordilleran group. There are 45 of these, 16 of which southern parts of their ranges. I have arranged them (Fig. 7A, C) range southward into the western United in two groups, carrying out the plan used in other States. The whole group can be separated on the categories. The first group contains 10 species which basis of its coastal extensions. First is a wide-rang- occur in the coastal strip as far south as southeastern ing series of 11 species, all of which occur on the Alaska, while beyond that area all but 3 retreat in- coastal mountains of southeastern Alaska, and miany land (Fig. 6A). One species of disjunct range is of them on the coast ranges farther south (Fig. 7A). included here, Lycopodiutm alpinum. Carex physocarpa, Luzula arcuata, Lloydia serotina, Woodsia ilvensis, Lycopodium alpinum, Calamagrostis Zygadenus elegans, Arabis lyrata var. kamchatica, Sedum *canadensis var. Langsdorfii, Cerastium Beeringianum, roseum var. integrifolium, Viola epipsila, Gentiana mon Ane e parviflora, Ancnmone Richardlsonii, Sibbaldia glauca, Aster sibiricus, Artemisia arctica s. 1., Petasites proca wnbens, A ndroineda Polifolia, Vaccinium micro- frigidus. ('carpumI, Vaccinqamn Wliginosum. Allied to these, but with their southern extensions A second group, of 21 species, avoids the coastal terminating in Yukon or the Alberta and British Strip except in southwestern British Columbia, where Columbia Rockies, are 15 species, all of which are iftour of them come out to Vancouver Island (Fig. coastal in southern or southeastern Alaska (13). Three species with discontinuous ranges are in (Fig 7B). this groul): Sagina Linnaei, Epilobium lactiflorum, Cryptograrnnia crispa vair. sitchensis, Arctagrostis arlun- ()xytropis ixodes. dinacea, Festuca altaica, Poa paucispicula, Carex loliacea,

FIG. 6. ' Timbher-line species.'' April, 1947 SOME NATURAL FLORISTIC AREAS IN BOREAL AMERICA 229

FIG. 7. Alaskan-cordilleran species.

Salix alaxensis s. 1., Salix arbusculoides, Salix pulchra, coast proper, either in Alaska or Mackenzie, and their Betula papyrifera var. humilis, Aconitum delphinifolium, eastward extensions into the northern plains are less Anemone narcissiflora, Saxifraga punctata s. 1., Bosch- numerous and less pronounced. All of them range niakki rossica, Campanula lasiocarpa, Arnica Lessingii. southward into the western United States. They fall Two groups are more strictly continental in the rather clearly into two groups: one with a strong region southeast of Prince William Sound, although coastal relationship and the other just as strongly a couple of species appear at the coast in the Puget continental. Sound area. Five species range southward into the Sixteen of the 27 follow the coastal mountains from United States (Fig. 7C). the Alaska Peninsula to western Washington, although a couple of them retreat inland a short dis- Delphiniumi glaucum, Potentilla uniflora, Myosotis al- tance in the latter region. They extend throughout pestris s. 1., Artemisia Tilesii s. 1., Senecio lugens. the northern Rocky Mountains, but most of them are The remainder, 14 species, extend from Alaska to confined to the southern ranges of Alaska. None of the northern Rocky Mountains, and avoid the coastal them reaches the lower Yukon valley (Fig. 8A), mountains excel)t west of Prince William Sound although two have been found at the Mackenzie River (Fig. 7D). delta. Our Mackenzie Mountain collections have con- stituted northeastern range extensions for most of Larix laricina va-r. alaskensis, Carex nesophila, Carex these species. podocarpa, Corydalis pauciflora, Draba longipes, Saxi- fraga radiata, Astragalus frigidus var. littoralis, Luzpinus Abics lasiocarpa, Carex nardina var. Hepburni i, Carex arcticus. Oytropis hyperborea, Oxytropis pygmaea, phacocephala, Carex pyrenaica, Juncus Dru int ondii, Gcntiana arctophila, Polentonium acutiflorum, Anten- Salix Barclayi, Salix commutata, Salix Scovelriana s. I., naria mvonocephala, Taraxacum alaskanum. Rn uncults Eschscholtzii, Draba nivalis var. clongata, Parnassia fimbriata, Saxifraga Lyallii, Hippuris nion- CORDILLERANSPECIES tav a, Phyllodoce glanduliflora, Senecio t riang ularis. Hieracintm gracile.. The last category of ranges to be considered in- volves 27 plants that are more strictly cordilleran The other 11 cordilleran plants avoid the coastal than any of the above. They do not reach the arctic strip except in the south, where four of them appeal

