ACTA PHYTOGEOGRAPHICA SUECICA 85 EDIDIT SVENSKA VAXTGEOGRAFISKA SALLSKAPET

Succession and zonation on mountains, particularly on volcanoes

- Dedicated to Erik Sjogren on his 65th birthday -

Edited by Eddy van der Maarel

UPPSALA 2000 2

ISBN 91-7210-085-0 (paperback) ISBN 91-7210-485-6 (cloth) ISSN 0084-5914

Editor: Erik Sjogren Guest editor: Eddy van der Maarel

Technical editor: Marijke van der Maarel-Versluys

© Respective author 2000

Edidit: Svenska VaxtgeografiskaSallskapet

Villavagen 14 SE-752 36 Uppsala

Lay-out: Opulus Press AB, Uppsala Printed in Sweden 2001 by Fingraf AB, Sodertalje.

Acta Phytogeogr. Suec. 85 3

Contents

Erik Sjogren 65 years 5 E. Rosen

Main types of vegetation zonation on the mountains of the Caucasus 7 N. Zazanashvili, R. Gagnidze & G. Nakhutsrishvili

Zonation and management of mountain forests in the Sierra de Mananthin, Mexico 17 M. Olvera Va rgas, B. LorenaFig ueroa-Rangel & F. Bongers

The distribution of rare plant species on Mount Prado, a northern Apennine diversity hot spot 23

C. Ferrari, G. Pezzi & A. Portanova

Succession and zonation of vegetation in the volcanic mountains of the Hawaiian Islands 31 D. Mueller-Dombois

Geographical determinants of the biological richness in the Macaronesian region 41

J.M. Ferndndez-Palacios & C. Andersson

Succession and local species turnover on Mount St. Helens, Washington 51 R. del Moral

Primary succession on lava flows on Mt. Etna 61

E. Poli Ma rchese & M. Grillo

Vegetation dynamics on Paricutin, a recent Mexican volcano 71 A. Veldzquez, J. Gimenez de Azcarate, M.E. We inmann, G. Bocco & E. van der Maarel

Hookeria lucens, a new record for the moss flora of Iceland - with some remarks on recent lava flows 79 A.H. Bjarnason

Acta Phytogeogr. Suec. 85 4

Above-right: Erik Sjogren at the 41st lA VS Symposium in Uppsala, 1998. Above-left: Misterio da Prainha, Island of Pica, Azores, 1995. Bottom: Madeira, 1971. Photo above-left; Christer Mattisson; others Eje Rosen.

Acta Phytogeogr. Suec. 85 Erik Sjogren 65 years

E. Rosen

Department of Plant Ecology, Evolutionary Biology Centre, Uppsala University,

Villavagen 14, SE-752 36 Uppsala, Sweden; Fax +4618 553419

This volume is dedicated to Docent Erik Sjogren at the epiphytic moss communities in the same woodlands on occasion of his 65th birthday. The idea to collect just these bland. In 1964 he visited Madeira and in 1965 the papers was born during the preparation of the 41th Sym­ Azores. From that time onwards field work has been posium of the International Association for Vegetation carried out during several periods on these volcanic Science held in July 1998. Most of the contributions are islands of Macaronesia. based on presentations on the theme 'Succession and As early as 1965 he initiated plans to preserve the zonation on mountains, in particular vulcanos', one of native vegetation of the Azores. In 1970 the protection Erik' s favourite topics. against exploitation of any kind of vegetation at altitudes Erik Sjogren was born in Kalmar, the old Swedish above 600 m was recommended. This recommendation coastal town in the western part of the Baltic Sea Region. was acknowledged by the Government in the 1980s. In Together with his parents he frequently visited the island the framework of these conservation studies several en­ of bland. Ever since he has been devoted to islands and demic Azorean plants and their habitats have been listed their vegetation. He studied Biology at Uppsala University as highly rare and in danger of extinction. Another aspect and became a staff member of the Department of Plant of this work was the use of rare bryophytes with narrow Ecology, then called Department of Ecological Botany ecological ranges as indicators of habitats suitable for the (Vaxtbiologiska institutionen). His doctoral thesis ( 1964) development of native forest vegetation with Fmmerous dealt with epilithic end epigeic moss vegetation in decidu­ endemic vascular plants. Most of these bryophytes are ous woodlands on bland. Prior to his dissertation he had hepatics which are preferentially epiphyllous. More than already published (1961) a comprehensive study on 350 bryophytes were recorded in the Azores between 1965 and 1997; several records were new to particular islands, and some were even new to the Azores or to Macaronesia as a whole. Acta Phytogeogr. Suec. 85: 5-6. ISBN 91-7210-085-0.

