VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA'

R. H. WHITTAKER Biology Department,Brooklyn College, Brooklyn10, N. Y.

TABLE OF CONTENTS

PAGE PAGE

I. INTRODUCTION ...... 279 East-West Gradation and the Prevailing Nature of the Study ...... 279 Climax Type ...... 308 Geology...... 279 V. PROBLEMS OF CLASSIFICATION ...... 311 The Central Relation of the Klamath Forests . . . 282 Formations ...... 311 Climate ...... 283 Dominance-Types and Sociations ...... 312 Culture and Disturbance ...... 284 Associations ...... 312 Literature ...... 284 Quantitative Approaches to Continuity and Species Grouping . .. 313 285 II. PROCEDURE ...... Distributional Groupings ... 314 Study Areas ...... 285 Vegetation Samples and Soil Data ...... 286 VI. FLORISTIC COMPARISONS ... 316 Arrangementof Samples in Transects .... . 286 Life-Forms and Growth-Forms ...... 317 Evaluation of Transect Techniques ...... 288 Species-Diversities ...... 319 Transect Tables ...... 289 Sample Similarities and Coenocline Differentiation ...... 320 III. VEGETATION DESCRIPTION ...... 291 Geographic Relations of Floras ...... 323 Low Elevations on Diorite ...... 291 VII. RELATIONS OF SPECIES POPULATIONS TO THE Low Elevations on Gabbro...... 297 THREE SOILS .325 on Serpentine and the Low Elevations Trees .325 Two-Phase Effect ...... 299 Undergrowth Species .. 326 Forest Vegetation of Higher Elevations General Results ...... 327 on Diorite ...... 302 Rare Species and Serpentine Indicators .328 Vegetationof Higher Elevations on Serpentine . . 305 VIII. CONCLUSION ...... 329 IV. CLIMAX INTERPRETATION ...... 306 SUMMARY ...... 330 Fire Effectsand Edaphic Climaxes.306 The Coenoclineand Climax Comparison.307 LITERATURE CITED . . . 332

I. INTRODUCTION ties to environmental gradients. Analysis and inter- pretation were based on the conception of the vege- NATURE OF THE STUDY tation as a multi-dimensionalpattern, and on study of The Klamath Region, betweenthe southernCas- the manner in which local patterns of vegetation in cade Range and the PacificOcean in southernOregon relation to topographic moisture gradients change and northernCalifornia, is an area of exceptional along climatic gradients and from one parent material ecologicalinterest. These old and geologicallycom- to another. The sections which follow include vege- plex mountainssupport an exceedinglycomplex pat- tation description, climax interpretation, community tern of natural communitiesin relationto steep cli- classification, floristic analysis, and consideration of matic gradients and diverse parent materials,and species distributions for the pattern of Siskiyou among these communitiesa prevailing climax, the forest vegetation in relation to four major environ- Mixed EvergreenForest, which has a centralrelation mental gradients-local topographic moisture, eleva- to otherwestern forest vegetation. One area of the tion, the diorite-gabbro-serpentineseries of parent region,the SiskiyouMountains along the California- materials, and the east-west climatic gradient from Oregon border,was selected for vegetationstudy. the Pacific Coast inland. Quantitativesamples were takenover a wide range of topographicsituations, climates, and parentmaterials GEOLOGY for gradientanalysis, seeking to relate distributions The dominatingtopographic features of the Pa- of plant populationsand characteristicsof communi- cific Coast states are the two chains of mountains 1 A contribution from the Department of Zoology, Washington which run parallel to the Coast from Canada to State University, and the Department of Biology, Brooklyn College. This study was supported in part by the funds for southernCalifornia, and the chain of valleysbetween biological and medical research of the State of Washington them. The higher,inner mountainchain is formed Initiative Measure No. 171. The author is indebted to M. Ownbey and A. Cronquist for determination of the plant by the Cascade Range and the Sierra Nevada; the collections, to D. I. Axelrod, R. W. Chaney, H. D. MacGinitie, and F. G. Wells for comments on the manuscript. lower,outer chain includesthe "Coast Ranges" in a 280 R. H. WHITTAKER Ecological Monographs Vol. 30, No. 3

1,0"EMT44,,,. IV R"

0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0

A/ 1 LEGEN A AR B L o~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~8l

? M~~~~~~~~~j~ ~ ~ ~ ~ iM

FIG. 1. Map of Siskiyou Mountains of California and Oregoal in relatioa to other, adjaceat ranges. Major study areas on three parent materials within the Mixed Evergreen Forest Region are outlined and marked with circled letters: (D) the diorite area, (S) the serpentine area, (G) the area of gabbro and hornblende diorite. Study areas for the west-east climatic transect from the Coast ialaad are marked with circled numbers: (1) Mill Creek State Park (Sequoia forest), (2) South Fork, Smlith River (coastal Pseu- dotsuga forest), (3) Siskiyou Fork, Smith River (mixed evergreea forest, snore oslesic phase), (4) Sturgis Creek (mixed evergreen forest), (5) Beaver Creek (mlixed evergreea forest, more xeric phase), (6) Emi- grant Creek (oak woodland, with Pseudotsuga forest inl most mlesic and valley grassland in most xeric sites). broad sense. In northern California and southern major mountain groups of the central Klamath Re- Oregon a complex of mountains, the Klamath Ranges, gion; like other major ranges of the Region, they arc extends from the Coast inland to the southern Cas- a complex area of mountains rather than a well- cade Mountains and interrupts the chain of valleys defined ridge. The Siskiyous extend in an east-west between the two mountain chains. Geographically direction along and on each side of the California- the Klamath mountains are part of the coastal chain; Oregon border, north of the Klamath River and south but they differ from the Coast Ranges north and of the Rogue River. Froma the Pacific Coast they south of them in age and history, in geological char- extend inland from low mountains, with elevations acter and complexity, in height and east-west extent. less than 1000 m near the coast, through the main Their closest geological relations are not with the area of mountains of intermediate elevations, in- Coast Ranges, but with the Sierra Nevada. and the cluding some peaks above 2135 m, to the low moun- older core of the Blue Mountain complex of Oregon tains which connect with the southern Cascade Range. (Fennenian 1931). The older geological history of the Klamath The Klaniath Region, as it was outlined, described, Mountains is essentially the same as that of the Sierra and named by Diller (1894, 1902, 1903, 1906, 1914), Nevada (Diller 1894, 1903, 1906; Fenneman 1931). extends north and south for about 390 km, and west- Extensive masses of sedimentary rocks were deposited ward from the Cascade Mountains about 120 kil, to in an inland sea in Devonian and Carboniferous time; reach the Pacific Coast between 410 and 430 north and, at the close of the Paleozoic, these and other latitude, with a total area of about 34,000 km2. The rocks were folded and raised into mountains. These Siskiyou Mountains are the northernmost of the mountains were worn down, and much of the area July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 281

submerged in Triassic and Jurassic time, while addi- (1949), Wells & Walker (1953), and Cater et al. tional sediments and volcanic materials were de- (1953). The Kerby and Grants Pass quadrangles posited. At the close of the Jurassic, extensive de- (Wells et al. 1949, Wells 1940) include the main formation and intr'usion occurred, accompanied by study areas in the central Siskiyous. In the Grants another uplift. A long cycle of erosion followed until, Pass quadrangle, the most extensive area is under- during the Cretaceous, the low mountains remaining lain by metavolcanic rocks of Paleozoic age; along subsided and were again largely submerged, with with these occur a number of larger and smaller out- deposit of extensive sedimentary rocks. crops of diorite, thought to represent a single great At the close of the Cretaceous the Klamath Moun- batholith of Jurassic or Cretaceous age. Of these, tains were again uplifted; and they have existed, as the larger outcrop of quartz diorite (about 120 kM2) at least low mountains, throughout the Cenozoic, with which includes Grayback Mountain served as one a complex history of uplift, subsidence, and erosion study area. Other rocks on which vegetation could (Diller 1902). Results of this history are to be seen be observed in the area included various metasedi- in wave-cut coastal terraces (Diller 1903, Fenneman mentary rocks-the marble in which the Oregon Caves 1931 :463, Dicken 1952) and peneplain remnants occur, slate, argillite, and quartzite-and serpentine. (Diller 1902) in the Siskiyou Mountains. One of In the Kerby quadrangle large areas of both serpen- these, the Klamath peneplain, is of major significance tines and gabbros occur, together with various meta- in the study area. Erosion during the early Cenozoic volcanic and metasedimentary rocks, granodiorite, reduced much of the region, by Miocene time, to a hornblende diorite, dacite porphyry, and amphibole surface of low relief above which scattered mountain gneiss. The unmetamorphosed serpentines are pre- ranges rose. This peneplain was later subjected to re- dominantly saxonite, with small areas of dunite and peated elevation and subsidence, and probably some pyroxenite. The fringes of the main peridotite mass tilting and deformiation. The Klamath peneplain now and the smaller outcrops outlying from it are largely appears as a dissected plateau, the surface of which in metamorphosed into serpentine rock in the narrower general rises from the Coast inland and from the sense. The peridotite and serpentine together form a north to the south. In the Siskiyous it is about 500 great sheet, thousands of meters, but probably not m near the coast, but rises to 1200 to 1350 m, 40 to over 4500 m, in thickness intruded along planes of 80 km inland; farther south in the Salmon and Yolla weakness into the other rocks of the area. The study Bolly Mountains remnants appear at 1800 to 2200 of low-elevation serpentine vegetation was concen- m (Diller 1902, Fenneman 1931). In the central trated in the peridotite and serpentine area from Siskiyou area, one may climb onto the peneplain by Eight Dollar and Josephine Mountains west to Chetco the Wimer Road and observe its surface in the serpen- Peak and south beyond Oregon Mountain into Cali- tine and gabbro uplands west of the Illinois Valley. fornia. This, with a north-south extent of 90 km Sighting across the valley one may observe remnants and an area exceeding 700 km2, is the largest body of the surface in the lower, metavoleanic mountains, of ultramafic rock in the United States and perhaps 1946 Cater et al. while the diorite monadnock of Grayback Mountain in North America (Wells et al. 36 rises above it to 2148 m. 1953). The olivine gabbro study area, of about of a belt of A great mass of ancient, elo~ely folded and faulted km2 including York Butte, is part larger 260 kmn2),also rocks, generally metamorphosed, and intruded by gabbro and hornblende diorite (over intruded in Jurassic or Cretaceous time, lying north- igneous rocks, form the Klamath Mountains (Fenne- to the serpentine area. man 1931). The consequence of their history of west of but parallel rock pattern and history of the sedimentation, vulcanism, and igneous intrusion, up- The complex have no well-defined lifting, folding, faulting, and erosion is an extremely Klamath Mountains produced and ridge direction; the complex mosaic of rock types, a melange of diverse trend in stream drainage of ". . . a parent materials. Peridotite and serpentine intru- Siskiyou Mountains give the impression and sions, characteristic of many mountains and island confusion of broken mountain ridges with steep chains (Hess 1955), are especially extensive and con- stony slopes . . ."' (Peck 1941). The principal rivers spicuous in vegetational effect. The deformation of of the Klamath Region cut transversely across it, these mountains was due to repeated compression running generally westward from the interior valleys, from the east; and the folds thus formed are curved through deep canyons in the mountains themselves, or crescent-like,striking southeast toward the Sierra to the ocean (Anderson 1902). Walls of these can- Nevada in the southern part of the Klamath Region, yons are long, and often steep, inclines from-the up- northeast toward the Blue 1Iountains in the northern land to the valley bottoms,interrupted in some cases part (Diller 1914, Fenneman 1931). A trend toward by terraces remaining from past erosion cycles (Diller the north-northeastis evident in the formations out- 1902, Fenneman 1931); valley bottoms are mostly cropping in the area of the Siskiyous studied (Wells narrow. Mountain slopes with an average grade of et al. 1949, Wells & Walker 1953). 300 or more through a range of 1000 m or more are Geologic maps for Siskiyou quadrangles have common; over most of the region rugged mountain been published by Diller (1903), Diller & Kay (1924), topography prevails. Although the Klamath pene- Maxson (1933), Wells (1939, 1940), Wells -et al. plain is represented by rolling uplands in some areas EcologicalMonographs 282 R. H. WHITTAKER Vol, 30, No. 3 of the Siskiyous, the mountains of lower elevations broadest statement,the historyof westernforests are in general stream-eroded to topographic maturity. from Miocene time to the presenthas been one of Topography of higher elevations of the Klamath progressiveshrinkage toward the Coast and higher Mountains has been shaped by many local alpine elevations,accompanied by progressivedifferentiation glaciers; the description of glacial effects in the in the differentareas of the West, while the diverse Trinity Alps by 'Hershey (1900) applies as well to floristicelements of the Madro-TertiaryGeoflorn the Siskiyou Mountains. Cirques, with steep rock were progressivelyexpanding and differentiatingin walls with scanty vegetation and floors with tarns the Southwestand the drierlowlands of the Interior and mountain meadows, occur in the Grayback area (Chaney 1947, 1948, Axelrod 1958, 1959). Forests and other higher mountains; but the fraction of the most suggestiveof the mixed forestsof the Miocene area showing glacial topography is small. The non- are now restrictedto the SouthernAppalachians in forest vegetation of the glacial topography is out- the East, the Klamath Region and coastal California side the concern of the present monograph. in the West. It is in thesetwo areas thatthe combi- nation of sufficientlyfavorable moistureconditions THE CENTRAL RELATION OF THE KLAMATH FORESTS withsufficiently warm temperatures permit the exist- To the ecologist familiar with both the Southern ence today of remnantsof the Areto-TertiaryGeo- Appalachians and the Klamath Mountains, there are flora. The significanceof the SouthernAppalachians a number of striking parallels between these regions. in relationto the easternforests has been extensively Both are old mountains, with land surfaces which developedby Braun (1935, 1938, 1947, 1950). have been continuously occupied by vegetation A second basis of the central relation of the throughout the Cenozoic, at least. Both persisted as Klamath Region is in its climatic,topographic, and monadnocks through major early Cenozoic cycles of edaphic diversity,together with the age of the moun- erosion, culminating in the Schooley peneplain of tains. Climaticvariation and topographypermit a the Southern Appalachians, the Klamath peneplain wide range of communitiesto exist in the regionand of this region. Both have been refuges for plant in differentsituations in a given area. The different populations destroyed in other areas by glaciation, parentmaterials also have strikingeffects on floristic submergence of coastal plains, climatic desiccation, and vegetationaldiversity. Many of the narrowly and, in the West, the great lava flows of the In- endemicspecies in whichthe regionis so rich occur terior. Both contain modern vegetation which is on serpentine,gabbro, or otherlocalized parent mate- most nearly related to the widespread Areto-Tertiary rials. Parent-materialdifferences make possible also forests of the earlier Cenozoic. Both are areas of much of the distributionaloverlap in the region of great vegetational diversity, and also areas of great species with diversegeographic relations. Thus, in floristic diversity and concentrations of narrowly the centralSiskiyous, Taxus brevifoliain the south- endemic species. Both have "central" relations to ern part of its distributionoccurs on diorite,Pinus the forest floras and vegetations of other, surround- ponderosanear thewestern limit of its distributionon ing areas. gabbro,and Pinus jeffreyinear the northernlimit of Figures are not available, but there is no doubt its distributionon serpentine;these three species ap- that the flora of the Klamath Region is extremely pear in the same limitedarea, but not in the same rich in numbers of species and numbers of narrow stands. con- endemics for its latitude. Many genera show The diversityof habitatshas been characteristic centration in- of high proportions of their species, of the area throughoutits long history,although cli- cluding endemics, in this area. Many of the en- matic were probablyless steep beforemid- demics appear to be relicts of formerly wider distri- gradients dle Pliocene time. Even at the maximumdevelop- butions; in other cases species have probably spread mountainsof diverse from the Klamath Region to other, areas of the mentof the Klamathpeneplain, parentmaterials existed in the area. The regionhas West. Thus in the genus Crepis, a number of diploid a mosaicof habitats,per- species are now reliet in the Klamath area, while at all timesoffered complex diverseenvironmental requirements genetic material from these has been used in apomictic mittingspecies of elimi- polyploids which have spread over semi-arid environ- to persistin the area while changingclimates to ments of the Interior (Babcock & Stebbins 1938). nated some species elsewhere,and caused others The region possesses also a greater diversity of forest migratenorth or southfrom the range,or to become east of the communities, in a more complex vegetation pattern, largely restrictedto the Interior range, than any comparable area of the West. With the while leaving relict populationsin the Klamath Re- exception of more typical forms of the Pigmy Conifer gion. Woodlands, all the plant formations dominated by Finally, it may be observedthat the location of trees of the western United States occur there, as the regionmakes it the meetingground of florasof they do in no other area. diverseclimatic relations. In it the florawhich may Three interrelated reasons for this central rela- be broadlytermed "Northwestern" meets part of the tion may be given on the basis of the vegetational "Californian"flora, with representationalso of the history of the West, the geological history and char- floras of the arid Interior and interiormountains. acteristics of the mountains, and their location. In The Mixed Evergreen Forest climax itself is the July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 283 link betweenthe coniferousCoast and MontaneFor- TABLE 1. Climatic data for stations at low elevations ests on the one hand, and the broad-selerophyllvege- in the Siskiyou Mountains, from the Coast inland in tationof Californiaand Oregonon the other. More California and Oregon. than any other area, the Klamath Region is central to the forestfloras and forestvegetation of the West. 0~~~~~ 0

0 CLIMATE 0 n 'e bD The locationand topographyof the Klamath Re- Happy Cam~ . ~470 ~ 3 -44 5 0 .96 o . -1 gion imply a wide range of climaticconditions, in- 4d-- S2 o cluding steep climatic gradients from the Pacific Montagu~~~e I . ...1130 771.71 2-64 411.92 California Coast inland (Engelbrecht1955). Coastal climates CrescentCity.... 16 38 11.3 7.7 1. 2 - 7 39 230 193 6.6 are stronglymaritime, with high precipitationsand HappyCamp . 64 700 3.2 22. 1 -14 46 169 106 3.9 66 humidities,abundant fog, limitedranges of tempera- ScottBar . 792 5649 63 6.0 tures,and low temperaturesfor theirlatitudes. The Yreka ...... 122 8601 10. 1.0 22.0 -22 44 129 44 9.6 42 Montague... 130 747 11.7 1.0 22.6 -26 43 141 31 14.9 24 mountainsparalleling the coast limitthe extentof the Oregon maritimeinfluences inland, and produce rapid cli- Brookings... 10 37 11.3 7.9 14.6 - 5 38 269 187 8.9 3 matic changetoward drier, warmer, and more conti- BuckhornFarm. 49 396 11.3 4.1 20.1 - 14 42 150 160 3.9 66 Waldo.. 47 603 10.3 2.4 19.6 - 17 43 166 126 6.6 64 nentalconditions in the interiorvalleys. Isoplethsof Williams 60 467 11.1 3.7 19.4 - 16 42 146 61 6.3 46 January and July temperatures,annual range of GrantsPass 866 267 11.6 3.9 21.1 -16 46 132 74 6.6 22 temperature,annual precipitation,and relative hu- Jacksonville... . 106 500 11.4 2.6 21.6 -18 41 190 67 8.8 32 parallel the Talent ...... 118 480 11.6 2.9 21.7 -17 42 170 42 13.6 21 midityand evaporation,all tend to coast Ashland... . 124 601 11.4 3.3 20.8 -17 41 182 50 12.8 45 (Visher 1954). Kendrew (1937) observes of one SiskiyouSummit. 130 1363 0.6 18.3 -21 38 83 9.8 aspect of these climaticgradients, that. the contrast in temperatureconditions between the coast at San Francisco and the Great Valley of Californiais as winter, December -and January, lightest in summer, great as that between Scotland and North Africa, July and August. The area is in the transition be- althoughthe distanceis only about 120 km. tween the two Pacific Coast types of rainfall regime Climaticdata availablefor the SiskiyouMountains (Ward 1925, Kendrew 1937), the North Pacific with themselves(Bowie 1934, Wells 1936, 1941, Sprague its limited summer rain and the California type with 1941) are summarizedin Table 1. The stationsare its practically rain-free summers. At Waldo the four arrangedin two sequencesfrom the coast inland in months from November to February include 65% of California and in Oregon. Mean annual tempera- the rainfall, the four months from June to Septem- turesfor low elevationsin thesemountains are rather ber 6%. Maritime fogs, or low stratus, are a promi- consistentlyaround 11.0-11.50C. Average tempera- nent feature of the coastal climate, where fog may ture conditionsof valley stationsare much the same be present 50% of the summer hours and where fog throughoutthe area; but the inland stationsare at and fog-drip from trees contribute significantly to higher elevations,and temperaturesat comparable the humid conditions of the redwood belt (Cooper elevationsare consequentlywarmer in the interior 1917, Byers 1953, Patton 1956, Oberlander 1956). thanon the coast. More markedcontrasts appear in Relative humidities may be assumed to decrease, and ranges of temperature.Mean monthlytemperatures evaporation, hours of cloud-free sunlight, and diurnal for January decrease fromabout 8.00 to 1.0-3.00C, temperature range to increase toward the interior; but while those for July increase from about 15.00 to data are not available. 21.0-23.0?C. The differencebetween these mean In general character the climates of low elevations monthlytemperatures, as an indicationof thegradient in the Siskiyous are warm-temperate, summer-dry, from more equable maritimetoward more variable "Mediterranean" climates; corresponding to these continentalclimates, increases from about 7.00 on the climates, plant communities in which broad-leaved coast to 19.0-21.50Cin the interiorvalleys. Similar evergreen or sclerophyllous trees are prominent pre- trendsmay be observedin the extremetemperatures vail over most of the range. The climates of these given in Table 1. forests, and especially those of the redwood belt, are Mean annual precipitation decreases strongly considerably more humid than those of Mediterranean from the coast inland, from values over 180 cm, Europe. Russell's (1926) application, with modifi- throughvalues of 80 to probably150 cm in the cen- cations, of the K6ppen (1900, 1923) system recog- tral Siskiyous,to values below 50 cm in the interior nized a mesothermalhumid belt with equable tempera- valleys. The data for Waldo, in the Illinois Valley tures near the coast (Csn), a belt of mesothermal betweenthe principalstudy areas, bear mostdirectly humid, Mediterranean selerophyll climates (Csb) in- on the workin the centralSiskiyous. SiskiyouSum- land from this, and warmer and drier woodland and mit is in the low mountainsconnecting the Siskiyous steppe (Csa and Bsh) climates in the inner valleys, with the southernCascade Range, and is thus not a while microthermal climates occur at higher eleva- part of the low-elevationsequence. tions. The Thornthwaite (1931, 1948) systems recog- All thestations have maritimepatterns of seasonal nize a sequence fromwet or perhumid conditions near distributionof precipitation,with heaviest rainfall in the coast through humid ill the main area of the range WHITTAKER Ecological Monographs 284 R. H. Vol. 30, No. 3 to subhumid in the interior; low elevations of the lages, or to fields with few vestiges of the old build- area are mierothermal by the criteria of the earlier ings. system, cooler mesothermal by those of the later. Lumber now forms the principal resource and in- The work in the central Siskiyous had as one ob- dustry of most of the Klamath Region. The coastal jective the comparison of vegetation patterns on three redwood belt contains immensely valuable timber parent materials in the same climate. So steep are lands; within the central Siskiyous Port Orford cedar the climatic gradients across the Siskiyous that the (Chamaecyparis lawsoniana), Douglas-fir (Pseudo- three study areas on diorite, gabbro, and serpentine tsuga menziesii), white fir (Abies concolor), sugar cannot have the "same" climate. They are all, how- pine (Pinus lambertiana), and to a lesser extent ever, in the same area of the central Siskiyous, repre- other species, form valuable timber stands. Lumber- sented approximately by the Waldo data. They are ing practice at low elevations generally involves re- thought to be close enough to be reasonably com- moval of the conifers of the mixed evergreen forests, parable, and the consistency of the vegetational con- leaving at least partial sclerophyll cover. With the trasts on these parent materials in differentareas of relative depletion of more available timber farther the Siskiyou Mountains supports the conclusion that north and consequent shift of lumbering activity to- effects of parent materials far overshadow those of ward the south, much of the -forestarea of the Siski- climate in the study areas. yous is being rapidly cut; and conservation and sustained-yield programs are little in evidence (Dick- CULTURE AND DISTURBANCE en 1952). The rugged topography and generally infertile In the dry-summerclimate of the Siskiyous, the soils of the Klamath Mountains have not offeredre- forests are easily set afire; and fires of widely vary- sources which would draw a large population into ing intensities have been frequent. The Klamath and the region. The largest city within the region is other Indians are believed to have set fires in connec- Medford, with a population somewhat over 19,000; tion with hunting and warfare. In Gold Rush days most of the region is sparsely settled, and iiiuch of and thereafter, fires were set prodigally, uninten- it is mountain wilderness. The greater part of the tionally from neglected camp-fires and intentionally area is in National Forest land. The central and to make travel easier, to clear the ground for prospect- western Siskiyous of Oregon, in which most of the ing, and for recreation. Miners are said to have set study was carried out, are within the Siskiyou Na- firesto enliven an evening's drinking with a mountain tional Forest, the eastern Oregon Siskiyous in the slope in flames. Fires were used to drive game, and Rogue River National Forest. Extensive wilder- there is a report of a successful hunt that bagged 18 ness areas have been set aside in the Yolla Bolly, Sal- elk through the destruction of 3 billion feet of timber mon and Trinity Alps, and Marble Mountains; and (Forest Service 1940). Probably all lower-eleva- small areas of the coastal redwood forests are pro- tion mixed evergreen forests have been affected by tected as parks. In the Oregon Siskiyous the Oregon less intense fires in the litter and undergrowth, at Caves National Monument includes typical mountain least; and most forests of higher elevations have been vegetation of the central Siskiyous, and an area of burned at some time (cf. Leiberg 1900). forests farther east is protected by the Mt. Ashland watershed. Two larger areas of the western Siskiyous LITERATURE are maintained by the Siskiyou National Forest as the There are apparently no published studies of the Illinois Canyon and Kalmiopsis Limited Areas. vegetation of the Klamath Region, except the author's Only limited farming is carried on in the Klamath (Whittaker 1954b) earlier account of low-elevation Region, although the upper Rogue River Valley is diorite and serpentine vegetation in the Siskiyou an important fruit-growing district. Parts of the Mountains. Brief descriptions of the Klamath Region Klamath Region are used as range land, although are given by Harshberger (1911), Peck (1925, 1941), the forests of most of the region provide little food and Munger et al. (1926). Jepson (1923-5, 1935), for stock. In the Siskiyou study areas, limited graz- Mason (1927), Peck (1941), and Detling (1948b) ing in the forest lands of the serpentine areas seems have commented on the concentration of narrowly not to have damaged their vegetation; limited grazing endemic species in this region. Vegetation descrip- in the higher elevations of the diorite area has severe- tions of other areas most pertinent to interpretation ly affectedsome of the drier mountain meadows. Gold of the Siskiyou vegetation include Cooper (1922) on was discovered in the Siskiyous in 1851 or 1852, and the broad-selerophyll vegetation of California, Munz the local gold rush brought thousands of miners into & Keck (1949, 1950, 1959), Jensen (1947), and the Galice-Kerby-Waldo area iwhich includes the Burcham (1957) on California vegetation types, serpentine and gabbro mountains studied. Diller Shreve (1927), Clark (1937), Bowerman (1944), and (1914, see also Diller & Kay 1909, Maxson 1933, Sharsmith (1945) on the California Coast Ranges, Shenon 1933, Wells et al. 1940, 1949) mapped and Baker (1951), Merkle (1951), and Detling (1953, described 52 placer mines and 67 lode mines and 1954) on the Oregon Coast Ranges, Grinnell & Storer prospects in this area. Little more than marginal (1924), Klyver (1931), and Oosting & Billings (1943) mining is carried on in the area now; and the mining on the Sierra Nevada, Grinnell et al. (1930) on Mt. settlements of Gold Rush times have shrunk to vil- Lassen and Merriam-(1899) and Cooke (1940, 1941, July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 285

1955) on Mt. Shasta, Becking (1956) on the Douglas- more typical soil parent-material in the central Siski- fir forests of the Northwest, and Hansen (1947) on you Mountains, (3) the parent-material series fronm northwesternvegetation and forest history. diorite through gabbro to serpentine, in the central The foreign studies most relevant to the present Siskiyou Mountains, and (4) the climatic gradient work are two studies dealing with vegetation of ser- from humid-maritimeto drier and more continental pentine and normal soils and the mafic soils which climates, from the Pacific Coast inland along the Cali- are intermediateto them (Socava 1927, Pichi-Sermolli fornia-Oregon border. Within each elevation belt, 1948). General characteristics of serpentine vegeta- parent material, and location along the maritime-con- tion as they appear in the Siskiyous and other parts tinental gradient, samples were taken to represent the of the world were reviewed by the author (Whittaker topographic moisture gradient. The study is based 1954a, 1954b); and this work includes (together with primarily on the comparison, not of individual com- additional European papers cited by Novak 1923, munities or community-types,but of moisture-gradient Pichi-Sermolli 1948, and Rune 1953) a bibliography patterns of vegetation as these change in relation to of world literature on serpentine vegetation. Cita- elevation, parent material, and the east-west climatic tions which should be added to it are discussions of gradient. Balkan serpentine by Beck (1901) and Adamovi6 The area of most intensive study was in the quartz (1909), an account of grass and herb communities on diorite mountains lying behind Oregon Caves Na- serpentine in an area of heath in Shetland (West tional Monument and including Grayback Mountain. 1912), descriptions of grassland and pine communi- This and the other study areas in the central Siskiyous ties on serpentine in Hungary (Soo 1934, Zdlyomi may be reached by side roads leading from the series 1936), mention of the effectof gabbro in Greenland of villages (Kerby, Selma, Cave Junction, and by B6cher (1933 :13), and observations in Switzer- O'Brien) along U. S. Highway 199 in the Illinois land by Lfidi (1937) and in the California Coast Valley in southernmost Oregon, back into the moun- Ranges by Sharsmith (1945). The extensive recent tains east and west of the valley. The diorite area is literature includes material on serpentine vegetation best reached from the Oregon Caves, at the end of a in Germany (Gauckler 1954) and Austria (Eggler road leading from Cave Junction, and a branch from 1954), Bosnia and Serbia (Ritter-Studnidka 1953, this road up Grayback Creek. About 290 vegetation 1956, Pavlovi6 1953, 1955, Krause & Ludwig 1956, samples, from elevations between 550 and 2100 m, 1957, Krause & Klement 1958), Scandinavia (Knaben were taken from this area in the summer of 1949. 1952, Rune 1954b, 1954c, Kotilainen & Seivala 1954), Study of effects of parent material was made Italy (Gismondi 1953), Cuba (Smith 1954), New possible by the existence, across the Illinois Valley, Caledonia (Baumann-Bodenheim 1956), Japan (Ya- 27 km west and 52 km northwest from the diorite manaka 1954, 1955, 1956, 1957, Hattori 1955, Taniguti area, of extensive outcrops of serpentine and gabbro. 1958), Great Britain (Steele 1955, Coombe & Frost These three rocks are part of a major gradient in 1956a, 1956b, Spence 1957), Quebec, (Rune 1954a), characteristics of parent materials, represented by and California (MeMillan 1956). The basis of ser- the sequence of intrusive, igneous rocks from acid or pentine infertility has been discussed by Hunter & felsic granite, through intermediate diorite, to basic Vergnano (1952), Vergnano (1953a, 1953b), Min- or mafic gabbro, and ultrabasic or ultramafic serpen- guzzi & Vergnano (1953), Martin et al. (1953), Wal- tine. No granite was available for study close to the ker (1954), Kruckeberg(1954), Walker et al. (1955), other areas. The main area of serpentine studied is Crooke (1956), Tadros (1957), and Krause (1958). accessible on the historic and scenic Wimer Road Taxonomic references used included Jepson (1923- leading southwest from O'Brien across Oregon Moun- 5), Peck (1941), McMinn & Maino (1937), McMinn tain, and other side roads leading in to Rough-and- Of these Peck (1939), and Abrams (1940-51). (1941) Ready Creek, Tennessee Mountain, and Eight-Dollar localized flora of applies most directly to the very Mountain. The gabbro area of York Butte is acces- special parent materials in southwestern Oregon, sible by a trail to York and Panther Creeks from the of the was carried out. Most plant where much study end of a minor road leading west from Selma to Oak names are in the form given in Peck's manual; names Flat. The serpentine sampling was carried out in not included in it, or given here in a differentform, the summers of 1949 and 1950, the gabbro samplinff are accompanied by authors' names where first men- in the summers of 1950 and 1951. These serpentine tioned. The more recent manual of Munz & Keck and gabbro areas are low mountains, mostly not ex- also includes most of the flora of the study (1959) tending above the level of the Klamath peneplain areas. (about 1200 m). The serpentine samples from this area were supplemented with samples from II. PROCEDURE various other, smaller outcrops from all elevations in different STUDY AREAS parts of the Siskiyou Mountains. The quantitative part of this study deals with re- For the study of vegetational gradation from the lations of plant populations, vegetation patterns, and coast inland, more limited sample series were taken at floras to: (1) topographic moisture gradients from low elevations on more typical parent-materials in six ravines to south-facing slopes, (2) elevation, on a areas from coastal redwood forest inland to an in- WHITTAKER EcologicalMonographs 286 R. H. Vol. 30, No. 3 terior valley. Locations of these areas and their environment, and evidence on fire-historywere also vegetation will be described in Part IV. Limited recorded. series of vegetation samples were taken also from a The most extensive work in gradient analysis in number of other parent materials, mainly metavol- this country, apart from the author's, has been that canic rocks, marble, slate, argillite, and quartzite in of the Wisconsin group (Cottam & Curtis 1949, 1955, the Oregon Caves area, from high-elevation forests 1956, Curtis & McIntosh 1951, Brown & Curtis 1952, in other areas of the Siskiyous, mainly near Ashland Hale 1955, Culberson 1955, Curtis 1959), using forest Peak and Preston Peak, and from successional com- samples based on random pairs of tree stems from munities. The study as a whole is based on 470 relatively large areas of more level terrain. The type formal vegetation samples, of the type to be described, of sample developed by the author for the present and supplementary field notes and plant collections. study is adapted to mountain topography, where a sample concentrated within a more limited area is VEGETATION SAMPIJES AND SOIL DATA needed. For the present study the sample based on the 50-in steel tape was considered effective,reason- the gradient analysis 60 vegeta- As the basis of ably rapid, and easily adapted to the variety of vege- moisture- tion samples, representing the full range of tation types studied. gradient conditions within elevation belts of 1000 ft An effort was made to collect herbarium speci- or 300 m, were sought from each of the three rock mens of all species dealt with in the study in sufficient belts on types at low elevations, and from all elevation numbers to represent differenthabitats in which the of relatively undisturbed diorite. Within a given area species occurred. The herbarium collection was de- field assistant walked over vegetation, the author and termined, except for some groups sent to other the mountain surface without set plan, but usually specialists, by M. Ownbey and A. Cronquist; and the line or ascending a slope either following a contour voucher specimens of mature plants were placed in vertically. In the former case samples were taken the Herbarium of Washington State University. from each new site or slope exposure which seemed Soil samples were collected from 15 stations, 5 favorable for sampling; in the latter case samples from different topographic situations within each taken at 200-ft (61 m) intervals of elevation. were rock-type, to compare inorganic nutrient conditions In case actual location of samples was subjec- either in soils developed from the three parent materials. chosen-only stands were sampled which seemed tively Analyses were carried out in the soils laboratory at of sufficientarea and homogeneity and in reasonably Washington State University; a summary of the re- undisturbed condition. After 40 of the 60 samples sults is given in Table 2. The "xeromorphic" char- for an elevation belt were taken, the last 20 samples were, if necessary, selected to obtain a reasonably even TABLE 2. Soil analyses for threeparent materials at representation of the differenttypes of topographic low elevationsin the central Siskiyou Mountains. Values sites. are averages of 5 samples representingdifferent topo- Vegetation samples were based upon a 50-m steel graphic situations on each parent material. tape, normally laid out perpendicular to the contour lines. All tree stems were recorded by diameters in a Quartz Olivine Serpen- strip extending 10 m on each side of the tape. This diorite gabbro tine tenth-hectaresample was supplemented by an addi- tional tenth hectare between 10 and 20 m on each Cationexchangeable ca- pacity,m.e./lOOg ...... 21.7 31.8 21.2 side of the tape, when stands were open or highly Exchangeablecations, mixed, but sufficientlyhomogeneous to permit such m.e./1Og: Calcium...... 9.8 7.9 3.3 expansion of the sample. Tree seedlings, shrubs, and Magnesium...... 2.6 2.6 13.0 Potassium...... 0 .56 0.48 0.12 herbs were counted in 25 one-meter-square quadrats; Sodium...... 0.29 0.28 0.30 these quadrats were the alternate square meters along Hydrogen...... 15.3 22.7 9.5 one side of the 50-m tape. To provide density as well Soil Acidity(pH) ...... 5.8 6.0 6.5 as frequency information, an effortwas made in all cases to count individual plants in the square meters, even though such counts have limited meaning for acter of serpentine vegetation (Whittaker 1954b) some plant species. When additional information on suggested to the author the desirability of comparing shrub or seedling populations was desired, a count levels of available soil moisture in soils of the ser- of individual plants was made in a strip 5 in on each pentine and diorite areas. On two occasions, in 1949 side of the tape. In stands with sparse herb growth and 1951, series of soil samples were taken from dif- the herb-layer samples were sometimes expanded to ferent topographic sites in the two areas to determine 50 sq in along the tape; and in all samples species moisture content by drying at 105'C. No significant observed outside the quadrats were recorded. Cover- differencesbetween the two areas were established. ages were determined by recording the presence or absence of herb and shrub (and in more open stands, ARRANGEMENT OF SAMPLES IN TRANSECTS tree) cover at the 100 points which marked the corners The study of Siskiyou vegetation patterns is based of the 25 sq m undergrowth quadrats. Location and on the premise that within each parent material, at July, 1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 287