Fic. 8. Cordilleran s1)cies. Ecological Monographs 230230 hUGHHUGH Al.M. RAUP ~~~~~~~~~~~~~~~~~~~~~~~~Vol.17, No. 2 in the Puget Sound region. A few of them (4) reach alpine, and alpine plants is about 128, or about 45 central Alaska, and two are found at the Mackenzie percent of the total vascular flora, and that of forest delta (Fig. SB). This group contains the very few species is 109, or about 38 percent of the whole. Rocky Mountain plants which have their known Although the flora as a whole is strongly conti- northern limits in the Mackenzie Mountains. nental in character, a great many of its species are able to live in the damp forests of the north Pacific Poa Buwkleyana, Carex albonigra, Carex atrata ssp. slopes or on the coastal mountains. The range maps atrosquama, Salix BRarrattiana, Draba McCallae, Draba praealta, Potentilla diversifolia s. 1., Erigeron jucundus, give a rough index to the degree of continentality Antennaria media, Arnica mollis, Arnica alpina ssp. which is achieved by the Brintnell Lake flora. There tomentosa. are about 140 species (49.4 percent) which avoid the coastal forests and mountains completely or nearly DISCUSSION so, and are the most rigorously continental plants in Table 1 gives a summary of the range patterns just our flora so far as the west is concerned. Another 56 described. Some generalizations from these figures (19.8 percent) avoid the British Columbia and Wash- will serve to bring out further the regional character- ington coasts, but occur in southeastern Alaska. The istics and affinities of the Brintnell Lake Flora. remaining 87 (30.7 percent) inhabit, in addition to Fully 70 percent or about 199 of the species, have the continental ranges, the coastal strip from Alaska wide ranges across the continent, most of them ex- to Washington. tending from Alaska to the Gulf of St. Lawrence. Table 2 will summarize the above generalizations. Approximately 25.4 percent are Alaskan or cordil- TABLE 2 leran, while the remaining 4.2 percent are endemic or local. No. of TABLE 1 spp. and Percent vars.

Endemics and of limited known 12 No. of plants range...... 4.2 spp. and Percent vars. Cosmopolitan species...... 3 1.0 of limited known ...... 12 4.2 Endemics and plants range Plants of wide range across the continent ...... 199 70.3 Plants with ranges mainly in Alaska and the Cordillera...... 72 25.4 Cosmopolitan in Canada and Alaska...... 3 1.0 Plants of the forested country...... 109 38.0 Wide-ranging forest species Arctic, arctic-alpine, or alpine species...... 128 45.0 38 13.4 Reaching coasts of Alaska, B.C., and western Washington. "Timber-line" ranges...... 31 10.9 Reaching coasts of southeastern Alaska...... 13 4.6 southeastern Alaska to ...... 48 17.0 Avoiding coasts, Washington Strictly continental with regard to S.E. Alaska, B. C., and Washington ...... 143 49.4 Wide-ranging arctic or arctic-alpine species Plants that avoid the coasts of B. C. and Washington, but 6.7 Reaching coastal mountains, Alaska to Washington...... 19 occur in S. E. Alaska ...... 6... 5 19.3 Alaska 6.4 Reaching coastal mountains of southeastern ...... 18 Plants whose ranges extend into the coastal strip of S.E. 10.2 Avoiding coastal mountains, S.E. Alaska to Washington .. 29 Alaska, B. C., and Washington...... 87 30.7

Wide-ranging "timber-line" species Plants that do not reach the western coasts of Alaska ...... 83 30.0 Reaching coast of S.E. Alaska...... 10 3.5 Avoiding coasts, S.E. Alaska to Washington.21 7.4

Alaskan-cordilleranspecies APPLICATION OF HULTtN'S HYPOTHESIS Reaching coasts, Alaska to Washington...... 11 3.9 Reaching coasts of southern and southeastern Alaska...... 15 5.3 In view of the geographic position of the Macken- Avoiding coasts, S.E. Alaska to Washington...... 19 6.7 zie Mountains in northwestern America, their recent Cordilleranspecies botanical exploration, the present isolation of their Reaching coasts, southern Alaska to Washington...... 16 5.6 alpine flora by the development of surrounding for- Avoiding coasts, southern Alaska to Washington...... 11 3.9 ests, and the simnplicity and apparently incipient endemism in their flora, it has occurred to me that this TOTALS...... 