Acta Phytogeogr. Suec. 85 6 E. Rosen

In 1965 Sjogren also started a sociological treatment students at Uppsala University and has been widely ac­ of the vegetation of Madeira and the Azores, after he had knowledged and appreciated for his wide knowledge and discovered that no such approach had been carried out for his active participation in the social life at the Depart­ before. In the establishment of the ecological ranges of ment of Plant Ecology, as Santaclaus, violonist and enter­ species and plant communities he used the climatic differ­ tainer. After his retirement Erik is still devoted to his ences on Madeira between the north and south side of the research and we are happy to see him around at the island. In the Azores there is both climatic variation Department. between the islands from east to west and altitudinal variation on each individual island. He also initiated the Major publications by Erik Sjogren on application of the age differences between the lava flows bryoecology, bryosociology and Macaro­ in order to suggest the time and direction of successional nesian vegetation phenomena. On this basis Sjogren investigated what he later called the 'telescopic behaviour of successional Sj ogren, E. 1961. Epiphytische Moosvegetation in Laubwaldern stages'. This was found to be especially typical for der Inselb1and (Schweden).- Acta Phytogeogr. Suec. 44: bryophyte communities; it is a feature of the development 1- 149. of such communities; it also. enables ·us to record the Sj ogren, E. 1964. Epilithische und epigaische Moosvegetation development at localities offering suboptimal to optimal in Laubwaldern der Insel bland (Schweden). - Acta habitat conditions. The different types of coexistence of Phytogeogr. Suec. 48: 1-184. successional stages were recorded as especially typical Sj ogren, E. 1972. Vascular plant communities of Madeira. - for epiphyllous groups of bryophytes. Bol. Mus. Municip. Funcha1 16: 45-125. From 1965 until now Erik Sjogren has published 23 Sj ogren, E. 1973. Recent changes in the vascular flora and vegetation of the Azores Islands. - Mem. Soc. Broteriana papers on the ecology and sociology of bryophytes and 22: 1-453. vascular plants of the Azores and Madeira. In several of Sj ogren, E. 1975. Epiphyllous bryophytes of Madeira.- Sven. these papers suggestions on the preservation of species Bot. Tidsk. 69: 217-288. and communities were included. Comprehensive reports Sj ogren, E. 1978. Bryophyte vegetation in the Azores Islands. on the detailed studies of the exceptional bryophyte veg­ - Mem. Soc. Broteriana 26: 1-283. etation of Madeira and the Azores were presented in 1975, Sjogren, E. 1993. Dry coastal ecosystems of Madeira and the 1978 and in 1997. Additional studies were performed in Azores. In E. van der Maarel (ed.) Dry Coastal Ecosystems, the Canary Islands, Portugal and Iceland. From the latter Vol. 2B- Elsevier, Amsterdam. the newly emerged island of Surtsey off the south coast of Sjogren, E. 1995. Changes in the epilithic and epiphytic moss Iceland caught his particular interest. cover in two deciduous forest areas on the Island of bland In the 1980s he activated a group of botanists at the (Sweden) - a comparison between 1958-1962 and 1988- 1990.- Stud. Plant Ecol. 19: 1-106. Universidade dos A\=ores to study the bryophyte vegeta­ Sjogren, E. 1997. Epiphyllous bryophytes in the Azores Islands. tion of the islands. A special bryophyte course was given in - Arquipelago Life Mar. Sci. 15 A: 1-49. Angra de Heroismo in 1992. In order to promote know­ Sj ogren, E. 2001. Azorean bryophyte communities- a revision of ledge on Macaronesian plants and vegetation he was in­ differential species.- Arquipelago Life Mar. Sci. (Submit­ volved in suggestions on the emission of stamps featuring ted.) endemic vascular plants of the Azores - which was later Sjogren, E. In press. Distribution of bryophytes in the Azores realized. Erikhas also organized excursions to the Azores Islands (- 1999) - with information on their presence on and to Madeira for the Swedish Phytogeographical Society, Madeira, inthe Canary Islands and in the world (for hepatics a society he has served as a Presidentfrom 198 1 until 2000, with the collaboration of R. Schumacker). - Bol. Mus. and as editor-in-chief of its journalActa Phytogeographica Municip. Funchal. Suecica since 1978. He also led several excursions for Ph. D. students at the Uppsala University Department of Plant Ecology. During the 1990s Sj ogren has also been Photo on p. 5 by Christer Mattisson, 1995. Locality on bottom of involved on a part-time basis in the updating of the Uni­ volcanic crater in the Misterio da Prainha, Island of Pico, Azores. versity's Bryophyte Herbarium. During two periods, 1958-1962 and 1988-1990, Erik Sj ogren has recorded the long-termdynamics and changes of epiphytic and epilithic moss communities in permanent plots in two deciduous woodlands on bland. Clear trends of increase in acidicline species were found in this unique study. Erik Sjogren has been involved in the teaching of ecology, biogeography and bryology for generations of