a given elevation in a limited area, one major en- are effectivelyshown by the composite topographic vironmental gradient is of such preponderant effect transects, however; and the distributional relations in determining differences of stable vegetation that thus indicated were the basis of the second ordination other environmental differences may be treated as technique, the composite weighted-average transects. secondary to it. This gradient is the topographic Species were grouped by the locations of their modes, "moisture gradient" leading from most mesic sites of or maximum population levels, along the gradient, as: ravines, through lower, or sheltered, mesic slopes, (1) Mesic-centered in step 1 or 2, (2) Submesic-in open north-facing and intermediate (east- and west- steps 3 to 5, (3) Subxeric-in steps 6 to 8, and (4) facing) slopes to open south- and southwest-facing Xeric-in steps 9 and 10 of the transect. Lists of xeric slopes. Along this gradient, the moisture con- species thus classified were prepared for each parent ditions affectingplants undoubtedly change, but also material and elevation belt and used as "ecological physiognomy of the vegetation, and other factors of groups" (Ellenberg 1948, Whittaker 1954c, 1956) soil and climate, change. The "moisture gradient" for the ordering of samples in that transect. Popu- is by no means simply a gradient of moisture, and it lation measurements for species were multiplied by is not a factor gradient in the usual sense; it is a weights (the numbers of the ecological groups above); gradient of whole environmental complexes, a "com- and the total of weighted values was divided by the plex-gradient" (Whittaker 1954c, 1956). It can never unweighted total, in the weighted-average technique tbe assumed that moisture itself is the cause of an ob- which was developed independently by Ellenberg served species distribution. The study is based, not (1948, 1950, 1952), Whittaker (1951, 1954c, 1956), upon any such-assumption, but on accepting the com- Curtis & McIntosh (1951), and Rowe (1956). Bi- plex-gradient as given, as a basis for ordering samples modal species and those for which the composite topo- into transects and comparing vegetation patterns be- graphic transect provided no significant indication tween differentparent materials, elevations, and cli- of moisture-gradientrelations were excluded from the mates. computations. Since the moisture gradient is not subject to direct Tree-stratum and undergrowth weighted averages measurement in the conditions of this study, other were computed separately; the values used for the means of arranging samples in sequence along the tree stratum were numbers of stems 1 cm dbh or over gradient must be sought. Techniques for ordering per 0.1 ha, for the undergrowthnumbers of apparent samples should, ideally, accomplish two -things: (1) in an effective They should express, and sensitive man- 4 I~oo 2.00l. 2490 4.00I93.00 36.5 ner, the relative positions of samples along the gra- dient in question. (2) They should indicate which 3:3.50- samples are deviant, departing from the main body of 0 samples along gradients of environment (or disturb- 247* ance) other than that being studied, gradients which FO3.2. 218 64 C also may be subject to later study (cf. Bray & Curtis D~~~~~~~~~~~ 1957). Formal statistical techniques may be adapted to ordering of samples and expression of their direc- (52.50O4 tions of interrelation (Goodall 1954a, Hughes & Lind- G)~~~~~~~~~~~~~~~~~~~C ley 1955). In the present study, however, the author -c si _ *ei0 has experimented with three less formal, quantitative 2.00 * but nonstatistical, approaches to sample arrangement or "ordination" (Goodall 1954b, Ordnung of Ramen- sky 1930). 1.50 .3 ' The first technique was based directly on topo- graphic relations of sites, and 10 types or groups of naio alon th 5osuegaint o-lvto .L.O sites from most mesic to most xeric: (1) deeper ra- 1.00 1.60 2.00 2.50 3.60 3.50 4.00 vines with flowing streams, (2) shallower or more Weighted Averages, Tree ____Mesic StratumXeiXrc open draws without streams, (3) lower, sheltered FIG. for ordi- slopes, (4) open N-, NE-, and NNE-facing slopes, 2. Weighted-averagetechnique sample nation along the moisture gradient. Low-elevation (5) open ENE and NNW slopes, (6) open E and samples from quartz diorite are plotted by weighted NW slopes, (7) open ESE and WNW slopes, (8) averages for both undergrowthand tree stratum, and open SE and W slopes, (9) open SSE and WSW grouped in 10 classes, or gradient-steps,along the mois- slopes, and (10) open S, SW, and SSW slopes. Sam- ture gradient. In this and Fig. 3, numberedsamples ples were grouped by these, into composite topo- outside the 10 transect steps are "deviant'" from the graphic transects for each elevation belt and parent main body of samples for reasons of parent material, material. elevation, geographic location, or disturbance. Samples in transect step 11 are sclerophyllstands withoutPseu- Topographic position alone is a relatively crude dotsuga,produced by fireson open south slopes, and more approach to moisture conditions of site. Over-all re- strongly xeric in compositionthan mature sclerophyll- lations of most plant species to the moisture gradient Pseudotsuga stands of comparable sites. 288 H. WHITTAKER Ecological Monograp.s R. Vol. 30, No. 3 plant individuals per 25 sq m. Fig. 2 illustrates the scatter-diagram which results when stands are plotted Moisture-GradientClasses 9 11 by weighted averages of the tree stratum on one axis, L.50 ;__ _,_1 2 31415 6 7 8 lo of the undergrowth on the other. The two values are necessarily correlated; a curvilinear,relation such 307 as illustrated results from the manner in which species Z1.00- are assigned to ecological groups. The oblique axis 4-G. 3._ Sample.-- ordination by Comp stan33with d- < 61 ~~~~128~~~ ~ 185 *32 *83 1.500 .0 10 20 .081 '5 30 of the scatter-diagram was taken as the best indi- 0 64 331 4)~~~ ~25 36 1 cator of relative position along the gradient; and in - .50- each case 10 segments of the scatter-figurecontain- E

ing 5-7 samples each were marked off,to provide the CO 10 steps of the transects. Sample deviance is some- times expressed in wide departure from the axis of 1.00 1.50 2.00 2.50 3.0 the scatter-figure,as indicated by the numbered sam- Mesic Moist ure-Gra di e nt I nde x Xeri ples and legend of Fig. 2. FIG. 3. Sample ordinationby comparisonwith stand- With deviant samples eliminated by this means ards. Low-elevation vegetation samples from quartz and judgment, the remaining 50 were grouped into dioriteare plottedby moisture-gradientindex (a weighted 10 sets of 5 each. Within each set, tree and under- comparison with samples representing mesic ravines, growth populations were tabulated for areas of 0.5 intermediateopen east slopes, and xeric south slopes) and hectare and 125 sq m, and coverages were averaged. by sample affinity(total of percentage similaritieswith Tree populations were also tabulated separately above these three comparisonstandards, as a measure of floris- tic consistencywith the main body of transect samples and below arbitrary size limits set to distinguish from diorite). Samples are grouped into 10 classes, or canopy and smaller trees. Compiled tabulations for transect steps, by position along the moisture-gradient undergrowth included, for each species in each step index. Samples with affinitiesless than 0.85 are con- of the transect, constancy, frequency, and density; sidered deviant from the main group of samples from from the data, presence, abundance, and an indica- low elevations on diorite. tion of contagion (Whitford 1949) could also be ob- tained. the greater stand-to-stand irregularityresulting from The third series of transects,the composite sample- fire. Quantitative indications of relative effective- comparison transects, were based on measurement of ness of differenttechniques of stand ordination are percentage similarity of samples (see Part VI). If possible, on the basis of certain assumptions about a stand at one extreme of the gradient is taken as a species distributions. It may be assumed, for exam- standard, then the degree to which other stands differ ple, that the more effectivethe sample ordination, the from it is in part an expression of their relative dis- more each species distribution will be concentrated in tances from it along the gradient. For a more sensi- part of the transect. Dispersions, or other indica- tive expression of stand position, all samples of a tions of relative spread of species populations in the transect were compared by undergrowth densities transects, may be used as an indication of effective- with three standards (each an average of 5 stands), ness of the ordination technique. Or, it may be as- (1) most mesic ravines, (2) intermediate, E-facing sumed that the more effectivelyis the natural distri- slopes, and (3) most xeric, S-, SW-, SSW-facing butional curve of the species represented in the tran- slopes. The resulting percentage similarities for a sect, the less will be the irregular, up-and-down dif- given stand were weighted from 1 to 3, and the ference of population measurements for that species weighted total divided by the unweighted total. The in successive steps of the transect. quotient, on the abeissa of Fig. 3, expresses the re- In application to the Siskiyou material, lists of lation of the stand to the moisture gradient and is species were selected for diorite, gabbro, and serpen- used to group the stands into transect steps. The un- tine, with the requirements that each species have weighted total, on the ordinate, expresses the affinity relatively high population levels for significance of of a sample to the main body of sample material as results, and have a distribution extending through the three low represented in comparison standards; several but not all 10 steps of the transects. The values provide an effective means of recognizing number of steps through which a species population deviant samples. extended was then taken as an indication of dispersion EVALUATIONOF TRANSECTTECHNIQUES, or spread; and these values were averaged for the on a soil. The three types of transects do not give equally species used to compare transects given satisfactory results, and none gave results as satis- The sums of signless differencesbetween successive factory as those of the weighted-average technique density values in steps of transects for the same in the Great Smoky Mountains (Whittaker 1956). species, also averaged for these sets of species, pro- The lower sensitivity of weighted averages as mois- vided an indication of relative smoothness vs. irregu- ture-gradient indicators in the Siskiyous is believed larity of the population curves in the transects. In to result from the smaller extent of vegetational each case, higher values indicate less effectivesample change along the moisture gradient (see Part VI) and ordination. The values obtained for relative spread July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 289

of species distributions in the topographic, weighted- These tables are intended to permit direct comparison average, and sample-comparison transects were: on of distributional relations of species, stand composi- diorite-8.1, 7.9, 8.2; on gabbro-8.3, 7.7, 8.0; on tions, and vegetation patterns on the three parent serpentine-8.7, 7.2, 7.6. Corresponding values for materials as the basis of the discussions which follow. relative irregularities of species distributions in the Only frequency values for undergrowth species can transects were: on diorite-32.0, 20.7, 23.0; on be given here from the density-frequency-constancy gabbro-17.6, 15.0, 20.0; on serpentine-34.9, 32.2, tabulations of the original transects; constancy and 34.7. density values for the transects as wholes are sum- TRANSECTTABLES marized in the last columns. Distributional relations to elevation on diorite were approached through a The main body of transect data for the central series of transects for differentelevation- belts. In Siskiyou Mountains are presented in the following the condensed form given here in Tables 12 to 14 tables, considerably compressed for publication. Dis- only average population values for whole transects tributions of tree, shrub, and herb species in relation are given for the various elevation belts-density and to the moisture gradient at low elevations on diorite, values of trees and per mille frequencies of gabbro, and serpentine are given in Tables 3 to 11. constancy shrubs and herbs. Distributions of grasses, sedges, TABLE 3. Distributions of trees in a moisture-gra- and rushes in relation to both parent materials and dient transectfor low elevations on quartz diorite in the elevation are summarized in Table 15. central Siskiyou Mountains. Based on 50 stand samples of 0.1 or 0.2 hectares each from elevationsbetween 2000 TABLE4. Distributionsof shrubs and seedlings in a and 3000 ft (610-915 m) arranged in 10 transect steps moisture-gradienttransect for lowAelevations on diorite of 5 samples each. All values in transect steps are in the central Siskiyou Mountains. Based on 50 under- numbersof stems over 1 cm dbh (i.e. fromthe 0.5-1.5 in growth samples, each of 25, 1-m2 quadrats, for eleva- class up) or, where indicated, 8 ia (20 cm) or 15 in tions between 2000 and 3000 ft (610-915 in) arranged (37 cm) dbh or larger in 0.5 hectare. Transect constancy in 10 transect steps of 5 samples each. All values ill is the per cent of 50 samples, 0.1 hectare each, in the transect steps are per cent frequencies in 100, 1-im2 transect in which the species occurred as a tree over 1 quadrats (based on 125, 1-m2quadrats per transectstep). cm dbh. Transect constancyis the per cent of 50 samples, each of 25, 1-im2quadrats, in the transectin whichthe species Sites occurred; total densityis the numberof apparent indi- viduals in an area of 1000 sq m, based on densitycounts Transectrstep 1 2 3 4 5 6 7 8 9 10X in 1250, -im2quadrats in the 10 steps of the transect. Observed presence in samples, outside the first25, -iM2 Conifers quadrats, is indicated by "x."' Chamaecyparislawsoniana . 102 75 2 22 15 in (37cm) dbh/and over 27 29 Pseudotsugamenziesei.. . 128 96 153 107 79 78 63 50 86 44 98 Sites 33u 15" and over...... 41 46 50 59 48 54 50 31 42 18 44 Pinus lambertiana..... 1 1 1 5 2 5 6 3 8 6 Transectstep 1 2 3 4 5 6 7 8 9 1OEtc;)E3 15" and over...... 1 2 1 6 2 5 1 Taxus brevifolia ...... 63 53 26 4 1 24 Coniferseedlings Abiesconcolor...... 1 1 2 2 12 Chamaecyparislawsoniana. . 2 2 6 5 Libocedrusdecurrens ... 1 12 4 Taxus brevifolia... . 7 2 2 2 14 15 Pinus ponderosa...... 2 Pseudotsugamenziesii. 2 6 4 9 3 10 12 8 56 70 Pinus lambertiana. .. . 2 2 1 1 1 10 6 Sclerophylls Libocedrusdecurren s 8 2 10 Lithocarpusdensiflora ...... 77 152 483 413 493 473 361 442 399 669 100 8 in (20cm) dbh/and over... 11 6 9 3 6 13 18 14 14 34 Broadleaftree seedlings Quercuschrysolepis .o . . . . 6 2 87 86 55 102 194 230 325 209 78 Acercircinatum ...... 5 2 6 2 18 22 8" and over ...... 2 7 21 34 24 12 18 Amelanchierflorida. .. 1 2 2 6 6 Arbutusmenziessi...... 2 2 27 28 50 75 93 124 79 322 68 Cornusnuttallii.... . 1 1 1 1 8 3 8" and over...... 1 1 4 9 17 40 27 48 13 21 Corylusrostratavar.californica 1 9 7 5 3 1 3 1 32 34 Castanopsischrysophylla ..... 23 34 89 79 44 52 11 20 43 162 74 Castanopsischrysophylla. 2 2 6 8 4 1 2 6 32 38 8" and over. . . . 1 2 1 3 2 1 3 1 Lithocarpusdensiflora. 31 18 27 26 39 28 30 21 34 30 100 374 Quercuschrysolepi . s . 3 8 6 13 5 20 21 14 14 64 148 Deciduoustrees Arbutusmenziesii.... 1 1 4 2 Alnus rubra...... 8 2 4 Alnusrhombifolia .9 1 4 Shrubs Acercircinatum ... 110 53 185 9 12 2 26 Rhododendroncalifornmcum 2 1 6 4 Acermacrophyllum 13 11 14 1 1 4 5 32 Vacciniumparvifolium 3 6 1 1 14 15 Salix sp...... 3 1 6 1 5 1 1 18 Caultheriashallon. 31 32 19 25 19 10 8 32 537 Cornusnuttallii ... 35 11 26 34 18 6 8 9 12 7 58 Rubusparviflorus.. . 1 4 2 1 1 2 16 14 Corylusrostrata var. californica . 141 146 58 34 16 11 2 2 4 52 Berberisnervosa ...... 42 45 21 26 25 20 14 9 9 76 632 Quercuskelloggii . .... 3 1 1 1 7 7 18 Berberispumila. .. 2 1 10 13 2 16 50 Amelanchierfiorida.. 6 3 4 Rubusvitfolius. . 27 26 6 7 5 20 10 11 19 32 70 271 Rosa gymnocarpa 8 14 22 32 24 38 27 30 37 43 98 497 Arborescentshrubs Rhusduversiloba. . 6 12 15 16 15 49 37 31 38 64 611 Philadelphuslewisoi 2 2 4 Holodiscusdiscolor. 1 2 2 10 6 Rhododendroncalifornocum ...... 15 4 4 Pachystimamyrsinites 1 1 4 2 Holodiscusdiscolor ...... 9 8 10 8 4 9 14 32 Symphoricarposhesperius ... 3 9 1 1 5 10 26 Rhusdiversiloba. . 2 2 Ceanothusintegerrimus x 0 0 Monographs 290 R. H. WHITTAKER Ecological

TABLE 5. Distributions (frequencies) of herbs in a TABLE 6. Distributions (stem densities per 0.5 hee- moisture-gradienttransect for low elevations on diorite tare) of trees in a moisture-gradienttransect for low ele- in the central Siskiyou Mountains. For basis of data see vations on olivine gabbro in the central Siskiyou Moun- heading of Table 4. tains. For basis of data see heading of Table 3.

Sites3 Sites~ .

Transectstep 1 2 3 4 5 6 7 5 9 1 EU6 E Transectstep 1 2 3 4 5 6 7 8 9 10 ca

Tiarellaunifoliata ...... 6 4 10 Conifers Saturejadouglasii ...... 2 2 6 Chamaecyparislawsoniana ...... 87 6 12 Asarumcaudatum ...... 1 2 4 15 in (37cm)dbh/and over . 6 4 Claytoniaparvifolia ...... x e e Pseudotsugamenziesii ...... 56 33 24 30 44 32 24 23 12 10 94 Tolmieamenzzesii ...... x O O 15"and over. . 11 13 16 14 14 10 6 8 2 3 Claytoniaspathulata ...... 2 1 4 2 Pinuslambera ....na . 38 4 9 15 21 20 46 36 11 21 90 Aradiacalifornica ...... 2 1 4 4 15"and over . . . 2 3 12 8 6 16 13 3 5 Boykiniaelata ...... 85 1 6 99 Pinusponderosa . . 46 1 6 1 1 18 11 6 36 Adiantumpedatum var. 15" andover ...... s.. 5 8 8 1 aleuticum.....,.,. 2 1 4 4 Libocedrusdecurrens . . 38 3 4 3 7 9 3 11 36 Smilacinastellata ...... 3 2 4 25 15"and over 2 2 1 2 2 3 Anemonedeltoidea ...... 7 4 2 2 16 40 Taxusbrevifoli a... 3 2 Cephalantheraauotinae ...... x x 6 Pinusattenuat. 1 2 Linnaeaborealis ...... 27 29 26 17 9 32 316 Trilliumovatum, ...... , 6 6 8 8 2 32 42 Sclerophylls Chimaphilamenzies . i . 2 2 1 2 10 14 Lithocarpusdensiflora.. 97 340308 243 280 79346 336 332 144 94 Pyrolapicta ...... ,,, . 2 2 2 1 10 10 8 in (20cm)dbh/ and over ... 12 28 26 8 11 5 10 7 Seneciobolanderi ...... 6 6 29 14 1 2 24 157 Quercuschrysolepis ...... 22 190232 271 179 326 339 230 339 266 96 Vancouveriahexandra ...... 1 6 4 2 2 2 22 38 8" andover ...... 2 30 35 60 41 46 22 14 12 Chimaphilaumbellata var. Umbellulariacalifornica ...... 17 69 37 22 26 31 43 17 28 53 54 occidentalis...... 2 4 24 14 13 6 8 1 32 249 8" andover ...... 3 Whippleamodesta ...... 8 6 28 22 12 10 3 10 4 44 340 Arbutusmenziesii ...... 1 49 12 31 15 21 13 5 42 Smilacinaracemosa ...... 13 1 4 9 x x 1 2 22 60 8" andover ...... 17 2 6 9 1 Polystichummunitum ...... 22 26 6 5 2 2 10 2 1 48 102 Arctostaphyloscinere .31 a 2 5 47 35 170397 289 805 56 Violasempervirens ...... 10 7 12 23 6 9 4 6 42 210 8" andover ...... 1 Galiumtriflorum ...... 3 6 3 6 2 1 1 2 2 32 34 Castanopsischrysophyll. . 1 2 79 6 Adenocaulonbicolor ...... 2 1 4 8 5 8 2 9 34 101 8" andover ...... 3 Irischrysophylla ...... 1 1 2 1 1 5 4 2 18 52 Achlystriphylla ...... 22 49 19 36 42 19 18 20 8 2 90 1561 Deciduoustrees Trientalislatifolia ...... 6 8 29 10 10 16 6 10 17 8 70 402 Alnusrubra ...... 17 6 Goodyearadecipiens ...... 2 2 7 9 10 7 1 2 1 1 42 81 Alnusrhombifolia ...... 6 4 Pteridiumaquilinum var. Slix ...p...... 15 4 pubescens...... 2 7 4 7 12 24 17 28 22 55 72 375 Amelaschzerflorida ...... 49 3 6 Apocynumpumilum ...... 5 2 4 10 11 14 22 15 10 26 66 203 Acermacrophyllum ...... 4 13 1 12 Disporumhookeri ...... 1 4 15 10 12 9 6 5 4 10 52 106 Cornusnuttallii .... 13 112 13 20 26 2 9 20 4 44 Lonicerahispidula ...... 1 6 8 2 6 13 8 15 2 48 145 Quercuskelloggii ...... 3 1 3 6 Hieraciumalbiflorum ...... 1 6 9 8 5 24 6 9 21 2 54 202 Allotropavirgata ...... x 1 1 x 4 2 Arborescentshrubs Heucheramicrantha ...... 1 2 3 Rhododendronoccidental...... 256 152 1 6 16 Equisetumhyemale var. Rhamnuscalifornicavar.occdentali.s 50 1 6 7 24 28 23 100 35 48 californicum...... 2 2 5 Garryafremontii ...... 8 2 2 1 12 2 2 5 26 Coptislaciniata ...... 14 4 42 Rhusdiversiloba ...... 2 1 17 20 12 Mitellaovalis ...... 1 2 1 Vacciniumovatum ...... 130 61 5 17 107 1 119 23 20 42 48 Clintoniauniflora ...... 1 2 2 Holodiscusdiscolor ...... 7 4 5 2 1 14 6 18 Corallorhizastriata 0...... x O Ceanothusintegerrimus ...... 1 4 8 8 Campanulascouleri ...... 2 1 1 4 1 12 41 Asarumhartwegi . : ...... 2 x 5 2 5 10 20 Liana Corallorhizamaculata ...... x 1 1 1 6 7 Vitiscalifsrnica...... 1 2 Phloxadsurgens ...... 2 3 4 7 2 5 1 10 36 78 Madiamadioides ...... 1 2 4 1 4 10 5 26 106 Lathyruspauciflorus ...... 2 1 1 1 1 2 12 10 Arenariamacrophylla . 2 2 2 These results are in accord with subjective judg- 1 2 1 Habenariaunalaschensis effectivenessof the three tech- Fradariavesca var. bracteata. 1 1 6 2 ment on the relative Tauschiakelloggii x 9 9 7 27 16 20 130 niques. The weighted-average technique is consistent- Collomiaheterophylla .. . . 1 2 6 3 2 18 81 ly superior to the other two in this application. It is 22 18 30 22 4S 404 Campanulaprenanthoides. . 1 1 6 10 consequently the transects based on this technique Pyrolasecunda ...... 1 2 1 Pleuricosporafimbriolata x 0 0 which are published here (Tables 3-11) and used as Boschniakiahookeri 1 1 x 1 6 6 the principal basis of describing and comparing Cyprzpediumfasciculatum. x 1 2 4 3 vegetation patterns. The other transects, however, Osmorhizachilensis . . .. 2 2 5 Psoraleaphysodes 4 1 1 13 2 8 31 have been used as checks on the weighted-average Cynoglossumgrande 3 2 10 transects in some respects, especially on the bimodality Claytoniaperfoliala var. of species populations (Part VII). The sample- 1 2 2 parviflora(Dougl.) Torr. . necessarily less sensitive Epilobiumminutum 2 2 2 comparison technique is not Galiumcalifornicum 2 2 5 and has some advantages over the weighted-average technique; bimodalities of a number of major species July, 1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 29i

TABLE 7. Distributions (frequencies) of shrubs and TABLE 8. Distributions (frequencies) of herbs in a seedlings in a moisture-gradienttransect for low eleva- moisture-gradienttransect for low elevations on olivine tions on olivine gabbro in the central Siskiyou Moun- gabbro in the central Siskiyou Mountains. For basis tains. For basis of data see heading oP Table 4. of data see heading of Table 4.

Sites~(UC Sites ._

Transectstep 1 2 3 4 5 6 7 5 9 Transectstep 1 2 3 4 5 6 7 8 9 10 cvo: Conifer seedlings Boykiniaelata .45 10 204 Chamaecyparislawsoniana .. 4 4 4 Boykiniamajor .6 4 11 Pseudotsugamenmiesii ...... 2 2 4 2 4 2 6 9 8 3 54 46 Rudbeckiacalifornica. 8 4 23 Pinuslambertiana ...... 1 2 2 3 2 2 5 7 8 2 56 36 Adiantumpedatum var. Libocedrusdecurrens ...... 1 1 1 2 4 14 8 aleuticum.5 8 13 Pinus ponderosa 1 1 2 6 5 Claytoniaparvifolia . 1 2 1 Goodyearadecipiens . 1 2 2 Broadleaftree seedlings Erigeroncervinus ...... 4 2 19 Alnusrhombifolia ...... 2 2 2 Luinahypoleuca . 2 2 Alnusrubra ...... 1 2 1 Tofieldiaglutinosa ssp. Amelanchierflorids ...... 2 2 1 occidentalis... .. 10 4 22 Lithocarpusdensiflora ...... 2 15 19 11 14 6 11 8 6 8 80 112 Darlingtoniacalifornica 5 2 18 Quercuschrysolepis ...... 2 13 12 19 10 6 15 6 3 6 74 96 Habenariasparsiflora. 1 2 1 Cornusnuttallii ...... 2 1 6 4 Peltiphyllumpeltatum . 6 4 11 Umbellulariacalifornica .... 2 1 1 2 2 x I 1 18 9 Galiumaparine .1 2 1 Corylusrostratavar.californica 1 2 1 Cypripediumcalifornicum 3 1 8 15 Arctostaphyloscinerea 1 2 2 2 6 8 12 36 34 Epipactisgiganta . 2 1 4 4 Castanopsischrysophylla.. . 1 2 1 Woodwardiafimbriata . 6 3 8 7 Querceuskelloggii 1 2 1 Schoenolirionalbum . 2 2 6 6 Lotusoblongifolius . 14 1 3 11 10 70 Shrubs Smilacinaracemosa . 2 2 6 1 2 2 2 4 18 26 Holodiscusdiscolor ...... 2 4 2 Galiumbolanderi .1 x 1 2 2 12 9 Rubuslitifolius ...... 2 6 2 Polystichummunitum var. Rhododendronoccidentale ..... 16 7 12 30 imbricans.2 23 49 29 26 41 11 7 l 60 290 Gaultheriashallone ...... 9 11 8 3 5 7 15 73 Trientalislatifolia 2 3 1 3 2 7 4 3 26 58 Rubusparviflorus ...... 3 1 1 2 8 13 Pyroladentata .I x 5 2 8 8 6 x 24 67 Garryafremontii ...... 1 x 2 6 2 Polygalacalifornica . 12 9 9 16 14 21 17 18 28 29 74 392 Rhamnuscalifornica var. Whippleamodesta . 10 31 34 42 22 50 27 31 21 10 88 769 occidentaiis...... 10 5 5 4 9 5 7 14 3 10 60 72 Lonicerahispidula . 4 27 39 42 38 34 30 7 15 12 86 508 Vacciniumovatum ...... 12 3 2 4 6 5 11 6 5 3 42 66 Irischrysophylla . 1 3 3 13 7 6 20 18 11 14 64 213 Rhusdiversiloba ...... 1 22 27 22 27 30 17 6 18 1 74 353 Campanulaprenanthoides. 1 2 4 2 11 1 14 6 1 36 146 Ribescruentum . . 1 x x 2 1 Vancouveriaplanipetala 3 2 17 Berberisnervosa ...... 3 11 4 38 Cheilanthesgracillima x 0 0 Symphoricarposhesperius ... 2 2 6 8 Achlystriphylla . . 6 2 6 43 Rosa gymnocarpa 4 1 2 6 10 Sedumoregonense . . 2 2 4 17 Holodiscusdumosus 2 1 3 2 8 9 Arenariamacrophylla 1 1 4 10 Ceanothusintegerrimus 2 1 2 6 5 Boschniakiahookeri . . 1 1 4 2 Amelanchiergracilis ...... 2 2 2 1 1 10 7 Asarumhartwegi . . 2 2 6 x 6 10 Berberispumila 2 x 2 6 Heucheramicrantha . . 7 1 3 2 2 14 24 Quercuschrysolepis var. Disporumhookeri . . 2 3 5 13 6 2 1 32 46 vaccinifolia 2 4 1 6 8 Apocynumpumilum. . x 4 1 1 6 8 Convolvuluspolymorphus .... 2 2 2 4 12 15 Chimaphilameniesi. . 2 6 2 5 1 1 1 24 29 Juniperusoibiric i x 0 0 Pteridiumaquilinum var. pubescens. . 2 2 2 2 1 4 10 13 12 28 76 Galiumambiguum . . 3 7 12 11 34 26 43 34 65 76 908 Hieraciumalbifflorum 4 5 1 12 x 14 46 (which could be excluded from the weighted aver- Pterosporaandromedea x 0 0 ages) are thought to be the reason for its relative Cheilanthessiliquosa 2 2 2 ineffectivenessin the Siskiyou transects. Eryoimumcapitatum 1 2 4 2 Chimaphilaumbellata var. occidentalis . . . 2 2 4 3 8 27 III. VEGETATION DESCRIPTION Madia madoides.. 2 x 2 4 9 Arnicaspathulata var. Low ELEVATIONS ON DIORITE eastwoodiae . . . 2 22 9 14 2 7 9 97 Hieraciumbolanderi 1 2 2 2 4 11 9 3 30 In 85 general character, the vegetation of low eleva- Pyrolapicta f. aphylla 1 2 1 tions on diorite is a forest of two tree strata-an Ceanothusprostratus x 2 2 2 upper stratum of evergreen needle-leaved trees and Lomatiumhowellii 1 4 .5 18 3 14 18 85 Anemone a lower quinquefolia 4 3 6 22 one of sclerophyllous broad-leaved trees- Arnicaparviflora ssp. forming together a closed canopy. In ravines, domi- parvllora. 1 2 12 21 8 15 12 106 nance of the upper tree stratum is shared by Pseudo- Viola lobata 3 2 5 14 14 6 24 126 Sidalcea malvaeflorassp. tsuga menziesii and Chamaecyparis lawsoniana. Three elegans 2 14 1 2 12 30 deciduous broadleaf trees-Acer macrophyllum,Alnus Allotropavirgata x 2 x 2 2 rubra Bong. (A. oregona Nutt), and A. rhombifolia Monardellaodoratissima var. glauca. 3 -occur in small numbers among the larger stems. 1 1 6 3 9 14 41 Linnaea borealis...... 1 38 4 57 Among the smaller trees, two deciduous species, Cory- Cordylanthusviscidus ...... 1 1 4 2 WHITTAKER EcologicalMonographs 292 R. H. Vol. 30. No. 3 (Table 8 Cont.) TABLE 10. Distributions (frequencies) of shrubsand seedlings in a moisture-gradienttransect for low eleva- > Sites 9$ Sn~ tions on serpentine in the central Siskiyou Mountains. For basis of data see heading of Table 4. Transectstep 1 2 3 4 5 6 7 8 9 O0s, st Sites~ Erigeronfoliosusvar.confinis. x I 1 4 10 Eriophyllumlanatum.. . a 1 1 4 8 5 Asterbrickellioides 2 2 4 6 Transectstep 1 2 3 4 5 6 7 5 9 10 BE E: Xerophyllumtenaz. 14 8 2 2 12 70 Coniferseedlings Ceanothuspumilus 2 7 15 14 20 62 Chamaecyparislawsoniana .. 6 7 Zygadenusmicranthus ...... 1 2 4 3 8 Pinusmonticola ...... 5 15 28 14 6 15 2 44 98 Cynoglossumoccidentale 1 2 1 Pseudotsugamenziesii ...... 2 14 15 10 11 6 3 46 71 Solidagoop ..... 5 1 8 11 Libocedrusdecurrens ...... 1 2 10 10 8 3 14 11 52 71 Monardellaojilosa var. Pinuslambertiana ...... 1 1 2 6 3 12 12 subserrata...... 7 9 6 25 Pinusjeffreyi . . . 2 1 3 2 2 4 1 1 24 16 Seneciofastigiatus ...... 2 2 2 8 12 Pinuscontorta var. murrayana 3 2 4 Eriogonumnudum ...... 1 2 4 9 Pinusattenuat 2...... 1 Phloxspeciosa ...... 1 2 2 Phlox 0 0 diffusa...... x Snrubs Pentstemonazureus ssp. Rosacalifornica C. & S.. 1 2 2 arureus ...... 2 2 2 Physocarpuscapitatus . 2 2 2 Salixsp .1...... 2 1 Rhododendronoccidentale.....43 2 12 77 TABLE 9. Distributions (stem densities per 0.5 hec- Lithocarpusdensifiora var. tare) of trees in a moisture-gradienttransect for low ele- echinoides.5 29 34 24 30 10 44 317 vations on serpentinein the central Siskiyou Mountains. Umbellulariacalifornica.....3 2 10 8 8 6 1 1 40 44 For basis of data see heading of Table 3. Vacciniumparvsfolium . 6 22 13 10 8 2 2 1 34 245 Amelanchiergracilis . 1 4 10 19 3 7 2 2 38 72 Rosagymnocarpa .2 12 15 4 6 2 1 2 36 66 Sites Berberispumila .2 25 8 14 8 18 9 6 44 211 Garryabuxifolia .7 8 26 12 6 15 16 10 48 154 Transectstep 1 2 3 4 5 6 7 8 9 10 E- Quercuschrysolepis var. vaccinifolia.13 54 47 46 27 38 22 15 76 646 Conifers Arctostaphylosnevadensis 4 6 2 10 10 10 5 3 2 28 205 Chamaecyparislawsoniana ...... 155 5 2 9 16 Convolvuluspolymorphus... 6 x 9 5 4 14 10 16 9 2 62 116 10in (25cn) dbhand over .... - 98 4 1 5 Rhamnuscalifornica var. Pinusmonticola ...... 152 93209 79 46 204 27 60 occidentalis. 14 4 12 18 18 4 8 2 4 4 50 110 10"and over ...... 8 4 20 7 3 30 Arctostaphyloscinerea and Pinuslambertiana ...... 2 23 9 11 25 7 17 1 2 56 canescens.1 2 2 2 x 1 12 14 10"and over ...... 2 9 6 9 11 2 3 1 1 Arctodtaphylosviscida ...... 3 6 4 14 6 10 7 3 6 50 82 Pinusattenuata ...... 12 11 16 48 17 23 37 3 2 38 Holodiscusdumosus . . 1 1 3 1 10 8 10"and over 1 1 Vacciniumovatum x 1 2 1 Pseudotsugameniesii ...... 43 61 54 41 52 15 87 21 13 84 Juniperussibirica ...... 10 3 7 9 7 16 99 10"and over ...... 21 43 26 12 20 4 22 9 2 Rubusvitifolius 6 2 15 Libocedrusdecurrens ...... 14 15 26 51 69 51 67 40 48 6 86 Castanopsischrysophylla var. 10"and over ...... 5 6 15 7 13 11 15 10 12 minor...... 4 2 7 Pinusjeffreyi ...... 34 10 34 24 28 65 95 86 108146 78 Rhododendroncalifornicum 1 2 1 10"and over ...... 2 2 10 8 3 18 14 39 36 38 Gaultheriashallon 7 20 4 145 Pinuscontorta var. murrayana 1 1 22 6 Quercusgarryana var. breweri 6 2 1 6 10 Ceanothuscuneatus ...... 3 2 10 Sclerophylloustrees Rhusdiversiloba 1 2 1 Arbutusmenziesii ...... 7 8 3 7 4 12 Cercocarpusbetuloides 1 2 2 Castanopsischrysophylla 15 9 4

Arborescentshrubs Rhododendronoccidentale...... 280 8 The principal shrubs are the evergreen species Physocarpuscapitatus ...... 10 2 Gaultheria shallon and Berberis nervosa; representa- Rhamnuscalifornicavar.occidentalis 7 4 Arctostaphylosviscida ...... 8 9 46 9 336184 28 tion of minor species is indicated in Table 4. Shrub Umbellulariacalifornica ...... 41 11 3 13 11 10 7 19 24 coverage ranges from less than 10% to more than Quercuschrysolepis var. vaccinifolza 9 2 1 10 40%, depending mainly on the occurrence of dense 3 2 Rhododendroncalifornicum patches of Gaultheria and Berberis; herb coverage Quercusgarryana var. breweri 4 15 6 is lower-5 to 20 %. The largest part of the herb stratum is made up of species with broad extent into lus rostrata var. californica A. DC. and Acer circi- sites other than ravines (see Table 5), which pre- natum, are most numerous (Table 3). These species dominate on slopes above the stream-side. The dis- and others form a small-tree stratum which is dense tinctive stream-side flora itself is not nearly so well (averaging around 1400 stems over 1 cm dbh per ha), developed as in the more open ravine stands on other physiognomically mixed (including needle-leaved soils; Boykinia elata is the principal riparian species evergreen, broad-leaved evergreen, and broad-leaved (see Table 5 and distributional grouping 1, Part V). deciduous species, the last predominant) and floristi- The selerophyllous small trees-Lithocarpus densi- cally rich (with 10 or more small-tree species in some flora, Quercus chrysolepis, Arbutus menziesii, and stands). Castanopsis chrysophylla-predominate in the lower July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 293

TABLE 11. Distributions (frequencies) of herbs in Table 11 (Cont.) a moisture-gradienttransect for low elevationson serpen- tine in the central Siskiyou Mountains. For basis of Sites data see heading of Table 4.

rransectstep 1 2 3 4 5 6 7 8 8 lo S Sites~ Castillejapruinosa ...... 4 4 Tauschiaglauca ...... 2 1 6 11 14 1 2 5 6 7 5 6 36 Transectstep 3 4 Haplopappusracemosus ssp. congestas...... Cypripediumcalifornicum.. 3 6 2 2 3 14 6 16 94 Sanicula peckiana Darlingtoia californica...... 6 4 25 ...... 6 8 10 20 7 3 22 170 Eriophyllumlanatum var. Rudbeckiacalifornica.. 23 10 67 achillaeoides Tofieldiaglutinosa spp...... 2 3 2 2 8 19 14 14 40 182 Horkeliasericata 2 occidentalis...... 10 6 44 ...... 5 27 14 2 22 27 28 675 Brickelliagreenei 1...... Castillejaminiata ...... 6... 5 4 18 2 1 Veratruminsolitum Lotusoblongifolius ...... 5 2 32 2 2 2 Microserisleptosepala Habenariasparsiflora ...... 1 2 2 (Nutt.) Habenariaunalaschensis .... 1 2 1 Gray .. 2 6 Pteridiumaquilinum Trilliumrivale ...... 1 2 1 var. pubescens . Lilium occidentale...... x O 0 ...... 7 7 4 24 Lupinuslatifolius var. Galiummultiflorum ...... 2 2 7 columbianus Adiantumpedatum var...... 3 10 8 14 Thlaspi 1 aleuticum...... X x O alpestreL... x 4 4 10 Achillealanulosa . .1 Fritillariaatropurpurea.... . 4 6 6 2 1 1 8 11 Eriogonumnudum Heleniumbigelovis ...... 40 x 10 208 ...... I x 1 2 4 8 15 Strepanthushoweliji 6 Lonicerahispidula ...... 2 1 6 2 x 6 Horkeliatridentata x 6 Goodyearadecipiens ...... x 2 4 5 0 Cordylanthusviscidus 16 1 Ligusticumapiifolium ...... 10 7 4 7 14 38 10 2 6 26 71 Erigeronfoliosus var. confinus Chimaphilaumbellata var. 1 3 9 5 3 1 24 102 Zygadenusmicranthus occidentalis...... 1 22 1 5 2 14 81 1 2 3 5 4 4 28 28 Calochortus Liliumhowellii ...... 2 1 2 3 1 14 8 howellii 2 1 2 6 8 50 Grindelsamaritima Smilacinaracemosa ...... 5 1 2 1 1 14 13 (Greene) Steyerm 1 Antennariasuffrutescens ..... 3 6 1 3 10 50 2 1 Lithospermumcalifornicum .. Angelicaarguta ...... 3 1 6 2 1 1 14 19 2 4 2 Phloxdiffuse Vancouveriachrysantha ...... 2 5 31 5 5 22 171 2 2 3 Eriogonumpendulum 6 Epilobiumrigidum ...... 2 2 6 14 12 109 x 0 Penstemonlaetus ssp. roezlii. Disporumhookeri ...... 8 16 11 6 1 26 84 2 2 2 2 x 14 20 Monardellaodoratissima var. Trientaloslatifolia ...... 17 18 35 26 31 11 14 8 62 838 glauca Pyroladentata ...... 4 14 6 2 4 6 x 2 42 106 1 6 2 1 10 26 Senecio Arnicaspathulata var. fastigiatus 3 3 1 6 2 14 28 Perideridiaoregana eastwoodiae...... 2 30 20 4 14 3 2 1 42 207 1 2 2 x 8 6 Cerastiumarsense Whippleamodesta ...... 16 16 29 7 17 10 18 1 50 413 3 2 4 12 Sidalcea Lomatiumhowellii ...... 2 10 15 8 12 17 4 1 40 158 malvaeflorassp. elegans Xerophyllumtenax ...... 16 18 27 28 20 39 24 2 1 62 562 1 7 8 1 10 46 Calochortustolmiei Galiumambiguum ...... 4 7 17 20 30 31 25 18 5 78 634 1 1 6 8 12 32 Lomatiummacrocarpum Polygalacalifornica ...... 1 1 1 1 6 x x 3 1 22 26 1 15 26 51 20 374 Phaceliadasyphylla var. Schoenolirionalbum ...... 16 2 10 3 1 7 16 32 115 Iris bracteata...... 5 31 49 46 29 39 31 22 5 1 86 944 ophitidis x 2 4 6 17 Arenariahowellii Phloxspeciosa ...... 2 15 24 9 14 5 10 11 5 I 58 266 6 2 22 Brodiaeahendersoni Cheilantheosiliquosa ...... 4 5 3 1 15 20 20 12 44 157 1 2 2 6 9 Allium Viola lobata...... 2 4 14 14 23 15 22 6 9 5 54 431 falcifulium 3 x 2 4 Eriogonumternatum Ceanothuspumilus ...... 1 3 15 3 5 13 22 36 71 53 72 869 x 11 4 30 Gilia capitata Lomatiumtriternatum var. 1 2 1 Erigeronbloomeri var. nudatus macrocarpum...... 1 6 6 6 4 208 26 26 4 62204 (Gray) Cronq Hieraciumcynoglossoides var. 5 2 8 Epilobiumpanniculatum var. nudicaule...... 1 13 5 6 8 9 12 5 11 3 60 180 hammonds 2 4 Apocynumpumilum . .... 2 2 4 5 Blepharipappusscaber Seneciobolanderi ...... 2 2 2 9 2 29 Polygonumspergulariaeforme. 6 Hieracium albifiorum...... 7 2 15 x 6 Monardellavillosa var. Allotropavirgata ...... 1 2 2 subserrata 6 Chimaphilamenziesii ...... x 1 2 2 x 6 Arenariamacrophylla ...... 1 1 4 6 Polystichummunitum var. . 2 2 1 10 4 imbricans ...... tree stratum Linnaea borealis ...... 14 4 6 89 in all stands but those of ravines, with Silenecampanulata var. average stem densities per hectare increasing from orbiculata ...... 1 2 4 6 1200-1400 in submiesic to 1600-2500 in xeric sites. ... 2 2 4 17 Campanulaprenanthoides Four major shrub species-Rosa Rubus Arnicaparvifiora ssp. gymnocarpa, parvifora ...... 2 4 2 4 4 14 38 vitifolius, Berberis nervosa, and Rhus diversiloba- Asterbrickellioides ...... 6 7 10 26 2 3 x 34 104 occur in submesic and subxeric stands. Shrub cover- . 1 2 1 x 6 4 Lotuscrassifolius ...... age in these stands varies between 10 and aver- Sedumlaxum ...... 1 2 6 1 1 1 6 18 30 40%, Balsamorhizadeltoidea .s. 6 8 11 18 8 38 13 4 x 40 186 aging about 30%; herb coverage ranges from 2 to Balsamorhizadeltoidea x 12%, averaging 6-7%. The herb stratum also is 4 3 2 6 39 platylepishybrids ...... dominated by a group of species which occur Balsamorhizaplatylepis ...... 12 6 10 2 2 4 1 16 162 along Dicentraoregana ...... 2 2 2 almost the whole length of the moisture gradient on 294 R. H. WHITTAKER EcologicalMonographs

TABLE 12. Distributions of trees in relation to elevation on quartz diorite in the central Siskiyou Mountains. Based on 6 moisture-gradienttransects of differentelevation belts; transects nos. 1 to 9 each included 50 stand samples of 0.1 or 0.2 hectare each in 10 transect steps, transectno. 11 included 16, 0.1 or 0.2 hectare samples in 8 transect steps. Density values are numbers of stems over 1 cm dbh (i.e. fromthe 0.5-1.5 in class up) or, where indicated, 8 in (20 cm) or 15 in (37 cm) or over in areas of 5.0 hectare per transect. Constancyvalues are per cents of the 50 (or 16) samples, 0.1 hectare each, of the transectin whichthe species occurredas a tree of the 1-in or larger diameter class.