283 99.8 flora might be used as a test of theories advanced some years ago (1937) by Dr. Eric Hulten concern- The wide-ralnging plants fall into two main groups, ing the origin and development of all our boreal aletic-alpine amid foijest species. In the first there biota. are 66 species, while in the second there are 99. A Whether or not there were ice-free land surfaces third group of 31 species (10.9 percent) are inter- ill the Mackenlzie Mountains during the Late Wiscon- mediate between the first two, and three (1 percent) Sill (W3) glacial episode is not certain, although are cosmopolitan. This is not a complete statement there is some evidence, both geological and botanical, of the pro portions of arctic-alpine and forest ele- that suggests then. If ice-free summnits were pres- inelts, however, because approximately 62 of the 72 ent, their flora miust have been composed of the Alaskan and cordilleranl plamits are of prevailingly aretic-alpimie renlilanlts of a wide pre-Late Wisconsin alpimie or arctic affinity and should be added to the dispersal of these plalts. A few species found at wide- angimig ones. Thus the total of arctic, arctic- Brintnell Lake suggest this. There are three that April, 1947 SOME NATURAL FLORISTIC AREAS IN BOREAL AMERICA 231 have extensive ranges in Eurasia but are not known Washington and radiate along the American coast elsewhere in North America, and there are eight or along the Rocky Mts. to Alaska. Of the plants rather poorly defined endemics or suspected endemics. discussed . . . no groups could be formed having Considering the small number of these unique plants, their centres in northern Europe or western Siberia, and the paucity of the Brintnell Lake flora as a or in North-Eastern America or in the country be- whole, it is hardly justifiable to set up the Mackenzie tween Yukon Valley and the Great Lakes" (Hulten Mountain area, at least so far as the eastern slopes 1937, p. 25). are concerned, as a refuge equivalent to those of Hulten immediately draws a correlation between Beringia, the North Pacific coasts, or the Yukon the distribution of his centra and the distribution of Plateau. It seems more reasonable to assume that ice during the Pleistocene. No centra could be found although a few alpine plants may have persisted at in areas that were covered by ice during the maximum Brintnell Lake through Late Wisconsin time, most of advance of the glaciers; and the existing floras have the flora has arrived during and since the retreat of all been derived from areas of refuge close to the the last valley glaciers. The findings of Porsild on ice. It is maintained, on genetical grounds, that the the western slopes of the mountains may require ability of the species to disperse themselves from modification of this view. their refugia has not been uniform. Those confined It is unnecessary to review in detail all of the rea- to small refugia, under difficult climatic conditions soning and implications of the theories advanced by and in small populations, are considered to have been Hulten, for this has already been done by several so depauperated of biotypes that they have been ex- students (see Stebbins 1942, Raup 1941, Halliday ceedingly slow to spread. Others had large areas and Brown 1943, Cain 1944). The following is a and populations, either within the generally glaciated brief outline of the broader geographic aspects, with regions or south of the ice, so that they retained their a few critical notes that have grown out of the pres- inherent variability and aggressiveness and could ent investigation. quickly invade lands freed of ice. In connection with his floristic studies of Kamt- It is thought that the so-called "Linnaean species" chatka and the Aleutian Islands, Hulten mapped, of the present boreal flora originated in the last great sometimes in detail and sometimes by limits, the interglacial or earlier, and that large numbers of them ranges of hundreds of arctic and boreal species. He achieved wide dispersals during that time. Their organized this factual material on the basis of what present areas are looked upon as reductions from he called "equiformal areas." That is, when large these wide ranges, with post-Glacial re-expansions numbers of ranges are superposed they fall into a that have been conditioned by the amount of depau- group of patterns which are more or less clearly de- peration suffered during the maximum ice advance. fined geographically, and "equiformal" within them- Hulten's arrangement of equiformal areas is open selves. Each equiformal area shows a region of to modification or criticism along three