Acta Phytogeogr. Suec. 85 Main types of vegetation zonation on the mountains of the Caucasus

N. Zazanashvili, R. Gagnidze & G. Nakhutsrishvili

Institute of Botany of the Georgian Academy of Sciences, WWF Country Office Georgia, Alexidze St. 11 , +995 380093 Tbilisi, Georgia; Fax 32 330190; E-mail nzazanashvili@ wwfgeo.org.ge

Abstract Introduction

Typological diversity of vegetation zonation on the moun­ The Caucasus region covers an area of ea. 500 000 km2 in tains of the Caucasus is defined frrst of all by ( 1) the Armenia, Azerbaydzhan and Georgia, the North Cauca­ geographic transitional position of the region between sian portion of the Russian Federation, NETurkey, and a temperate deciduous broad-leaved forests and subtropical small part of NW Iran. As to its origin, the Caucasus latitudinal zones; (2) the location of different phytogeo­ isthmus forms part of the huge mountain belt, the Alpine graphical provinces (Mediterranean, Minor Asian, Ira­ Orogene, which embraces the whole of Eurasia from the nian) in the contact area; (3) the evolutionary history of Pyrenees and Atlas Mountains in the west to the Malay the Caucasian native flora (during the ice ages there were Peninsula and Vietnam in the east. The Caucasus region two refugia of the Tertiary flora in the region). Basically, was transformed into the present high mountain structure there are four types of vegetation zonation on the Cauca­ during the Neogene. sus mountains: West Caucasian (Colchic), East Cauca­ The Caucasus is a region of natural contrasts which sian, South Caucasian (Front Asian) and Southeast Cau­ includes several prominent elements (Fig. 1). These in­ casian (Hyrkanic). clude the North Caucasian Plain (the easternpart of which is below sea level), the Great Caucasus Range (with Mt. El'brus as the highest peak at 5642 m), the Transcaucasian Keywords: Altitudinal profile; Refuge; Vegetation zone. (South Caucasian) Depression, theSmall Caucasus moun­ tain chain (to 3500 m), and Transcaucasian Uplands (with Mt. Ararat as the highest point at 5165 m). There is a Nomenclature: Cherepanov (1995). considerable relief, with erosion-tectonic and accumula­ tion forms bordered by volcanic, glacier and karst forms. Not surprisingly, the climate is also variable. Mean