Transectno. 1 3 5 7 9 11 Elevationsin feet 1500-2500 2500-3500 3500-4500 4500-5500 5500-6300 6300-7000 Elevationsin meters 460-670 670-1070 1070-1370 1370-1680 1680-1920 1920-2140 Dens. Cons. Dens. Cons. Dens. Cons. Dens. Cons. Dens. Cons. Dens. Cons. Conifers Chamaecyparislawsoniana ...... 93 13 499 46 439 54 370 32 15" and over...... 42 167 160 121 Taxus brevifolia...... 215 16 185 26 10 12 54 12 Pinus ponderosa...... 1 2 3 2 Pseudotsugamenziesii ...... 904 96 767 92 674 100 435 90 5 10 15" and over...... 378 506 604 330 4 Libocedrusdecurrens ...... 1 2 15 6 8 6 72 24 11 12 15" and over 2 2 37 6 Abiesconcolor...... 5 10 215 40 905 100 1602 100 1241 58 15" and over 11 143 477 404 Pinus lambertiana...... 35 42 20 22 7 12 29 8 1 2 15" and over...... 14 13 7 9 Abiesnobilis 2 2 71 40 1260 92 2400 100 15" and over 1 22 421 693 Tsuga mertensiana 2 2 796 54 1338 75 15" and over 205 368 Pinus monticola 1 2 3 6 Sclerophylls Lithocarpusdensiflora ...... 4236 100 2007 76 54 6 8" and over...... 109 66 1 Quercuschrysolepis ...... 1422 90 463 48 59 8 8" and over...... 171 57 Arbutusmenziesii ...... 856 82 341 56 72 16 8" and over...... 135 142 23 Castanopsischrysophylla ...... 470 82 712 62 339 50 13 6 8" and over...... 18 103 24 Deciduoustrees Fraxinusoregano...... 4 2 Cornusnuttallii ...... 201 60 48 12 Quercuskelloggii...... 22 18 11 8 Acercircinatum ...... 642 46 198 22 32 10 Acermacrophyllum ...... 61 24 32 24 17 4 15. 8 Salix sp...... 29 24 2 4 1 2 4 4 Alnusrubra and A. rhombifolia...... 34 10 23 12 4 2 3 .2 Amelanchierflorida ...... 25 10 1 2 33 8 3 2 Corylusrostrata var. californica...... 497 56 246 38 109 12 78 30 4 2 Acerglabrum var. douglasii 14 6 190 24 23 10 Sorbusamericana ...... 5 4 Arborescentshrubs Rhus diversiloba...... 2 2 Philadelphuslewisii...... 8 6 Rhododendronoccidentale...... 2 Ceanothusintegerrimus ...... 4 4 15 10 Vacciniumparvifolium ...... 3 2 1 2 1 2 Holodiscusdiscolor ...... 171 38 33 20 55 18 53 18 5 2 Rhododendroncalifornicum ...... 290 24 318 18 164 10 Quercussadleriana 1 2 24 6 17 2 Cornusstolonifera 3 2 1 2 Loniceraconjugalis ...... 3 2 x 0 Ceanothusvelutinus 3 2 Prunusemarginata 1 2 July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 295 TABLE 13. Distributions of shrubs and seedlings in TABLE 14. Distributionsof herbs in relation to ele- relation to elevation on quartz diorite in the central vation on quartz diorite in the central Siskiyou Moun- Siskiyou Mountains. Based on 6 moisture-gradienttran- tains. For basis of data see heading of Table 13. sects for differentelevation belts; transectsnos. 1 'to 9 each included 50 undergrowthsamples of 25, -iM2 quad- Transectno. 1 3 5 7 9 1t rats each in 10 transect steps, transect no. 11 included -Elevationin feet 1500- 2500- 3500- 4500- 5500- 6300- 16 samples 2500 3500 4500 5500 6300 7000 of the same size in 8 transect steps. All Elevationin meters 460- 760- 1070- 1370- 1680- 1920- values are per mille frequenciesin transects (the num- 760 1070 1370 1680 1920 2140 ber of 1-M2 quadrats, among 1000 such quadrats, in C&noglo.ssumgrande. 3 which a species was observed,based on 1250, -iM2 quad- Claytontaperfoliata ....1.. . I rats in the ten steps of transects1 to 9, 400, -iM2 quad- Epi7lbiumminutum ...... 2 ' rats in transect no. 11). Observed presence in one or Saturejadouglasii ...... 2 more samples of the transect,outside the 25 m2 samples, Lathyruspauciflorus ...... 6 1 is indicated by "x." Cwpripeaiumfasciculatum .... 2 3 Boykiniaelata ...... 2 12 Adiantumpedatum var. Transectno. 1 3 5 7 9 11 aleuticum. .... 4 3 Elevationin feet 1500- 2500- 3500- 4500- 5500- 6300- Aralia californica... 4 2500 3500 4500 5500 6300 7000 1 Elevationin meters 460- 760- 1070- 1370- 1680- 1920- Lonicerahispidula ...... 83 14 760 1070 1370 1680 1920 2140 Tauschiakelloggii ...... 51 23 Campanula Coniferseedlings prenanthoides 106 56 7 Madia madioides .... 22 7 3 Chamaecyparis lau oniana. . 2 13 7 9 Boschniakiahookeri ..... 2 2 2 Taxus brevifolia...... 12 12 5 5 Allotropavirgata ...... 1 2 1 Pseudotsuga men2iesii ...... 64 46 35 18 Whippleamodesta ...... 140 35 212 21 Pinu8 lambertiana...... 7 4 2 1 Linnaea borealis...... 179 128 96 30 Libocedrusdecurrens ...... 2 11 1 4 2 Asarumcaudatum ...... 1 1 8 17 Abiesconcolor 7 90 68 28 2 Tolmieamenziesii ...... x 5 2 Abiesnobilis 8 64 100 Collomiaheterophylla ...... 22 6 13 4 T8ugamertensiana 18 22 Psoralea 1 Sclerophyllseedlings physodes...... 28 1 12 Cephalantheraaustinae ..... x x 1 2 Lithocarpusdensiflora ...... 291 174 4 Listeracaurina ...... 2 1 9 1 Arbutusmenziesii ...... 2 1 Anemonedeltoidea ...... 23 24 126 163 62 Castanopsischrgsophylka..... 31 43 13 1 Achlystriphylla ...... 193 307 805 542 128 Quercuschrysolepis ...... 130 47 3 x 1 Trientalislatifolia 161 Deciduoustree seedlings ...... 77 205 296 37 Pteridiumaquilinum var. Cornusnuttali ...... i - 2 1 pubescers...... 141 163 53 63 Acermacrophyllum ...... 2 9 14 Viola sempervirens...... 57 154 248 Acercircinatum ...... 18 10 4 1 78 4 Apocynumpumilum ...... 102 66 58 68 45 Corylusrostratavar. californica 36 14 5 2 Seneciobolanderi ...... 105 12 40 62 1 Amelanchierflorida ...... 6 2 4 2 4 Tiarellaunifoliata .... . 5 27 78 197 12 Acerglabrum var. douglasi.. 1 2 Shrubs Galiumtriflorum...... 34 45 165 190 64 Goodyearadecipiens ... 40 91 49 Rhododendronoccidentale ..... 1 ..... 38 1 Adenocaulonbicolor ...... 34 79 Symphoricarposhesperius .... 32 6 1 289 268 14 Vancouveriahexandra .. . 23 11 140 311 Berberispumila ...... 17 15 2 50 Disporumhookeri ...... 48 79 152 307 7 Gaultheriashallon ...... 132 263 77 Iris chrysophylla...... 10 14 8 20 Berberisnervosa ...... 197 264 455 160 11 Coptislaciniata ...... 32 Pachystimamyrsinites ...... 6 9 13 1 .. 5 6 11 Asarumhartwegi ...... 18 14 17 58 6 Rhus diversiloba...... 226 85 1 Clintoniaunifiora ...... 5 21 81 172 15 Holodiscus 11 2 11 16 discolor...... Habenariaunalaschensis. . . 1 x 2 Symphoricarposrivularis ..... 8 1 4 32 19 Smilacinaracemosa ...... 47 14 11 21 51 22 Vacciniumparvifolium ...... 14 48 22 1 Polystichummunitum ... . . 88 53 42 17 2 Rubusvitifolius ...... 126 204 177 38 1 Trilliumovatum ...... 38 48 62 126 50 10 Rosa gymnocarpa...... 273 199 102 126 24 Chimaphilamenziesii... 7 42 46 56 66 38 Rubusparviflorus ...... 19 3 2 15 2 Chimaphilaumbellata var. Ceanothusintegerrimus x occidentalis...... 96 160 258 92 12 2 Rhododendroncalifornicum 65 30 14 Pyrolapicta ...... 5 4 8 16 55 47 Rubusnivalis 8 70 37 Hieraciumalbiflorum ...... 93 58 160 172 202 197 Quercussadleriana 21 22 8 18 Phloxadsurgens ...... 67 70 181 352 146 65 Vacciniummembranaceum. .. 1 13 28 8 10 Campanulascouleri ...... 10 29 171 328 204 35 Ribeslacustre 2 Arenariamacrophylla ...... 2 48 183 268 234 Ribesmarshallii 1 58 211 157 Corallorhisamaculata ...... x 9 5 13 5 7 Ribeslobbii x 2 Fragariavesca var. bracteata.. 6 11 59 102 70 45 Cornusstolonifera x Osmorhizachileneis ...... 4 4 110 284 219 82 Loniceraconjugalis x 6 x Equisetumhyemale var. Ribesviscosissimum 25 36 13 californicum Sambucusracemosa var. cal- ...... 2 Heucheramicrantha. . 1 licarpa (Greene)Jeps 2 Galiumbifolium ...... 1 Pleuricosporafimbriolata 1 diorite (see Cornuscanadensis 2 distributionalgroupings 5 & 6, part V). Mitellaovalis . . 14 82 Important grasses include Festuca occidentalis and F. Listeracordata. 1 4 ovina, Melica harfordii, Bromus suksdorfii,and Tri- Pyrolauliginosa ...... 3 2 setum, canescens. The herb stratumon Gaultheriaovatifolia. . 1 1 diorite in- Curallorhizastriata x 2 x cludes a group of non-greenvascular plants-Cepha- Pyrolabracteata 2 x x WHITTAKER EcologicalMonographs 296 R. H. Vol. 30, No. 3 Table 14 (Cont.) Table 14 (Cont.)

Transectno. 1 3 5 7 9 11 Transectno. 1 5 7 9 11 Elevationin feet 1500- 2500- 3500- 4500- 5500- 6300- Elevationin feet 1500- 2500- 3500- 4500- 5500- 6300- 2500 o500 4500 5500 6300 7000 2500 3500 4500 5500 6300 7000 Elevationin meters 460- 760- 1070- 1370- 1680- 1920- Elevationin meters 460- 760- 1070- 1370- 1680- 1920- 760 1070 1370 1680 1920 2140 760 1070 1370 1680 1920 2140

Osmorhizaoccidentalis 2 1 x Sanicula nevadensis. 2 7 Vicia californica... . 1 32 121 30 Anemonequinquefolia . 1 2 Lathyrus 5 74 126 24 polyphfllus. Erysimumcapitatum 1 7 Pyrolasecunda. . . . 1 25 63 122 112 Sedumoregonense .. 23 10 Claytoniaspathulata 2 7 41 74 2 Astersiskiyouensis .. . .8 242 Smilacinastellata 9 94 251 98 5 Ligusticumgrayi 26 42 Prunellavulgaris. I Lupinusalbicaulis . 10 2 Liliumcolumbianum 2 Gayophbtumnuttallii 2 Sarcodessanguine a . 1 Castillejapruinosa.. .. 2 Trautvetteriacarolinensis Eriophyllumlanatum.. .. 2 (Walt.) Vaill... . . 8 Epilobiumangustifolium. 15 Lotusoblongifolius. I Polygonumdavisiae . . . 25 Mitellacaulescens... . 14 Calptridiumumbellatum Orobanche . 3 uniflora.. (Torr.) Greece . . 5 areunculusuncinatus var. Phaceliaprocera 2 parviflorus(Torr.) Bens 1 1 Claytoniasibirica L x 5 Habenariasparsif ora 2 2 1 Streptopusamplexifolius. x lanthera austinae, Allotropa virgata, Corallorhiza Csrcaeaalpina var. pacifica 22 46 Ath riumfelix-femina. . 12 9 mnaculata and C. striatex,Pyrola picta f. aphylla Lewisialeanas .. 2 x (Smith) Camp, Pleuricospora fimbriolata,and Bosch- ,4ctaeaspicata. . . . 2 135 58 niakia hookeri-which are rare or absent on the other Nemophilaparviflora. 16 37 15 Rubuslasiococcus. 2 39 157 130 soils. Dicentraformosa 3 5 14 25 In xeric stands (Fig. 4) the coverage of Pseudo- Viela glabel. . 39 171 58 tsuga (and the smaller numbers of Pinus lembertiana) Claytoniaparvifolia. .. 2 1 12 22 Veratruminsolitum 1 3 5 13 is relatively low, generally well below 50%. In most Mstelladiversifolia. . 1 22 18 5 stands Lithocarpus is most numerous among smaller Corallorhizamertensiana x 1 6 2 stems, but Arbutus and Q. cherysolepisshare domi- Beo.kiniamajer . . 10 Heracleumlanatum. . . 4 nance with it. Rose gymnnocarpaand Rhlus diversi- Listeraconvallarioides... 1 lobe are the principal shrub species, as in subxeric Cardaminebrewers. 2 sites; these and the prostrate Rubus vitifolius are 1 Pterosporaandromedea. .. often the only shrub species present, apart from tree Pentstemonanguineus. . . x 4 Hackeliajessicaee . . . 5 20 seedlings. Shrub coverage averages around 30%, Monardellaodoratissima. 2 14 but with tree seedlings-making up most of this. Herb Agastacheurticefolia 2 11 coverage is consistently low (1-7%), with Pteridiurn Aquilegiaformosa . 1 9 7 Senecsetriangularis . . . 23 43 22 aquilinum var. pubescens the major species of the Polygonumphytolacceaefolsum 4 2 5 herb stratum (see Table 5, distributionalgroups 6 and Viola sheltonii 2 20 13 8). Arnicalat7folia 79 365 253 Phaceliamagellanica .. . . 15 42 22 Stands which are floristicallysimilar to these, but Erigeronaliceae ...... 2 32 13 without large trees of Pseudotsuga or P. lambertiana, Valerianasitchensis ssp. occur on some slopes, especially south-facing ones. 42 265 .?08 s7tchens7s...... more severe Angelicaarguta . . 1 5 10 These stands are considered products of Hydrophyllumoccidentale. . 67 76 10 fires, and fire-effectsare reflected both in size-distri- Mertensiabella. . 2 26 7 butions of the sclerophyll canopy and in the fact that Erythroniumpregonumand E. Pseudotsuga is present as seedlings, but absent as a grandiflorum ...... 1 3 7 Mitellabreweri. . 5 39 25 tree. Pentstemonnemorosus. . 2 128 254 Some trends along the moisture gradient in low- 200 Pedicularisracemosa . . 9 90 elevation diorite vegetation have been commented on Artemesiadouglasiana 3 Epilobiumpaniculatum. 6 (Whittaker 1953 :49, 1954b). Among the trees, the Pentstemontolmisi Hook 11 (leciduous broad-leaf growth-form predominates in Crypt.9grammacrispa var. numbers of stems of all sizes in mesic sites and de- acrostichoides(R. Br.) C. B. Clarke ..1...... I clines to 1-2% of stems in xeric sites. Character of Collinsiaparviflora 2 the canopy changes from evergreen needle-leaf doii- Delphiniumsonne i 7 2 nance in mesic sites, with relatively small numbers of Saxifragaferruginea 4 10 Pentstemonnewberrbi. 2 38 selerophylls, to a two-level canopy of Pseudotsuga Castillejaminiata.. . 2 13 and the sclerophylls in submesic and subxeric sites, Orogeniafusiformis. . 1 -7 to a closed canopy of sclerophylls with an overgrowth Epilobiumhornemanni . 6 18 Vsolapraemorsa ...... 2 5 of large Pseudotsuga in xeric sites. Shrub coverage July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CAUFORNIA 297

TABLI 15. Distributionsof grasses and grass-like plantsin relationto elevationon quartzdiorite (tran- sects 1-11), and in relationto parent-materialsat low elevations(transects 1 and 3 on diorite,18 on olivine gabbro,15 on serpentine). Observedpresence in the samplesof a transectis indicatedby "x." Per mille frequencies,and densitiesof apparentindividuals, for 1000, -rM2quadrats are givenfor all grass,sedge, and rushspecies together,based on 400, -rM2quadrats in transect11, 1250 quadratsin othertransects.

Transeetno. 1 3 5 7 9 11 18 15 Elevationin feet 25000 2500 MO05600* 3300 1500- 2000- 2500 3500 4500 5600 600 7000 2500 3000 Elevationin mete 460- 760- 1070-1370- 160- 1920- 460- 610- 760 1070 1370 1680 1920 2240 670 915 Parentmaterial diorite...... gabbroserpentine

P crnm...... x x a rubr...... x M*U=Arfordii.... x x x LJdao wpestris.... X S x x x Trisnucmnaunew.... x z x x x x Brousukaa.o.rfi.... z z x x x Malicanubula ...... x x x x x x Putwadoiow dis. . x x x x x x x Fuatums ubulat.... x Cars buladeri x rahw,udwi ,ora x x x Broomuoreaiaus. . x Cares ." i x x Cares usrkai..i x x Lwun parunitJml x x Pkvropopourvfau. x x Gberiasrida. x Lhuda x x x Cara tra. . x x 4. View withina sclerophyll-Pseudotsugafor- Carmroni ...... x x Pio. Tri= _x x est on dioriteon an open SE slope,inclination 300, at Poa baudori. x x 670 m, OregonCaves area, SiskiyouMts., Ore. Paeu- Jumucparri. x dotauga menzieii, Quercu8 chrypolepi8,and few Litho- Poia ms.sa x carpus denaiflora; Rhue diverailoba, Rosa gymnocarpa, EinusuaiAus. . x x x Lonicera hiapidula, Polyatichum munitum, Hieraciun Bromsuearinatue x albiflorum,Whipplea modeata. July 19, 1950. Jurci iWfmaliu x

Cars sndi vriao x continuum,three community-types are distinguished- Erehrscrisiueu S X Chamaecyparis-Pseudotsugaforest with few sclero- Caredis.s i .. x x phyllsand manydeciduous stems in mesicsites, Pseu- Agro~ags& x x BSipaklxm . .i x x dotsuga-sclerophyllstands withtwo-level canopies in Ca= aagmstiwar x intermediatesites, and sclerophyll-Pseudotsugastands C-. withquite open growthof conifersin xericsites. kosireide. x Pee rAiamata. X Mdivagwi...... x Low ELEVATIONSON GABBRO Brouwabreviaridatus x The general descriptionof the dioritevegetation Dtfwinia0.liforiica. x Koekaiecraidda ... as "mixed evergreenforest" can apply also to the Xitaxiefujubam.... x gabbro vegetationpattern at low elevations. Apart fromover-all physiognomic similarity and the sharing AUgraminoid speces Frequency/1000m2.. 110 96 269 307 304 227 182 728 of somespecies, however, the two vegetationpatterns Density/1000mn2.... 269 211 731 789 845 490 451 6066 are quite different.The gabbro vegetationis much I __ moreopen than that on diorite (cf. Figs. 4 and 5). Average densitiesof large stemsof conifers(37 cm is moderate,averaging 20-35% throughoutthe tran- dbh and over) wereless than half as greaton gabbro sect; herbcoverage is low, decliningfrom an average (8.4 stems/havs. 20.5 on diorite); and densityof of 12% in mesic to 6% in xeric sites. The whole larger sclerophyllstems is muchlower on gabbro in vegetationpattern, dominated as it is by largerever- morexeric sites. It is consequentlypossible to stand green needle-leavedand sllmallerevergreen sclerophyll on one hillsidein the gabbro area and look through trees,may be describedas Mixed EvergreenForest thecanopy to thesoil on another,nearby hillside (Fig. (Munz & Keck 1949); but the gradual changesalong 6); in the diorite area one cannot similarlylook the moisturegradient make mesic and xeric stands throughthe dense evergreencanopies. The under- very unlike one another. Within this -vegetational growthstrata on gabbroare in generalof lowercover- 2938 R. H. WHITTAKER EcologicalMonographs Vo.80,No. 3

FIG. 5. View withina more open Pinus-Paeudotauga- sclerophyllforest on gabbro, on an open east slope, in- clination 350, at 370 m near Panther Creek, Siskiyou Mts., Ore. Pinus lambertiana,Arbutus menziesii, Quer- c'us chrysolepi8,Rhu8 diversiloba.,Polygala californica, Whipplea modesta, Chimaphila umbellata var. occiden. talis. July15, 1951.

age than on diorite,and of quite differentfloristic FIG. 6. View into a Pinus-Pseudotsuga-Quercusstand composition. on gabbro fromacross a ravine; the stand photographed is on a NW slope, 320 inclination,400 Stands of ravines in the gabbro m, York Creek, area are domi- &iskiyouMts., Ore. Pinus lambertiana,Pacudotsuga ?twn- nated by Pseudotsuga menziesii and Chamaecyparis siesii, Quercue chryaolepi8,Lithocarpue densiflora,Um- lawsoniana; but significantnumbers of Pinus lam- bellularia californica. Kalmiopsia leachiana occurs in bertiana,P. ponderosa,and Libocedrusdecurrens also this stand and on ridge above. July 8, 1950. occur. Below the quite open canopy is an open stratumof sclerophylls;deciduous tree species are conifers,80% for the sclerophylls. The relativeim- muchless numerouson gabbrothan on diorite(Table portanceof P. lambertianais muchgreater than on 6). Shrub coverage in ravines is patch-likeand diorite; the ratio of larger stemsof P. lambertiana variable, but generallylow; Rhododendronocciden- to Pseudotsuga was about 2:3 on the average on tale is the principal species. The stream-sideherb gabbroand only 1:20 on diorite. Threeof the major florais muchricher in species than on diorite(Table sclerophylls-Lithocarpitsdensiflora, Quercus chryso- 8). Like the rest of the gabbro flora,it includes lepis, and Arbutus menziesi-are shared with the some species whichare shared withthe dioriteflora, diorite vegetation; but the sclerophyllstratum on some shared with the serpentineflora (see distribu- gabbroincludes two major species, Umbellilariacali- tionalgroup 1, Part V), and some (Luina hypoleuca, fornicaand Arctostaphyloscinerea Howell, whichare Peltiphyllumpeltatum, Epipactis gigantea,Engeron essentiallyabsent from the dioritevegetation. cervinus) encounteredonly in the gabbro transect. The principal shrub species on gabbro are Rhus A numberof sedges and rushes are conspicuousin diverailoba-theusual shrub dominant-, and Rhain- the herbstratum of ravineson gabbroand serpentine nu8 californicavar. occidentalisand Vacciniurmovat- (see distributionalgroup 1). um,both of whichare rare or absenton diorite.Shrub On mesic and submjiesicslopes, the stands are coverageincreased from averages of 10-12% in sub- dominatedby Pseudotsugaand P. lambertiana,in a mesic sites to 20-32% in xeric ones; herb coverage quite open tree stratum,together with a denser,but was low, averagingbetween 2% and 7% throughout not closed,lower stratum of sclerophylls. Estimated the transect. The rare endemic shrub Kalmiopsis coveragesfor the two strata averaged 40% for the leachianaoccurs in some,more xeric stands. Grasses July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 299 of broad distributionin the gabbro transectinclude low-elevationserpentine stands have a characteristic Melica harfordii, Festuca occidentalis, F. californica, sparse and xerophyticappearance. In the tree and Trisetum cauescens, together with the rush Lu- stratum,pines are more numerousthan Pseudotsuga zula campestris. and otherconifers, and deciduousbroad-leaved trees In most xeric stands on gabbro (transect steps are absent. The sclerophylloustrees are also virtually 7-10) Pseudotsugais stronglyoutnumbered by pines absent, as trees, from serpentine. Some of these (P. lambedtina and P. ponderosa). These stands sclerophyllsare verymuch in evidenceon serpentine, are unlike the 1Pseudotsuga-dominatedforests on however,as shrubs. Quercus chrysolepisis repre- diorite,and like the serpentinevegetation, in the sentedon serpentineby var. vaccinifolia(Kell.) Ein- mixtureof pines with I'seudotsuga and Libocedrus gelm.,the most abundantsingle shrubspecies there; in the upper treestratum and in the low coverageof Lithocarpusdensiflora is representedby var. echi- this stratum(below 50%f and in some stands below noides, Umbellularia californica by an unnamed 20%). The xeric stands on gabbro differalso from shrubbyvariant, and Castanopsis chrysophyllaby those on dioritein the occurrenceof Arctostaphylos var. minor(uncommon in the studyarea). Quercivs cinerea as the most numeroussmall tree and in the garryana occurs on serpentineas the shrubbyvar. low coverageof the selerophyllstratum (10 to 40%). bretveri(Engelm.) Jeps. Among other trees aind The gabbro vegetationis thus in general more shrubs a series of congenericpairs appear in non- open, less stronglydominated by Pseudotsuga,with serpentineand serpentinefloras, with the serpentine greaterimportance of pines and smallerof deciduous species in each case of smallerstature: Amelan:hier trees,than that on diorite. Libocedrusand P. pon- floridaand A. gracilis Heller, Garryafremontii and derosa are absent from low-elevationdiorite in this G. buxifolia,Rhamnus purshiana and R. californica area, except as rare large individuals presumably var. occidentalis,Holodiscus discolor and H. damosits survivingfrom past disturbance. On gabbro they (Nutt.) Heller, Ceanothusintegerrimus and C. pumi- are presentalong mostof the moisturegradient, with lus, Berberisnervosa and B. pumila. The impression all size classes includingcurrent reproduction repre- given by the serpentinevegetation is that, as the sented;presumably the more open structureof gabbro physiognomicchanges evident from diorite to gabbra vegetationmakes possible their reproductionthere. vegetationare carrieda step furtheronto serpentine. Physiognoinictrends along the moisturegradient the coniferoustree stratumis reduced to an open are less strikingon gabbro than on diorite. Cover- stand mostlyof smallerpines, the sclerophylloustree age of conifersdeclines somewhat, and proportionof stratumshrunken to a shrub layer, and the other pines among themincreases, along the gradientfrom broad-leavedtrees and shrubs replaced by smaller mesic to xeric. Sclerophyllcoverage declinesalong shrubs and dwarf shrubs. the gradientfrom submesic to xeric; but sclerophyll Stands of ravines on serpentine (Whittaker stem numbersincrease toward more xeric sites be- 1954b: Fig. 2) are still more open than those on cause of the largernumbers of small stemsof Arcto- gabbro, with Chamaecyparis lawsoniana and Pintus staphylos. monticola each contributingabout 3/8of theconiferous stems (Table 8). P. ,,onticola,however, is a small SERPENTINE LOw-ELEVATIONSON treeon serpentine,and amonglarger stems Chamaecy- AND THE TWO-PHASE EFECT paris is stronglydominant. As on gabbro,there are The vegetation of serpentine is still more open numeroussmall stems of Rhododendronoccidentale than that of gabbro (Fig. 7); even in mesic sites the along the streams; and most other shrubs of the serpentineflora may occur in ravines at low cover- ages. In the herb stratuminthe (listinctivecharacter of theserpentine flora is apparent; rare and nari'owly endemic-species (Cypripedium californicuim.lul- beckia californica,Darlingtonia californica, Trilln rivale,Lilium howelliiand L. occidentale)are mixed withmore widespread species of bogs and otherimoist situations(see Table 11, distributionalgroups 1 mind 4). Tree strata on submesicand subxeric sites are typically mixturesof several conifers-Libocedrus, Pseudotsuga, Pinus jeffreyi,P. monticola,P. lamu- FIG. 7. A serpentinelandscape at low elevations in bertiana, and in some stands P. attenuata. Tree the Siskiyou Mts., Ore.; view of Eight-Dollar Mountain, coverageof thesestands is low, generallybelow 50c/7, west-facingslopes, from Oak Flat Road, July 9, 1950. but shrub coverageis high,often over 80% in suh- Open Pinus jeffreyiwoodlands occur on the most open mesicstands, giving these a mostdistinctive physiog- with or W- and SW-facingslopes, mixed conifers patchy nomyof smallconifers in open growthabove a d(lense, 2-phase shrub undergrowthin somewhatless xeric situa- tions, and denser (but still rather open) stands of low selerophyllousshrub stratum (Fig. 8). Shrub Chamaecyparis lawsoniana, Pinus monticola, and Pseu- coveragesincrease along the moisturegradient from doteuga mensiesii on most mesic slopes and in ravines. 20-50% in ravines to 50-90% in submesicstanl:;, 300300 R. H.WHITTAKERH. WHITTAKER Ecological. Monographs ~~~~~~~~~~~~~~~~~~~~~~Vol.80, No. 3 leroides,and an abundant hut undeterminedsmall Carex sp. In moststands of intermediatesites on serpentine, the shrubs show a patch-wisedistribution, giving undergrowthof these stands a two-phasecharacter of essentiallyclosed shrub cover alternatingwith shrub-lessopenings with grasses and other herbs (Whittaker1954b). Since shrub cover is high in submesicsites but declinestoward more xeric ones, thereis a reversalof phase along the moisturegradi- ent.In submesicstands the shrubs form the continuous phase withdiscontinuous openings (Whittaker1954b: Fig. 3); but as shrubcover falls below 50% in sub- par.~~~~~ xeric sites, the grassy openingsform the continuous AM~~~~~~' phase with discontinuousshrub patches (Fig. 9). Diametersof patches of the dliscontinuousphase in the samples were mostly 2-5 mn,with occasional larger or smnallerones. In such two-phasevegetation the distributionsof shrub populations are nesesawrilynon-random; but thetendency toward correlation with one of thephases applies to populationsother than thoseof the domi- nant shrubs. Ten samples from vegetationof two-