lines. First, concentration in number of species, which is called it is possible to make additions to his lists, at least its "centre." Since it is assumed that each species in the American boreal flora. The Brintnell Lake has acquired its present range by dispersal from a region is found to have approximately 45 species not point of origin or survival, the regions of concentra- discussed by him at all. When these are sorted into tion within the equiformal areas are regarded as patterns they have a distribution among the equi- fundamental centers of origin for the various major formal areas, however, that makes no serious modi- elements in the flora; and the "equiformal areas" fication in the latter. Second, there are apparent become "equiformal progressive areas" that are gaps in the ranges of boreal American plants which thought to indicate the general patterns of dispersal are gaps in exploration rather than in the actual among the major elements. The species are termed ranges. Lacunae in our knowledge of the more uni- "radiants" from the various "centra." form floras of eastern glaciated regions are not so The geographic arrangement of centra as worked troublesome; but the gaps in northern British Co- out by Hulten is as follows: They are " . .. in North- lumbia and Yukon are more serious, for the distribu- Eastern Siberia and in the Amur-Manchurian region. tion of species into equiformal areas sometimes de- Another occurs in the Altai-Sajan region, sending out pends upon their behavior in this region. Some of radiants towards the Arctic shore. A third centre Hulten's dispositions of species are therefore open to is northern Japan, whence numerous plants radiate to modification due to range extensions discovered since the north and to the coast of the Asiatic Continent. his study was made. Third, Hulten's original sorting A centre of great importance is the region around of species among his equiformal areas can be ques- the northern part of the Bering Sea. It sends out tioned in many cases. This is particularly notable in progressive radiants reaching symmetrically as well his treatment of wide-ranging forms. to the west into arctic Asia and Europe as to the I have found it necessary to make a number of east to Eastern America, and also often extends arms changes in Hulten's lists along the three lines just along both the Asiatic and American Pacific coast. mentioned. The changes are in many cases matters In America radiants proceed from the Yukon valley of judgment, and even when all are taken together along the Arctic American coast, others centre around they make no serious modification in his general sort- the Arctic Archipelago, and others again have the ing of ranges. centre of their progressive figures in the State of If the general thesis outlined by Hulten is tenable, 2.32 HUGH AI. RAUP Ecological Monographs 2ol. 17. No. 2 thein the flora of the Mackenzie Mountains should forests. Another small part of the Mackenzie Mouni- show affinities with his various centra which, within tain flora should be derived from the coastal radiants, the g( neral limits of accessibility, are directly pro- for they could not be expected to spread far outside portional to the size of the relic populations, and the habitats of the relatively warm and wet north pi)esuinably inversely proportional to their degree of Pacific shores. It is true that Hulten allows manv biotype dlepauperation. of these coastal species to spread widely across the It should be possible to draw up a sequence show- continent, but they are more southern types that iin- the relative availability of plants for the coloni- would not be expected at high latitudes. zaItion of the Brintnell Lake area during the retreat In Table 3 I have compared Hulten's floristic ele- of the glaciers. It is presumed that for a time after ments of the Brintnell Lake area in two ways. First the disappearance of the ice the mountains and ad- the percentages of the total know flora which are de- jacenit valleys and plains had free access to the alpine rived from the various sources are compared, and and tnn(lra floras of the Rockies, the Cassiars, the second, the percentages of the total numbers of radi- Yukon Plateau, the Richardson Mountains, and the ants (as revised) from the various sources that are aretic lowlanids. With the amelioration of the climate represented in the mountains. Approximately the conifelros forests