Abbreviations: ZAP = Zonation-altitudinal profile. annual precipitation in its southwestern (South Colkhic) part exceeds 2000 mm in the coastal area of the Black Sea (to 4000-4500 mm on the 1300-1400 m); in con­ trast, in the southwestern (lowland) part of the Caspian Sea coast, it rarely exceeds 150 mm. Mean annual air temperature in the Transcaucasian part of the Black Sea coast and Caspian Sea coast is 15 °C declining from south to north and with increasing altitude (with average fall 0.65 oc per 100 m). The flora and vegetation of the Caucasus are very diverse, and they depend on both the physical features discussed above and the evolutionary history of the eco­ systems. First of all, there are two refugia of the Tertiary flora in the region, the Colchic in the catchment basin of the Black Sea and the Hyrkanic in the extreme southeast­ em end of the Caucasus. Even now, many relict forms still appear as dominant or eo-dominant in a number of plant

Acta Phytogeogr. Suec. 85:7- 16. ISBN 91-72 10-085-0. communities.

Acta Phytogeogr. Suec. 85 8 N. Zazanashviliet al.

Fig. 1. Main types of mountain zo­ nation in the Causasus. I. West­

Caucasian (Colchic); 11. East Cau­ casian: Ill.South Causasian (Front Asian); IV. Southeast Caucasian (Hyrkanic).

In terms of vascular plant diversity, over 6300 species and Colchic partsof Georgia (N akhutsrishvili1999). Some have been recorded in the Caucasus, and of these, more publications contain generalized schemes (Stanukovich than 1600 are endemic. In addition, there are 17 endemic 1955, 1973; Gagnidze 1970). genera in the Caucasus. In the presented paper the experience of a new gener­ It is noteworthy that the high biodiversity of the Cau­ alization of mountain zonation in the Caucasus is used. It casus is under strong human impact: the Caucasus is is hoped that a discussion on the phytogeographical posi­ identified as one of the 25 biologically richest and also tion of these zonation types will be stimulated. most endangered terrestrialecoregions of theWorld (Zaza­ nashvili et al. 1999; Myers et al. 2000). The first scheme of mountain zonation for the Cauca­ Some concepts and terms sus (i.e. its southwestern part) was elaborated by Wagner (1848). At the end of the 19th century and the beginning First some concepts and terms will be explained. A veg­ of the 20th century researchersoften returnedto zonation etation fo rmation is defined in its traditional, ecological­ items (e.g. Kuznetsov 1890; Albov 1896; Busch 1898; physiognomic meaning. A mountain vegetation zone is Radde 1899; Sosnovsky 1915). Dolukhanov et al. (1942) defined as a combination of primary/climaxfo rmations, analysed and summarized these studies for the Caucasus which (1) has altitudinallclimatic limits (altitudinal eco­ high mountains. Later Shiffers (1953) made the detailed logical optimum) and therefore more or less evident scheme of,mountainzonation for the northernslope of the altitudinal limits of distribution, and (2) is common on the Great Caucasus range. During preparations for an Interna­ regional/sub-regional level. A zone may consist of belts tional Symposium 'Alps-Caucasus' (Grebenschikov & (subzones), which differ from each other by the propor­ Zimina 1974; Golubev et al. 1974; Kolakovski et al. tion between characteristic formations of a zone (i.e. by 197 4) information on mountain zonation for western and fo rmation spectra). For instance, in the Colchic zonation northeastern parts of the Great Caucasus mountain range type Castanea and Fagus-Castanea forests are character­ was collected by Russian and Georgian scientists with istic of the lA1 belt as well, but in the formation spectrum four river valleys as examples. Other zonation schemes ofiA2 they dominate; or, in the upper subalpine belt (ID2) for different parts of the Caucasus include three detailed elfin woods with Betula litwinowii and meadows are schemes of altitudinal distribution of vegetation in eastern widespread, while other types of elfin woods, open wood-