4'.~~~~~

Fio. 8. A stand of a iniesicslope on serpentineat low elevations;SE-facing slope above CedarCreek, near theWinier Road, at 650 in,Siskiyou Mts., Ore. An open standof Chamaecyparialawsoniana, Pinus monticolaand Pseudotsuga menziesii above a dense evergreenshrub layer,90% coverage,dominated by Lithocarpus densi- flora var.echinoide8 and Quercue chrysolepi8var. vaccini- folia. October10, 1951. ,5 s .'e a ? and then decreaseto 20-50% in subxericand 0-20% in xericstands. Quercus ehrysolepis var. vaccinifolia, Lithocarpus densiflora var. echinoides, Vaccinizum parvifolium, Ocairryabuxifolia, and Umbellularia cali- fornica are the iprineipalspecies and formthe canopy of shrubs of minddleheight-generallv below, but approaching1 m, withUinbellularia extending some- what higher. Herb coverages in submesicsites on serpentine were relativelylow, ranging fromn1 to 20% and Fia. 9. A mixed-coniferstand wvith2-phase under- averaging 11%, but increased along the moisture growthon serpentine,open ENE slope, 250 inclination, gradientto 10-40%k,averaging 30%, in subxericsites. 740 m, on Tennessee Mtn. in the Siskiyou Mts., Ore. Compositionof the floristicallyrich herb stratum Pinus jeffreyi,P. monticola,P. lambertiana,Libocedruix decurrens,Pseudotsuga menziesii, in open stand (about maybe judged fromTable 11. Grassesand grass-like 40% coverage), scatteredshrub patches (24/. coverage) plants included J'ou rhizomnata,Trisetuin canescen9s, dominated by Quercus chrysolepisvar. vaccinifolia,andl Melica geiyeri. Liv-ulac(ampestris, Calarnagrostis kne- sparse grass (29%o coverage). July 18, 1950. July, 1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 301 phase character were selected, evenly divided between Arctostaphylos viscida is taller, and Berberis pumila, submesic and subxeric stands and numbers of herb Aretostaphylos nevadensis, and Convnolvulus poly- and shrub quadrats; and plant populations were tabu- morphus are lower than the shrub canopy, and OCcur lated by their occurrence in quadrats assigned, by independently of it. their predominant character, to either the shrub or Relations of herb species to the phases are sig- herb phase. Table 16 summarizes results for species nificantly correlated with their moisture-gradientre- lations. Whipplea and Trientalis, the only species con- centrated in the shrub phase, are the only species popula- TABLE 16. Two-phase relations of species whose distributional centers are in mesic or mesie- tions in serpentine vegetation of the central Siskiyou submesic sites; and they are also the only herb species Mountains. listed which occur in the well-shaded herb stratum FREQUEN- on diorite. The next 8 species listed, Xerophyllum CONSTANCY DENSITY FREQUENCY CY RATIO to Galium ambiguum are centered in submesic, or sub- mesic and subxeric, sites; the last 13 species are cen- No. out of 125 1-mi tered in subxeric or xeric sites. In general, the more a No. ofin iivi- quadratsin duals in whiehob- species population is concentrated toward the imesic 125 m2 served end of the gradient, the more it is concentrated also in the shrub phase; the more a species is concentrated to- (10 samples) open shruD open shrub open/shrub ward more xeric sites, the more it is concentrated in Shrubs the open phase. Grasses in general show marked Quercuschrysolepis var. concentration in the open phase, but significant data 10 20 186 16 85 0.19 vaccinifolsa data Umbellularsacalsfornica. . 5 4 8 2 8 0.25 are not available for individual species. Such Garryabuxifolia ...... 6 9 34 7 22 0.32 suggest differentresponses of species populations to Amelknchiergracilts 5 5 12 4 10 0.40 the somewhat differentmicroclimatic and soil condi- Lithocarpusdensifiora var. tions underneath the shrubs and in the grassy open- echinoides...... 4 12 31 9 17 0.53 Vacciniumparvifolium. 3 27 47 7 12 0.58 ings. It appears that these micro-environmentsmay Arctostaphylosviscida ... 5 12 13 10 11 0.9 also influence tree distributions through effects on Rhamnuscalifornica var. seedlings. Pseudotsuga, P. monticola, and P. attenu- occidentalis.. . . 5 14 13 12 11 1.1 Rosa gymnocarpa 5 16 8 10 7 1.4 ata tended to be concentrated in the shrubs; Libo- Berberispumsla ...... 6 39 27 17 12 1.4 cedrus to be indifferentor less strongly concentrated Arctostaphylssnevadensis. 3 25 11 10 5 2.0 in the shrubs; P. jeffreyi to be concentrated in the 6 19 4 14 4 3.5 Convolvuluspolymorphus..'. open. Herbs The two-phase character is developed in quite 48 15 34 0.44 Whippleamodeota ...... 8 131 differentdegrees in differentstands. The condition Trientalislatifolia 8 67 98 13 24 0.54 Xerophyllumtenax 6 53 36 18 22 0.82 is most easily recognized in those stands in which Phloxspeciosa ...... 7 22 34 14 12 1.2 the proportions of shrub cover and opening are rela- Iris bractea . ta . 10 159 157 48 47 1.0 tively even. Comparison of samples with regard to Viola loba ta . . 5 73 45 17 17 1.0 Asterbrickellioides ...... 4 22 14 8 7 1.1 disturbance and succession suggests that the two- Lomatiumhowellii 4 10 8 7 6 1.2 phase effect is characteristic of climax, rather than Balsamorhizadeltoidea 6 46 32 24 19 1.3 successional, stands. The two-phase effectseems best 29 18 1.6 Galiumambiguum . . 10 122 62 developed in some stands with well-developed soil, Ceanothuspumilus 6 110 33 27 13 2.1 Lomatiumtriternatum 6 39 11 15 6 2.5 but poorly developed in stands with more rocky, clear- Hieraciumcynoglossoides ly immature soils. It seems best developed also in var. nudicaule . . 6 22 8 12 4 3.0 those stands least disturbed by fire, but poorly de- Eriyeronfoliosus var. confinis . . 4 34 9 6 2 3.0 veloped in stands with more severe fire effects. Carexsp ...... 4 117 28 19 6 3.2 Two-phase vegetation, with reversal of phase along Horkeliasericata . . 2 72 9 12 3 4.0 an environmental gradient, is known for other cir- Eriophyllumlanatum var. of and krummholz achillaesides . . 2 8 1 4 1 4.0 cumstances-as the break-up taiga Polygalacalifornica 4 8 1 6 1 6.0 into tundra in the far North (Rousseau 1952), the Seneciofastigiatus 2 16 1 7 1 7.0 transition of deciduous chaparral into semi-desert in 9.0 Cheilanthessiliquosa 5 10 1 9 1 the Wasatch Mountains (Hayward 1948), and the Cordylanthusviscidus 4 20 0 9 0 Zygadenusmicranthus 4 12 0 7 0 aspen groveland (Coupland & Brayshaw 1953, Perideridiaoregana 2 4 0 3 0 Lynch 1955). In these circumstances the two-phase character seems transitional, an "ecotone" of the "mosiac-insular" type considered by Nytzenko (1948) represented by significant numbers in 125 quadrats the most common kind of ecotone. In a broad sense, of each phase. Of the shrub species, the first six the two-phase vegetation of the Siskiyou serpentine are necessarily concentrated in the shrub phase, for is a transition between closed forests of high eleva- they form the shrub canopy of this phase. Of the tions and the open pine steppes of low elevations and other species one (Rosa gymnocarpa) is often part most xeric sites (Fig. 12). The two-phase veget?- of the shrub canopy but occurs also in the openings; tion is an extensive community-type,diverse within 302 WHITTAKER Ecological Monographs R. H. Vo.30,No. 8 itself,however, with a distinctivephysiognomy, and grassyfloor, are unlikeanything yet described;they many character-specieswhich are centeredwithin it are of the widespreadphysiognomic type whichmay rather than in forest or pine steppe. Species be termedpine steppe or pine woodland. populationsshow all degreesof correlationwith one Although pine woodlands are of very wide or anotherof the phases; the phases are by no means occurrence,the two-phasestands of intermediatesites distinctand relativelyimmiscible communities in mo- on Siskiyou serpentineare most distinctive. Their saic arrangement. physiognomy-mixedconifers in open growthabove Too facile a descriptionof this vegetationas "an a dense scrub-oakstratum with grassy openings-is ecotone betweentwo communities"may do more to scarcelyduplicated elsewhere to the author'sknowl- obscurethan clarifyits significance,for the meanings edge, althoughpine heath and inixed coniferswith of "ecotone"and "(colrnmunity"in this descriptionare heathor chaparral are reportedfrom serpentine in scarcelyunderstood by ecologists. The authorwould otherareas (Ritter-Sudnicka1953, Yamanaka 1954, prefer a differentdescriptive interpretation. The 1956, 1957, McMillan 1956). Neitherphysiognomic open treestratum on serpentinemakes possiblewell- type occurs on dioriteor gabbro in the study area. developedundergrowth communities of highcoverage. Comparing the three vegetationpatterns described Along the wholeof the inoisturegradient the under- here, there is progressivereduction of tree-stratum growthis of mixed shrubs,and grasses and other coverage and biomass from diorite throughgabbro herbs; but the shruh populations are concentrated to serpentine;while undergrowth strata are betterde- in submesic sites, the glasses in more xeric ones. velopedon serpentine.Compared with the other vege- Along the moisturegradient from submesic to xeric tations,the serpentinevegetation suggests a shiftof sites,there is gradual shiftof balance betweenthese the concentrationof plant life fromthe tree stratum two major componentsof the undergrowth,as the downwardtoward the shrub and herb strata (Whit- extentof the woody strata whichthe sites can sup- taker 1954b). Viewing serpentinevegetation as a port declines. Since thereis also a tendencyfor the whole, a comparable shift of vegetationalbiomass shrubsto grow in closed patches,the vegetationas- towardthe lowerstrata along the moisturegradient sumes a two-phasecharacter with the relative im- is suggested-fromthe trees in mostmesic sites to a portanceof the phasesshifting continuously along the highlydeveloped shrub stratumin intermediateand moisturegradient. herb stratumin xeric sites. Within the serpentine In stands of xeric sites shrub coverage ranges vegetationthe herb and shrub strata show a dis- downwardto zero. Undergrowthof these stands is tinct inversecorrelation, such as is encounteredin predominantlyof grasses, in rathersparse growth; some vegetationconditions (Whittaker 1956) but grass-like plants include Stipa lemmoni, Sitanion not in others,and not on the othersoils in the Siski- jubatum, Melica geyerii Elyotis glaucus, Festuca yous. ovina, Broinus breviuristatts, Danthonia californica, Withinthe Siskiyou Mountainsat low elevations, Agrostis hallii, Koelewis' cristata, Carex sp., and Lu- serpentinevegetation shows a general consistency zula campestris. With these occur many broad- withwhat has been described;but climaticvariation, leaved herb species (Table 11, distributionalgroup- the extremelocalization of some of the species, and ings 7 and 9). In sonie xeric stands P. jeffreyi is probable variationin chemistryof serpentinerocks the only tree species lplesent,in very open stand imply differencesin florasfrom one outcropto an- (Fig. 10). Liboeedrusoccurs in smallernumbers in other. Smaller outcropsof serpentine,such as may bc observed west and northwestfrom the diorite A and which *;' v I'~ studyarea, oftensupport vegetation floras are intermediateto those of serpentineand more Al typical soils. Extent of developmentof the serpen- tine floraon one of these outcropsseems in general to be directlyproportional to its size and inversely proportionalto its distancefronm one of the larger serpentineareas. Departing also from what has been describedare thedense stands of youngstems of Pinus attenuataencountered in many areas, and the Jeffreypine woodlands on private land near the valleys,with shrub strata dominatedby Ceanothus lmueatus. These conditions are thought to he prod- woodlands with scat- Fio. 10. Jeffreypine widely ucts of disturbance,the formerby fireand the latter tered Arctoataphylo8viscida and grassy floor, on dry flats and lower slopes on serpentine. Oak Flat Road, by grazing. ;Ibout 400 In, Siskiyou Mts., Ore., October 9, 1951. FOREST VEGETATIONOF HIGHER most, and l'seulotsuga in sonme stands. Arcto- ELEVATIONSON DIORITE staphylos riscidta forms a very open high-shrub Althoughintensive study was concentratedin a stratum in some stands. Physiognomically these single area, serpentinevegetation was observedat all stands. with their pines in scattered growth above a elevations in other areas of the Siskiyou Mountains. July, 1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 303 In the diorite patternthe sclerophyllsare con- SISKIYOU MOUNTAINS centratedin the low-elevation,xeric sites, the lower FEET, METERS VEGETATIONPATTERN ON DIORITE rightcorner of Fig. 11. With increasingdeparture Tsuga mertensiana fromthese sites,toward either more mesic situations 2000 SUBALPINE FORESTS or higherelevations, the proportionof selerophyllsin Abies nobili the stands decreases. This distributionalrelation, 6000o and the complexrelations of moistureconditions to Abies concolor both topographyand elevation, imply an oblique boundaryfor the sclerophyll-Pseudotsugatype. Be- 5000esoo MONTANE yond this boundarythe Pseudotsuga stands form a stands at Chamaecyparis FORESTS pSe' transitionto Chamaecyparis-Pseudotsuga Pseudotsugo Co lower and Pseudotsuga-Abies concolor stands at higherelevations.

1000 The Chamaecyparis-Pseudotsugaforests, which are rathernarrowly restricted to ravinesand mostmesic slopes at lower elevations,expand towardhigher ele- vations onto many shelteredslopes and some open MIXED EVERGREEN 20 northerlyones. The character of these forests FORESTS 500 changesgradually as theyare followedfrom the low Scierophyll-Pseudotsuga elevationsalready describedto elevationsabove 1200 1000IO m. Representationof the sclerophyllsand all other i 2 3 4 5 6 7 8 9 10 Ravines RLowerSlopes* Ridges.Summit broad-leavedtrees decreasestoward higher elevations, Draws N_ E W S 0D e n S I o D e s NNENE N ENE NNWE NWESWNW SE SSE WS Sws 5 whileAbies concolor becomesincreasingly important. Mesic Xeric Shrubcoverage is low; the low-elevationspecies Rosa FIG. 11. Mosaic chartof vegetationon quartz diorite, gymnocarpa, Berberis nervosa, and Rubus vitifolius central Siskiyou Mountains, Oregon. are joined by otherspecies largely restricted to forests of higherelevations-Quercus sadleriana, Ribes mar- shallii, Vaccinium membranaceum. The change in SISKIYOU MOUNTAINS ELEVATION the herb stratumtoward higher elevationsis more FEET, METERSVEGETATION PATTERN ON SERPENTINE conspicuous;for herb coverageincreases to 20-60%, Tsuga mertensiona oE withmany mesophytic species including most of these X2000 SUBALPINEFORESTS 7F of low-elevationmesic stands and montane forests 6000- Abies nobilis E00 o Pinusjeffreyi, (distributionalgroups 3, 6, 10, and 11). a0x At elevationsbetween 1200 and 1850 m (4000- Abies concolor, Pseudotsuga, Grass 6000 ft) "Montane Forests" occur, dominatedby

500& 500o Xerophyllum tenax Pseudotsugaand Abies concolor at lower,by A. con- -> Pseudotsuga- color at higher,elevations. Chamaecyparisis present . o RP monticola, _ > Pseudotsuga- A.nevadensis- in some stands of more mesic situations,Libocedrus 40 u R monticola, I Xerophyllum 4000.Offi : X? Q.vaccinifolia I decurrens in some stands of more xeric situations. . FOREST-SHRUBCOMPLEX The only broad-leavedtrees of significanceare Acer . L >50% SHRUBS <50% glabrum var. douglasii and Corylus rostrata var. cali- - 3000 Pseudotsuga- / fornica; sclerophyllsare whollyabsent above lower E P monticola- , ? Libocedrus, / PINE elevationsof the montaneforests. Shrub coverageis O Q.vaccinifolia-, Pseudotsuga-Libocedrus L.montanus/ Rmonticola-RjeffreyirWOODLANDS low, generally0-10%, with Rosa gymnocarpa, Ber- 2000 /Q.vaccinifolia-G.buxifolia beris nervosa, Rubus vitifolius, and Holodiscus dis- / Pinusjeffreyi, .500 color the principal species. The herb stratumis / / Grass floristicallyrich and of high,though variable coverage 1000 as and 2 3 4 5 6 7 3 9 10 -30-80% in moststands, but as high 90% Ravines Lower Slopes Ridges.Summits Draws W S Open Slopes below 10% in some stands. These forestsshow, in NNE NE N ENE NNWE NW ESE WNWSE W SSE WSWS SW SSW less markeddegree, the gradationfrom rich to poor Mesic Xeric herb strata to be describedfor the subalpineforests. FIG. 12. Mosaic chart of vegetationon peridotiteand Herb species includeAchlys triphylla, the species of serpentine,central Siskiyou Mountains,Oregon (L. mon- greatestdensity and coverage,and many others of tanas = Lithocarpus densifloravar. echinoides). distributionalgroupings 6, 10, and 11, togetherwith some species of groupings3 in moremesic sites and The diorite vegetation was intensively sampled at all 12 toward higher elevations. Average composition elevations in the principal study area. Figs. 11 and of theseforests is indicatedby data for the 1370-1680 12 present "mosaic charts" (Whittaker 1956) for the m elevationbelt in Tables 12-15. two patterns in the central Siskiyous, based on 270 At elevationsapproaching 1850 m the A. concolor vegetation samples on diorite, 70 vegetation samples forestsare usuallymixed with Abies nobilis Lindl. (A. and 90 other field records of serpentine stands. procera Rehd.), sometimeswith Tsuga mertensiana. 304 R. H. WHITTAKER EcologicalMonograph Vol. 80, No. a At elevationsabove 1850 m, subject to differencesin coverage,Achlys triphylla, Campanula scouleri,Are- exposureand topographicsituation, the latterspecies naria macrophylla,Anemone deltoidea, Valeriana becomethe dominantsof subalpine forests. At still sitchensisssp. sitchensis,Phlox a(lsurgens,Osmorhiza higherelevations, above about 1920 m, T. mertensiana chilensis,Hieracium albiflorum, Claytonia spathulata, Viola glabella, Actaea spicata, Galium triflorum, Trilliumovatum-with high densitiesand with con- stancies of 80-100% in such stands. In stands of low herb coverage some of these,especially Arnica latifolia,Campanula scouleri,Hieracium albiflorum, and Arenariamacroplylla, are stillpresent, but repr(e- sentedby few individuals. Along withthese, the herb strata includeChimaphila menziesii, Pyrola secunda. Pyrola picta, Corallorhiza maculata, and toward lower elevations, Chimaphila umbellata var. occi- dentalis; the evergreenground heaths are a conspicu- ous part of the sparse herbstrata in such stands but are of low constancyin stands with well-developed herbstrata. One may recognizeat least threedistri- butional groupings or unions among undergrowth species of theseforests (cf. Qosting& Billings1943), although these groupings intergrade: (1) Species with maximum population levels in stands with rich herb strata, and limited occurrencein stands with poor herb strata and in meadows or other open communities,(2) Species of extensive oc- currence in meadows and other open cominuni- ties, which also occur in stands with rich herb strata (see distributionalgroupings 11 and 12), and (3) Species which are of greatest importancein stands with poor herb strata. The low coveragesof the shrub stratum,in which Ribes marshalliiis the only importantspecies, are correlatedwith the herb FIG. 13. A bies nobilia forestof high elevationsin coverages; in stands with herb coveragesover 70%, the SiskiyouMts., Ore. A NE-facingslope on Dutch- shrub coverages averaged 5%, in those with herb man Peak, 1920 m, July14, 1950. coveragesof 2% or less, 0%. Average undergrowth coveragesdecrease along the moisturegradient from and A. nobilis are the only tree species represented N- to S-facingslopes and fromthe 1370-1680to the in the stands; these are subalpine forests without 1920-2140m elevationbelt; and at any elevationand birches or any other broadleaf associates. Most exposure are inverselycorrelated with tree-stratum stands contain both dominants. Proportionsof the density. two are not stronglycorrelated with elevation or The general characterof the diorite vegetation site, suggestingthat stand compositionis to some shownby the mosaicchart is a patternof mixedever- extent determinedby chance in the newer stands green forestsat low elevations which is gradually whichhave developedfollowing fires, or in the pri- transformedtoward higherelevations into montane marysuccessions on glacial topography. Withinthe and subalpine coniferousforest. Apart from the subalpineforest pattern, however, there is some tend- Chamnaecyparis-Pseudotsugastands, the vegetation ency for A. nobilis to predominatetoward lower ele- types forma series of belts or zones towardhigher vationsand more xeric sites, T. mertensianatoward elevations;in life-zoneterminology (Hall & Grinnell higherelevations and moremesic sites. Canopy cover- 1919, Jepson 1923-5,Grinnell 1935, Cooke 1941) the ages average somewhatless than in the montanefor- Mixed EvergreenForest is presumablyto be identi- ests,and in some standsare below50%. fiedwith the Upper Austral,the forestsof Pseudo- Average compositionof the subalpine forestsis tsuga and Abies concolor with the Transitionand indicatedby data for the 1920-2140and 1680-1920m thoseof A. nobilisand T. mertensianawith the Cana- elevationbelts in Tables 12-15. These forestsshow dian and Hudsonian zones. a very wide range of herb-stratumcoverages-from A kind of relative discontinuity,defined by less than 1% to morethan 90%-and great variation growth-formsand not species distributions,between in appearance fromnrelatively open stands with lush subalpine forestsand those of lower elevationswas anal floristicallyrich herb strata to denser stands describedby the authorfor the Great SmokyMoun- with almost bare floorsof needles and twigs. The tains (Whittaker1956). No such relative discon- formerare characterizedby a group of herbs- tinuitycould be detectedin the Siskiyoupattern; the A rnica latifolia, the species of highestdensity and Subalpine and MontaneForests are continuouswith July, 1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 305

one another, and the latter with the Mixed Ever- snevadensis the principal shrub and Xerophyllum green Forests. Abrupt forest-edges occur between tenax the principal herb. high-elevation forests and meadows; but within the At highest elevations, Pinus iaonticola occurs in forests themselves the vegetation on diorite is a very open stand above an undergrowthstrongly domi- single, continuously intergrading pattern from most nated by Xerophyllum tenax. Physiognomically such mesic sites to most xeric, and from lowest elevations Xerophyllum-richstands have the appearance of pine to highest. Low-elevation vegetation patterns on "steppes." Floristically they are much like the high- serpentine and gabbro are correspondingly con- elevation Jeffreypine woodlands to be described, but tinuous; and the more limited samples from higher contain also some species of the A. concolor forests. elevations on serpentine indicate, though they are In a stand of the type which was studied, at 2040 m insufficientto demonstrate, the full continuity of the on a west-facing slope of Big Red Mountain, the serpentine pattern. coverage of P. monticola was below 30%, with few stems of A. concolor and P. jeffreyi. Shrub coverage VEGETATION OF HIGHER ELEVATIONS ON SER?-ENTINE was 24%, predominantly Arctostaphylos nevadensis, but with Amelanchier gracilis present; Xerophyllum The serpentine vegetation, described for eleva- predominated in a floristically rich herb stratum of tions of 610-920 m, has much the same character up 57% coverage. to elevations of 1100 and 1200 m, and (lown to eleva- At elevations mostly between 1370 and 1770 m, tions of 300 m and less. Toward lowest elevations, montane forest stands dominated by A. concolor and however, there is striking expansion of the Jeffrey Pseudotsuga with P. monticola, P. jeffreyi,and Libo- pine woodlands so that these, rather than the two- cedrus often present, occur. These are more open phase stands, become the prevailing vegetation type stands than the montane forests on diorite, with on serpentine slopes of low elevations (Figs. 10, 12). coverages generally between 40% and 70%. Shrub The Chamaecyparis-P. monticola stands are more coverages were rather low and herb coverages high, narrowly restricted to ravines than the comparable 13 and 48% in a representative sample, with Arcto- Chamaecyparis-Pseudotsuga stands on diorite. Some staphylos nevadensis and Xerophyllum tenax the expansion toward higher elevations occurs. and stands stratal dominants. The latter forms,in many of these dominated by Chamaecyparis with P. mionticolaand stands, an apparently "grassy" undergrowth. Flora Pseudotsuga were encountered on slopes landflats up of the undergrowth includes both some species char- to 1400 m. Such stands had high shrub coverages, 65 acteristic of serpentine and some of those of high- and 83% in two samples, with Quercus chrysolepis elevation diorite vegetation. At higher elevations var. vaccinifolia, Lithocarpus densiflora var. echi- closed stands of Abies nobilis were encountered in noides, and Vaccinium parvifolium the dominant the Observation Peak and Big Red Mountain serpen- shrubs, and low herb coverages (1-2%), with various tine areas near Ashland Peak. Physiognomically mesic and subinesic stands at lower of the species of they differedlittle from comparable stands on diorite, elevations. but many herb species characteristic of high-elevation The vegetation types which prevail on intermediate forests on diorite were not recorded in the stands ob- sites and intermediate elevations, and thus have a served. pattern, central position in the Siskiyou serpentine Most xeric sites on serpentine are occupied by have been grouped together as the "forest-shrubcom- pine steppes or woodlands at all elevations. Floristic plex." The diverse stands of this grouping are in composition of these gradually changes upward from general characterized by open canopies of mixed that already described at low elevations to the quite conifers, and well-developed undergrowths in which differentflora of high-elevation stands. P. monticola dominance is shared in varying proportions between and A. concolor occur in some of the high-elevation evergreen shrubs and grasses or grass-like plants. Jeffreypine steppes, and Libocedrus is often absent. In stands of higher elevations, 1100-1400 in, Pseudo- Shrubs included Chrysothamnus nauseosus var. occi- tsuga and P. mnonticolaare the dominant trees, with dentalis, Amelanchier gracilis, Ceanothus cuneatus, Chamaecyparis toward more mesic and P. jeffreyi Holodiscus dumosus, and Quercus chrysolepis var. toward more xeric sites, Libocedrus toward lower and vaccinifolia, with coverage below 10%. Grasses were Abies concolor toward higher elevations. Under- Melica subulata, Sitanion hystrix, Festuca orina, growths are of the same general character and com- Bromus breviaristatus and B. carinatus, and Trisetum position described for lower elevations, with Quercus canescens. Xerophyllum tenax was the principal herb chrysolepis var. vaccinifolia the principal shrub and species; other herbs shared with the low-elevation Xerophyllum tenax most conspicuous in the herb serpentine flora included Cheilanthes siliquosa, Mo- stratum. Toward still higher elevations, above about nardella odoratissimna var. glauca, Eriophyllurn 1400 m, stands of the complex extend upward in lanatum var. lanceolatum, Eriogonum nudum, Cas- subxeric sites between the A. concolor forests and tilleja miniata, Phlox diffusa,Achillea lanulosa, Pha- pine woodlands. In these stands Pseudotsuga and celia dasyphylla var. ophitidis, and Silene campanu- P. monticola are dominant, and the two-phase under- lata var. orbiculata Rob. Other species of the high- growth is replaced by lower strata of quite different elevation pine steppes were Iris chrysophylla, Vicia appearance and composition, with Arctostaphylos californica, Arenaria macrophylla, Aster siskiyouensis, Monographs 306 R. H. WHITTAKER Ecological

Pedicularis racemosa., Anemone quinquefolia, Erig- of charcoal in the soil, bark-burns or burned stumps, eron aliceae, Astragalus whitneyi Gray, Orthocarpus or in stand composition; it is believed that no low- copelandi, Anemone drummondii, Sedum obtusatum elevation stand unaffectedby fire was observed. All ssp. boreale Clausen, Linum lewisii, Pentstemon this vegetation is consequently "disturbed" in this azureus ssp. parvulus (Gray) Keck, Eriogonum urn- sense; but it does not follow from this that the vege- bellatumn, Calochortus elegans var. nanus Wood., tation is to be interpreted as simply secondary. Senecio integerrimus var. exaltatus (Nutt.) Cronq., Self-maintaining, all-age forest stands show a Erysimurn capitatum, Polystichumnscopulinum (Ea- characteristic J-curve for the relation of numbers of ton) Maxon, Aster shastensis var. eradiata, and Lu- tree stems to diameters of stems. Plotted with num- pinus leueophyllus. bers of stems on a logarithmic scale, these curves be- Sample coverage is inadequate to prepare a mosaic come somewhat convex upward (Whittaker 1956) in chart for gabbro vegetation. In the low elevations the form illustrated in Fig. 14 by no. 8, Pinus jeffreyi studied, the Chamaecyparis-Pseudotsuga type is more narrowly restricted to ravines than on diorite. Be- 60 cause of the greater openness of the gabbro stands, -)40 the sclerophyll-Pseudotsuga is in immediate contact with Chamaecyparis-Pseudotsuga, without an inter- vening Pseudotsuga-sclerophyll type. The more xeric 4)0 stands on gabbro, with pines predominant and a quite a)2 open sclerophyll stratum, represent a third physiog- nomie type within the Mixed Evergreen Forests (see

Part V). 30 South of the main study area, in the area of Youngs Peak, Sanger Lake, and Preston Peak, a dis- tinctive high-elevation forest pattern occurs, with 60o varied mixtures of Picea breweriana, Abies concolor C40 and A. nobilis, P. monticola, Pseudotsuga, and Libo- cedrus in rather open stands. In more mesic sites C10 2- - 0 5 3 141 7 02 3 near 1500 m stands of Chamaecyparis with P. brew- eriana and P. monticola occur; and in more xeric ones Co 2-4 5-7 80- 29-37 34-46 17-19 50-64 stands of A. concolor with Pseudotsuga and Libo- E30 23- cedrus. Toward higher elevations true spruce-fir stands occur, dominated by P. breweriana and A. nobilis, with P. monticola also present. Shrub and Co 20 herb coverages were generally low, with Vaccinium \8 N .0 membranaceum and Quercus sadleriana the principal 8:~~~~~~~~N shrubs, Achlys triphylla and Chimphilla umbellata var. occidentalis the principal herbs. This high-eleva- Z tion pattern is in some respects intermediate to those 3~~~ . 2-10 Il-I9 20-28 29-37 38-46 47-55 56-64 65- for of diorite and serpentine. Since observations Diameter classes in middle elevations are lacking, its continuity with the inches low-elevation pattern described for gabbro cannot be FIG. 14. Stand curves of stem numbersvs. diameters established. for principal trees at low elevations in the central Siski- you Mountains,with numbersof stems on a logarithmic IV. CLIMAX INTERPRETATION scale. Top-Pseudotsuga menriesai on quartz diorite (1), olivine gabbro (2), and serpentine (3). Middle of the raises The complex vegetation Siskiyous three selerophylls on diorite, Arbutus menriesii (4), Those to several problems of climax interpretation. Quercus chrysolepis(5), and Lithocarpus densiflora(6). be discussed are: (1) The role of fire, (2) The ques- Bottom-two pines onl serpentine, Pinus lambertiana tion of edaphic climaxes, (3) The basis of comparing (7), and P. jeffreyi(8). climax vegetation for differentparent-materials and climates, (4) Gradation of climax vegetation along on serpentine soils. A wide variety of all-age and climatic gradients, and (5) Prevailing or "climatic" presumably climax stands have been found to have climax types for this area. curves of this form, though it is by no means the only possible one for climax forests. When stand FIRE EFFECTS AND EDAPHIC CLIMAXES curves for differentspecies on differentsoils in the The forests of the Siskiyous exist in a summer- low-elevation Siskiyous are plotted, these curves (with dry, "Mediterranean" climate. During the dry season the exception of P. jeffreyi) show an interesting and the forests are inflammable,and the means of igniting consistent departure from this singly-convex form, as them have not been lacking. Essentially all the low- illustrated in Fig. 14. elevation vegetation described has been burned re- Increment borings permit some interpretation in cently enough for fire-signsto be detected in the form relation to tree ages and history of the area. Three July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 307 phases of the curves may be recognized in relation to to the same climax even in the absence of fire (Whit- three historic periods: (1) A pre-white period, ex- taker 1954b). The problem of edaphic climaxes, tending back from somewhat more than a century to however, goes beyond the fact that there are three several centuries ago, which is reflected in an essen- differentclimax patterns in the same climate. If so tiallv normal, convex curvature for the largest size- marked a differenceof climax occurs on these three, classes in the curves. (2) The historic period of a and gabbro is intermediate to the other twco,it is century or somewhat more, during which the forests reasonable to suppose that rocks intermediate to the were more frequently burned. The resulting reduced diorite and gabbro, or gabbro and serpentine ob- reproduction, or heavier mortality among younger served, would support still different, intermediate trees, is reflected in the flattened intermediate sec- vegetation. Granite, the fourth member of the series, tions of the curves. (3) A more recent period of is known to support vegetation differentfromn diorite protection from fire, during which the frequency of in some areas (Williams 1933). Moreover, the geo- fires has been less, and the rate of tree reproduction logical diversity of the Siskiyous involves many rock and survival of younger trees greater, as indicated in types other than the granite-serpentineseries. Close the concave, uppermost sections of the curves. by the diorite area one may observe vegetations of In relation to fire-disturbanceand the climax state marble, quartzite, slate, and argillite, which differ of these forests, the curves suggest: (1) The older from the vegetation of diorite and gabbro, especially trees are several centuries old; they have survived in more xeric sites. There may be, within the same cli- the more intensive burning by white man and still mate, many climax vegetation patterns differingin dominate the upper levels of the tree strata and differentways and differentdegrees in response to characterize the vegetation. (2) The general simi- differentparent materials. If one sets aside the vege- larity of the curves implies that the differentspecies tation of serpentine as highly anomalous, there still have been affected in similar wavs by increase and remain the climiaxpatterns, differingin many details, decrease of fire frequency. (3) The mixed conifer- of these other rocks. If one regards these climaxes sclerophyll dominance of these stands is not a prod- as part of "the same" climax formation, Mixed Ever- uct of more recent, more severe burning; the balance green Forest, one must recognize that they are "the between the two groups in communitydominance has same" in membership in a man-made class of forest been similar, though not necessarily identical, through- communities, but not vegetationally "the same" in out the history represented. (4) The principal effect plant populations or details of physiognomy. Any of less severe fires in this vegetation seems to be not statement that one of these vegetations represents in radical change in community composition, but in the true, climatic climax surely represents an arbitrary reducing the seedling survival, and hence eventually choice of what is to be thought a "normal"' or "typi- reducing stand density, for both conifers and sclero- cal" parent material and vegetation pattern. phylls. It was one of the assumptions of the monoclimax If this vegetation is obviously not, on the one theory that the climax, given time for development hand, climax vegetation unaffected by fire, neither, of a mature soil, should be independent of soil parent on the other hand, is it a vegetation of fire-successions. material. The diversity of climaxes in the Siski- It may be regarded, rather, as a fire-adapted vege- yous seems to contradict this assumption. The ex- tation of a summer-dry climate, in which fires of tent to which differencesof parent material are re- varying frequencies and intensities and varying flected in differencesof vegetation must be influenced sources-white man, Amerind, and lightning-have by various factors other than properties of the rocks for a very long time been part of its environment. themselves-climate, physiography, and fire, the gen- If the term "climax" is to be applied in such circum- eral character of the vegetation and soils, floristics stances, it seems supposititious to apply it to the non- and the kinds of plants available to form and domni- existent vegetation which might develop after cen- nate communities. Two parent materials may well turies of complete fire protection. The climax may support different climax vegetations in one set of better be regarded as that reasonably stable and self- circumstances, but in another support vegetation maintaining vegetation which exists in this area, in which seems not significantly different. But it is adaptation to fires and other factors of environment. suggested that the statement that climax is inde- It may be understood in this case that the climax, or pendent of soil parent material mnaybe replaced by fire-climax, condition embodies a degree of popula- understanding that a significant differencein parent tion instability and irregularity resulting from fires material may usually imply some degree of difference affecting different areas in a patch-wise fashion at in climax vegetation. irregular intervals. AND CLIMAX CO.MPARISONT It further appears that vegetation patterns of THE COENOCLINE differentparent materials have been similarly affected Within each parent material and elevation belt by fire; no one of them is fire-successionalto another. there is conspicuous vegetational gradation front It seems inconceivable that the three vegetation pat1 riavines to southwest slopes; the vegetatiomnforlas a terns of diorite, gabbro, and serpentine, with their continuous pattern, or vegetatiomnalspectruam, in re- different physiognomies and floras in relation to lation to the topographic moisture gradient. So different nutrient economies, should ever converge far as can be determined, in stands not recently and Ecological Monographs 308308 R. IIH. WHITTAKERWHITTAKER ~~~~~~~~~~~~~~~~~~~~~~Vol.30, No.3 severely disturbed, vegetation and soil are in the have been suggested by the author (Whittaker 1956). climax condition along the whole of the moisture Average climax composition is the average composi- gradient, at least in the sense that they are in steady- tion of climax stands in a given area, or within a state in relation to the mountain-slope environments given coenoeline. Intermediate clinax stands are in which they exist. "The climax" on a given parent intermediate, or median, to the extremes of a coeno- material is consequently not a particular kind of cline, as determined by the midpoint of a transeet, by stand, but the whole climax pattern in relation to the percentage similarity or coefficientof communitycom- moisture gradient (ef. Whittaker 1951, 1956). Con- parison with the extremes, or other means. In some sidering also elevation and the three soils, climax cases, as in the kind of topography dealt with here, vegetation in the central Siskiyous may be conceived open east-facing slopes are approximately intermnedi- as a highly diversified, generally continuous, multi- ate to the extremesof the moisture gradient. The pre- dimensional pattern comprising all the stabilized yiailing climax type is the community-type,wvhatever vegetation described in relation to the mo sture, ele- the criteria chosen to define it as a type, which eoiii- vation, and diorite-to-serpentinegradients, and other prises the majority of the climax stands in a given vegetation undescribed. area or a given coenocline. A principal concern of the present work is with These concepts underly the present work in var- comparison of vegetation patterns of different eli- ious ways. Vegetation has been sampled, the sam- mates and parent materials. Comparison for dif- ples arranged in composite transects, the transects ferent climates, at least, can be based on "climatic interpreted, and the vegetation itself described, in climaxes"; but there are limitations to the effective- terms of moisture-gradient coenoclines. Tables 3-11 ness of such comparison, for the comimunity-typesto summarize community composition and the distribu- be designated "climatic climax" must be chosen from tions of plant populations in coenoclines in such a the vegetational gradient by an ecologist. Much that way that these may be directly compared from one is significantabout vegetational expression of climate parent material to another. Tables 12-14 summarize may be lost sight of if a single climatic climnaxtype relations of diorite vegetation to elevation in terms is selected, to the neglect of the rest of the vegetation of average climax composition, and Tables 3-11 per- pattern. Comparison of vegetation fromnone area mit a similar summary in relation to the three parent to another may be most effectiveif based, so far as materials at low elevations. The floristic analysis of possible, on the whole moisture-gradientpatterns for Part VI is based primarily on species lists for coeno- each area. elines, and the comparison of these in terms of life- The expressions community gradient, or pattern, forms and growth-forms,species-diversities, and areal or spectrum, or ecological series have been used in the types. The problem of prevailing climax types for preceding sections of this work, but a shorter term the Siskiyous will be considered in the section which may be desirable. The term ecocline was suggested follows. by Clemnents (1936), apparently for sequences of - climax communities along environmental gradients, EAST-WEST GRADATIONAND THE but has been little used. The term may be a useful PREVAILINGCLIMAX TYPE one; conceived as a grada- and an ecocline may be In order to supplement the work in the central of an en- tion in characteristics ecosystems along Siskiyous with information on low-elevation cli- vironmnentalgradient, a gradation which maaybe un- niaxes, at least, for other areas of the range, a series derlain or caused by a particular environmental gra- of small-scale studies were made from the Coastal dient, but is usually expressed in complexly interre- Sequoia forests at the western end, to the vegetation lated changes in all aspects of ecosystems. Ecocline of the interior at the eastern end of the Siskiyou has been applied also (Huxley 1943) to the genetic Mountains. At each study area six vegetation sam- gradients or elines within species populations along ples were taken at low elevations, in differenttopo- ecosystemic gradients. For the environnental aspect graphic positions representing the moisture gradient of the ecosystemicgradient the term comtplex-gradient -ravines, mesic lower slopes, and open NE, E, SE, has been suggested (Whittaker 1954c, 1956), and the and S or SSW slopes-along with other notes on the term catena is in use for the more strictly edaphic character of moisture-gradient coenoclines. Since aspect (see also terms suggested by Major 1951). each study was a limited transect, this phase of the For the gradient of natural communities in an eco- work was designed as a transect of transects, repre- eline, the term coenocline is here proposed; the coeno- senting change in low-elevation vegetation patterns dline and complex-gradient together constitute the along the climatic gradient from the more humid and ecosystemic gradient or total ecoeline. maritime to the drier and more continental extremii- If the climax on mature, streamn-erodedtopog- ties of the Siskiyou Mountains. Some features of the raphy is conceived as a "coenocline," then comparison climatic changes along this gradient are indicated in of climaxes for differentclimates and parent ma- Table 1. The study areas (see Fig. 1), and vegeta- terials may be facilitated by means of further ab- tion patterns, were as follows: straction, to express average or general character of 1. Mill Creek State Park, Del Norte County, Calif., coenoclines. Three approaches to such abstraction in hills or low mountains, about 8 km fromnthe coast, July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 309 at elevations of 120-180 mn,in coastal Sequoia forest. Chamaecyparis; of the four coastal conifers only Stands were strongly dominated by Sequoia semper- Tsuga heterophylla occurred, in small numbers in virens in the upper tree stratum and Tsuga hetero- mesic sites. Lithocarpus was the principal small tree, phylla in the middle tree stratum in all topographic averaging 600 stems/ha; and the other principal small situations. Smaller numbers of Pseudotsuga men- trees of the mixed evergreen forests were present. ziesii, Abies grandis, and Chamaecyparis lawsoniana Vaccinium ovatum, V. parvifolium, Rhododendron also occurred, and, in more mesic sites, Thuja plicata californicum, and Gaultheria shallon were major and Picea sitchensis. The lower tree stratum of shrubs; but the number of Vaccinium stems was much broad-leaved species included Lithocarpus densiflora, lower than in the preceding transects, reaching 250- Corylus rostrata var. californica, Acer circinatum, 750 small stems/ha in drier sites. Shrub coverage and Umbellularia californica, but was _of very low increased from 25-30% to 50-75%, herb coverage coverage and density in all sites (averaging only 80 decreased from 30-40% to 0-5% along the moisture stems/ha). The high shrub stratum, dominated by gradient. In general -characterthis coenocline is one the arborescent coastal ecotypes of Vaccinium ovatum of mixed evergreen forests, gradating from Chamae- and V. parvifolium was, in contrast,highly developed; cyparis-Pseudotsuga stands in ravines to selerophyll- and in the S-facing stands this stratum formed a Pseudotsuga stands on xeric slopes, as in the diorite dense underbrush, with 3000 and 300 stems/ha re- vegetation of the umainstudy area. spectively for the two species. More important un- 4. Low-elevation vegetation already described- dergrowth plants, besides these, included Gaultheria, that of diorite (and metavolcanic rocks) in the Oregon m, shallon, Rhododendron californicum, Oxalis oregana, Caves area, Josephine County, Ore., at 460-760 Polystichum munitum, Blechnum spicant (L.) J. E. about 75 km from the coast-may serve as the fourth Smith, Viola sempervirens, Galium triflorum,Trilli- transect of the series. Selerophyll density was still urn ovatumn,Disporum smithii, Whipplea modesta higher than in the Siskiyou Fork transect-averag- and, in ravines Rubus spectabilis, Rhododendron ing 800 stems/ha for Lithocarpus and 600 stems/ha occidentale, and Adiantum pedatum var. aleuticum. for other species. Deciduous small trees averaged Herb coverage (20-85%) was much higher than in about as dense (250 stems/ha) as at Siskiyou Fork. the main Siskiyou study area, shrub coverage was Vaccinium ovatum, V. parvifolium, and Rhododen- variable from 1-10% in mesic to 70-95% in xeric sites. dron californicum, on the other hand, were infre- In stands of mesic flats the shrub cover approached quent. Herb coverage decreased from 8-20% to 0-8% zero and the herb coverage 90% or more, dominated along the moisture gradient, and shrub coverage was b)y Polystichum munitum and Oxalis oregana. between 10 and 40% in most stands. 2. In the hills above the Smith River near its 5. Above Sturgis Creek, near Steve Peak, Jose- junction with the South Fork, at elevations of 180-210 phine County, Ore., 15 km east from the Oregon m, about 14 km from the nearest point on the coast, Caves area, on metavolcanic rocks at elevations of at the inland edge of the Sequoia range, Del Norte 850-980 m. The vegetation was also mixed ever- County, Calif. Few, widely scattered, large Sequoia green forest in predominant character, but showed occurred in stands strongly dominated in the upper effects of somewhat drier and more continental cli- tree stratum by Pseudotsuga. The lower tree stratum mate in a number of features. Three dominant spe- was dominated by Lithocarpus and (in less xeric cies of the preceding vegetation patterns-Chamae- stands) Corylus, averaging around 400 stems/ha cyparis lawsoniana, Lithocarpus densiflora,and Gaul- each. Arbutus menziesii, Quercus chrysolepis, and theria shallon-were absent. Selerophyll density, Cornus nuttallii also occurred, predominantly in more which had risen in the preceding series of transects, xeric sites. As in the coast forest, Vaccinium ovatum fell to an average of 500 stems/ha (from about 700 dominated the high shrub stratum, reaching 2500- at comparable elevations in the Oregon Caves area), 4000 small stems/ha; Vaccinium parvifolium and predominantly Quercus chrysolepis with smaller num- Rhamnus purshiana also occurred, along with nuiner- bers of Arbutus menziesii and Castanopsis chryso- ous stems of Rhododendron californicum. Shrub phylla. Numbers of deciduous tree stems were higher coverages increased from 15-25% to 70-95%, herb -250 stems/ha in contrastto about 120 at comparable coverages decreased from 60-70% to 0-10% along elevations in the Oregon Caves area-and included the moisture gradient. In general character this more numerous stems of Quercus kelloggii Newb. in coenocline is one of Pseudotsuga forests with limited more xeric stands. Important shrubs included Ber- sclerophyll and heavy evergreen shrub undergrowth, beris nervosa, Rhus diversiloba, Rubus vitifolius, and physiognomically very similar to the Sequoia forests Rosa gymnocarpa; they included also species rare in of slopes, but with a somewhat greater range of or absent from the Oregon Caves diorite transect- physiognomic and floristic variation along the mois- Rubus leucodermis and B. parviflorus,Rosa spaldingii, ture gradient. Berberis piperiana, and in xeric sites Ceanothus in- 3. Along the Smith River at its junction with the tegerrimus and Arctostaphylos viscosissimum Peck. Siskiyou Fork, about 34 km from the coast, Del The herb stratum included most of the species familiar Norte County, Calif., on lower slopes with elevations in the Oregon Caves area and some additional ones. between 350 and 450 m on slate soils. Forest stands Herb coverages decreased from 7-18% in more mesic were dominated by Pseudotsuga and, in mesic sites, to 0-5% in more xeric sites; shrub coverage was gen- Ecological Monographs 310310 R. II.H. WHITTAEKERWHITTAKER ~~~~~~~~~~~~~~~~~~~~~~Vol.30, No.3 erally between 10 and 40%. Difference from the and open S and SW slopes supportedgrassland with- Oregon Caves diorite vegetation was most conspicu- out or with widely scatteredoaks. The grassland, ous in the south-facing slopes, where Pseudotsuga severelydisturbed, was dominatedby Elymus caput- shared dominance with Pinus ponderosa and P. lam- medusae whereanalyzed. This finalvegetation pattern bertiana in open stands with well-developed shrub representsshrinkage of coniferforest into most mesic strata. Pine dominance appeared to be at least part- sites,reduction of the sclerophyllelement to insignifi- ly a consequence of fire; but such stands are not en- cance, and appearance of the oak woodland and countered even as a consequence of burning in the valley grasslandcommunities of the Interioras part Oregon Caves area on metavolcanic rock or diorite. of the coenocline,at the easternend of the Siskiyous. 6. Near Beaver Creek, 105 km from the coast, on The transitionsfrom the Siskiyouvegetation to that granite at elevations between 640 and 850 m, Jackson of the Cascades and Interiorare more complexthan County, Oregon. More mesic stands were dominated this single transectindicates; but the pattern de- by Pseudotsuga with a poorly developed sclerophyll scribedis that characteristicof low elevationsin the layer; the (fire-affected)stands of south-facing slopes interiorvalleys of the Rogue and Umpqua Rivers were dominated by Pinus ponderosa and Pseudotsuga (Peck 1941). in open growth. Representation of sclerophyll spe- Of the conclusionsfrom this climatictransect, a cies (Q. chrysolepis and Arbutus) was still lower-200 series of trendsmay firstbe observed: sterns/ha; representation of deciduous trees other a) Stature and coverage of the coniferoushigh- than oaks only 170 stems/ha. Quercus kelloggii and tree stratumdecreases progressively from the great Q. garryana, however, were in this area major small- Sequoia forestseastward. In inverserelation to this, tree species, contributing about 220 stems/ha on the the densityof deciduousbroadleaf trees increases to- average, with populations of 270 and 550, 150 and ward the east. Maximum densitiesof sclerophylls 390 stems/ha in the south-slope samples. Important (and of all broadleaf trees taken together)were in shrubs included Berberis nervosa and B. piperiana, the Oregon Caves area near the middleof the tran- Rubus vitifolius and R. leucodermis,Rhus diversiloba, sect, and the scierophyllsdecrease toward both ends Rosa gymnocarpa and, in xeric sites, Ceanothus inte- of the transect. gerrimus, Arctostaphylos viscosissimum and A. vis- b) Both shrub and herb coveragesdecrease from cida. Herb stratum coverages decreased from 12- the Sequoia forests to the middle of the transect, 25% to 5-12%, shrub coverages increased from 7-25% thenincrease again to the easternend of the transect to 20-50% along the moisture gradient. In general (though these trends are complicatedby variations character the coenocline is one of Pseudotsuga forests with site in a given area). Undergrowthcoverages with sclerophylls gradating into Pinus-Pseudotsuga- thus show some inversecorrelation with development deciduous oak forests in xeric sites. The transition of the sclerophyllstratum. from mixed evergreen forestto oak woodland is repre- c) Floristictrends in representationof life-forms, and sented in approximate equality of sclerophylls growth-forms,and areal types will be discussedin deciduous oaks in the pattern as a whole, and pre- Part VI. xeric dominance of the oaks over sclerophylls in more d) Distributionsof species change in relationto sites. the topographicgradient. (1) There is a general Emi- 7. The final transect of the series was near narrowingof species amplitudes toward the east. Ore- grant Creek, near Steinman in Jackson County, Many species occupy the whole of the topographic on ande- gon, at elevations between 820 and 1070 m, moisturegradient at Mill Creek, apparently none Stands of site, 140 km from the Pacific Coast. more than part of the gradientat EmigrantCreek. were dominated by ravines and mesic lower slopes (2) Distributionsof species (and community-types) P. and Libocedrus pres- Pseudotsuga with ponderosa shiftalong the moisturegradient in such a manner species in these coniferous forests ent. Deciduous as to compensatefor more or less humid climates. included Acer macrophyllum,Alnus rhombifolia, and Species of wide amplitudesnear the coast are in- Fraxinus oregana. On the open NE slope Pseudo- creasinglyrestricted to the mesicend of the gradient shared dominance with P. ponderosa and Libo- tsuga towardthe interior;species occurringin the interior cedrus in open stand, with a well-developed lower transectsare increasinglydisplaced towardthe xeric tree stratum (50% coverage) dominated by Quercus end of transectsnearer the coast. (3) Ecotypicpop- garryana. The conifers were absent from less mesic ulations show complexrelations to parent materials stands (open E and SE slopes), except for scattered and the climaticgradient. Many species largelyre- P. ponderosa; these stands were oak woodlands of strictedto "special" parent materials (serpentine, Q. garryana and Q. kelloggii with moderate tree (65 slate, argillite,marble), generally and 45%) and grass (20 and 60%) coverage. Sclero- gabbro, quartzite, phylls were unimportant in either coniferous forest with more open or xeromorphicvegetation in one or oak woodland; only a feewstems of Arbutus were area, occur in the vegetationof "normal" parent recorded in the former. Ceanothus intergerrimus, materials (granite, diorite, andesite, metavolcanic Arctostaphylos viscida, and Cercocarpus betuloides rocks) in a drierand morecontinental climate toward were important high shrubs. Oak coverage decreased the east. In contrastto this, several species (Vac- while grass coverage increased toward most xeric sites, cinium parvifolium, V. ovatum, Vancouveria chry- July, 1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 311