eventually covered the valleys of same results are reached by both methods. The the Mackenzie, upper Liard, Peel, and Wind Rivers, Arctic-Montane group is present in largest numbers. anod finally the divides between these streams and the Over 42 percent of all the plants listed in the group Yikon system. In so doing they effected the isola- have been found at Brintnell Lake, and thev make tion of the alpine and arctic elements of the Macken- up about 35 percent of the local flora there. Next zie Mountain flora. are the Continental Western American plants, and The time intervals are conjectural, but probably third, the Boreal Circumpolar group. The Beringian can be brought within reasonably definite limits. groups, taken together, and the western American There is fairly good evidence that forests did not Coast Radiants have each supplied only 8.1 percent appear in the central Mackenzie basin until after the of the flora. The most strictly coastal group set up final retreat of the ice from Great Slave Lake and by Hulten, of Atlantic-Pacific plants, is not repre- until after the last of the post-Glacial lakes in the sented at all. Athabaska-(Creat Slave Lake region had been drained. It will be seen at once that these results correlate Forests probably did not enter southwestern Macken- very well with the series expected on the basis of zie until the period of the climatic optimum, as late Hulten's theory. They lend considerable support to as 7,000 years ago. They can hardly have reached his general concept. It should be noted that the re- the higher, interior valleys and divides until some- sults from similar comparisons made with Hulten's what later. unrevised figures are not far different from the above. According to Hulten's theory the first aggressive The positions of the Boreal Circumpolar and Con- floristic elements to invade the glaciated lands were tinental West American groups are reversed, but they the AlctiC-Montane plants, and such parts of the are not far apart in any case. Beringiani and other groups as are supposed to have Hulten's method of handling his data, however, reached refugia south of the ice before the maximum leaves a great deal to be desired. The success achieved advance. In both cases these plants were able to by him probably can be attributed to his perspicacity maintaini large populations during the last great ice in "sensing" the geographic relationships of Ameri- advance, fand to maintain their genetic plasticity. can plants rather than to an objective treatment of The most lplastic of all Hulten's groups, and certainly his information. This is best shown in his sorting the most wide-ranging species, are among the Boreal of wide-ranging species into his several groups. It Circumpolar plants, but since a very large proportion is virtually impossible, for instance, to say whether of this group is of forest species, it cannot be ex- a wide-ranging arctic-alpine species should be placed p)cted to show so high a representation in the Mac- among the arctic-montane, boreal circumpolar, or kenzie Mountains as the arctic-alpine group. continental western American radiants. The equi- Next to the Arctic-Montane plants in a decreasing formal progressive areas are clear so long as their scale of availability would come the Continental West individual plant ranges are relatively small. The American Radiants. These are thought to have sur- larger ones disappear in the great "pools" of species vivedl in the Yukon valley and the northern Rocky whose ranges extend all the way across the continent. Mlountains, many of them below the ice boundary. A few doubtful decisions of this sort would not be Since they contain a large arctic-alpine element they serious, but Hulten had to make hundreds of them can be expected to be well represented in the Macken- in order to relate all the ranges to his equiformal zie M\Iounitains. Furthermore a part of their refuge areas. areas are directly adjacent to the Mackenzie Moun- With these difficulties in mind I have tried to sort tains. the actual ranges of the species, so far as they are The various Beringian radiants of Hulten should known, into natural patterns without regard to hy- be rather poorly represented, not only because many pothetical centers of dispersal. Using the maps pre- of them are thought to have lost their spreading sented here and the genetic reasoning employed by capacity, hut also because they, like other alpine ele- Hulten, it is possible to arrange a sequence of floristic meats, were everitually stopped by the advance of the influences similar to that described above. The size April, 1947 SOME NATURAL FLORISTIC AREAS IN BOREAL AMERICA 233