Acta Phytogeogr. Suec. 85 -Main typesof vegetation zonation on the mountains of the Caucasus - 9

lands and tall herbaceous vegetation predominate in the Om 14.0°C 2718mm lower belt (ID1). A typeof mountain zonation is determined by a stable, altitudinally regular combination of vegetation zones (or zonal altitudinal profile- ZAP). which is repeated at least on the sub-regional level with certain variation and de­ fines regional/sub-regional phytogeographical individu­ ality. The concrete appearance of a zonation type is re­ lated to the geographical location of the area, regulation of natural macro-ecological factors (first of all to the re­ gional circulation of air masses) and features of evolu­ tionary development of ecosystems. A zonation type is a more or less general feature and encompasses the most characteristic and commonaspects of sub types. The char­ acteristics of the latter are mainly determined by special meso-climatic, geomorphologic and lithologic features. I Besides, anthropogenic subtypes for the regions with Fig. 2. Climate diagram forthe southern Colkhis. consequential level of human impact on the natural eco­ systems may be considered. The main types of vegetation zonation clearly exposed on the Caucasus mountains are described below. Besides, precipitation is mostly more than 2000 mm, in certain we take into consideration that the characteristic forma­ places it even exceeds 4000 mm (Fig. 2). This is clearly tions of certain zones often intrude into the adjoining indicated by the concentration of Colchic relicts (Fig. 3). zones. Thetransitional zones frequently cover large areas; The main characteristic of this type is a wide distribution indicators of altitude of relatively wide distribution outside of Colchic relicts along the whole zonal altitudinal pro­ the optimal zones and belts are given in brackets (the file, almost from sea level up to 2300 m. Colchic relicts Latin names of dominant and eo-dominant species char­ either form a 2 - 4 m tall dense understorey in different acteristic to the formations are given in the corresponding forest types, or they occur as independent shrub commu­ brackets as well). nities in certain habitats. Here, in the Transcaucasus (lower The following material has been used for the determi­ subalpine belt) endemic oak and birch elfm woods are nation of the zonation types: found, with Quercus pontica, Betula medwedewii and B. - published articles and monographs, part of which has megrelica; other endemic relicts include Rhododendron been mentioned above; ungernii,R. smirnowii,Ep igaea gaultherioides and Corylus - published and non-published cartographic information colchica. (including a map scale 1: 200 000 prepared by us for ea. A characteristic ZAP of this type is well-developed in 7% of the Caucasus area); the south Caucasian part of the Black Sea catchment - cosmic coloured photographs (1989; scale 1: 200000); basin. - data from our field expeditions conducted in almost every region of the Caucasus. lA. Humid thermophilous Colchic broad-leaved fo rest zone. This zone occurs up to 1000 (1200) m.

Zonation typology !Al. Mixedbroad-leave dforest belt. This belt occurs up to 500 (600) m and includes Castanea sativa, Carpinus Four main zonation types and some subtypes are outlined caucasica, Fagus orientalis, Quercus hartwissiana and (Fig. 1). Zelkova carpinifolia, with a Colchic understorey including Rhododendronponticum, lnurocerasus officina/is andRuscus I. West Caucasian (Colchic) type colchicus as well as the lianas Hedera colchica, H. helix, Dioscorea caucasica and Vztis sylvestris (Fig. 4). In rela­ This type is characteristic of the western sections of the tively dry habitats thermophilous hombeam-oak forests oc­ Great Caucasus range and of the Small Caucasus moun­ cur with Quercus iberica, Carpinus caucasica and C.

tain chain, mainly where the Caucasus embraces the orientalis. In southern Colkhic (from 200 m upwards) we Black Sea catchment basin, i.e. of that region, where one findpine-oak forests with Quercusiberica, Q. dshorochensis of the refugia of hygro-thermophilous representatives of and Pinus kochiana as well as Colchic thickets with the Tertiary flora existed during the ice ages. This type Rhododendron ponticum, /lex colchica, Laurocerasus was formed under humid conditions - the mean annual officina/isand Ruscus colchicus.

Acta Phytogeogr. Suec. 85 10 N. Zazanashvili et al.