santha, Galium ambiguum, Polygala californica) had 1957). Four formations-Coast Forest, Mixed Ever- ecotypic populations widely distributed in the humid green Forest, Oak Woodland, and Valley Grassland coastal forests, and other populations largely re- -appear in the climatic transect; three other forest stricted to serpentine and gabbro in the drier climate formations-Montane Forest, Subalpine Forest, and of the main study area. Pine Woodland-occur in other environments. The Second, the results indicate that the prevailing Montane Forests and Subalpine Forests are in these climaxes of the Siskiyous at low elevations are: (1) mountains physiognomically continuous with one an- Sequoia forest near the Coast, (2) a belt of coastal other. If they are to be separated as formations,as Pseudotsuga forest within this, (3) Mixed Evergreen is customary,it must be on the basis of environmental Forests over the greatest part of the range between relations, primarily elevation, and broad floristicand these coastal Forests and the interior valleys, (4) geographic relations, to the western Montane Forests Oak Woodlands in these valleys at the eastern end of and eircumboreal taiga, not physiognomy. The yel- the range. The Oregon Caves study area on diorite, low pine woodlands or pine steppes of the western within the area of maximum development of the states are often treated as part of the Montane Forest sclerophylls, is thought as "representative" of the Formation. Their physiognomy is fundamentally mixed evergreen forests at their best development in different,however; and they may well, even though the Siskiyous as any area known to the author. In they intergrade with the closed Montane Forests, be relation to the Mixed Evergreen Forest region as a regarded as a separate formation. Pine woodlands whole, however, it represents not average or median appear both on serpentine in the central Siskiyous conditions, but those of maximum stand density to- (dominated by P. jeffreyi) and in drier situations of ward the northern and mesophytic limits of these middle and higher elevations in the eastern part of forests. the range (dominated by P. ponderosa). A final result of significance is the continuity of The Mixed Evergreen Forests have affinitieswith vegetational change along the climatic gradient studied the sclerophyll vegetation complex of California. Two (cf. Whittaker 1956). This continuity can scarcely major community-typeshave been recognized in this be perfect, considering the irregular topography and complex-the chaparral of shrubs and the broad effectsof parent materials. But with allowance for selerophyll forest dominated by trees (Cooper 1922, these, the climatic gradient was expressed in gradual Oosting 1956). Among the associations of the broad- transformation of communities in both physiognomy selerophyll forest formation Cooper (1922) recog- and floristic composition, in shifts of position along nized one, the Lithocarpus densiflora-Quercus chryso- the moisture gradient for both species populations lepis-Arbutus menziesii association, as characteristic and community-types,and not in abrupt transition. of lower altitudes of the north Coast Ranges of Cali- This continuityis the more impressive in that four for- fornia. He further observed that this association is mations-the Coast Forest, Mixed Evergreen Forest, itself somewhat a transition between broad-sclero- Oak Woodland, and Valley Grassland-are involved. phyll and coniferous types; for it rarely occurs with- One can divide the continuity by prevailing cliniaxes out a sprinkling of conifers, especially Pseudotsuga, and major community-types,but these approaches and its principal species occur commonlyas an under- do not clarify the underlying quantitative relations. story in the Pseudotsuga and Sequoia forests. Coo- The approach through comparison of moisture-gradi- per's Lithocarpus-Quercus-Arbutus association is part ent patterns makes possible some quantitative analy- of the Mixed Evergreen Forest, as designated by sis of the gradual transformationof community-types Munz & Keck (1949, 1950,. 1959) and understood in and vegetation patterns along a climatic gradient. this paper. The sharing of dominance between ever- green needle-leaved and selerophyllous trees, and the V. PROBLEMS OF CLASSIFICATION transitional relation to the sclerophyll vegetation on of There are many ways classifying vegetation; the one hand, the Coastal and Montane Forests on and most studies are influencednot merely in forii of the other, are essential features of this community- presentation, but in the way vegetation is observed type. and sampled, by the author's particular approach The Mixed Evergreen Forest is unrecognized as to classification. One object of the present study a major community-typein its own right in many ac- was to take samples which could be used to compare counts & Clements differentapproaches to classification. The preceding (Weaver 1938, Oosting 1956). It is evidently transitional between and description embodies the author's own approach selerophyll conifer but in the sense that through physiognomy and dominance. The same forests, only many between other material may be considered, however, in relation to major community-typesare transitional community-types. More critical than its transitional several major vegetation units-the formation, (1o011i- nance-type, sociation, association, and union. relation in evaluating the Mixed Evergreen Forest are the facts that: (1) It is itself a grouping of a FORMATIONS number of dominance-types or associations. (2) It A formation is a grouping of communities which is the regional climax of an extensive area of the are similar in physiognomy and in broader environ- Klamath Region and north California Coast Ranges. mental relations (cf. Beadle & Costin 1952, Whittaker (3) It possesses, especially on some parent materials, EcologicalMonographs 312 R. H. WHITTAKER Vol.30, No. 3 a wealth of character-specieslargely or whollyre- be grouped into three more broadly defined domi- strictedto it. The author advocates recognitionof nance-types: Chamaeeyparis-Pseudotsuga forests of the Mixed EvergreenForest as one of the major com- mesie sites, with deciduous and selerophyll lower-tree munity-typesof westernNorth America and, because strata, forming closed stands on diorite and more its definitionis physiognomiesprefers to regardit as open ones with Libocedrus and P. lambertiana on a formation. The characterizationsuggested for it gabbro. (2) Pseudotsuga-Lithocarpus-Arbutus for- is: mixedforest with an upper treestratum of needle- ests, with other selerophylls present and in some cir- leaved evergreenor coniferousand a lowertree strat- cumstances dominant and with wide variation in cov- um of broad-leavedevergreen or selerophyllousspe- erage of the tree strata, occurring along iiost of the cies, but with thesevarying widely in coverageand moisture gradient on diorite and in less xeric sites on densityin differentstands. Either one or both of gabbro. Pseudotsuga-selerophyll and selerophyll- these strata may be eitherrelatively dense or quite Pseudotsuga physiognomic subtypes are distinguished open,but as bothbecome open the formationgradates within this dominance-type (Part III). (3) Open into pine-oakwoodland. Pinus-Pseudotsuga-Q. chrysolepis-Aretostaphylosfor- A much more complex patternof physiognomic ests of more xeric sites on gabbro and eastward from types of vegetationexists in the westernstates than the main study areas. in the eastern,and in manycases the formationsseem Vegetation types defined by dominant species of to intergradeas parts of longer physiognomniccon- the various strata, or by coinbinations of stratal tinua. These circumstancessuggest that the ap- unions, are properly termed sociations (Du Rietz proachdeveloped by Beard (1944,1955) in thetropics 1930, 1936, Nordhagen 1937, 1943, Braun-Blanquet may be appropriateto westernvegetation. The ap- 1951, Hanson 1953, Bocher 1954); such units permit proach permitsthe conceptionof physiognomicre- a finer division of the vegetation into more numerous lationsof vegetationthrough larger numbers of more types than dominance-types based on the uppermost narrowlydefined formations, arranged in formation- stratum alone. An effortwas also made to classify series complexlyintersecting one anotherin relation the Siskiyou samples by stratal dominants, with re- to major gradientsof environment.The author is sults similar to those on dominance-types: a very not prepared to erect a systemof formationsand large number of sociations, intergrading with one formation-seriesfor the whole of the Klamath Re- another through various combinations of dominance gion. The materialdiscussed suggests, however, these in the differentstrata, differingwidely in extensive- major relationson more typical soils: (1) At low ness or importance, many of thein of no observable elevations,from the Coast inland, with decreasing significance in relation to environmental difference, precipitation and increasing continentality-Coast resulted. Much of the Siskiyou vegetation is relative- Forest,Mixed EvergreenForest, Oak Woodland,and ly rich in species and shows no well-defined domi- Valley Grassland. (2) At low elevations,in mari- nance in the lower strata. Certain community-types time environmentsnear the Coast, with increasingly with well-marked stratal dominance (e.g. Abies con- humidclimates-Chaparral, Broad-Selerophyll Forest, color-Arctostaphylos nevadensis-Xerophyllum tenax Mixed EvergreenForest, and Coast Forest. (3) In forest and Pinus jeffreyi-Arctostaphylos viscida- the central SiskiyouMountains, from low elevations Stipa lemmoni woodland on serpentine, Sequoia sem- to high-Mixed Evergreen Forest, Montane Forest pervirens-Tsuga heterophylla-Vacciniumovatum-Poly- (and ponderosapine woodlandin some drier sites), stichumnmunitum Coast Forest) are conveniently de- Subalpine Forest. fined in the manner of sociations. But the experi- DOMINANCE-TYPES AND SOCIATIONS ments with classification of Siskiyou vegetation sup- the European experience, that the sociation is Withinthe formationsaccepted by the authorfor port best suited to communities with marked stratal the SiskiyouMountains, various dominance-types de- a unit in floris- finedby species dominancemay be recognized.With- dominance such as occur, for the most part, in the Subalpine Forests Abies nobilis and Tsuga tically poor vegetation. merten'siana dominance-types,within the Montane Forests Abies concolor and Pseudotsuga menziesii, ASSOCIATIONS dominance-types,within the Coast Forests Sequoia The principal vegetation unit of phytosociology is sempervirens and Pseudotsuga menziesii dominance- the association, defined primarily by character-species types may be distinguished.Such types definedby -species of -narrower ecological amplitudes which a single dominant species are "consociations"in are largely or wholly restrictedto or, at least, centered Clementsian terminology (Clements 1928, 1936, in, stands of a given community-type (Braun-Blan- Weaver & Clements1938). quet 1932, 1951). The samples taken from the Siski- The author attemptedto classify samples from you vegetation were intended to be relatively complete mixedevergreen forests by combinationsof dominant floristicallyand suitable for classification into associa- species,to see whetherthe standswould fall naturally tions, although. they include no data on bryophytes into groups representingwell-defined dominance- and lichens. The transects provide an effectivepic- types. They did not, and a large numberof types ture of the distributional relations which underlie intergradingwith one anotherresulted. These may problems of floristic classification. July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 313

The transects show that population centers or character-species groupings by which the author's distributional modes, and also the limits of distribu- community-typesmay be defined as associations (or, tions, are scattered along the various gradients more probably, higher units of this system) are indi- studied. Species populations do not form well-de- cated in the following section on distributional group- fined groups of character-species with similar or con- ings. gruent distributions. Character-species groupings must be created by the phytosociologist,and there are QUANTITATIVE APPROACHES TO CONTINUITY a number of ways in which the species of a given AND SPECIES GROUPING vegetation pattern may be grouped without violating The work on classification included also experi- their distributional relations (cf. Whittaker 1956). ments on quantitative methods by which relative dis- It is consequently often possible to choose character- continuities and natural clusters of species might be species groupings which fit into community-types objectively revealed from the transect tables. The recognized initially by physiognomy, dominance, and first of these methods is designed as an aid to the environmental relations. It cannot be assumed that recognition of different rates of change between these units, secondarily defined by floristic composi- community-types. Percentage similarities (see Part tion, will be the same as the units which would be VI) were computed between successive steps of the recognized by a phytosociologist who sought to base transect tables, for each of the three strata on each his work primarily on floristic composition. It is at of the three parent materials. The resulting values least possible, however, to bring these two major ap- are plotted in line graphs (upper halves of Figs. proaches to vegetation classification into some accom- 15-17). Most of the graphs are of the form to be ex- modation to each other. Full discussion of the flo- pected in continuously gradating vegetation-points ristic approach to classification of this vegetation is generally at the same level, but with some zig-zag but outside the scope of the present monograph; up and down due to chance variations in stand com- position. Low points appear between the first and __0(1) 80 second steps on gabbro and serpentine. The change indicated is that from riparian communities of wet E) rsTrees 2o E (X 80 Trees P U) 80 60 01.40 60 Shrubs E 2o 40 U 20 80 Shrub - 60 80 Herbs a 40 60 C 20 o40 0 V) 20 o60Herbs 60 60 o 40 ATrees En 20 40 en . O 20

aD 80 Shrubs 60 a. 60 C)40 140 O 20 20 C80 Shrubs T80 Herbs 60 0 I40 20 0 40 ()20 Hoerbs -1 2 3 4 5 6 7 8 9 10 )_60 MoistureGradient 040 D Mesic Xeric Uf) 20 FIG. 15. Sample and species associations in a tran- 1 2 3 4 5 6 7 8 9 l0 sect of the moisturegradient at low elevationson diorite. Moisture Gradient Above-percentage similarities of successive steps of G Mesic Xeric the transect,for the tree, shrub,and herb strata. Below a tran- -percentage similarities of distributions of pairs of FIG. 16. Sample and species associations in species, for the three strata. sect of the moisture gradient at low elevations on gabbro. 314 R. H. EcologicalMonographs 314 R. H. WHITTAKERWHITTAKER ~~~~~~~~~~~~~~~~~~~~~~Vol.30, No. 3 of species, by the formula PSd E min (a,b), in 80s Trees which a is the per cent of the total population of - 60 species A in the transecttable which occurs in a C~k40 givenstep of the table,and b the same for species B E 2oi~~ ~ (Whittaker& Fairbanks 1958). Percentage simi- V) 80 Shrubs laritiesof distributionwere computed for all possible pairs of speciesin a given stratumfor a given tran- 60 sect table, and the resultingvalues were arrangedin 40 matricesor Kulezyn'skitriangles. Species were then c 20 arrangedin sequenceon thebasis of their"mean posi- 0 Herbs tions"; and for each pair of species contiguousin 60 thesequence, a pointwas enteredin thespecies-associ- tion diagramsindicating the distributionalassociation 040 V) 20 value fromthe matrixon the ordinate,the mid-point betweenthe mean positionsof the two species on the C0 80 Trees abscissa. 60 Natural clusters of species should appear as C)40 clustersof points in the diagramwhich are both of O 20 high values for distributionalsimilarity, and close togetherin mean positionalong the 80Shrubs abscissa. Indi- vidualistic 60 scatteringof species distributionsshould 1 40 be reflectedin pointswhich are neithervery high nor 20 very low but moderate and variable, formingan irregular,zig-zag line. It is the latterpattern which D Herbs 80 appears in theresults, in the lowerhalves of Figs. 15- 60 17. The species of the mostdistinctive union of the C.40 vegetationdescribed, the shrubs of the two-phase U)20 serpentinevegetation, do not show very strongdis- tributionalassociation with one another,or generally 5 6 7 8 9 IC I 2 3 ~4 higherassociation with one anotherthan with other MoistureGradient shrubs with distributionsoverlapping theirs. The S Mesi Xeric graphs do indicate distinctclusters of undergrowth FIG. 17. Sample and species associations in a tran- species in ravines (points 1 on the transects); these at low elevations on ser- sect of the moisture gradient clusters again representthe distinctivestream-side pentine. communities.The species whichthese clusterscom- prise are easily shownto be individualisticin their mesic soils along stream-sides to the communities of distributionalrelations to elevation, the three soil as a slopes above the ravines. It may be interpreted parent materials,and geography;they are clustered in re- steeper gradient of community composition in relationto the range of environmentsrepresented rather sponse to a steeper environmental gradient, by a giventransect, but not in theirover-all distribu- a than as a vegetational discontinuityoccurring along tions. Results thussupport conclusions suggested in other low points-in the uniform gradient. Two a previous study (Whittaker 1956:22). Natural shrub strata between steps 8 and 9 of the serpentine clustersare weaklydeveloped in, or absentfrom, most to transect, 6 and 7 of the diorite transect-appear of the vegetationstudied, but may be presentin soine without vegetational result from sample irregularity circumstances.Where theyoccur, as in the stream- discontinuity, and suggest caution in the interpreta- side communities,they are not exceptions to the of such data. tion principleof species individuality;and theirrecogni- is designed to reveal natural The other technique tion is dependenton the range of conditionsover in the transects. For clusters of species populations whichdistributional relations of are observed. numbers of species each species represented by significant What is a natural clusterin relationto one transect two values were computed: (1) The mean individuals, or one local vegetationpattern may not be a natural of the in relation to the transect position population clusterin a broader range of vegetationalobserva- Inx tion. gradient, computed fromii in which a is the numn- DISTRIBUTIONALGROUPINGS ber of individuals of the species in a given step of Species occurring in significant numbers in the the transect, n the number of the step in the transect. principal study area, at low elevations on serpentine These values may bear little relation to population or gabbro or at any elevation in forests on diorite, centers or modes, but provide an indication of rela- have been classified into distributional groupings. tive positions along the gradient. (2) The per- Despite the arbitrariness of such groupings, they may centage similarity of species distributionsfor all pairs have real usefulness; they are used here to summarize July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 315

distributional relations of the Siskiyou species dealt folia. Shrubs-Gaultheria shallon*, Philadelphus with, to avoid repetitious listing of species in the lewisii, Vaccinium parvifolium. Herb-Polystichum vegetation descriptions, and to provide somt defini- munitum. tion of major community-typesby character-species 3. Mesic forest species of lower and middle eleva- groupings. A distributional grouping, within a given tions on diorite, or diorite and gabbro (*), centered stratum or life-form,is regarded as a union (Lippmaa in and character-speeies for Chamaecyparis-Pseudo- 1935, 1939, Du Rietz 1936, Oosting & Billings 1943, tsuga forests, but occurring also in more mesic Pseu- Billings 1945, Braun-Blanquet 1951:47-49, Dauben- dotsuga-selerophyll forests at lower, and montane mire 1952, Whittaker 1956). The groupings given forests at middle, elevations, and in many cases in here are based on broad relations to topographic lower subalpine forests: Trees-Acer macrophyllum*, moisture gradients, elevation, and parent material as Corylus rostrata var. californica. Shrubs-Holo- represented in the transects; they are not intended to discus discolor*, Pachystima myrsinites, Rhododen- reflect any distributional relation outside the range dron californicumr. Herbs-Cephalanthera austinae, of environments covered by these transects. Most Coptis laciniata, Heuchera miicrantha,Linnaea borea- species can be found outside the indicated distribution lis, Listera caurina, Mitella ovalis, Pyrola bracteata, patterns, which are intended to represent more fre- Senrecio bolanderi, Viola sempervirens. quent occurrence in communities,and to exclude very 4. Mesic species of low elevations on serpentine scattered occurrence at transect constancies of less (character-species of the forest-shrubcomplex) or on than 6 or 8%. Certain species with complex distri- serpentine and gabbro ( ) Tree-Pinus monticola. butional relations, or morphologically distinct popu- Shrubs-Amelanchier gracilis", Berberis pumila, lations in different habitats, are listed more than Garrya buxifolia, Juniperus sibirica, Lithocarpus once. densiflora var. echinoides, Quercus chrysolepis var. 1. Riparian species, occurring predominantlyon or vaccinifolia, Rosa gymnocarpa, Umbellularia cali- near stream-banks in ravines at lower (and middle) fornica (shrubby population), Vaccinium parvifoli- elevations, in forests in which Chamaecyparis law- um. Herbs-Angelica arguta, Antennaria suifrutes- soniana occurs as a dominant. cens, Arnica spathulata var. eastwoodiae': (Rydb.) On diorite (character-species of Chaimiaecyparis- Mag., A. parviflora ssp. parviflora", Aster brickelli- Pseudotsuga forests) at lower, or at middle (i) ele- oides Greene, Epilobium rigidumn,Haplopappus race- vations: Herbs-Aralia californica, Athyrium felix- mosus ssp. congestus (Greene) Hall, Ligusticum apii- femina0, Carex laeviculmis*, C. mertensii*, Claytonia folium, Lilium howellii, Lomatium howellii*, Lotus spathulata, Listera cordata, Luzula parviflora-, Pleu- oblongifoliust, Lupinus latifolius var. columbianus, ropogon refractus*, Satureja douglasii, Tolmiea men- Sancicula peckiana, Tauschia glauca, Vancouveria ziessi*. chrysantha. On diorite and " gabbro (character-species of 5. Species with broad mnoisture-gradientdistribu- Chamaecyparis-Pseudotsuga forests): Trees-Alnus tions at lower elevations on diorite, or diorite and rhombifolia, A. rubra. Herbs-Boykinia elata, B. gabbro (*), character-species of Mixed Evergreen major. Forests in a broad sense, occurring also in Charnaecy- On gabbro (character-species of Chaiinaecyparis- paris-Pseudotsuga forests: Trees-Arbutus men- Pseudotsuga forests) : Herbs-Epipactis gigantea, ziesii*, Castanopsis chrysophylla, Cornus nuttallii , Erigeron cervinus, Juncus ensifolius, Peltiphyllum Lithocarpus densiflora*,Quercus chrysolepis*. Shrubs peltatum, Woodwardia fimbriata. -Berberis pumila, Rhus diversiloba*, Symphori- On gabbro and serpentine (occurring in Chamae- carpos hesperius G. N. Jones, Herbs-Allotropa vir- cyparis-Pseudotsuga and Chamaecyparis-P. monticola- gata*, Lathyrus pauciflortis, Lonicera hispidtla*. Pseudotsuga ravine forests of these two parent mate- 6. Species with broad moisture-gradient distribu- rials): Shrub-Rhododendron occidentale. Herbs- tions at lower and middle elevations on diorite, or Carex debiliformis, Cypripedium californicum, Dar- on diorite and gabbro (*), occurring in mixed ever- lingtonia californica, Eriophorum crinigerum (Gray) green and Chamaecyparis-Pseudotsuga forests of Beetle, Habenaria sparsiflora, Rudbeckia californica, lower elevations and montane forests of middle ele- Tofleldia glutinosa ssp. occidentalis (Wats.) C. L. vations, -and in many cases extending into lower Hitche. subalpine forests: Shrubs-Berberis nervosa, Rosa On serpentine (character-species of Chamaecy- gymnocarpa, Rubus parviflorus*, B. vitifolius. paris-P. monticola-Pseudotsuga forests) : Shrubs- Herbs-Achlys triphylla, Adenocaulon bicolor, Apo- Physocarpus capitatus, Rosa californica. Herbs- cynum pumilum, Bromus suksdorfii,Galium triflorum, Carex angustior, Castilleja miniata, Galium multi- Festuca occidentalis', Goodyeara decipiens, Iris chry- florum,Helenium bigelovii, Lilium occidentale, Trillt- sopthylla, Melica harfordii. um rivale. 7. Species with broad moisture-gradientdistribu- 2. Lower-elevation mesic species of diorite, or tions at low elevations on gabbro and serpentine: diorite and gabbro (*), centered in Chamiiaecyptiris- On gabbro: Trees-Pinlis p)onderosa, Umbellilaria Pseudotsuga forests but occurring also in more Inesic californica. Shrubs-Garry a fremontii, Vaccinium Pseudotsuga-selerophyll forests: Trees-Acer circi- ovatum. Herbs-Carex vnulticanlis, Festuca cali- natum, Amelanchier florida', Salix sp., Toxus breivi- fornica, Galium bolanderi. EcologicalMonographs 316 R. H. WHITTAKER 11Vol. 30, No. 3 On gabbro and serpentine: Tree-Libocedrus de- liflora, Habenaria sparsiflora, Hackelia jessicae , currens. Shrubs-Convolvulus polymorphus, Holo- Lathyrus polyphyllus, Nemophila parviflora, Osmo- discus dumosus, Rhamnus californica var. occidentalis. rhiza occidentalis*, Tiarella unifoliata, Vancouveria Herbs-Galium ambiguum, Polygala californica, Poly- hexandra, Viola glabella*, Vicia californica. stichum munitum var. i'mbricans, Pyrola dentata, 11. Species of middle- and higher-elevationforests, Schoenolirion album, Viola lobata, Xerophyllum occurring in both montane and subalpine forests, or tenax. both these forests and more open situations (*) On serpentine: Tree-Pinus attenuata. Shrub- Shrubs-Quercus sadleriana, Ribes marshallii, R. vis- Arctostaphylos nevadernsis. Herbs-Balsamorhiza cosissimum, Symphoricarpos rivularis,Vaccinium mem- deltoidea, B. platylepis, Calamagrostis koeleroides, branaceum. Herbs-Agastache urticifolia*, Are- Cordylanthus viscidus, Hieracium cynoglossoides var. naria macrophylla, Arnica latifolia, Artemesia doug- nudicaule Gray, Iris bracteata, Lotus crassifolius, lasiana*, Campanula scouleri, Claytonia spathulata Phlox speciosa, Sedum laxum. Dougl., Corallorhiza maculata, Dicentra formosa, Fra- 8. Species of lower elevations in more xeric sites garia vesca var. bracteata, Hydrophyllum occiden- on diorite and gabbro, character-species of sclero- tale*, Luzula spicata, Mitella diversifolia, Monar- phyll-Pseudotsuga forests (including Pinus-Pseudo- della odoratissima*, Osmorhiza chilensis, Phacelia tsuga-Quercus-Aretostaphylos forests on gabbro): magellanica' Coville, Polygonum phytolacceaefolium*, On diorite: Herbs-Collomia heterophylla, Cypri- Pyrola secunda, Rubus lasiococcus, Senecio triangu- pedium fasciculatum, Festuca rubra, Madia madioides, laris*, Smilacina stellata (L.) Desf. Psoralea physodes, Tauschia kelloggii. 12. Species of forests of high elevations, centered On diorite and gabbro: Tree-Quercus kelloggii. above 1800 m, character-species of subalpine forests Herbs-Boschniakia hookeri, Campanula prenan- dominated by Tsuga mertensiana and Abies nobilis, thoides, Pteridium aquilinum var. pubescens. and species of both these forests and more open situ- On gabbro: Tree-Arctostaphylos cinerea. Shrub ations (*): Trees-Abies nobilis, Tsuga mertensiana. -Ceanothus integerrimnus.Herb-Hieracium bolan- Shrub-Lonicera conjugalis*. Herbs-Anemone deri. quinquefolia, Angelica arguta, Aquilegia formosa*, 9. Species of more xeric sites at lower elevations Aster siskiyouensis, Bromus carinatus*, Carex rossii, on serpentine, or on serpentine and gabbro (*). The C. tracyi*, Castilleja miniata*, Claytonia parvifolia former are character-species for the Pinus jeffreyi Moe., Corallorhiza mertensiana, Delphinium sonnei hornemanni, woodlands on serpentine; the latter occur also in Greene, Elymus glaucus*, Epilobium Erigeron aliceae", Erysimum capitatum (Dougl.) Pinus-Pseudotsuga-Quercus-Aretostaphylos forests on Greene, Ligusticum grayi', Lupinus albicaulis*, Mer- gabbro: Tree-Pinus jeffreyi. Shrubs-Aretostaphy- tensia bella, Mitella breweri, Orogenia fusiformis, los viscida, Quercus garryana var. breweri. Herbs- Pedicularis racemosa, Penstemon anguineus, P. ne- Achillea lanulosa, Agrostis hallii*, Bromus brevia- morosus, P. newber'ryi Gray, Poa bolanderi*, Poly- ristatus, Calochortus howellii, C. tolmiei, Ceanothus gonum davisiae*, Sanicula nevadensis*, Saxifraga pumilus*, Cheilanthes siliquosa, Danthonia califor- ferruginea, Smilacina racemosa var. glabra, Trisetum nica, Elymnus glaucus, Epilobium pa iculatum var. spicatum, Valeriana sitchensis ssp. sitchensis, Vera- hammondi, Erigeron foliosus var. confinus', Erio- trumrinsolitum', Viola praemorsa, V. sheltonii. nudum>, E. ternatum Howell, Eriophyllum gonum 13. Parent-material ubiquitists, occurring at low Horkelia sericata, Koe- lanatum var. achillaeoides:, elevations on diorite, gabbro, and serpentine: Trees- L. triternatum, leria cristata, Lomatium macrocarpum, Chamaecyparis lawsoniana, Pinus lambertiana, Pseu- var. glauca*, Melica geyeri, Monardella odoratissima dotsuga menziesii. Herbs-Adianturn pedatum var. M. villosa var. subserrata*, Pentstemon laetus ssp. aleuticum, Chimaphila umbellata var. occidentalis, roezlii (Regel) Keck, Perideridia oregana, Phacelia Disporum hookeri, Festuca ovina, Luzula campestris, dasyphylla var. ophitidis, Senecio fastigiatus*, Sidal- Smilacina racemosa, Trientalis latifolia, Trisetum cea malvaeflora ssp. elegans* (Greene) C. L. Hitch., canescens, Whipplea modesta. Sitanion jubatum, Stipa lemnmoni*,Zygadenus mi- 14. Elevation ubiquitists, occurring in forests of cranthus. all elevations on diorite: Herbs-Campanula scou- 10. Species occurring at or centered in middle leri, Chimaphila, menziesii, Chimaphila umbellata var. elevations, 1200-1800 mn,on diorite, character-species occidentcZis, Festuca, occidentalis, Hieracium albi- for the montane forests dominated by Pseudotsuga florum, Melica subulata, Phlox adsurgens, Pyrola menziesii and Abies concolor. Many of these species picta, Smilacina racemosa, Trillium ovatum. extend into low elevations in Chamaecyparis-Pseudo- tsuga forests, and some (*) were recorded frequently VI. FLORISTIC COMPARISONS from open situations as well as forests: Trees-Abies Floristic comparisons for differentvegetation pat- concolor, Acer glabrum var. douglasii. Shrub- terns and communities in the Siskiyou Mountains Rubus nivalis. Herbs-Actaea spicata L., Anemone are summarized in three tables (17, 18, 20). These deltoidea, Asarum caudatum, A. harqtwegi,Circaea tables are intended to show the manner in which life- alpina var. pacifica (Asch. & Mag.) M'. E. Jones, form composition, growth-fornicomposition and spe- Clintonia uniflora, Corallorhiza striata, Festuca ,sobu- cies-diversities, and geographic relations of floras July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 317