TABLE 3 ORIGINS OF THE BRINTN ELL LAKE FLORA By way of summary of the foregoing we may pic- Percentage Pereentage ture Hultln's Found at of total of Hulten's the Mackenzie Mountains as having been one of totals Brintnell Brintnell amended the latest montaine areas to lose its glaciers (rem- L. totals (amended) Lake plants nants of them still exist). While the glaciers were Local plants, or plants of limited known range 12 4.2 retreating there was a period of time during which the lower slopes were tundra-covered, with the tundra Radiants.248 23 8.1 9.3 Beringia more or less continuous across the neighboring val- Atlantic-Pacific plants...... 16 0 0 0 leys and to neighboring mountain ranges. The length W. American Coast Radiants.... 267 23 8.1 8.6 of the period is problematical. There is some evi- Continental W. American Radiants.190 73 25.8 . 38.4 dence, however, that forests did not come into the

Aretic-Montane plants. 235 100 35.3 42.6 lower valleys until about 7,000 years ago, and that they did not cover the divides until somewhat later. Boreal Circumpolarplants...... 262 52 18.3 19.9 In any case the tundra period must have been one TOTALS.1218 283 in which arctic and alpine plants of all the northwestern refugia had open routes of migration to the Mackenzie Mountains. There is evidence that they and character of existing ranges can be used as indi- availed themselves of these routes in differing pro- cators of the success with which plants have migrated portions, depending first upon their broad habitat from the Pleistocene refugia. preferences, and second upon their inherent abilities Considering first the alpine flora, by far the largest to migrate. Their success at colonizing the moun- single group (66 spp.) is derived from wide-ranging tains was conditioned also by the remoteness of the arctic-alpine plants-51.5 percent of the alpines, or various refugia. Plants of continental habitat pref- about 23 percent of the total flora. These are the erences were of course most successful; and of these species that are thought to have maintained compara- the wide-ranging Arctic-Montane group were out- tively large populations during the Pleistocene, and standing because of both accessibility and lack of to have been able to migrate during that time. biotype depauperation. Survivors in the Yukon val- Second in extent of range are the species that have ley and the northern Rockies were second in numbers, occupied wide areas in Alaska or the northern Cor- for they were only partially depauperated and were dillera or both, some of them extending eastward to- near at hand. Northern Beringian and Coastal ra- ward Hudson Bay (maps, Fig. 7A, B, C; Fig. 8). diants were next in importance, probably held in There are 49 of these about 38.2 percent of the check by distance and lack of plasticity. alpines. Finally there are 13 species of far north- The process of alpine colonization appears to have western range (map, Fig. 7D), only a few of which been checked, at least for the Brintnell Lake region, extend south of the 60th parallel. They comprise before it had reached anything like completion. Evi- about 10.1 percent of the alpine plants at Brintnell dence for this is to be found in the general paucity of Lake. species in the alpine plant cover. The "partial" na- The plants of wide range on the mountains of ture of the flora is to be seen in all the common arc- Alaska, Yukon, British Columbia, and Alberta may tic families and genera. Such groups as Potentilla, well have had access to the large refuge of the Senecio, Arnica, and Pedicularis, all represented in Yuikoni Plateau and to possible refugia in the north- surrounding regions by half a dozen or so species, erin Rockv Mountains and eastern foothills, as well have only 2 to 4 species each in the Brintnell Lake as to those of Beringia. The smallest ranges are district. The colonization appears to have been those of the Alaskan species which were probably stopped by the advance of forests into the surround- limited to Beringian refugia. ing valleys. These forests brought a new element to Nearly all of the forest plants at Brintnell Lake the mountain