..-·· .• >......

5

.· · ..... )

Fig. 3. Distribution of semi-prostrate Colchic relicts in the Causasus: Laurocerasus o.fficinalis, Rhododendron ponticum, Rh. luteum,

Rh. ungemii, Rh. smimowii, Va ccinium arctostaphylos, Ep igaea gaultherioides, Viburnum orientale, Ruscus colchicus, /lex colchica, I. stenocarpa, I. hyrcana- includingRh. x sochadze but excluding Rh. causasicum. a = areas with 1 - 2 species; b = 3 - 4 species; c = 5 - 8 species; d = 9 - 11 species (Dolukhanov 1980). 1 = Russian Federation; 2 = Georgia; 3 = Azerbaydzhan; 4 = Armenia; 5 = Turkey; 6 = Iran.

IA2. Chestnut fo rest belt. This belt is characterized by ID. Subalpine elfin wood and meadow zone. This zone forests with Castanea sativa and Fagus orientalis, occur­ occurs from1800 ( 1600) - 2400 (2700) m. ring from 500-1000 (1200) m with a Colchic understorey; further by thermophilousoak forests and Colchic thickets !DJ. Lower subalpine belt. This belt occurs from 1800 (see IA1 with Va ccinium arctostaphylos). (1600)to2100 (2200)m and includes: - beech, oak and birch elfin woods (Fagus orientalis, lB. Humid beechfo rest zone. A zone between 1000 (800) Quercuspontica, Betula medwedewii, B. megrelica), often

-1400 (1800) m with: with a Colchic understorey; - Fagus orientalis forest often witha Colchic understorey; - tall herbaceous vegetation (Heracleum ponticum, Ligus­ - dark coniferous and mixed beech-dark -coniferous forests ticum physospermifolium, Senecio cladobotrys); (Abies nordmanniana, Picea orientalis, Fagus orientalis), - dark coniferous and beech-dark coniferous forests often partly with a Colchic understorey; with a Colchic understorey; - Colchic thickets (see lA 1, with Rhododendron ponticum, -Rhamnus imeretina, Sorbus subfuscaor Coryluscolchica Rh. ungernii, I..aurocerasus officinalis, flex colchica, Ruscus thickets; colchicus, Va ccinium arctostaphylos, Viburnumorientale) . - Colchic thickets (Rhododendronponticum, Rh. ungernii, Laurocerasusofficinalis, !lex colchica, Ruscus colchicus, !C. Nemoral humid coniferous forest zone. Thiszone, from Va ccinium arctostaphylos); 1400 (1000) - 1800 (2100) m, includes forests withAbies - subalpine meadows ( Calamagrostis arundinacea, Poa nordmanniana, Picea orientalis and Fagus orientalis, partly iberica, Geranium platypetalum). with Colchic understorey. Colchic thickets occur as in I B.

Acta Phytogeogr. Suec. 85 -Main types of vegetation zonation on the mountain of the Caucasus- 11

Fig. 4. Mixed broad-leaved Colchic fore t outhern Colchic, we tern Cauca u .

ID2. Upper subalpine belt. Thi belt, from 2100t o 2400 IE2. Upper alpine belt, from 2750 to 2900 (3000)m, wit h: (2700) m, include : - gra land (Festuca supina, Kobresia schoenoides and - birch/ ash-birch elfin wood withBetula litwinowii, Sorbus Geranium gymnocaulon); caucasigena; - mat (Cerastium cerastoides, Ranunculus svaneticus, - Rhododendron caucasicum thicket ; Potentilla crantzii); - ubalp ine meadow on lime tone (Calamagrostis - rock and cree veget at ion. arundinacea, Festuca djimilensis);

- Woronowia speciosa, Carex pontica. IF. Subnival zone. Thi zone occurs from 2900- 3700(4000) m and include op enpl ant communities (Cerastium polymor­ /E. Alpine grassland and thicket zone. Thi zone occur phum, Minuartiatrautvetteriana Saxifragascleropo da). between 2400 and 2900 (3000) m. !El. Lower alpine belt. Thi belt occur from 2400 to !G. Nival cryptogam zone. Thi zone occurs > 3700 m. 2750 m and include :

- gra land (Nardus stricta, Festuca djimilensis, Agrostis ll. East Caucasian type lazica, Geranium gymnocaulon); - Rhododendron caucasicum thicket . Thi type i characteri tic of the ea tern ect ion of the

Fig. 5. Aridwoodland withluniperus pp. and Pistacia mutica in the ub­ montane belt, ea tern Cauca u .