change in relation to four major environmental gra- TABLE 17. Relations of life-forms to environments in dients-the diorite-gabbro-serpentineseries of parent the Siskiyou Mounitains, per cents of species in floras of materials, local topographic moisture gradients on transects and site groupings. each of these, elevation on diorite, and the east-west climatic gradient. The basic unit of comparison is the transect flora-the list of species (with presence values of 4% or more) in a given transect of 50 samples. One list (1920-2140 in on diorite) included all species in a 16-sample transect; and the lists for the east-west climatic gradient are based on all species A. CentralSiskiyou Mtns., by parent materials, 610-915m of 6-sample transects. The comparisons for different 24 2 84 Quartzdiorite ...... 32 12 30 moisture-gradientconditions and community-typesin 2 Olivinegabbro ...... 32 14 32 19 101 15 3 1 16 the central Siskiyous are based on lists of species Serpentine...... 6.. 19 44 occurring more than once in the inesic, intermediate, B. CentralSiskiyou Mtns., by sites, and xeric ranges of the transects (steps 1-2, 3-7, 8- 610-915m On diorite-Mesic3...... 35... 14 28 21 2 72? 2 10). The lists for mesic, intermediate,and xeric sites Intermediate... 333120 1 79 correspond to community-types:on diorite-Chainae- Xeric ... . 3 8 42 13 47 11 3;3 21 1 7d cyparis-Pseudotsuga, Pseudotsuga-sclerophyll, and On gabbro-Mesic...... 34 Intermediate30...... 30. 22 35 12 1 72 selerophyll-Pseudotsuga forests; on gabbro-Chamae- Xeric ...... 29 2;3 38 8 2 66 18 1 88 cyparis-Pseudotsuga, sclerophyll-Pseudotsuga, and On serpentine-Mesic27...... 27.... 19 . 35 Pinus-Pseudotsuga-Quercus-Arctostaphylos forests; Intermediate.5...... 25..... 24 39 12? 1 1()2 Xeric...... 15 17 43 14 11 76 on serpentine-Chamaecyparis-P. monticola-Pseudo- C. CentralSiskiyou Mtns., by tsuga forest, mixed conifer stands with two-phase elevationbelts on diorite 3 undergrowth (forest-shrub complex), Jeffrey pine 460-760i ma ...... 3 10 32 19 86 760-1070m ...... 203 12 33 21 3 911 woodland. 1070-1370m30...... 30 ..9 37 20 4 99 1370-1680m ...... 25 . 10 40 21 4 106 1680-1920m ...... 18 . 14 44 18 6 98 LIFE-FORMS AND GROWTH-FORMS 1920-2140m ...... 10...... 17 54 12 7 72 In the life-formispectra for lower elevations in the D. West-Eastclimatic gradient acrossSiskiyous, low elevations central Siskiyous (Table 17), floristic dominance is Mill Creek...... 48 5 34 id 0 41 shared equally by phanerophytes and hemicrypto- SouthFor k ...... 38 10 29 22 2 58 0 phytes, with smaller, but substantial, numbers of SiskiyouForkmediat...... 3..3.40 9 31 22 56 33 12 37 16 2 94 chamaephytes and geophytes and few therophytes. SturgisCreek ...... B eaverCreek ... t...... 30 14 37 15 4 102 2 These spectra represent the mixed evergreen forests EmigrantCreek. 27 15 35 10 81 and express the temperate, moderately humid forest E. Comparablespectra from other climate of the region. Very similar spectra have areas Cove forest,Great Smoky Mts., 26 3 l13 been obtained for eastern deciduous forests, as repre- Tenn. (Cain 1945) 3...... 36. 4 30 sented in the spectra for Appalachian cove forest Mixedeationpbetic forest,Cin- (mixed mesophytic) and Long Island oak-chestnut cinnati,Ohio (Withrow1932)12 34 4 2033 4 627 Chestnut Oak forest, Long Is- forest in Table 17. The no indication of 1 spectra give land, N.Y. (Cain..... 1936) .. 35 11 33 21 92 the wide differencesin climate and physiognomy of Oak-hickoryclimax, Piedmont, these forests on opposite sides of the continent; life- N. Car. (Oostingeek 1942) . 48..60 0 36 4 0 89 form spectra are relatively insensitive to some dif- Mediterranean live-oak forest, 0-500m elev. (Braun-Blanquet ferences of environment and community character, 1936) ...... 54 9 24 9 4 highly sensitive to others. The "normal spectrum" Beech forest,Cevenne Mts., for the flora of the world computed by Raunkiaer 1000-1500 m elev. (Braun- Bl anquet1936) e...... 27 4 52 40 2 (1918, 1934) is often used as a standard of compari- Subalpineforests, Great Smoky son. For detailed comparisons among temperate-zone Mts. (Cain 1945)...... 21 2 57 17 3 011 forests a differentkind of standard may be appropri- Subalpineforests, Olympic Mts., Wash.(Jones 1936)...... 2 3 7 49 3 11 ate-one representing mesophytic, floristically rich Subalpineforests, Mt. Rainier. forests of unglaciated areas, midway between the Wash. (Jones 1938)...... 11 8 47 31 3 phanerophyte dominance of the tropics and warm- Serpentineand gabbroin Tus- cany (PiOhi-Sermolli1948) ...... 601 9 40 1526 4 05 temperate forests and the heimicryptophytedominance "Normalspectrum" (Raunkiaer of cooler-temperateforests. The very similar spectra 1918,1934) .5. 4 46 9 26 6 13 for Arcto-Tertiary remnants on opposite sides of Mid-temperate mesophytic forests...... 34 8 3 2 North America suggest such a spectrum. Averaging two spectra for the Mixed Mesophytic Forest of the East (Withrow 1932, Cain 1945) and two for the eSalized rnid-tesGperateSmeo ophktispeytruiii forest" Mixed Evergreen Forest in the West (South Fork entered in Table 17. and Beaver Creek, for the transition to redwood Effents of parentTs given-elinateaterial within a forest and a more continental phase) giyes the gen- ale indicated by the first threespehtc a of this table. WHITTAKER EcologicalMonographs 318 R. H. Vol. 30, No. 3 All three spectra have the general pattern of the TABLE 18. Relations of growth-formsand species- mixed evergreen forests; but from diorite through diversities to environmentsin the Siskiyou Mountains. gabbro to serpentine the phanerophytes and geo- Numbersof species infloras of transects Average number of spe- phytes decrease, while chamaephytes and hemicrypto- andsite groupings cies in vegetation samples phytes increase. The effectof the topographic mois- ture gradient at low elevations on diorite, as repre- sented in the second series of spectra, appears in 0 reduction of chamaephytes, increase in hemicrypto- X3 O ,P X nX phytes and therophytes toward drier sites. Some- XQ ce 0 c what comparable results appear in the contrast of ______~~~~ H.P 0 5 E north and south-facing slopes within the same climate A. CentralSiskiyou Mtns., by in the eastern forests by Oosting (1942), Cantlon parentmaterials, 610-915 m Quartzdiorite ...... 17 1346 8 84 43 7.9 4.0 10.7 0.9 23.5 (1953), and Miller & Buell (1956). Generally similar Olivinegabbro ...... 1619 54 12 101 45 7.0 3.4 11.0 1.4 22.8 trends in relation to site moisture appear also in the Serpentine...... 918 73 16 116 59 4.5 6.2 16.5 2.9 30.1 low-elevation gabbro and serpentine floras. The xerie B. CentralSiskiyou Mtns., by sites,610-915 m serpentine spectrum, representing Jeffreypine wood- On diorite-Mesic...... 15 1142 4 72 42 9.7 4.5 11.7 0.7 26.6 land, is the lowest in phanerophytes of the low-eleva- Intermediate.. 15 1148 5 79 50 7. 6 3.9 11.0 1.2 23.7 tion Siskiyou forests, and is similar in character to Xeric. 9 6 27 5 47 53 6.5 3.7 9. 3 0.7 20.2 the spectrum for ophiolitic soils (serpentines and Ongabbro-Mesic. .. . 16 1040 7 73 41 7.0 4.0 10.3 0.8 22.1 Intermediate.. 10 1344 5 72 52 7.0 2.8 9.9 1.5 21.2 gabbros) in a drier Mediterranean climate as given Xeric...... 1011 40 5 66 48 7.0 3.513.3 2.025.8 by Pichi-Sermolli (1948, see Table 17). Changes in On serpentine-Mesic... 9 1856 5 88 59 5.3 8.3 18.4 2.2 34.2 life-formrepresentation with elevation in mountains, Intermediate9 19 68 6 102 72 5.2 6.6 16.8 2.5 31.1 Xeric...... 6 749 14 76 72 2.7 2.6 14.0 4.0 23.3 such as have been shown in other areas by Guyot C. CentralSiskiyou Mtns., by (1920), Hansen (1930), Raunkiaer (1934), Gelting elevationbelts on diorite (1934), Braun-Blanquet (1936, see Table 17), Jones 460-760m ...... 1813 47 8 86 42 7.8 3.811.3 1.124.0 760-1070m...... 19 1550 8 92 44 6.1 4.2 11.9 0.6 22.8 (1936, 1938), Allan (1937), Davidsson (1946), Cain 1070-1370m. 15 1262 10 99 55 3.8 4.4 20.2 1.0 29.4 (1950), Dansereau (1957), and Costin (1959), are 1370-1680m1.3 1370 10 106 39 3.5 3.4 24.5 1.4 32.8 indicated in the third series of spectra. The decline 1680-1920m.6 8 76 8 98 36 2.5 1.7 18.4 1.7 24.3 of the phanerophytes and increase of the chamae- 1920-2140m. 3 3 58 8 72 50 1.8 0.7 14.0 1.5 18.0 D. West-Eastclimatic gradient phytes and hemicryptophytestoward higher elevations acrossSiskiyous, low are consistent with observations in other areas; the elevations slight increase in therophytesis in contrast to results MillCreek . . 11 8 17 5 41 74 5.0 3.7 6.5 0.2 15.4 SouthFork.... 1111 33 3 58 46 5.2 5.0 8.6 0.7 19.5 in most other areas. The spectra for subalpine for- SiskiyouFork. 13 1030 3 56 57 7.3 5.0 9.4 0.3 22.0 ests in the Siskiyous (1670-1920 and 1920-2140 m) SturgisCreek ...... 17 1648 12 93* 42 7.8 6.7 17.5 1.7 33.7 resemble those of other subalpine or Canadian zone BeaverCreek ...... 13 1757 15 102 37 7.2 7.2 15 0 2.5 31.9 Creek. 12 forests (Table 17) in proportions of phanerophytes Emigrant 10 44 15 81 33 4.2 3.8 12.0 3.7 23.7 and hemicryptophytes,but differ in other respects. As suggested by the larger per cent of therophytes, subalpine climates in the Siskiyous are probably drier forms in these floras are evident. Numbers of tree than in the other ranges. species may be observed to decrease (1) from diorite The fourth series of spectra, finally, represent through gabbro to serpentine, (2) from low eleva- the change in climates from the coastal redwood tions toward higher ones, (3) from the central Siski- forests at Mill Creek to the interior valley at Eini- yous (Oregon Caves and Sturgis Creek) toward the grant Creek; the central Siskiyou (Oregon Caves Pacific Coast, (4) from the central Siskiyous toward area) diorite sample fits into this series between the drier interior, and (5) froni mesic sites toward Siskiyou Fork and Sturgis Creek. Since life-form xeric within the central Siskiyous. Numbers of shrub spectra give little expression of the difference be- species (1) increase somewhat fronmdiorite to gabbro tween maritime and continental climates, the gradient and serpentine, (2) decrease toward higher eleva- in question is primarily one of decreasing climatic tions, (3) increase fronl the Coast to the eastern humidity. The four gradients studied parallel one Siskiyous (Beaver and Sturgis Creeks), and (4) another to some extent in the kind of departure from decrease from mesic sites toward xeric ones within the spectra of the mixed evergreen forests, or the the central Siskiyous. The decline in shrub species generalized mid-temperate miesophytic forest, they toward higher elevations parallels that for trees and produce. In this area an approach toward a more is part of the very general decrease in floristic di- "extreme" environment,whether of parent material, versity of woody strata toward the north and to- elevation, or topographic or climatic dryness in al- ward higher elevations (ef. Whittaker 1956 :11). most all cases implies a decrease of phanerophytes The increase in numbers of shrub species from the and geophytes and an increase of therophytes. the Coast toward the Intel-ior is paralleled amiong, Table 18 summarizes the distribution of growth- the herbs. The data for Emigrant Creek suggest forms over some of the saiiie range of environments. that these trends do not extend bevond the floristically A number of trends in the representation of growth- rich comlmunitiesof the eastern Siskivous into those July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 319 of the continental Interior. Because of the dis- bers of species toward higher elevations occurs among turbance at Emigrant Creek, however, the numbers woody plants, but not herbs; and the diversity rela- of shrub and herb species there cannot be relied upon tions of the latter are such that total numbers of as comparable with those of the other transects. The species increase into middle elevations before de- increase in numbers of graminoil species along this creasing into the highest elevations sampled. same gradient into the drier and more open vegeta- These data do not fit into any simple pattern of tion of the Interior is marked. interpretation in relation to environmental favor- Within each of these growth-forms,trends in the ableness. It is difficultto frame any definition of representation of evergreen vs. deciduous forms ap- "favorableness" that does not amount to the observa- pear, notably along the climatic gradient from the tion that those environments are most favorable to Coast inland. Evergreen forms are strongly domi- high species-diversity where high species-diversities nant in all strata in the redwood forests at Mill are encountered. Floristic diversity appears to oc- Creek, they are predominant in the woody strata, at cur as it occurs, a complex resultant of a number least, of the mixed evergreen forests of the central of influences which are at present scarcely under- Siskiyous, but in the oak woodlands at Emigrant stood. The limited interpretations which may be Creek deciduous forms are strongly dominant in all offeredfor the present material are as follows: (1) strata. The most striking expression of continentali- There is some correlation between species-diversity ty in these floras is not in life-form spectra, but in of a stratum and the extent of its development as the representation of evergreen forms both in num- expressed in coverage or biomass. This correlation, bers of species in floras and in community and however, is conspicuously a partial one to which ex- stratal dominance. ceptions may be observed; one author (Yount 1956) has even suggested the reverse relation. (2) Diversity SPECIEs-DIVERSITIES relations to environment are clearly different for The data in Table 18 bear on another problem of differentcommunity fractions and groups of organ- communityrelations to environment-that of species- isms (cf. Whittaker 1956:18). In the Siskiyous, a diversities, or the relative richness of communities in general inverse relation appears between diversities numbers of species. Studies of species-diversity of tree strata and undergrowth. (3) In this vegeta- (Fisher et al. 1943, Williams 1945, 1947, 1951, 1953, tion the herb strata in all cases include half or miore 1954, Preston 1948, Simpson 1949, Whittaker 1952, of the total vascular flora; and communitydiversity MacArthur 1957, Hairston 1959) have dealt mostly is in large part an expression of herb stratum diver- with animal populations, in which measurements can sity. (4) The conditions which are "favorable" for be based on numbers of species vs. numbers of indi- maximum diversity of herb strata, and thereby for viduals. Although measurements of this type can be community diversity, are not the conditions most applied to plants (Williams 1950, Margalef 1949, favorable for development of the tree stratum and 1951, 1958, Black et al. 1950, Patrick et al. 1954, community biomass. Maximum herb-stratum di- Hopkins 1955, Whittaker 1956, Dahl 1957, Curtis versities are encountered in the more open vegeta- 1959) the author has preferred, because of the prob- tions of drier climates and "atypical" parent mate- lem of what constitutes an individual among many rials, in which lower tree coverage permits greater plants, to deal with numbers of species in fixed development of the undergrowth. And, from low to sampling areas-whether the individual community- middle elevations, herb-stratum and community di- samples or the combined areas of the samples for a versity also increase as the density of tree cover (de- moisture-gradienttransect. termined primarily by the density of selerophylls) de- There may be a general tendency for species-di- creases, versity, communityproductivity and biomass, develop- Perhaps the most significant diversity trend en- ment of the woody strata, and differentiationof the countered is that from maritime to continental cli- communityinto strata, to decrease from environments mates. Average numbers of vascular plant species which are "favorable" in the sense of being warm, per sample increased inland from 15.4 in the Mill moist, and equable, toward environments which are Creek redwood forests, to 33.7 in the imiixedever- "unfavorable" or "extreme" in their low tempera- green forests at Sturgis Creek. Biological meaning tures, or aridity, or instability, or low nutrient levels, of this trend is quite uncertain. The greater variety or high salinities. As indicated in Table 18, species- of temporal niches, of life-history timings avoiding diversities do decrease along the moisture gradient in direct competition,in the more widely variable seasons the mixed evergreen forests on diorite, and in a less of the continental climate may contribute to it, along clear-cut fashion on gabbro and serpentine. Species with the more open structure of the inland forests. diversities increase, however, from the relatively There is also a marked decrease in average ampli- humid and equable climates of the Coast toward the tudes of species distributions along topographic drier and more variable climates of the Interior. moisture gradients, from the coast inland. The Even more striking is the increase in floristic diver- corollary of this decrease in species amplitudes is an sity from the "normal" diorite to the "abnormal" inerease in floristic differentiation of communities serpentine soils. In relation to elevation, finally, it along the moisture gradient, from the coast inland. may be observed that the expected decrease in num.- It is as if only one flora occupied the whole of the .. 320 R. H. WHITTAKER EcologicalMonographs ~~~~~~~~~Vol. t~~~. WHITTAKER 30, No. 3 moisture gradient in the redwood belt, but two or ber of authors (Gleason 1920, Kulezynski 1928, Ren- three floras in the inland environments. Because the konen 1938, Dyksterhuis 1949, Odum 1950, Whit- distributions of species overlap more broadly in the taker 1952; Bray 1956, Curtis 1959) compares two maritime pattern, the community types there seem samples in terms of percentage composition: to intergrade and interpenetrate one another more PS-1-0.5J I a-b I-Emin (a,b), broadly. Beeause of the narrower amplitudes in the continental climates, a larger number of distinguish- in which a and b are the percents which a particular able community-types appear along the moisture species includes of the total samples A and B. The gradient, and these comimunity-typesseem more dis- computation, and some of the limitations inherent in tinct. the measurement, have been discussed elsewhere The increasing species-diversities from the coast (Whittaker 1952, Whittaker & Fairbanks 1958). inland are thus the product of two phenomena- It has been observed that olivine gabbro is cheiii- diversity of individual stands, and relative differen- ically intermediate to quartz diorite and serpentine, tiation of vegetation patterns along topographic and correspondingly supports vegetation which is moisture gradients-, both increasing inland to pro- intermediate and connects the very anomalous ser- duce the observed contrasts. The extent to which pentine communities with the "normal" ones of dio- these contrasts characterize continental and maritime rite. Physiognornicallythe gabbro vegetation is inter- climates in general is unknown; but they are con- mediate, but closer to that of diorite. It is conse- sistent with conditions of western Europe observed quently of interest to determine the underlying flo- by RObel (1927) and with the author's material on ristic similarities of these vegetation patterns, coni- the central Siskiyous and the more strongly conti- puting coefficientsof communityfor lists of 81, 101, nental Great Smoky Mountains. The two phenomena and 116 species in the transects. Resulting values involved suggest the need for distinguishing three were-diorite vs. gabbro 34.4%, gabbro vs. serpen- aspects or levels of species-diversity in natural comn- tine 38.2%, diorite vs. serpentine,14.8%. The gabbro munities: (1) The richness in species of a particular flora shares species in almost equal proportions with stand or community,or a given stratum or group of diorite and serpentine. organisms in a stand. Fisher's (1943) alpha index Percentage similarities also were used to compare is one means of measuring this, which may be desig- the three vegetations in terms of quantitative conm- nated primary or "alpha" diversity. (2) The extent position of stands. The three comparison standards of change of community composition, or degree of (see Part II) for mesic sites (ravines), intermediate community differentiation,in relation to a complex- slopes (open E-facing), and xeric slopes (open SW- gradient of environment, or a pattern of environ- facing) were compared in all possible combinations, ments, which may be designated secondary or "beta" with separate computations for the tree strata and diversity. (3) The species-diversity of a number of undergrowth. Relations of interest in the resulting communitysamples, for some range of environments, sample-association matrices include the following: which have been combined, so that the diversity value 1) In general, the gabbro samples are clearly in- is a resultant of both alpha and beta diversities of termediate to the comparable samples from other these samples. The transect diversities of Table 18 parent materials; but in detail, complex interrelations are of this tertiary or "gamma" diversitytype, as are appear. The gabbro ravine samples were more siumii- many of those for animal collections to which diversity lar to serpentine than to diorite ravine samples in measurements have been applied. The same types of both strata; among E-slope samples gabbro was more as to measurements may be applied to "gamma" similar to diorite than to serpentine; among SW-slope "alpha" diversity; "beta" diversity represents a dif- samples the gabbro undergrowth was equally similar ferent problem, to be discussed in the following sec- to diorite and serpentine, but the tree stratum was tion. closer to diorite. Averaging all three samples to- gether, the undergrowth of gabbro was closer to ser- SAMPLE SIMILARITIES pentine (21%) than to diorite (15%), while the tree AND COENOCLINEDIFFERENTIATION stratum of gabbro was closer to diorite (36%) than to There are many possibilities for expressing the serpentine (20%). relative similarity or dissimilarity of two communi- 2) In general, the contrast between the different ties; but two most simple and direct ones have been vegetation patterns increased from mesic sites to xeric most widely used-coefficient of community and per- ones. Averaging percentage similarities for the strata centage similarity (Whittaker & Fairbanks 1958). and parent materials (diorite vs. gabbro and gabbro The coefficientof community, first used by Jaccard vs. serpentine) gave: 30% for ravine samples, 27% (1902a, 1902b, 1908, 1912, 1932; Koch 1957) is: for east slopes, and 13% for southwest slopes. 3) In all cases the SW-slope samples were mnore number of species in both samples or floras = of the next number of species in one or both similar to E-slope than SW-slope samples parent material in the series-diorite, gabbro, seI- Percentage similarity, which, with variations, has pentine. This fact is to be correlated with the "shift been independently discovered and used by a nuiii- toward the mesic" of species distributions (Part VIT). July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 321

4) Although percentage similarities for the tree stratum and undergrowth are necessarily correlated, 80: lack of any strict consistency of these is evident in 30 the data. In extreme cases contrasts of these values 20 were as high as 67 and 34, 30 and 8, 9 and 1%. This 0 enso8 _ fact is consistent with what has already been ob- 6- served on the lack of strong correlation between dif- -4' ferent strata and community-fractions. 0 3 A further problem to which such measurements 0 27 may be applied is that of degree of differentiationof (.0 80: the communities along an environmental gradient. (.0 60. 040- "Half-changes" in a rather crude form were suggested 30. a)80 in a previous study (Whittaker 1956) as a basis of 020- such measurement. A quite differentapproach has 10

been suggested as the "index of biotal dispersity" of 6: 04 Koch (1957). .3- The simplest measurementof beta diversitymay be 2 0 the relation, ,B- ,in which mis the diversity value 80: (.0 60: 40- C30- for an individual sample and y that for the sample 420 resulting from merging a number of individual sam- 20 ples from a communitypattern or coenocline. Thus, 0)8. comparing Fisher alpha measurements for average 4 numbers of tree species and individuals in individual 3 samples, with those for tree species and individuals 2- in whole transects,gives "beta" diversities of 1.95 for 1 2 3 4 56 7 8 9 11 diorite, 2.24 for gabbro, and 2.33 for serpentine at Intervals of transects low elevations, and 1.60 for high-elevation forests (1920-2140 m) on diorite. An alternative approach, FIG. 18. Change in community composition, as when a particular gradient is in question, is direct measured by coefficientsof communityand percentage comparison of the extremes of the gradient by either similarities for the tree strata, iin relation to environ- coefficient of community or percentage similarity. mental gradients in transects. In each case the upper curveis for coefficientsof community,the lower for per- Results have been found generally unsatisfactory; centage similarities; both these values are plotted on for such values are often zeroes, or so low as to per- logarithmic scales. Top-along the elevation gradient mit no effective comparison. In the search for a in the Great Smoky Mountains; transect intervals are more effectivemeasurement, a number of transects 400 ft (122 m). Middle-along the topographicmois- from the Great Smoky and Siskiyou Mountains have ture gradient between 460 and 760 m, Great Smoky been studied to determine the relations of percentage Mountains. Bottom-along the topographic moisture similarity and coefficientof community to distance gradient, 610-915 m on diorite, Siskiyou Mountains. along ecological gradients. Curves are smoothedby averaging values for given num- bers of intervalsalong the transects; thus the firstpoint When these values on a logarithmic scale are plotted is an average of comparisons of samples one plotted against distance along a gradient, the curves transectstep apart, the second point an average of com- take the form illustrated in Fig. 18. The striking parisons of transect steps two intervalsapart, etc. feature of the curves is the apparent straight-line and log of relation between environmental gradients the case for, though species reach their limits gradu- sample similarities along much of the length of the ally, one by one, along the gradient, a point is ulti- gradients. Three qualifications of this straight-line mately reached beyond which no species present in relation must, be observed: (1) These however, the first sample occurs. (3) In. some curves of this straight lines, extrapolated back to zero intervals type, steepness gradually increases beyond the first along the gradient, strike the ordinate not at 100% 2-3 intervals of the transect. but at 70% or 80% (and values as low as 40% and For of the as high as 95% have been obtained on other tran- the straight portions curves, change in sects). This is consistent with the fact that coeffi- community composition in relation to distance along cient of community and percentage similarity mmieas- an environmentalgradient may be expressed in rela- urements for two samples from the same stand yield tively simple form: y - bx, in which y is sample not 100%, but values of these magnitudes (cf. Whit- similarity (coefficient of community or percentage taker 1952 :12 on "internal association," Bray & Cur- similarity), x is the number of intervals along the tis 1957). (2) Beyond a certain point, the lower gradient, a is the value of y for zero distance, and b parts of the curves slope off with rapidly increasing is the constant ratio which (letermines the slope of steepness toward zero similarity. This also must be the line on log-linear plot and expresses rate of com- 322 R. H. WHITTAKER EcologicalMonographs Vol. 30, No. 3 munity change in relation to the scale of the x axis. terial on the Great Smoky and Siskiyou Mountains, the following observations are suggested: Or: y = ae Xx in which )X also is a constant for 1) Extent of change of community composition, slope in the form of the decay constant. This rela- or "coenocline differentiation,"as measured in half- tion may provide, with due allowance for its limita- changes along the moisture gradient, increases from tions, a unit for "ecological distance" (Whittaker the diorite to the serpentine vegetation pattern. The 1952, 1956) -relative distance along an environ- lower coenocline differentiationon gabbro may re- mental gradient, as expressed in change in commu- flect its location closer to the coast. nity composition in response to that gradient. The 2) Extent of population change along the moisture "half-change" unit may be defined as the distance gradient tends to increase from the tree, through the along an environmental gradient necessary to reduce shrub, to the herb stratum (Table 19). This contrast sample similarityto one-half that of the zero distance. in rates of change of differentstrata becomes con- When coefficient of community or percentage spicuous in some subalpine forests (Whittaker 1956). similarity values are available for a transeet, the 3) There is some decrease in coenocline differentia- number of half-changes along the gradient may be tion with increasing elevation in the forest vegeta- determined: tion of the Smokies and the central Siskiyous. This tendency does not apply equally to all strata and ele- z=aX2-5 vations, and is most marked in the tree strata of highest elevations. log a - log z 4) There is a very impressive increase in coeno- ,3= log 2 dline differentiationin all strata from the Coast in- land in the Siskiyous, as indicated above. Coenocline in which is the number of half-changes, a is the P differentiationis still higher in the lower-elevation sample similarity at zero intervals, and z is the sam- forests of the Great Smoky Mountains, in a more ple similarity for samples from the extremes. This continental climate, than in any of the mixed ever- measurement, however, gives a disproportionate green forest areas. weight to the sample comparison for the extremes. A 5) Coenocline differentiationand numbers of spe- measurement which better expresses community cies in individual stands, beta and alpha diversities, change along the gradient as a whole can be derived appear to increase and decrease in parallel over from the slope of the first,nearly linear portion of most of the range of environmentsrepresented in the the curve. The line formed by this, extrapolated to study. the number of environmental intervals represented The results discussed do not fit into any very sim- by the extremes of the transect, indicates a value of ple pattern of interpretation. The greater humidity z, from which a value for 3 may be determined. and more equable temperatures of the maritime cli- Table 19 gives percentage similarity half-changes by mate may imply smaller contrast of moisture-balance both methods. From this, and the author's other ma- conditions in most xeric and most mesic sites, and thereforea smaller contrast of moisture-balance con- TABLE 19. Change in communitycomposition (per- ditions along the topographic moisture gradient in a centage similarities) along the topographic moisture maritime, as compared with a continental, climate. gradient in low-elevation transects in the Siskiyou The more open vegetation structure of serpentine, Mountains,by parent materials and strata. with greater exposure to insolation and evaporation, may imply greater contrast in moisture-balance condi- -tions along the moisture gradient than on diorite. In general, extent of change in community com- Cd co Cd~+ Cm a position seems correlated with extent of environmental Transect ' S contrast along the topographic moisture gradient. Coenocline differentiationis not a direct expression of range of environments, however. In the Great Smoky Mountains there is striking contrast in the degree of change in tree-stratumcomposition between Diorite Trees .... 90 17.8 2 .5 43 1. 1 the Picea-Abies forests above 4500 feet on the one Shrubs.... 90 14.2 2.7 22 2.0 hand, and the mostly deciduous forests outside the Herbs 82. 5.8 3.7 15 2.4 range of spruce and fir at the same elevations, on the Gabbro Trees. 73.. 27.9 1.4 40 . 86 other. Coenocline differentiationmay thus be af- Shrubs.... 72 25.0 1 .5 38 .92 fected by the kinds of communities that develop in Herbs ...... -70 14.3 2.3 24 1.6 a range of environments,as well as by that range of Serpentine Coenocline Trees. 5.... 0 7 .0 3 .5 23 1.8 environments itself. differentiation,like Shrubs ..72 7.0 3.4 21 1.8 species-diversity of samples, is presumably a corn- Herbs ... 70 3 . 8 4 . 2 18 2 .0( - munity characteristic complexly related to character- istics of environments and of the communities which July,1960 VEGETATION OF THE SISKIYOIJ MOUNTAINS, OREGON AND CALIFORNIA 323

TABLE 20. Geographic Relations of Floras.

PARENT MATERIAL AND ELEVATION IN THE CEN- EAST-WEST CLIMATIC TRANSECT OF THE TRAL SISKIYOU MTS. SISKIYOU MT'. AT Low ELEVATIONS

Low-Elevation High-Elev. West ...... East Forests Forests Summits Coastal ...... Continental

1:5 9_2 66U~~~~~~~~~~~~0~J 0rO~ Ctn+2 0o S cVzuXv 0Z- U ; V, X

Total nos. of spp. in transect 41 58 56 93 102 81 fioralns..of....i.transect 84 101 116 81 87 54 48

Floristic Groups (per cent of transectfloras) SiskiyouEndemic .....1 7 16 5 2 8 2 2 2 1 KlamathEndemic ...... 3 9 4 6 7 10 2 1 3 Sierran...... 4 9 10 6 10 7 6 2 3 2 3 8 6 6 Sierra-Cascadian ... . 8 5 3 7 2 7 2 7 9 6 5 3 Cascadian.1 1 6 5 6 6 2 2 2 1 4 Sierra-Coastal 1...... 6 3 9 1 7 4 2 2 3 5 6 5 Coastal . 12 20 15 6 11 6 6 7 9 11 11 10 5 7 Coastal-Northern .. . 17 9 5 6 2 2 39 28 27 12 11 48 Western.3...... 6 25 21 38 34 46 48 27 34 30 39 37 23 Nearctic...... Tota . 15 10 11 25 17 13 10 17 10 12 19 19 FloristicVectors Tree: North. 33.4 3.1 3.0 4.2 2.9 3.6 3.5 3.3 3.4 3.5 4.0 East. 2.3 2.3 2.3 2.8 2.7 2.4 2.4 2.2 2.5 2.9 2.8 South....3.0 3.4 3.3 2.2 3.4 2.2 3.0 2.8 3.2 3.5 3.8 Total . 8.7 8.8 8.6 9.2 9.0 8.2 8.9 8.3 9.1 9.9 10.6 Shrub:North ....3.7 2.6 1.5 3.0 2.2 3.0 4.1 4.6 4.0 3.2 3.4 3.2 East... .342.5 2.0 1.6 2.8 2.5 2.8 1.8 2.5 2.4 2.6 2.8 3.2 South . 2.6 2.6 2.1 2.0 3.3 3.0 2.4 2.5 2.7 2.7 2.6 3.2 Total ...... 58.8 7.2 5.2 7.8 8.0 8.8 8.3 9.6 9.1 8.5 8.8 9.6 Forb:North ....3.6 2.5 1.9 3.6 2.5 2.8 2.1 4.3 3.6 3.8 3.4 3.2 3.4 3.2 East. .. . 2.9 2.2 1.9 3.4 3.0 3.0 2.7 3.2 2.8 3.0 3.2 3.2 South. 2.8 2.5 1.8 2.5 2.3 2.4 2.0 3.1 2.9 2.9 2.8 2.5 3.0 Total. 9.3 7.2 5.6 9.5 7.8 8.2 6.8 10.6 9.3 9.7 9.4 8.9 9.6 4.2 Graminoid: North . 4.4 3.4 3.3 4.6 4.1 3.5 3.6 4.6 4.0 3.3 4.0 4.5 East...... 4. 2 3.3 3.8 4.6 4.4 3.8 4.1 2.8 2.3 3.3 5.1 4.6 4.9 South. 2.9 2.6 2.9 3.2 3.1 3.5 3.4 2.0 2.3 2.7 3.1 2.7 2.8 Total . . 11.5 9.3 10.0 12.4 11.6 10.8 11.1 9.4 8.6 9.3 12.2 11.8 11.9 3.6 All together: North. 3.6 2.7 2.1 3.8 2.7 2.9 2.3 4.1 3.8 3.8 3.5 3.5 East...... 2.8 2.3 2.2 3.5 3.1 3.1 2.9 2.7 2.6 2.7 3.2 3.3 3.5 South...... 2.8 2.7 2.1 2.6 2.6 2.6 2.3 2.6 2.8 2.8 2.9 2.7 3.1 Total.....9.2 7.7 6.4 9.9 8.4 8.6 7.5 9.4 9.2 9.3 9.6 9.5 10.2 Maximumextents (per cent of traset fors)North ....45 29 21 37 22 22 10 56 47 45 34 33 30 East...... 13 10 15 23 22 22 25 17 9 12 20 18 18 South....13 27 23 9 20 19 12 12 12 14 16 12 15 Indeterminate29 35 41 31 37 37 52 15 33 29 30 37 37 develop in those environments,and may be very dif- four of the high peaks on acid or intermediate rocks ferently expressed in differentstrata and groups of (Lake Mtn., Mt. Elijah, and Grayback Peak, and Mt. organisms in the same communities. Ashland) and two high peaks of serpentine (Big Red Mtn. and Observation Peak). GEOGRAPHICRELATIONS OF FLORAS Geographic relations of all species, and subspecies All were Another approach to analysis of relations of plant and varieties, were codified in two ways. of state- life to environments in the Siskiyou AMiountainsis classified first by areal types on the basis nients of distribution in floristic manuals (primarily that dealing with geographic relations of possible, Areal floras. The comparisons given in Table 20 are based Peck 1941, Abrams 1940-51, Jepson 1923-5). been on transect floras; data are given also for species in types in relation to plant coimmunitieshave 6 samples of high-elevation serpentine vegetation much less studied in North America than in Europe ranging from Abies nobilis forest to Pinus jeffreyi (cf. Hansen 1930, B6cher 1938, 1954, Meusel 1939, woodland at elevations between 1900 and 2100 m on 1943, Schmid 1950, Ellenberg 1950, Walter 1954), Big Red and Little Red Mountains and Observation and the information on which to base them is less Peak. Summit lists are based on collections from adequate; but a system was developed with the fol- Ecological Monographs 324324 RR. 'LH. 'WH3ITTAKERWUITTAKER ~~~~~~~~~~~~~~~~~~~~~~~Vol.30,No. 3 lowing major types: (1) Siskiyou Endemics-species grouping including the Slerran, Coastal, anid Sierra- (or, when distinguished, subspecies or varieties) re- Coastal species from diorite to serpentine. These re- stricted to the narrower area of the Siskiyou Moun- lations seem interpretable on the basis that the gabbro tains, (2) Klamath Endemics-restricted to the and serpentine vegetation is more open and more larger Klamath Region, (3) Sierran-extending xerophytic in appearance, with greater exposure of south from the Klamath Mountains and having the the undergrowth to light and evaporation. A larger principal area of occurrence in the Sierra Nevada, proportion of the species of the closed forests on (4) Cascadian-extending northward and having the diorite extend into the relatively dense mesophytic principal area of occurrence in the Cascade Range, forests to the north of the Siskiyous; a larger pro- (5) Sierra-Cascadian-extending both north into portion of the gabbro and serpentine floras extend the Cascades and south into the Sierra Nevada, (6) into the more open and xerophytic communities of Sierra-Coastal-extending south from the Klamath California (Whittaker 1954b). The data suggest Region both in the Sierra Nevada and the Coast that differentparent materials in the same climate Ranges, (7) Coastal-extending along the California may not only have marked effectson extent of nar- and Oregon coasts (including coastal mountains), but row endemism in their floras, but may act as differen- not north along the coast to Washington, (8) Coastal- tial filters,admitting differentproportions of species Northern-extending northward along the coast to from the whole. spectrum of geographic relations into Washington and beyond with extensive occurrence in their floras. the area of mesophytic forests surrounding Puget Various trends in floristic relations appear when Sound, (9) Western-widespread in the western In- floras of low-elevation forests on diorite and ser- terior, especially the Great Basin and interior moun- pentine are compared with those of high-elevation tain ranges, (10) Nearctic-of wide distribution in forests and high summits on these same parent mate- North America, extending eastward beyond the rials. The proportion of narrow endemiicsis higher Rocky Mountains. in the higher-elevation floras on diorite, and repre- The second approach was through actual extent sentation of wide-spread Western species (of interior of species areas in a given direction-north (and mountains, primarily) increases toward higher ele- west to Alaska), east, and south. Because of the vations, especially on serpentine. The contrast be- difficultyof translating statements of distribution in tween serpentine and diorite floras decreases toward floristic manuals into numbers of kin, arbitrary 7- higher elevations, though it is by no means absent at point scales were used to express distances of extent, the highest elevations available for study. This trend not beyond: 0-the Siskiyou Mountains themselves is consistent with the more general observation that in a given direction; 1-the Klainath Region north floristic distinctiveness of serpentine is less in far- and south, and the Cascades east; 2-the Oregon- nori;herin environments (and especially glaciated Washington border northward, the Oregon-Idaho areas) than in warm-temperate and tropical ones border and central Nevada eastward, and middle (Whittaker 1954b). California (Monterey) southward; 3-the Washing- Floristic trends in the east-west climatic transect ton-British Columbia border northward, the Idaho- include: (1) Decreasing proportions of coastal spe- Wyoming border and central Utah eastward, and cies, in the broad sense, toward the interior. (2) southern California (Los Angeles) southward; 4- Among the different coastal groups, the Coastal- Prince Rupert, British Columbia northward, eastern Northern species decline in numbers into the drier Wyoming and the Colorado Front Range eastward, climates of the interior, but the Coastal and Sierra- northern Baja California southward; 5-the north- Coastal species increase in numbers fronmMill Creek south border of Alaska and Yukon territory and its to Sturgis Creek. There is no paradox in this in- projection to the coast northwestward, west-central crease of "Coastal" species toward the interior since Kansas and Nebraska (100th meridian) eastward, most of them are species of the California Coast Guaymas southward; 6-the Alaska Peninsula and Range mountains, rather than the more narrowly Bering Sea northwestward,the Mississippi River east- maritime environmentsof the coast itself. (3) Repre- ward, Mexico City southward; 7-northwestward sentation of Sierran and Sierra-Cascadian species into Siberia, eastward to the Atlantic Ocean, south- increases toward the interior. (4) Representation of ward into Central America or beyond. the widespread Western and Nearetic species also Representation of areal types in the low-elevation increases inland. These trends seem clearly correlated floras of the three parent materials in Table 20 is with climates; for the more mesophytic species of the very significantlydifferent. There are few narrowly Coastal-Northern grouping decrease toward the in- endemic species in the diorite flora (Tauschia kelloggii terior while species of the somewhat drier and more is the only such species in the list) ; but the number continental climates of the California mountains and of species endemic to the Siskiyou Mountains and the the interior increase. Representation of narrowly Klamath Region increases to 10% of the gabbro and endemic species is small throughout; for representa- 25% of the serpentine floras. The three parent ma- tion of these is generally high on special parent ma- terials show marked differencesalso in the decreasing terials, moderate at high elevations on any parent representation of Coastal-Northern species and in- material, but low on the more "normal" parent ma- creasing importance of a southern or "Californian" terials at low elevations. July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 325