flora, derived from refugia south of the are derived from wide-ranging species in boreal ice or on the slopes of the more southern mountains. America. I have already noted that only ten of the Whether any of it came from the Yukon valley is western plants represented are typically forest types. uncertain. It was derived from Hulten's Boreal Cir- Lake forest flora is clearly derived from The Brintinell cumpolar, Continental western American and possibly lands to the southeastward and has undoubtedly come in part from his western coastal groups. IIi terms up from the Liard valley. That of the central Mac- of range patterns outlined in this paper it was drawn kenzie regiolo, in turn, as I have suggested in an chiefly from wide-ranging forest species and in earlier paper, has probably developed from the amal- part "timber-line" By nature it is gamatiort of populations that persisted through Late from the group. ag- Wisconsin time in the East and in the valleys and gressive, but probably shows so sniall a portion of the foothills of the Rocky Mountains. All of these spe- total Brintnell Lake flora as it does onlv because of cies could have maintained fairly large ranges during the short time available for its invasion, and because that time, and have no doubt increased their spread- of the subarctic situation of the Mackenzie Moun- in, capacity by sllbsequent fusion of populations. tains. Ecological Monographs 234 HUGH Al. RAUP Vol. 17, No. 2 GENERAL CONSIDERATIONS kenzie and Yukon we probably are seeing the last I would be bold indeed if I were to attempt a de- stages of this achievement, with the amalgamation of tailed correlation of these geobotanical events with eastern and western relic elements, and the advance those in other boreal and temperate parts of North of newly aggregated forests into the mountain val- America. It is impossible, however, to resist the leys and adjacent plateaus. When prairies were temptation to do a little speculating. developing in Ohio, therefore, and when there was, As I have said previously, there is good evidence we suppose, a warmer climate in southern New Eng- that there were no forests in most of Alberta, north- land, forests were just beginning to cover the upper ern Saskatchewan or Mackenzie until the last of the Saskatchewan and southern Mackenzie watersheds, great post-Glacial lakes of the central Mackenzie and the alpine areas of the Mackenzie Mountains still basin were drained. Using Antevs' chronology, this had floristic access to the Rockies, Coast Ranges, and did not occur until approximately the advent of the the Arctic. post-Glacial optimum, or toward the close of his LITERATURE CITED "younger Late Glacial" period. Similar evidence indicates that there were no forests in the Saskatche- Antevs, Ernst. 1931. Late-Glaeial correlations and ice recession in Manitoba. Geol. Surv. Can. Memoir 168. wan River basin until Glacial Lake Agassiz had reached nearly the modern level of Lake Winnepeg. Cain, Stanley A. 1944. Foundations of plant geography. New York. The latter event may have occurred somewhat earlier than the drainage of the last Mackenzie basin lakes, Halliday, W. E. D., & A. W. A. Brown. 1943. The distribution of some important forest trees in Can- but Antevs places it also in the "younger Late Gla- ada. Ecology 24: 353-373. cial." He suggests that the period of Lake Agassiz Hulten, Eric. 1937. Outline of the history of arctic may have been 10,000 to 15,000 years in length, and and boreal biota during the Quaternary period. Stock- that it corresponded to the period of ice recession holm. from northern New England north to beyond Lake Porsild, A. E. 1945. The alpine flora of the east slope Timiskaming. of the Mackenzie Mountains, N. W. T. Nat. Mus. Can. If these chronologies are reasonable then it seems Bul. 101, Biol. Ser. 30: 1-35. necessary to recognize a wide gap in the continuity Raup, Hugh M. 1941. Botanical problems in boreal of the northern coniferous forest, beginning with the America. Bot. Rev. 7: 147-248. advance of the Mankato or Wisconsin3 ice, and last- 1947. The botany of southwestern Mackenzie. Sar- ing at least until 7,000 or 9,000 years ago. The wide gentia VI. 275 pp. + PI. I-XXXVII. ranges of the forest species described in this paper Stebbins, G. L. 1942. The genetic approach to problems must have been achieved since that time. In Mac- of rare and endemic species. Madrofio 6: 241-258.