Acta Phytogeogr. Suec. 5 12 Zazanashvili et al. N.

Fagus orientalis Fig. 6. fore t, ea tern Cauca us.

Great Cauca u range and of the Small Cauca u moun­ J. fo etidissima, Celtis caucasica); ee Fig. 5; tain chain. The climate ha continental feat ure over mo t - hibliak (Paliurus spina-christi, Rhamnus pallasii, of the area: mean annual pr ecipitation varie from 600 to Atraphaxis spinosa, Ephedra procera). 1000 mm. Be ide , the norther n lope of the Ea tern Great Cauca u and the Small Cauca u ar e drierth an the out h­ IIA2. Lower mountain belt. Thi belt, from 500- 1000 em lope of the Ea tern Great Cauca u , which i re­ ( 1 200) m, includes oaklhom beam fore ts (Que reus iberica, fl ect ed on the zonation sub-type level. Furt hermore, in Carpinus caucasica); beech and hombeam-beech fore t comp ari on withth e humid Colkhic, corre ponding zone (Fagus orientalis, Carpinus caucasica). ar e located 100-200 m higher her e. Due toth e ab ence of refugia the zonation i relatively imple. liB. Mesic beech fo rest zone. This zone occursfr om 1000 - 1800 (2000) m. See Fig. 6. IIAJ. Riverside and fo othill fo rest belt. Thi belt, it uated at < 500 -600(1 000)m, includes: /JBJ. Middle mountain belt. At 1000- 1500 m thi belt - river ide oak and poplar-oakfo re t (Que reus pedunculi­ include Fagus orientalis and Pinus kochiana forest . flora, Populus hybrida, P. nigra with Acer velutinum, Ulmus suberosa); IIB2. Upper mountain belt. This belt, at 1500 - 1900 - ther mophytic-hemixeric hornbeam-oak fore t on the (2000) m, includes Fagus orientalis fore t ; Quercus lope (Quercus iberica, Carpinus orientalis); macranthera fore t ; Pinus kochiana fore t and wood­ - ar id woodland (Pi tacia mutica, Juniperus polycarpos, land (Fig. 7).

Acta Ph togeogr. Suec. 85 -Main types of vegetation zonation on the mountains of the Caucasus - 13

7. Pinus kochiana Fig. fore t, ea tern Caucasu .

IIC. Subalpine elfin wood and meadow zone. Thi zone 11C2. Upper subalpine belt. Thi belt , at 2200 - 2500 occur between 1900 (2000) and 2500 (2700) m. (2600) m, i characterized by: - birch and a h-birch elfin woods ( ee ITCI); IICJ. Lower subalpine belt. Thi belt, at 1900 - 2200 m, - Rhododendron caucasicum thicket ; includes: - ubalp ine meadows (Festuca varia, Geranium ibericum, - oak, pine and maple woodland , including op en wood­ Betonica macrantha). land (Quercus macranthera, Pinus kochiana, Acer trautvetteri); liD. Alpine grassland and thicket zone. Thi zone occur - birch and a h-birch elfin wood (Betula litwinowii, B. bet ween 2500and 3000 (3200)m. raddeana, Sorbus caucasigena);