The data on floristic "vectors," or relative extents populations are distributed in relation to the three in differentdirections, are compiled in two forms in soils. Table 20. Distance values for the differentdirections have been averaged for species within growth-forms, TREES and for whole floras; and numbers of species with Pseudotsuga menziesii is the major tree species greatest extent in a given direction, according to the of low elevations; and on all three soils it extends scale values, are tabulated also for whole floras. Di- along most or all of the moisture gradient with ap- rection of greatest extent is determined in this case parently bimodal populations (Fig. 20). Average by the largest of the three values for directional ex- densities of Pseudotsuga stems are considerably lower tent. "Indeterminate" species are those in which the on gabbro and serpentine; densities of stems from two highest scale values were equal; and many of the the 1-in. class up per hectare were 177, 60, and 78 narrow endemics are included among these. on the three soils. Effects of serpentine and gabbro A very marked decrease in average extent of spe- are more striking when numbers of large stems are cies areas appears in the series of floras from diorite compared; densities of stems 37 cm (15 in.) dbh and through gabbro to serpentine. This decrease applies, over on the three soils were 88, 19, and 23. The ratio in essentially parallel fashion, to the shrub and forb between the two density values, expressed as num- groupings; but it does not apply to trees and grami- bers of stems 37 cm and over among 100 stems over noids; for the conifers and grasses of serpentine are 1 cm dbh, provides an indication of slope of the stein predominantly widespread Western or Nearctic spe- number-diametercurve (Part IV), and average sur- cies. In directions of maximum extent also, the dio- vival of trees into larger size classes. These values rite and serpentine floras are in contrast. A high for diorite, gabbro, and serpentine were 49.7, 32.6, proportion of species have maximum extent to the and 6.0, indicating an appreciable reduction of average north in the former, maximum extent to the south Pseudotsuga size on gabbro, as compared with diorite, in the latter. On both soils average distances of ex- and a striking reduction on serpentine. Increment tent appear to increase from low-elevation into high- borings are available for diorite and serpentine; the elevation forests, but to be lower in the summit floras, lower rate of diameter growth on the latter is illus- with their greater representation of alpine endemics, trated in Fig. 19 (see also Fig. 14). than in the high-elevation forests. Other approaches to floristic generalization were experimented with. Correlations of geographic rela- tions and topographic moisture gradients were C l studied. On diorite and gabbro some correlation in the expected direction appeared; average extents to- ward the north were greater, and a higher propor- tion of species had maximum extents toward the north, in mesic sites than in xeric. No such correla- tion was observed in the more strongly endemic ser- pentine flora. Analysis of dispersion values for the distances of extent, as an approach toward measure- AAge in ye ars a t br e ast hei ght ment of relative floristic was heterogeneity, attempted. FRIG. 19. Grlowthin diameterin relationto age for No significant differenceswere obtained, apart from Pseudotsuga mnenziesiion diorite (D) and serpentine those resulting from degree of endemism. Measure- (S), based on 11 ineremelltborings for lowvelevations ments of relative difference in extent to the north on diorite, 7 inecremlelltborings from serpentine,choseil and the south were attempted as an approach to meas- to representall sites and average out site differenceso ureinent of the degree to which the differentfloras a given parent material. were intraneous or extraneous to the Klamath area. Again, no significant differencesappeared which did Chamsaecyparis lawvsonian~awlas of about equal did not result from the proportion of narrow en- abundance on diorite and serpentine but less abundant demics. These and other approaches to floristic ab- in the gabbro transect. The reduction in tree size oil straction encounter a fundamental limitation in the serpentine which is so evident in Pseudotsuga did not fact that all floras are mixtures of species with hetero- appear in Chamaecyparis, which is in consequence a geneous, "individualistic" relations to climate and smaller tree than Pseudotsuga on dliorite,but a larger geography. Yet the differentcommunities within a one on serpentine. Whereas the densities of Pseudo- limited area do differ significantlyin geographic re- tsugal and Chamlaecyparis were loosest ill the gabbro lations of their floras; and these differencesare svste- transect, that of Pinuls lantbertiairawas highest therem mnaticand interpretable, and subject to measurement. Relative survivals into larger size classes were similar on all three soils, but P. lamtbertiana showed very VII. RELATIONS OF SPECIES POPULATIONS differentrelations to the moisture gradient on the TO THE THREE SOILS differentsoils (Fig. 20). On diorite the population The results of the study which remain to be dis- is highest in xeric sites, on gabbro in subxeric ones eussed are autecological-the manner in which species (sigllificance of the mesic population peak is doubt- WHITTAKER EcologicalMonographs 326 R. H. Vol. 30. No. 3

1500 300 D Pseudotsug a Lithocarpus D menzie si 1000 densiflora 200 S 100 T C -= = 500 __ _

60 Pinus 6600 Quercus G. . \ lambertiana chryso p __ 40 S N 400 200 4200-c

150 -Libocedrus 600 Arbutus 100 decurrens -- (L) menziesii a) 400 X__200G

a Castanopsis FIG 20 lawsonisna CX200 chrysophylla ,200. \\ (n D Eioo ts E 100

0300 m`\S.~ onio m 100-,Umbellularia californica ponderosa (G) 200 jeffreyi (S) 5 0 100 . -G J.

300 Pinus 1000 Arctostaphylos Q 200o n monticota (in cinerea (G) N*attenuata (a) 500 viscida (S) A 100 a S

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 M oistur e Gradient Moisture Gradient Mesic Xeric Mesic Xeric FIG. 20. Population distributionsof conifersin rela- FIG. 21. Population distributionsof selerophyllson tion to topographicmoisture gradients at low elevations diorite (D), gabbro (G), and serpentine (S). on diorite (D), gabbro (G), and serpentine (S). In this and following figures,populations on diorite are represented by continuous lines, those on gabbro by californica is almost absent from diorite; it is an broken lines, and those on serpentineby dot-and-dash abundant small tree oIn gabbro and a common shrub, lines. with habitus very differentfrom that on gabbro, oil serpentine. ful), on serpentine in submesic ones. Through the UNDERGROWTHSPECIES series of soils the population center or mode is dis- Figs. 22-25 summarize distributional relations for placed progressively toward the imesicend of the gra- a number of major undergrowth species. Almost all dient. Such "shifts toward the mesic," as they will show the shift toward the mesic, though there are a be called, are a most general feature of the popula- few exceptions (e.g. Whipplea modesta from diorite tion relations to be discussed. to gabbro, Fig. 23) presumably implying occurrence Among the sclerophyllous trees, both Lithocarpus of ecotypic populations with greater drouth tolerance densiflora and Quercus chrysolepis have apparently on the more mafic soils. In some undergrowth spe- bimodal populations on both diorite and gabbro (Fig. cies, as in a number of trees, there is evidence of 21). Both species are represented on serpentine by genetic complexity in population bimodality along the shrubby varieties with submesic modes (Lithocarpus moisture gradient. The last four species, in Fig. densifloravar. echinoides and Quercus chrysolepis var. 25, are bimodal in relation to the parent-material gra- vaccinifolia). Castanopsis chrysophylla occurs on dient, with higher populations on diorite and serpen- diorite wvitha large and apparently bimodal popula- tine, lower on the intermediate gabbro. tion; its occurrence on the other soils is scattered and Further complexities appear in Polystichum, irregular, but the serpentine population again is pre- mnnitum and Trientalis latifolia (Figs. 23 and 25). dominantly shrubby (Castartopsis chrysophylla var. P. munitumvar. munitumis an important mesic pop- minor). Arbutus rnenziesiioccurs on diorite, with an ulation on diorite, with an apparent secondary mode apparently bimodal population centered in more xeric in subxeric sites. The population on gabbro (pre- sites, and on gabbro with a lower population cen- dominantly var. imbricans) is centered in submesic tered in submesic sites; there is no shrubby form on sites; allowing for the shift toward the mesic this may serpentine, where the species is represented by trees correspond to the subxeric-xeric population on diorite of uncommon and irregular occurrence. Umbellularia (in which var. imbricans or its characteristics appear July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 327

10RhusD 750 G~~~ hispidula 50 25 D

400 Achlys D t rip hyllIaD 200 E. E. Gaultheria shalicin o 75 Polystichum G o m u nit um. - 50 - = N, 50

G 2- . _ Pteridium aquilinum D >% 20 a;_. ._ Rhamnus californica 100 var. pubescens var. occidentalis

C50 s G

o Campanula D 75 Polygala 100 prenanthoides californica G._-- 5 0 50 G 2 5

300300 if 40 IrisQ Galium G chrysopylla 0.0 200i ambiguum S 7 20 ?D

1 2 3 4 5 6 7 8 9 10 1 2 4 1 10 Moisture Gradient Mo i s ture Gradient Mesic Xeric Mesic Xeric FIG. 22. Population distributions of undergrowth FIG. 23. Population distributions of undergrowth species on diorite (D) and gabbro (G). species on diorite (D), gabbro (G), and serpentine(S).

exceptionally). Only var. imbricans appears on ser- nature are observed along the east-west gradient pentine, with a shift toward the mesic compared with from humid to drier climates (Part IV). The inter- the gabbro population. Trientalis, with two modes relation of local topographic and broader climatic in relation to moisture on diorite and a third on ser- distributions of species, with the tendency of topo- pentine, has a triangular pattern of population cen- graphic distributions to shift in such a manner as to ters in relation to the moisture and parent-material compensate for differencesin climate, has twice been gradients. Trientalis is also one of a number of un- formulated as a distributional "law"-the law of geo- dergrowth species (Rosa gymnocarpa, Symphori- ecological distributionof Boyko (1945, 1947) and the carpos rivularis, Whipplea modesta, Apocynum pu- law of relative site constancy of Walter & Walter miltum,Smilacina racemosa, Senecio bolanderi) show- (1953). ing evidence of population bimodality in relation to The three parent materials studied are in similar elevation (Tables 13 and 14). general climates. To the extent that the shift may represent a response to climate, it is a response not GENERAL RESULTS to general climate outside communities,but to micro- climates within communities as affected by the kinds The shift toward the mesic is one of the general of vegetation developing on the three soils. Equiva- features emerging from such study: In general, a lence of moisture-balance conditions in such different species population which is observed on two or more relations between soil mois- soils of the diorite-gabbro-serpentine series in the ecosystems,with different conditions, as these three soils same climate, will show a shift of its population dis- ture and evaporative The tribution toward more mesic sites on the more mafie may be very difficultto define. average magni- mesic, measured in soil. Comparable shifts in relation to climate are tudes of the shifts toward the indi- more familiar. In transects of different elevation terms of steps of the ten-step transects, may moisture-balance belts on diorite, high-elevation species shift toward cate something of what equivalent the mesic as their populations are followed down into conditions on the three soils are. Among sets of 8 the warmer and drier climates of lower elevations; and species which could be compared on two of the three low-elevation species show parallel shifts toward the soils, the average shift toward the mesic was 3.0 xeric toward higher elevations. Shifts of the same transect steps from diorite to gabbro, 4.2 transect 328 H. WHITTAKER Ecological R. Vnl-Monographs R Nn R

400 Ceanothus 40 Amelanchier pumilus 200- 20 G

75 Eriophyllum Saws 40. jS\ Pyrola dentata 50 lanatum / \ 20/ \ N20- E__ a 20 Convolvulus S o 75- Chimaphila umbellata .o I polymorphus a 50 D var. occidentalis - 10A 25- S

>400 Lomatium 75^ Rosa D howellii .o 2, 5 gymnocarpa u)200- N 25_ 2

75 Viola 200 Trientalis latifolia

25 I-

75, As.Xerophyllum _.~ 75t s S Vaccinium parvifolium *-*- * tenox 50 50 7 25_ . 25-yIG. 05 ~ ouaindsrbtoso~ N negot 0 I 2 3 4 5 6 7 8 9 10 I 2 3 ~4 5 6 7 8 9 1 Moisture Gradient Gra d ie n Mesic Xeric escMoisture tX FIG. 24. Population distributions of undergrowth FIG. 25. Population distributionsof undergrowth species on gabbro (G), and serpentine(S). species on diorite (D'), gabbro, (G), and serpentine(S). steps from gabbro to serpentine. Among 8 species tine soils, has been described in the European litera- compared from diorite to serpentine, several showed ture (Vilhelm 1925, Pichi-Sermolli 1936, 1948, Rune the 7.0 to 7.5-step displacement suggested by these 1953, Ritter-Studnicka 1956). It has been observed values; the average was 6.5. that most woody species occurring both off and on A second generalization suggested by the data is serpentine are of reduced stature on serpentine, or the individuality in species responses to the sequence are represented there by smaller varieties. Even of soils. Different species show all possible patterns casual observation of herb species occurring on ser- of distributional response or "preferences" among the pentine and non-serpentine soils permits recognition three parent materials. The only relations to the of serpentinomorph differencesin leaf size, texture, three soils which were not observed were the ideal and color, in pubescence, in stature and branching ones which an ecologist might like to expect-sets of pattern-differences which in most cases suggest those ecologically equivalent varieties or congeners replac- which would be expected in a drier or more open en- ing one another at comparable population levels on vironment. The environments on serpentine are in the three soils, or sets of clearly "associated" species fact more open; a species whose population shifts with closely similar distributional relations to the toward the mesic 7 transect steps from diorite to three soils. The principle of species individuality serpentine is still in a communityof less tree coverage (Ramensky 1924, Gleason 1926, 1939), which is and greater exposure to sunlight and evaporation on familiar in application to geographic areas of species the latter. Genetic differentiationof serpentine eco- and distributional relations to moisture and eleva- types is known to be frequent among species occur- tion gradients, also appears clearly in relations to ring both on and off serpentine (Kruckeberg 1951, soil parent-materials in the Siskiyous. 1954). A third conclusion suggested is the general oc- currence of differences in stature or morphology RARE SPECIES AND SERPENTINE INDICATORS when species populations occur on more than one of Restriction to special soil materials is relative; the soils-especially serpentine and non-serpentine Siskiyou species may be arranged in a continuous soils. The phenomenon of "serpentinomorphism,"the sequence from those observed on only one soil to those occurrence of morphologically distinctive forms on with substantial populations on all three. Expansion serpentine in species which also occur on non-serpen- of the field of observation further illustrates the rela- July, 1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 329

tiveness of parent-material restriction. Among spe- serpentine communities of the Siskiyous are cli- cies which might be thought "serpentine species" maxes in the sense of stability or self-maintenance from their occurrence in the low-elevation transects, a (Whittaker 1954b), and are very old. Serpentine whole series appear on diorite at higher elevations communities have existed in the Klamath Region (Pinus monticola, Arctostaphylos nevadensis, Quer- through the Cenozoic, and the rich floras of these cus garryana var. breweri, Eriophyllum lanatum, communitiesrepresent the product of millions of years Silene campanulata var. orbiculata, Eriophyllum la- of species evolution into and in climax communities natum, Erysimum.capitatum, Monardella odoratissima on serpentine. Rare species occur in most varied var. glauca, Epilobium. paniculatum, Biabenaria circumstances; they occur where, according to the sparsiflora, Gilia capitata, Phlox diffusa, Eriogonum population characteristics of and environmental limi- umbellatum, Holodiscus dumosus, Castilleja miniata, tations for particular species, they occur. So far as Angelica arguta, Lomatium macrocarpum). Most of concentrations of considerable numbers of rare spe- these occur on diorite,in high-elevation meadows or cies are concerned, however, the Siskiyou material is on rocky summits, in communities as open and well- in accord with the hypothesis of Detling (1948a, lighted as serpentine vegetation, rather than in for- 1948b) that concentrations of narrow endemics are ests. Conversely, a few low-elevation "diorite species" correlated with concentrations of environmental ex- appear in the denser serpentine stands of higher ele- tremes, and the emphasis by Mason (1946a, 1946b) vations (Arenaria macrophylla, Hieracium albiflorum, of edaphic factors. In relation to succession and Iris chrysophylla). Others of the "serpentine spe- climax, it is suggested that such concentrations may cies" could be observed on soil materials (quartzite be associated not with unstable successional com- and argillite) that had little to -do with serpentine munities, but with relatively stable or climax com- and gabbro, but which supported vegetation some- munities of environments which are distinctive or what more open and xerophytic-looking than that "extreme" and spatially restricted. of diorite (Silene campanulata var. orbiculata, Crepis The group of species largely restricted to serpen- pleurocarpa, Eriophyllum. lanatum, Convolvulus poly- tine, but widely distributed on it, form a nucleus of morphus, Erysimum capitatumr,Monardella odoratis- "characteristic" serpentine species which might serve sima var. glauca, Lotus crassifolius, Calochortus tol- as indicators of serpentine soil. In practice, observ- miei, Gilia capitata, Eriogonum umbellatum, Poly- ing serpentine outcrops scattered through the Siski- stichum munitum var. imbricans, Galiumnambiguum, you Mountains, they were found to have little indi- Lomatium macrocarpum). Still other species large- cator value. Stands containing several of them were ly restricted to serpentine within the study area oc- usually within the major serpentine areas and ob- curred on non-serpentine soils in other parts of the viously serpentine vegetation; stands of smaller and Siskiyous (Vancouveria chrysantha, Balsamorhiza sometimes unmapped outcrops, for which serpentine deltoidea., Eriophyllum lanaturn, Hieracium cyno- indicators were most needed, might contain one or glossoides var. nudicaule, Arctostaphylos viscida, two of these species, but usually none. The most use- Rhododendron occidentale, Calochortus tolmiei, Epi- ful indicators were found to be some of the stratal lobium paniculatvtm,Phlox speciosa, Ceanothus cune- dominants of serpentine (Pinus jeffreyi,P. attenuate, atus, Rhamnus californica var. occidentalis, Galium P. monticola, Quercus chrysolepis var. vaccinifolia, ambiguum, Lomatium triternatum, Perideridia ore- Xerophyllum tenax, Ceanothus cuneatus, Arctostaphy- gana, Polygala californica). For such species rela- los nevadensis) and other species of more frequent tive serpentine restriction, like relative community- occurrence in smaller serpentine areas (Galium am- type fidelity, is clearly dependent on extent of the biguum, Pyiola dentata, Lomatitunmmacrocarpum, field of observation of the species' distribution. Cheilanthes siliquosa, and in wetter sites Rhododen- When these are eliminated from the list of "ser- dron occidentale and Darlingtonia californica). Xero- pentine species" there remain a good number which phyllum tenax seems the most useful single indi- both were rarely or never observed off serpentine, or cator for small serpentine outcrops. Most of these serpentine and gabbro, and are fairly abundant and are species of extensive occurrence on non-serpentine widely distributed in the main serpentine area (Hor- soils. In other areas they are wholly meaningless kelia sericata, Balsamorhiza platylepis, Iris bracte- as serpentine indicators but within the Siskiyous their ata, Lomatium howellii, Epilobium rigidum, Calochor- presence may suggest, though not indicate, serpen- tvts howellii, Antennaria suffrutescens,Cordylanthus tine soil. viscidus, Erigeron foliosus var. confinis,Aster brick- ellioides, Schoenolirion album, Tauschia glauca, Sani- VIII. CONCLUSION cula peckiana). Almost all these are endemic to the One object of the present study was the analysis Siskiyou or Klamath Mountains. These and other and description of a pattern of vegetation of much species largely restricted to, but less widely distrib- intrinsic interest; another was the further develop- uted on, serpentine are truly rare species in the ment of techniques for quantitative analysis of vege- sense of restricted occurrence. tation patterns. In the study as it has been developed It has been suggested by Griggs (1940, cf. Rune in the preceding sections, various aspects of commu- 1954a) that rare species tend to occur in unstable, nity relations have been brought into a system of successional communities rather than climaxes. The quantitative treatment, In some cases the quantita- WHITTAKER EcologicalMS nographs 330 R. H. Vo.30, No. 3 tive analysis may seem only to provide numerical as a multi-dimensionalpattern of vegetationalgra- expressionsfor relationswhich an observantecolo- dients in relationto these environmentalgradients. gist would detectwithout them, and yet may be well Thus the distributionalrelations of species, floristic worthwhile for theadditional clarity with which these groups,and community-types,and the interrelations relationsare expressed. In othercases relationsare of communitiesmay be conceivedthrough an abstract, revealed which are impossibleto determineby field conceptual"pattern" which is not simple,but brings observation. Quantitativetechniques can, when in- into comprehensibleform much of the still greater eptlyor mechanicallyused, obscure important ecologi- complexityof the vegetationitself. cal relations(Braun 1956). They can neversubsti- The fact that the Southern Appalachians and tutefor effectiveobservation, judgment, intuition, and Klamath Mountains bear comparable "central" re- scope of understanding;for only these can suggest lations to easternand westernforests has been com- which quantitativeanalyses are worth undertaking mentedon. One other parallel betweenthese two and provideevaluation and interpretationof the re- mountainareas may be observed-theircombination lationswhich emerge. Yet, in thosestudies for which of unusual geological and biological interestwith the time-consumingsampling and analysis are feasi- relatively undisturbed conditions. The Klamath ble, techniquesof gradientanalysis may be productive Mountainsare now, as the SouthernAppalachians for problemsof both species autecologyand com- were some decades ago, relativelyremote and little munityrelations. known. Partly because they are remote,extensive A centralconcept of the treatmentin thiswork is areas remainwhich are wild and natural. Both areas the community-gradientor "coenocline"in relation have exceptionalscenic attraction; in theKlamath Re- to the topographicmoisture gradient; these commu- gion the coast of southernOregon and northernCali- nity-gradientsare the units of comparisonfrom one fornia,the Rogue and other canyons,and some of climateor parentmaterial to another. It is feltthat the mountainlandscapes themselvesdeserve to be forsome uses thisapproach has significantadvantages morewidely known-and, to the extentthat is feasi- overthose centered on climaticclimaxes or community ble, protected. Because the Klamath Region is not classification.Any "system"for the studyof natural well known,it may be appropriate to observethat communitiesis an integrationof selectedaspects of this is an area of biologicalinterest as great as that speciesand communityrelations, which can be investi- of theSouthern Appalachians: a comparablecenter of gated througha. given, coherentset of concepts,to survivalof ancient,Arcto-Tertiary forms in the Coas- the neglectof otherrelations which cannot,and is tal Redwood and Mixed EvergreenForests, an even necessarilymore appropriate and productivein some greaterconcentration of rare species and remarkable circumstancesthan others. The limitationsinherent floras,together with an unusuallydiverse range of in any approach or systemmay suggestthe desira- floras and communitieswithin a limited distance bilityof applyingmore than one approachto a given fromthe Coast inland, and as dramatican expres- studywhen this is feasible-of supplementinga study sion of relationsof natural communitiesto geologi- in gradient.analysis with considerationof classifica- cal formationsas is to be found anywherein the tion as in the presentwork, a studybased on classi- world. Thereare manyareas of biologicaland scenic ficationwith considerationof gradientrelations. value in the West, but among these the over-allin- Differentiationin relation to topography,and terestof the Klamath Region is high,and suggests patterningin relationto patternsof topographicand continuedthought on long-rangepolicies of utiliza- otherenvironmental gradients, are generalcharacter- tion and preservation. isticsof vegetation;and the conceptof such pattern- ing should be part of the ecologist'sapproach to in- SUMMARY terpretationof a presentvegetation or a fossil flora. The termsecocline -andcoenocline are suggested,not I. Introduction to name theobvious, but becauseof theneed for these The Klamath MountainRegion, lying betweenthe gradient-conceptionsin the interpretationof relations southernCascade Mountains and Pacific Coast in of naturalcommunities to environment.The vegeta- northernCalifornia and southernOregon, is an area tion of the Siskiyous can be conceivedin termsof of great climatic,geological, and vegetationaldiver- climaxregions and vegetationunits. One may regard sity. One major range of the Region, the Siskiyou the Mixed EvergreenForests as forminga vegeta- Mountainsalong the California-Oregonborder, was tional matrix for lower elevations in the central chosenfor a studyof relationsof mountainforests to Siskiyous,a prevailingclimax type in whichthe ser- climatesand parent materials. pentineand otherlocalized communities are dispersed, For reasonsof vegetationalhistory, the age of the and which is replaced toward higherelevations by Klamath Mountains,and the diversityof habitatsin Montane Forest and Subalpine Forest climaxes,to- them,the KlamiathRegion is a "center" for forest ward the west by Coast Forest, and towardthe east florasand forestvegetation of the West. by Oak Woodland,climaxes. But it is profitablealso to conceiveof the vegetationin termsof coenoclines II. Procedure changingin composition,physiognomy, and floristic An area of quartz dioritein the centralSiskiyou relationsin relationto climatesand parentmaterials, Mountainswas chosenfor intensivestudy of elevation July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 331

and moisture-gradientrelations of vegetationon a cline is here suggested; for the gradientof eco- moretypical parent material. Low-elevationvegeta- systemscomprising both of these the term ecocline tion of olivinegabbro and serpentinewas compared may be used. The principalbasis of the studywas withthat of dioriteto studyeffects of parent mate- the comparison of moisture-gradientpatterns, or rials. The change of vegetationfrom the humid coenoclines,from different climates and parentmate- coastal forestsinland at low elevations,and serpen- rials. Along a gradientfrom humid,maritime cli- tine vegetationat all elevations,were studiedless in- mates near the Coast to drier and more continental tensively. ones inland,the characterof low-elevationcoenoclines Gradientanalysis techniqueswere applied to re- changesfrom Sequoia sempervirentsand coastal Pseu- lationsof species and communitiesto major environ- dotsuga forests,through mixed evergreenforests in mental gradients. Within each 1000-footelevation the centralpart of the range,to a patternof Pseudo- belt,or parentmaterial, 60 vegetationsamples repre- tsuga forest,oak woodland,and valley grassland in sentingall parts of the topographicmoisture gradient the Interior. were taken. Three techniquesfor arrangementof these samples into composite transects-by topo- V. Problemsof Classification graphicclasses of sites,by weightedaverages of com- Formations,dominance-types, sociations, associa- munitycomposition, and by comparisonwith stand- tions,and unions in the Siskiyouvegetation are dis- ards representingmesic, intermediate, and xeric sites cussed. The Mixed Evergreen Forests of conifers -were used and evaluated. The weighted-average and selerophyllsare a prevailing climax of the techniquegave best resultsin this study. Klamath Region and northCalifornia Coast Ranges. These forestsare regardedas a major community- III. VegetationDescription type of the West, and interpretedas a formation Low-elevationdiorite vegetation gradates from connectingthe coniferousCoast and Montane For- Chamaecyparis lawsoniana-Pseudotsuga menziesii for- ests withthe Californiabroad-selerophyll vegetation. ests in mesic sites,through Pseudotsuga forests with selerophylltrees in intermediatesites, to selerophyll VI. FloristicComparisons forestwith scattered Pseudotsuga in xericsites. Low- Areto-Tertiaryforest remnants from opposite elevationgabbro vegetation gradates from more open sides of North America,the Mixed Mesophyticand Chamaecyparis-Pseudotsugastands, through more Mixed Evergreen Forests, have very similar life- open selerophyll-Pseudotsugastands, to open, xeric formspectra, approximating: P 34, C 8, H 33, G 23, Pinus-Pseudotsuga- Qtercus-Arctostaphylos stands. and T 2 per cent. The transformationsof such "mid- Low-elevationserpentine vegetation gradates from temperatemesophytic" spectra along gradients of still moreopen Chamaecyparis-Pinus monticola-Pseu- parentmaterial, elevation, topographic moisture, and dotsuga mesicstands, through very distinctiveforest- east-westclimatic change in the Siskiyousare tabu- shrubstands with several conifersand two-phaseun- lated and discussed. The principaleffect of the mari- dergrowthof selerophyllshrubs and grass, to Pinus time-continentalclimatic gradient is in decreasing jeffreyi woodlands. representationof evergreenforms. Towardhigher elevations on diorite,the forestsof Species diversitiesof plant communitiesincrease Pseudotsugaand selerophyllsgradate into montane markedlyfrom maritime into continentalclimates, in- forests dominatedby Pseudotsuga and Abies con- crease fromdiorite to serpentine,decrease along the color above about 1200 m, and these into subalpine topographicmoisture gradient on diorite,and in- forestsdominated by Abies nobilis and Tsuga mer- crease, then decrease, toward higher elevationson tensiana above 1800 m. Mosaic charts(Figs. 11 and diorite. Such relationsare not simplyinterpreted on 12) show relations of vegetationto elevation and the basis of environmental"favorableness." In the topographyon both dioriteand serpentine. Siskiyous,community diversities tend to be higher in the more open communities,with more highly de- IV. Climax Interpretation veloped herb strata,of drier sites and climatesand All low- and most high-elevationvegetation of special parent materials. the Siskiyou Mountainshas been affectedby fires. Three aspects of species-diversitymay be dis- Evidencefrom stand-curve analysis suggests, however, tinguished:richness in species of individualstands, thatthe effect of frequent,less severefires during the degree of floristicchange of communitiesalong en- past centuryhas been to reduce stand densitywith- vironmentalgradients, and species-diversityof a out changingessential character of moststands. Dif- communitypattern or broadly definedcommunity- ferentpatterns of climax vegetationdevelop on dio- typewhich results from both of these. Measurements rite, gabbro, serpentine,and other parent materials of "coenoclinedifferentiation" along the topographic in theSiskiyou Mountains. Any significantdifference moisturegradient, as one approach to the second of in parent materialmay imply differencein climax these, are discussed and applied. Coenocline dif- vegetation. ferentiationincreases from maritimeto continental For a gradientof environmentalcomplexes the climatesin the Siskiyous. term complex-gradient has been suggested; for the Measurements(coefficients of communityand per- correspondingcommunity-gradient the term coeno- centage similarities) show that the vegetationof 332 R. H. WHITTAKER EcologicalMonographs Vol. 30, No. 3 gabbro is floristicallyintermediate to that of diorite Calapooya Range, Oregon. Amer. Midland Nat. 46: and serpentine, though physiognomically closer to the 132-173. former. Baumann-Bodenheim,M. G. 1956. tber die Beziehungen Geographic relations of floras were studied der neu-caledonischenFlora zu den tropisehenand den through two approaches-representation of areal sfidhemisphiirisech-subtropiselhenbis -extratropischen types, and average extents of distribution in different Floren und die giirtelmassigeGliederung der Vegeta- tion von Neu-Caledonien. Geobot. Forschungsinst. directions. Marked correlations of these geographic Riibel Zurich Ber. 1955: 64-74. relations with local environmentswithin the Siskiyous Beadle, N. C. W. & A. B. Costin. 1952. Ecological appear. The serpentine flora has a much higher rep- classificationand nomenclature. With a note on pas- resentation of endemics, smaller average extent in ture classificationby C. W. E. Moore. Linn. Soc. N. S. all directions, but stronger southern or Californian Wales Proc. 77: 61-82. affinitythan the diorite vegetation. Beard, J. S. 1944. Climax vegetationin tropical Amer- ica. Ecology 25: 127-158. VII. Relations of Species Populations to the Three . 1955. The classificationof tropical American Soils vegetation-types.Ecology 36: 89-100. Some Siskiyou species show evidence of popula- Beck von Mannagetta, G. 1901. Die Vegetationsver- tion bimodality in relation to the topographic mois- hlItnisse der illyrischenLander. Die Vegetation der ture gradient or elevation. Most species show a Erde, ed. Engler & Drude 4: 1-534. Leipzig: Engel- "shift toward the mesic," or displacement of their mann. distributions toward less xeric sites, from diorite to Becking,R. W. 1956. Die natlirlichenDouglasien-Wald- gabbro and gabbro to serpentine. Rare species in gesellsehaftenWashingtons und Oregons. Allg. Forst- und Jagdztg. 127: 42-56. this area are concentrated in stable or climax com- Billings, W. D. 1945. The plant associations of the munities on special parent materials, especially ser- Carson Desert Region, western Nevada. Butler Univ. pentine. Bot. Studies 7: 89-123. Black, G. A., T. Dobzhansky,& C. Pavan. 1950. VIII. Conclusion Some attempts to estimate species diversityand population Quantitative techniques of gradient analysis per- densityof trees in Amazonian forests. Bot. Gaz. 111: mit this complex vegetation to be analyzed and under- 413-425. stood as a multi-dimensional pattern in relation to Bocher, T. W. 1933. Studies on the vegetation of the topographic moisture gradients, climates, and parent east coast of Greenland between Scoresby Sound and materials. The biological interest and scenic appeal Angmagssalik (Christian IX.s Land). Meddel. om of the Klamath Region equals that of the eastern Gronland 104(4): 1-132. center of vegetational and floristic diversity, the . 1938. Biological distributional types in the Southern Appalachians, and suggests further con- flora of Greenland. A study on the flora and plant- sideration of preservation of parts of the region for geographyof South Greenlandand East Greenlandbe- the future; tween Cape Farewell and Scoresby Sound. (Danish summ.) Meddel om Grenland 106(2): 1-339. LITERATURE CITED 1954. Oceanic and continental vegetational complexes in Southwest Greenland. Meddel. om Abrams, L. 1940-60. Illustrated flora of the Pacific Gr0nland148(1): 1-336. States, Washington, Oregon, and California. Stan- ford: Stanford Univ. Press; London: Oxford Univ. Bowerman, Mary L. 1944. The floweringplants and Press. Vol. I, 1940, 538 pp. Vol. II, 1944, 635 pp. ferns of Mount Diablo, California, their distribution Vol. III, 1951, 866 pp. Vol. IV, 1960 (L. Abrams & and association into plant communities. Berkeley: Roxana S. Ferris), 732 pp. Gillick Press. 290 pp. Adamovid, L. 1909. Die Vegetationsverhiiltnisseder Bowie, E. H. 1934. Climatic summaryof the United Balkanlinder (M6sische Lander). Die Vegetation der States. Section 15-Northwestern California. U. S. Erde, ed. Engler & Drude 11: 1-567. Leipzig: Engel- Dept. Agric., Weather Bureau, Washington. 24 pp. mann. Boyko, H. 1945. On foresttypes of the semi-aridareas Allan, H. H. 1937. A considerationof the "biological at lower latitudes. Palestine Jour. Bot. R. Ser. 5: spectra" of New Zealand. Jour. Ecol. 25: 116-152. 1-21. Anderson,F. M. 1902. The physiographicfeatures of . 1947. On the role of plants as quantitative the Klamath Mountains. Jour. Geol. 10: 144-159. climate indicators and the geo-ecological law of dis- tribution. Jour. Ecol. 35: 138-157. Axelrod, D. I. 1958. Evolution of the Madro-Tertiary Geoflora. Bot. Rev. 24: 433-509. Braun, E. Lucy. 1935. The undifferentiateddeciduous . 1959. Geological history. In A California flora, forest climax and the association-segregate. Ecology by P. A. Munz & D. D. Keck, pp. 5-9. Berkeley & 16: 514-519. Los Angeles: Univ. Calif. Press. . 1938. Deciduous forest climaxes. Ecology 19: Babcock, E. B. & G. L. Stebbins, Jr. 1938. The Ameri- 515-522. can species of Crepis: their interrelationshipsand dis- 1947. Development of the deciduous forests tribution as affected by polyploidy and apomixis. of easternNorth America. Ecol. Monog. 17: 211-219. Carnegie Inst. Wash. Pub. 504: 1-199. 1950. Deciduous forests of eastern North Baker, W. H. 1951. Plants of Fairview Mountain, America. Philadelphia: Blakiston. 596 pp. July, 1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 333

1956. The development of association and Cooper, W. S. 1917. Redwoods, rainfall and fog. Plant climax concepts; their use in interpretation of the de- World 20: 179-189. ciduous forest. Amer. Jour. Bot. 43: 906-911. 1922. The broad-selerophyllvegetation of Braun-Blanquet, J. 1932. Plant sociology, the study of California. An ecological study of the chaparral and plant communities. Transl. by G. D. Fuller & H. S. its related communities. Carnegie Inst. Wash. Pub. Conard. New York: McGraw-Hill. 439 pp. 319: 1-124. 1936. La Cheiaie d'Yeuse mediterraii6ene Costin, A. B. 1959. Vegetation of high mountains in (Quercion ilicis), monographie phytosociologique. Sta. Australia in relation to land use. In Biogeography Internatl. de G6obot. M6diter. et Alp. Montpellier and ecology in Australia, ed. A. Keast, R. L. Crocker, Commun.45: 1-147. & C. S. Christian,pp. 427-451. Den Haag: W. Junk. . 1951. Pflanzensoziologie. Gruidziige der Vege- Cottam,G. & J. T. Curtis. 1949. A methodfor making tationskunde. 2nd ed. Wien: Springer. 631 pp. rapid surveys of woodlands by means of pairs of Bray, J. R. 1956. A study of mutual occurrence of randomlyselected trees. Ecology 30: 101-104. plant species. Ecology 37: 21-28. 1955. Correctionfor various exclusion angles Bray, J. R. & J. T. Curtis. 1957. An ordinationof the in the randompairs method. Ecology 36: 767. upland forest communities of southern Wisconsin. -. 1956. The use of distance measures in phyto- Ecol. Monog. 27: 325-349. sociological sampling. Ecology 37: 451-460. Brown,R. T. & J. T. Curtis. 1952. The upland conifer- Coupland, R. T. & T. C. Brayshaw. 1953. The fescue hardwoodforests of northernWisconsin. Ecol. Monog. grassland in Saskatchewan. Ecology 34: 386-405. 22: 217-234. Crooke,W. M. 1956. Effect of soil reaction on uptake Burcham, L. T. 1957. California range land. An his- of nickel from a serpentinesoil. Soil Sci. 81: 269- torico-ecological study of the range resource of Cali- 276. fornia. Sacramento: Calif. Dept. of Nat. Resources Culberson,W. L. 1955. The corticolouscommunities of Div. of Forestry. 261 pp. lichens and bryophytes in the upland forests of Byers, H. R. 1953. Coast redwoods and fog drip. northernWisconsin. Ecol. Monog. 25: 215-231. Ecology 34: 192-193. Curtis, J. T. 1959. The vegetation of Wisconsin; an Cain, S. A. 1936. The composition and structure of an ordination of plant communities. Madison: Univ. oak woods, Cold Spring Harbor, Long Islalnd, with Wisconsin Press. 657 pp. special attention to sampling methods. Amiier. Mid- T. & R. P. McIntosh. 1951. An forest land Nat. 17: 725-740. Curtis,J. upland continuumin the prairie-forestborder region of Wis- . 1945. A biological spectrum of the flora of the consin. Ecology 32: 476-496. Great Smoky Mountains National Park. Butler Univ. Bot. Studies 7: 1-14. Dahl, E. 1957. Rondane; mountainvegetation in South Norway and its relation to environment. Norske . 1950. Life-forms and phytoclimate. Bot. Rev. Vidensk.-Akad.Oslo, I. Mat.-Naturv.Kl. Ski. 1956(3): 16: 1-32. 1-374. Cantlon, J. E. 1953. Vegetation and mieroclimates on Dansereau, P. 1957. Biogeography; an ecological per- north and south slopes of CushetunkMountain, New spective. New York: Ronald. 394 pp. Jersey. Ecol. Monog. 23: 241-270. Daubenmire, R. 1952. Forest vegetation of northern Cater, F. W., Jr., & F. G. Wells 1953. Geology and Idaho and adjacent Washington and its bearing on mineral resources of the Gasquet quadrangle, Cali- concepts of vegetation classification. Ecol. Monog. fornia-Oregon. U. S. Geol. Survey Bull. 995(C): 79- 22: 301-330. 133. Davidsson, I. 1946. Notes on the vegetation of Airsk,;- 1947. centers and Chaney, R. W. Tertiary migration gsstrondNorth Iceland. Acta Nat. Island. 1(4): 1-2(0. routes. Ecol. Monog. 17: 139-148. Detling, L. E. 1948a. Environmentalextremes and en- . 1948. The ancient forests of Oregon. Eugene: demism. Madroflo 9: 137-149. Oregon State System of Higher Education (Condon Lectures). 56 pp. 1948b. Concentration of environmentalex- Clark, H. W. 1937. Association types in the north tremes as the basis for vegetation areas. Madrofio coast ranges of California. Ecology 18: 214-230. 9: 169-185. . 1953. Relict islands xeric Clements, F. E. 1928. Plant successioil and indicators. of flora west of the Cascade Mountains in A definitive edition of plant succession ard plant indi- Oregon. Madrofto12: 39-47. eators. New York: Wilson. 453 pp. 1954. Significant features of the flora of 1936. Nature and structure of the climax. Saddle Mountain, Clatsop County, Oregon. Northw. Jour. Ecol. 24: 252-284. Sci. 28: 52-60. Cooke, W. B. 1940. Flora of Mount Shasta. Airier. Dicken, S. N. 1952. The Rogue River countryof Ore- Midland Nat. 23: 497-572. gon: A study in regional geography. Assoc. Pacific Coast 1941. The problem of life zones on Mount Geog. Yrbk. 14: 3-18. Shasta, California. Madrofio6: 49-56. Diller, J. S. 1894. Tertiary revolutionin the topog- 1955. Fungi of Motlit Sbastai (19,36-19.51). raphy of the Pacific Coast. U. S. Geol. Survey 14th Sydowia, Ann. Mycol. Ser. 2, 9: 94-215. Ann. Rept. 1892-93 (Part IT): 397-434. Coombe, D. E. & L. C. Frost. 1956a. The heatlis of the 1902. Topographic developmentof the Kht- Cornish serpentine. Jour. Ecol. 44: 226-256. math Mountains. IJ. S. Geol. Survey Bull. 196: 1-69. . 1956b. The nature and origin of thie soils over -- . 1903. Description of the Port Orford (qlua(l- the Cornish serpentine. Jour. Ecol. 44: 605-615. rangle. IJ. S. Geol. Survey Folio 89: 1-(6. .. 334 R. H. WHITTAKER EcologicalMonographs Vol. 30, No. 3 . 1906. Description of the Redding quadrangle. . 1954b. Vegetational classificationand vegeta- U. S. Geol. Survey Folio 138: 1-14. tional continua. (Germ. summ.) Angew. Pflanzensoz. 1914. Mineral resources of southwestern [Wien], Festschr. Aichinger 1: 168-182. Oregon. U. S. Geol. Survey Bull. 546: 1-147. Griggs,R. F. 1940. The ecology of rare plants. Torrey Diller, J. S. & G. F. Kay. 1909. Mineral resourcesof Bot. Club. Bull. 67: 575-594. the Grants Pass quadrangle and bordering districts, Grinnell,J. 1935. A revisedlife-zone map of California. Oregon. U. S. Geol. Survey Bull. 380: 48-79. Univ. Calif. Pub. Zool. 40: 327-330. 1924. Description of the Riddle quadrangle. Grinnell,J., J. Dixon, & J. M. Linsdale. 1930. Verte- U. S. Geol. Survey Folio 218: 1-8. brate natural history of northernCalifornia through Du Rietz, G. E. 1930 (1932). Vegetationsforsehung the Lassen Peak region. Univ. Calif. Pub. Zool. 35: auf soziationsanalytiseherGrundlage. Handb. Biol. 1-594. Arbeitsmeth.ed. Abderhalden 11, 5: 293-480. Grinnell,J. & T. I. Storer. 1924. Animal life in the . 1936. Classificationand nomenclatureof vege- Yosemite. An account of the mammals,birds, reptiles, tation units 1930-1935. Svensk Bot. Tidskr. 30: 580- and amphibians in a cross-sectionof the Sierra Ne- 589. vada. Berkeley: Univ. Calif. Press. 752 pp. Dyksterhuis,E. J. 1949. Conditionand managementof Guyot, H. 1920. Le Valsorey. Esquisse de botanique range land based on quantitativeecology. Jour. Range geographique et 6cologique. Schweiz. Naturf. Gesell. Mangt. 2: 104-115. Pflanzengeog. Komm., Beitr. Geobot. Landesaufn. 8: Eggler, J. 1954. Vegetationsaufnahmenund Bodenun- 1-155. tersuchungenvon den Serpentingebietenbei Kirchdorf Hairston, N. G. 1959. Species abundance and com- in Steiermark und bei Bernstein im Burgenland. munityorganization. Ecology 40: 404-416. Naturw. Ver. SteiermarkMitt. 84: 25-37. Hale, M. E., Jr. 1955. Phytosociologyof corticolous Ellenberg,H. 1948. Unkrautgesellschaftenals Mass fur cryptogamsin the upland forests of southernWiscon- den Sduregrad, die Verdichtung and andere Eigen- sin. Ecology 36: 45-63. schaften des Ackerbodens. Be-. liber Landtechnik 4: Hall, H. M. & J. Grinnell. 1919. Life-zoneindicators in 130-146. California. Calif. Acad. Sci. Proc. Ser. 4, 9: 37-67. 1950. LandwirtsehaftlichePflanzensoziologie. Hansen, H. M0lholm. 1930. Studies on the vegetationof I. Unkrautgemeinschaftenals Zeiger fur Klima und Iceland. The botany of Iceland, ed. Rosenvinge & Boden. Stuttgart: Eugen Ulmer. 141 pp. Warming, 3 (10): 1-186. Copenhagen: Frimodt. 1952. LandwirtschaftlichePflanzensoziologie. Hansen, H. P. 1947. Postglacial forest succession,cli- II. Wiesen und Weiden und ihre standbrtlicheBewer- mate, and chronologyin the Pacific Northwest.Amer. tung. Stuttgart: Eugen Ulmer. 143 pp. Philos. Soc. Trans. N. S. 37(1): 1-130. Engelbrecht,H. H. 1955. The climatologyand ecology Hanson, H. C. 1953. Vegetation types in northwestern of the Pacific Coast. Natl. Shade Tree Conf. Proc. Alaska and comparisons with communitiesin other Ann. Meeting 31: 7-24. Arctic regions. Ecology 34: 111-140. Fenneman, N. M. 1931. Physiography of western Harshberger,J. W. 1911. Phytogeographicsurvey of United States. New York: McGraw-Hill. 534 pp. North America. (Germ. summ.) Die Vegetation der Fisher, R. A., A. S. Corbet, & C. B. Williams. 1943. Erde, ed. Engler & Drude 13: 1-790. Leipzig: Engel- The relation between the number of species and the mann; New York: Stechert. numberof individuals in a random sample of an ani- Hattori, S. 1955. Hepaticae occurringon serpentineon mal population. Jour. Anim. Ecol. 12: 42-58. Mt. Apoi, Hokkaido. Bot. Mag. [Tokyo] 68: 320-323 Forest Service. 1940. Map of Siskiyou National Forest, (not seen). Oregon and California. U. S. Dept. Agric., Forest Hayward, C. L. 1948. Biotic communitiesof the Wa- Service. satch Chaparral, Utah. Ecol. Monog. 18: 473-506. Gauckler,K. 1954. Serpentinvegetationin Nordbayern. Hershey, 0. H. 1900. Ancient alpine glaciers of the Bayer. Bot. Gesell. Ber. 30: 19-26 (fide Krause & Sierra Costa Mountains in California. Jour. Geol. 8: Ludwig 1957). 42-57. Gelting, P. 1934. Studies on the vascular plants of Hess, H. H. 1955. Serpentines,orogeny, and epeirog- East Greenland between Franz Joseph Fjord and eny. In Crust of the earth, ed. A. Poldervaart. Geol. Dove Bay (Lat. 73?15'-76?20' N). Meddel. om Soc. Amer. Spec. Paper 62: 391-407. Gronland 101(2): 1-337. Hopkins, B. 1955. The species-area relations of plant Gismondi,A. 1953. Intorno ad una particolare fito- communities. Jour. Ecol. 43: 409-426. cenosi comune al terreni serpentinosied al caleare Hughes, R. E. & D. V. Lindley. 1955. Application of dolomitico in, Liguria. Nota preliminare. Accad. biometric methods to problems of classification in Ligure Sci. e Lett. Atti 9: 240-241. ecology. Nature [London] 175: 806-807. Gleason, H. A. 1920. Some applications of the quadrat Hunter, J. G. & 0. Vergnano. 1952. Nickel toxicityin method. Torrey Bot. Club Bull. 47: 21-33. plants. Ann. Appl. Biol. 39: 279-284. . 1926. The individualisticconcept of the plant Huxley, J. S. 1943. Evolution: The modernsynthesis. association. Torrey Bot. Club Bull. 53: 7-26. New York: Harper. 645 pp. . 1939. The individualisticconcept of the plant Jaccard, P. 1902a. Gesetze der Pflanzenvertheilungiic association. Amer. Midland Nat. 21: 92-11(). der alpinen Region. Auf Grund statistiseb-floristise Goodall, D. W. 1954a. Objective methodsfor the classi- Untersuchungen. Flora 90: 349-377. ficationof vegetation. ITH. An essay in the use of - 1902b. Lois de distributionflorale dans le zone factor analysis. Austral. Jour. Bot. 2: 304-324. alpine. Soe. Vaud. des Sci. Nat. Bull. 38: 69-130. July, 1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 335

. 1908. Nouvelles recherchessur la distribution Math. et Nat., Bull. Internatl., S6r. B. 1927 (Suppl. florale.Soc. Vaud. des Sci. Nat. Bull. 44: 223-270. 2): 57-203. 1912. The distribution of the flora in the Leiberg, J. B. 1900. Cascade Range and Ashland forest alpine zone. New Phytol. 11: 35-50. reserves and adjacent regions. U. S. Geol. Survey, . 1932. Die statistisch-floristischeMethode als Ann. Rept. 21, 1899-1900, Part 5-Forest Reserves, Grundlage der Pflanzensoziologie. Handb. Biol. Ar- pp. 209-498. beitsmeth.ed. Abderhalden 11, 5(1): 165-202. Lippmaa, T. 1935. La methode des associations unis- Jensen, H. A. 1947. A systemfor classifyingvegeta- trates et le systame6cologique des associations. Inst. tion in California. Calif. Fish & Game 33: 199-266. et Horti Bot. Univ. Tartu. Acta 4(1/2, art. 3): 1-7. Jepson,W. L. 1923-5. A manual of the floweringplants . 1939. The unistratal concept of plant commu- of California. Berkeley: Unliv.Calif. Assoc. Students nities (the unions). Amer. Midland Nat. 21: 111-145. Store. 1238 pp. LUdi, W. 1937. Berichtfiber den Kurs in Alpenbotanik, veranstaltetin Davos vom 20.-29. Juni 1936. Geo- . 1935. Centersof plant endemismin California bot. Forschungsinst. Riibel in relation to geological history. Zesde Internatl. Bot. Zurich Ber. 1936: 14-35. Congr. AmsterdamProc. 2: 82-83. Lynch, D. 1955. Ecology of the aspen groveland in Glacier Montana. Ecol. 25: 321-344. Jones, G. N . 1936. A botanical surveyof the Olympic County, Monog. Peninsula, Washington. Univ. Wash. Pub. Biol. 5: MacArthur,R. H. 1957. On the relative abundance of 1-286. bird species. Natl. Acad. Sci. Wash. Proc. 43: 293- 295. * 1938. The floweringplants and ferns of Mount Rainier. Univ. Wash. Pub. Biol. 7: 1-192. McMillan, C. 1956. The edaphic restrictionof Cupres- sus and Pinus in the Coast Ranges of central Cali- Kendrew, W. G. 1937. The climates of the continents. fornia. Ecol. Monog. 26: 177-212. 3rd ed. Oxford: Clarendon. 473 pp. McMinn, H. E. 1939. An illustrated manual of Cali- Klyver, F. D. 1931. Major plant communitiesin a fornia shrubs. San Francisco: Stacey. 689 pp. transectof the Sierra Nevada Mountains of California. Ecology 12: 1-17. McMinn, H. E. & Evelyn Maino. 1937. An illustrated manual of Pacific Coast trees. 2nd ed. Berkeley: Univ. Knaben, Gunvor. 1952. Botanical investigationsin the Calif. Press. 409 pp. middle districts of western Norway. Bergen Univ. Aarbok, Naturv. rekke 1950(8): 1-117. Major, J. 1951. A functional, factorial approach to plant ecology. Ecology 32: 392-412. Koch, L. F. 1957. Index of biotal dispersity. Ecology 38: 145-148. Margalef, R. 1949. Una aplicaci6n de las series loga- ritmicasa la fitosociologia. (Engl. summ.) Inst. Biol. Koppen, W. 1900. Versuch einer Klassifikation der Aplicada Barcelona Pub. 6: 59-72. Klimate, vorzugweise nach ihren Beziehungen zur Pflanzenwelt. Geog. Ztschr. 6: 593-611. . 1951. Diversidad de especies en las comuni- dades naturales. (Engl. summ.) Inst. Biol. Aplicada . 1923. Die Klimate der Erde. Grundriss der Barcelona Pub. 9: 5-27. Klimakunde. Berlin & Leipzig: de Gruyter. 369 pp. 1958. Informationtheory in ecology. General Kotilainen, M. J. & 0. Seivala. 1954. Observationson Systems 3: 36-71. Transl. from R. Acad. Ci. Art. serpentine flora at Sunnmore. A serpentinicolous Barcelona Mem. 23: 373-449, 1957. form,Cerastium alpinum L. var nordhageniiKotil. & Martin,W. E., J. Vlamis, & N. W. Stice. 1953. Field Seivala, var. nova. Nytt Mag. Bot. 3: 139-145. correctionof calcium deficiencyon a serpentinesoil. Krause, W. 1958. Andere Bodenspezialisteii. In Hand- Agron. Jour. 45: 204-208. buch der Pflanzenphysiologie,ed. W. Ruhland, 4: 755- Mason, H. L. 1927. Fossil recordsof some West Ameri- 806. Berlin: Springer. can conifers. Carnegie Inst. Wash. Pub. 346: 139- Krause, W. & 0. Klement. 1958. Zur Kenntnis der 158. Flora und Vegetation auf Serpentinstandortendes . 1946a. The edaphic factor in narrowendemism. Balkans. 3. Felsflechten-Gesellsehaftenim Gostovic- I. The nature of environmentalinfluences. Madrofio8: Gebiet (Bosnien) und Zlatibor-Gebirge (Serbien). 209-226. Vegetatio 8: 1-19. . 1946b. The edaphic factor in narrow endem- Krause, W. & W. Ludwig. 1956. Zur Kenntnis der ism. II. The geographic occurrence of plants of Flora und Vegetation auf Serpentinstandortendes highly restricted patterns of distribution. Madrouio Balkans. 1. Halacsya sendtneri (Boiss.) Dorfl. 8: 241-257. Deutsch. Bot. Gesell. Ber. 69: 417-428. Maxson, J. H. 1933. Economic geology of portions of * 1957. Zur Kenntnis der Flora und Vegetation Del Norte and Siskiyou Counties, northwesternmost auf Serpentinstandortendes Balkans. 2. Pflanzenges- California. Calif. Jour. Mines & Geol. 29: 123-160. ellsehaften und Standorte im Gostovic-Gebiet (Bos- Merkle,J. 1951. An analysis of the-plant communities nien). Flora 145: 78-131. of Mary's Peak, western Oregon. Ecology 32: 618- Kruckeberg,A. R. 1951. Intraspecificvariability in the 640. response of certain native plant species to serpentine Merriam,C. H. 1899. Results of a biological surveyof soil. Amer. Jour. Bot. 38: 408-419. Mount Shasta, northern California. North Amer. 1954. The ecology of serpentine soils. III. Fauna 16: 1-179. Plant species in relation to serpentinesoils. Ecology Meusel, H. 1939. Die Vegetationsverhailtnisseder 35: 267-274. Gipsberge im Kyffhduserund im sUdlichenHarzvor- Kulczyfiski, S. 1928. Die Pflanzenassoziationen der land. Ein Beitrag zur Steppenheidefrage. Hercyiiia Pieninen. Acad. Polon. des Sci. et Lett., Cl. de Sci. 2: 1-372. 336 CR.H. WHITTAKER E ogc MoNOgah Vol. 30, No. 3 1943. Vergleichende Arealkunde. Berlin: Pichi-Sermolli,R. 1936. Osservazioni sulle principali Borntraeger. 2 vols., 466 & 180 pp. (fide Walter morfosidelle piante del serpentino. Nuovo Gior. Bot. 1954). Ital. N. S. 43: 461-474. Miller, Harriet C. E. & M. F. Buell. 1956. Life-form ---. 1948. Flore e vegetazione delle serpentine e spectra of contrastingslopes in Itaska Park, Minne- delle altre ofioliti dell 'Alta Valle del Tevere (Tos- sota. Bot. Gaz. 117: 259-263. cana). (Engl. summ.) Webbia 6: 1-380. Minguzzi, C. & 0. Vergnano. 1953. Il contenuto di Preston, F. W. 1948. The commonness,and rarity,of elementi inorganici delle piante della formazione species. Ecology 29: 254-283. ofiolitica dell 'Impruneta (Firenze). (Engl. summ.) Ramensky,L. G. 1924. Die Grundgesetzmiissigkeiten Nuovo Gior. Bot. Ital. N. S. 60: 287-319. im Aufbau der Vegetationsdecke. (Russian) Wjestn. Munger,T. T., W. E. Lawrence, & H. M. Wight. 1926. opytn. djela Woronesch.37 pp. (Bot. Centbl. N. F. 7: Oregon. In Naturalist's guide to the Americas, ed. 453-455, 1926). V. E. Shelford et al., pp. 181-193. Baltimore: Wil- 1930. Zur Methodik liams & Wilkins. der vergleichendenBear- beitung und Ordnung von Pflanzenlistenund andere Munz, P. A. & D. D. Keck. 1949. California plant Objekten, die durch mehrere,verschiedenartig wirk- communities. El Aliso 2: 87-105. ende Faktoren bestimmtwerden. Beitr. z. Biol. der 1950. California plant communities-supple- Pflanz. 18: 269-304. ment. El Aliso 2: 199-202. Raunkiaer, C. 1918. tber das biologische Normalspek- . 1959. A California flora. Berkeley & Los An- trum. K. Danske Vidensk. Selsk. Biol. Meddel. 1(4): geles: Univ. Calif. Press. 1681 pp. 1-17. Nordhagen,R. 1937. Versuch einer neuen Einteilung . 1934. The life forms of plants and statistical der subalpinen-alpineniVegetation Norwegeus. Ber- plant geography; being the collected papers of C. gens Mus. Aarbok Naturv. rekke 1936(7): 1-88. Raunkiaer, translated into English by H. G. Carter, . 1943. Sikilsdalen og Norges Fjellbeiter. En A. G. Tansley, and Miss Fausboll. Oxford: Clarendon. plantesosiologiskmonografi. Bergens Mus. Skr. 22: 632 pp. 1-607. Renkonen,0. 1938. Statistisch-okologischeUntersuch- Nov&k, F. A. 1928. Quelques remarques relatives au ungen uiberdie terrestrischeKiiferwelt der finnischen probldmede la vegetation sur les terrains serpentin- Bruchmoore. (Finnish summ.) Ann. Zool., Soc. Zool.- iques. Preslia 6: 42-71. Bot. Fenn. Vanamo 6(1): 1-231. Nytzenko, A. A. 1948. Boundaries of plant associa- Ritter-Studnicka,Hilda. 1953. Das Calluneto-Ericetunm tions in nature. (Russian) Bot. Zhurnal S.S.S.R. 33: in Bosnien. osterr. Bot. Ztschr. 100: 80-90. 487-495 (Biol. Abstr. 24: 17407, 1950). . 1956. Beitrag zur oikologieder Serpentinflora Oberlander,G. T. 1956. Summer fog precipitationon in Bosnien. Vegetatio 7: 89-98. the San Francisco Peninsula. Ecology 37: 851-852. Rousseau, J. 1952. Les zones biologiques de la penin- Odum, E. P. 1950. Bird populations of the Highlands sule Quebec-Labradoret l'hemiarctique. Canad. Jour. (North Carolina) Plateau in relation to plant succes- Bot. 30: 436-474. sion and avian invasion. Ecology 31: 587-605. Rowe, J. S. 1956. Uses of undergrowthplant species in Oosting, H. J. 1942. An ecological analysis of the forestry. Ecology 37: 461-473. plant communitiesof Piedmont,North Carolina. Amer. Ruibel, E. 1927. Einige skandinavische Vegetations- Midland Nat. 28: 1-126. probleme. Geobot. Inst. Ruibel Zurich Ver6ffentl.4: . 1948. The study of plant communities:An in- 19-41. troductionto plant ecology. San Francisco: Freeman. Rune, 0. 1953. Plant life on serpentinesand related 440 389 pp. 2nd ed. 1956, pp. rocks in the North of Sweden. Acta Phytogeog. Sue- Oosting,H. J. & W. D. Billings. 1943. The red firfor- cica 31: 1-139. est of the Sierra Nevada: Abietum magnificae. Ecol. - . 1954a. Notes on the flora of the Penin- Monog. 13: 259-274. Gaspe sula. Svensk Bot. Tidskr. 48: 117-136. Patrick, Ruth, M. H. Hohn, & J. H. Wallace. 1954. A new methodfor determiningthe pattern of the diatom . 1954b. Arenaria humifusa on serpentine in flora. Notulae Nat. Acad. Nat. Sci. Philadelphia 259: .Scandinavia. Nytt Mag. Bot. 3: 183-196. 1-12. . 1954c. Soil requirementsof arctic and alpine Patton, C. P. 1956. Climatologyof summerfogs in the plants as indicated by their occurrenceon ultrabasic rocks. San Francisco Bay area. Univ. Calif. Pub. Geogr. 10: VIIIe. Congr. Internatl. Bot. Paris 1954. Rap- 113-200. ports et Commun.Sect 7: 161-162. Pavlovi6, Zagorka. 1953. Prilog poznavanju serpen- Russell, R. J. 1926. Climates of California. Univ. tinskeflore Ozren planine kod Sjenice. (French summ.) Calif. Pub. Geogr. 2: 73-84. Mus. d'Hist. Nat. du Pays Serbe Bull. Ser. B 5-6: 3-19. Schmid, E. 1950. Zur Vegetationsanalysenumidischer -. 1955. Prilog poznavanju serpentinskeflore i Eichenwiilder. Geobot. Forschungsinst.Ruibel Zurich vegetacije Ozrena kod Sjenice (II). (French summ.) Ber. 1949: 23-39. Mus. d'Hist. Nat. du Pays Serbe Bull. Sdr B 7: 1-45. Sharsmith,Helen K. 1945. Flora of the Mount Ham- Peck, M. E. 1925. A preliminarysketch of the plant ilton Range of California. (A taxonomnicstudy and regions of Oregon. I. Western Oregon. Amer. Jour. floristicanalysis of the vascular plants). Amer. Mid- Bot. 12: 33-49. land Nat. 34: 289-367. 1941. A manual of the higherplants of Oregon. Shenon, P. J. 1933. Geology and ore deposits of the Portland: Binfords & Mort. 866 pp. Takilma-Waldo district, Oregon, including the Blue July,1960 VEGETATION OF THE SISKIYOU MOUNTAINS, OREGON AND CALIFORNIA 337

Creek District. U. S. Geol. Survey Bull. 846(B) New York: McGraw-Hill. 520 pp. 2nd ed. 1938, 601 141-194. Pp. Shreve,F. 1927. The vegetation of a coastal mountain Wells, E. L. 1936. Climatic summaryof the United range. Ecology 8: 27-44. States. Section 3-Western Oregon. U. S. Dept. Simpson, E. H. 1949. Measurementof diversity. Na- Agric., Weather Bureau, Washington. 49 pp. ture [London] 163: 688. 1941. Climate of Oregon. In Climate and Smith,E. E. 1954. The forestsof Cuba. Maria Moors man, U. S. Dept. Agric. Yrbk. 1941: 1075-1086. Cabot Found. Pub. 2: 1-98. Wells, F. G. 1939. Preliminarygeologic map of the Socava, V. 1927. Botanische Beschreibungder Wdlder Medford quadrangle, Oregon. Oregon Dept. of Geol. des Polar-Urals vom Flusse Nelka bis zum Flusse & Mineral Indust. Hulga. (Russ. with Germ. summ.) Trav. Mus. Bot. . 1940. Preliminarygeologic map of the Grants Acad. Sci. URSS 21: 1-78. Pass quadrangle, Oregon. Oregon Dept. of Geol. & So6, R. von. 1934. Vasmegye szociol6giae es fLorisz- Mineral Indust. tikai nbvenyfldrajzhhoz. (Germ. summ.) Vasi Wells, F. G., F. W. Cater, Jr., & G. A. Rynearson. 1946. Szemle, Folia Sabariensia (Dunntfili Szemle) 1: 105- Chromite deposits of Del Norte County, California. 134. Calif. Div. Mines Bull. 134(Pt. 1, Ch. 1): 1-76. Spence, D. H. N. 1957. Studies on the vegetation of Wells, F. G., P. E. Hotz, & F. W. Cater, Jr. 1949. Pre- Shetland. I. The serpentinedebris vegetationin Unst. liminarydescription of the geology of the Kerby quad- Jour. Ecol. 45: 917-945. rangle, Oregon. Oregon Dept. Geol. & Mineral In- Sprague, M. 1941. Climate of California. In Climate dust. Bull. 40: 1-23. and man, U. S. Dept. Agric. Yrbk. 1941: 783-797. Wells, F. G., L. R. Page, & H. L. James. 1940. Chro- Steele, B. 1955. Soil pH and base status as factors in mite deposits in the Sourdough area, Curry County, the distributionof calcicoles. Jour. Ecol. 43: 120-132. and the Briggs Creek area, JosephineCounty, Oregon. U. S. Geol. Survey Bull. 922(P): 461-496. Tadros, T. M.' 1957. Evidence of the presence of an Wells, F. G. & G. W. Walker. 1953. Geology edapho-bioticfactor in the problem of serpentinetol- of the Galice quadrangle, Oregon. U. S. Geol. Survey, Geol. erance. Ecology 38: 14-23. Quad. Maps. Taniguti, M. 1958. Phytosociologicalstudy of serpen- West, W. 1912. Notes on the flora of Shetland, with tine area of Mt. Asama, Mie Pref. (Japan. with Engl. some ecological observations. Jour. Bot. [London] summ.) Acta Phytotax. Geobot. 17: 122-127. 50: 265-275, 297-306. Thornthwaite,C. W. 1931. The climates of North Whitford,P. B. 1949. Distributionof woodland plants America according to a new classification. Geog. Rev. in relation to succession and clonal growth. Ecology 21: 633-655. 30: 199-208. . 1948. An approach toward a rational classifi- Whittaker, R. H. 1951. A criticism of the plant as- cation of climate. Geog. Rev. 38: 55-94. sociation and climatic climax concepts. Northw. Sci. Vergnano,0. 1953a. L'azione fisiologicadel nichelsulle 25: 17-31. piante di un terreno serpentinoso. (Engl. summ.) 1952. A study of summerfoliage insect com- Nuovo Gior. Bot. Ital. N. S. 60: 109-183. munitiesin the Great SmokyMountains. Ecol. Monog. - . 1953b. Effetti di alte concentrazionidi nichel 22: 1-44. su alcune piante da cultura. (Engl. summ.) Nuovo . 1953. A considerationof climax theory: The Gior. Bot. Ital. N. S. 60: 189-196. climax as a population and pattern. Ecol. Monog. 23: Vilhelm, J. 1925. Serpentinomorfosybryophyt z had- 41-78. covychskal u MariAnskychLazni v 6echAch. (French 1954a. The ecology of serpentine soils. I. summ.) Fac. des Sci. de l'Univ. Charles Prague Pub. Introduction. Ecology 35: 258-259. 33: 1-16. . 1954b. The ecology of serpentinesoils. IV. 1954. atlas of United Visher,S. S. Climatic the States. The vegetational responseto serpentinesoils. Ecology Cambridge,Mass.: Harvard Univ. Press. 403 pp. 35: 275-288. Walker, R. B. 1954. The ecology of serpentinesoils. 1954c. Plant populations and the basis of II. Factors affectingplant growthon serpentinesoils. plant indication. (Germ. summ.) Angew. Pflanzensoz. Ecology 35: 259-266. [Wien], Festschr. Aichinger 1: 183-206. Walker, R. B., Helen M. Walker,& P. R. Ashworth.1955. 1956. Vegetation of the Great Smoky Moun- Calcium-magnesiumnutrition with special referenceto tains. Ecol. Monog. 26: 1-80. serpentinesoils. Plant Physiol. 30: 214-221. . 1957. Recent evolution of ecological concepts Walter, H. 1954. Einfiihrungin die Phytologie. III. in relation to the eastern forests of North America. Grundlagender Pflanzenverbreitung.II. Teil.: Areal- Amer. Jour. Bot. 44: 197-206. kunde. Stuttgart: Eugen Ulmer. 246 pp. Whittaker,R. H. & C. W. Fairbanks. 1958. A study Walter, H. & E. Walter. 1953. Einige allegemeine of plankton copepod communitiesin the Columbia Ergebnisse unserer Forschungsreisenach Suidwestaf- Basin, southeasternWashington. Ecology 39: 46-65. rika 1952/53: Das Gesetz der relativen Standortskon- Williams, C. B. 1945. Recent light trap catches of stanz; das Wesen der Pflanzengemeinschaften.Deutsch. Lepidoptera in U. S. A. analyzed in relation to the Bot. Gesell. Ber. 66: 228-236. logarithmicseries and the index of diversity. Entom. Ward, R. D. 1925. The climates of the United States. Soc. Amer. Ann. 38: 357-364. Boston: Ginn. 518 pp. . 1947. The logarithmicseries and its applica- Weaver, J. E. & F. E. Clements. 1929. Plant ecology. tion to biological problems. Jour. Ecol. 34: 253-272. 338 WHITTAKER EcologicalMonographs R.R. H.WHITTA:KER H. ~~~~~~~~~~Vol.30,No. 3 . 1950. The application of the logarithmicseries raga, K6chi Pref. (Japan. with Engl. summ.) Fac. of to the frequency of occurrence of plant species in Educ. K6chi Univ. Bull. 5: 47-53. quadrats. Jour. Ecol. 38: 107-138. 1955. Sociological studies on the serpentine . 1951. Diversity as a measurable character of vegetation. IV. The vegetation on Mt. Higashi- an animal or plant population. Annee Biol. Ser. 3, akaishi, Ehime Pref. (Japan. with Engl. summ.) Fac. 27: 129-141. of Educ. KUMhiUniv. Bull. 8: 49-57. . 1953. The relative abundance of differentspe- 1956. Sociological studies on the serpentine cies in a wild animal population. Jour. Anim. Eeol. 22: 14-31. vegetation. V. Namekawa and Yusuhara in Kochi Prefecture. (Japan. with Engl. summ.) Fac. of Educ. 1954. The statistical outlook in relation to K6chi Univ. Bull. 9: 35-39. ecology. Jour. Ecol. 42: 1-13. Williams, R. F. 1933. An ecological analysis of the 1957. On the Chamaecyparis obtusa forest in plant communitiesof the "Jarrah" Region occurring Shikoku, southernJapan. (Japan. with Engl. summ.) on a small area near Darlington,with special refer- Japan. Jour. Ecol. 6: 149-152. ence to the indicator value of species. Roy. Soc. West. Yount, J. L. 1956. Factors that control species num- Austral. Jour. 1931-2, 18: 105-124. bers in Silver Springs, Florida. Limnnol.& Oceanogr. Withrow, Alice P. 1932. Life forms and leaf size 1: 286-295. classes of certain plant communitiesof the Cincinnati Z6lyomi, B. 1936. tubersichtder Felsenvegetation in region. Ecology 8: 12-35. der pannonischenFlorenprovinz und dem nordwestlich Yamanaka, T. 1954. Sociological studies on the ser- angrenzendenGebiete. (Magyar summ.) Mus. Natl. pentine vegetation. III. The vegetation on Mt. Shi- Hungar. Ann. 30(Pars Bot.): 136-174. October, 1960 Peromyscus leucopus noveboracensis AND P. maniculatus gracilis 407

Lack, D. 1933. Habitat selection in birds with special Petersen, C. G. J. 1896. The yearly immigration of reference to the effects of afforestation on the Breck- young plaice into the Limfjord from the German land avifauna. Jour. Anim. Ecol. 2: 239-262. Sea, etc. Rept. Danish Biol. Sta. for 1895, 6: 1-48. 239-262. Priddy, R. B. 1942. Population levels of small mam- . 1945. The ecology of closely related species mals in various habitats of central New York. Un- with special reference to cormorant (Phalacrocorax published thesis, Cornell Uaiv. carbo) and shag (P. aristotelis). Jour Anim. Ecol. Rhoads, S. N. 1903. The mammals of Pennsylvania 14) 12-16. and New Jersey. Philadelphia: privately published. Lindeborg, R. G. 1952. Water requirements of certain 266 pp. rodents from xeric and mesic habitats. Contrib. Lab. Shelford, V. E. 1913. The reactions of certain animals Vert. Biol. Univ. Mich. 58: 32 pp. to gradients of evaporating power of air. Biol. Bull. Manville, R. H. 1949. A study of small mammal pop- 25: 79-120. ulations in northern Michigan. Univ. Mich. Mus. . 1945. The relative merits of the life zone and Zool. Misc. Pub. 73: 83 pp. biome concepts. Wilson Bull. 57: 248-252. Merriam, C. H. 1894. Laws of temperature control of Spaeth, J. N. & C. H. Diebold. 1938. Some inter- the geographic distribution of terrestrial animals and relationships between soil characteristics, water ta- plants. Natl. Geog. Mag. 62: 229-238. bles, soil temperature, and snow cover in the forest and adjacent open areas in south-central New York. Odum, E. P. 1944. Water consumption of certain mice Cornell Univ. Agr. Exp. Sta. Mem. 213. 76 pp. in relation to habitat selection. Jour. Mammal. 25: 404-405. Stickel, L. F. 1946. Experimental analysis of methods for measuring small mammal populations. Jour. --. 1949. Small mammals of the Highlands (North Wildlife Mangt. 10: 150-159. Carolina) plateau. Jour. Mammal. 30: 179-192. . 1954. A comparison of certain methods of . 1953. Fundamentals of ecology. Philadelphia: measuring ranges of small mammals. Jour. Mammal. Saunders. xii + 384 pp. 35: 1-15. Oosting, H. J. & D. W. Hess. 1956. Microclimate and Stinson, R. H. & K. C. Fisher. 1953. Temperature se- a relic stand of Tsuga canadensis in the lower Pied- lection in deer mice. Canadian Jour. Zool. 31: 404- mont of North Carolina. Ecology 37: 28-39. 416. Osgood, F. L. 1938. The mammals of Vermont. Jour. Swanson, G. A., T. Surber & T. S. Roberts. 1945. The Mammal. 19: 435-441. mammals of Minnesota. Minn. Dept. Cons. Tech. Osgood, W. H. 1909. Revision of the mice of the Bull. 2. 108 pp. American genus Peromyscus. N. Amer. Fauna 28: Weaver, J. E. & F. E. Clements. 1938. Plant ecology. 285 pp. New York: McGraw-Hill. xxii + 601 pp.

ERRATA Two errors were introduced into the paper on 2) Page 321. The equation in the second para- "Vegetation of the Siskiyou Mountains, Oregon and should = California" by R. H. Whittaker by incorrect entry graph read, P - of galley proof corrections: A note from the author on uncertain taxonomic 1) Pages 289-290. The paragraph beginning status of the population designated Arctostaphylos "These results. . ." on page 290 should follow the first cinerea on gabbro was not received in time for in- paragraph on page 289 ending ".... on serpentine clusion. -34.9, 32.2, 34.7" and precede the section on "Tran- sect Tables."