- tall herbaceous vegetation (Heracleum sosnowskyi, liD I. Lower alpine belt. This belt, found bet ween 2500 and Aconitum orientale); 2800 m, include : - low juniper op en communitie (Juniperus hemisphaerica) - alp ine gra land (Festuca varia, Carex tristis, Kobresia mainly on rock and scree · capilliformis, Nardus stricta); - Rhododendron caucasicum thicket ; - Rhododendron caucasicum thicket . - ubalpine meadow (Agrostis planifolia, Bromopsis IID2. Upper alpine belt. Thi belt,fr om 2800-3000 (3200) variegata, Hordeum violaceum, Geranium ibericum); m, include : - meadow tepp e (Festuca ovina, Carex humilis, Bro­ -alp ine gra land (Festuca varia, Carex tristis Kobresia mopsis variegata, Thymus collinus). schoenoides); alpine mat (Sibbaldia parviflora, Carum caucasicum, Campanula biebersteiniana).

Acta Phytogeogr. Suec. 85 14 N. Zazanashviliet al.

liE. Subnival open zone. In this zone, at 3000 -4000 m, liiD. Subalpine woodland and grassland zone. In this open plant communities with Cerastium kasbek and zone, between 2300 and 2800 (2900) m, we find: Tripleurospermum subnivale occur, with fragments of - Quercus macranthera woodlands; mats and grasslands (up to 3300 m). - steppes (Festuca valesiaca, Koeleria cristata, Sesleria phleoides); liF.Nival cryptogam zone. This zone is found at > 4000 m. - subalpine meadows (Bromopsis variegata, Phleum nodosum, Koeleria caucasuca); Ill. South Caucasian (Front-Asian) type -meadow steppes (Festuca valesiaca, F. ovina, Bromopsis variegata, Sesleria phleoides); This type is characteristic of the uplands and mountains of - thorn-cushion communities (Astragalus aureus, A. the Southern Caucasus mainly composed of volcanic lagurus, Onobrychiscomuta). sediments. Here, representatives of the Caucasian relict flora do not occur: Anatolian-Iranian components pre­ lifE. Alpine grassland zone. This zone, at 2800 -3400 dominate in the plant communities floristic composition; (3600)m, includes: the typical forest zones are not characteristic of the zonal - alpine grasslands (Festuca varia, F. chalcophaea, altitudinal profile which is formed in xerothermic, conti­ Alopecurus aucheri, Carex tristis) and mats (Sibbaldia nental conditions: the mean annual airprecipitation varies parviflora, Alchemilla erythropoda ). within 250 -500 mm limits and increases in high-mountain regions. Incomparison with humidregions in theCaucasus //IF. Subnival open zone. Between 3400 and 4200 (4400) the corresponding zone limits are situated 300 - 400 m m open plant communities occur with Draba araratica, higher. The typical zonal altitudinal profile is as follows: Poa araratica and Saxifraga hirculus.

IliA. Desert zone. This zone, at<800 m, includes: IliG.Nival cryptogam zone. This zoneis found at > 4200 m. - dwarf semi-shrub deserts (Artemisia fragrans, Salsola spp. with ephemeroids. e.g. Poa bulbosa, Catabrosella IV.Southeast Caucasian (Hyrkanic) type humilis); - deserts with Halocnemum strobilaceum and Suaeda This type is characteristic of the extreme southeasternpart microphylla on saline soils and salt marshes; of the Caucasus, southeast Azerbaydzhan and the north­ - thorn-cushion communities (Artemisia microcephalus, west Iranian mountains along the Caspian Sea coast. A. vedicus, A. karjaginii). Here, the other refugium from the Tertiary flora, the Hyrkanic refugium, occurs. There is more differencethan /liB. Xeric grass and semi-shrub zone. This zone, at 800 similarity between the Colchic and Hyrkanic refugia. In (1200) - 1600 m, includes: the Hyrkanic area evergreen species are less widely dis­ - tomillares (Thymus kotschianus, Scutellaria spp., tributed and are of less phytocoenotic importance. Bes­ Stachchys injlata); ides, if relicts range from sea level to alpine belt in - friganoids (Ambliopogon spp., Caccinia rauwolfii, Hedy­ Colkhic, communities in the Hyrkanic area, where relicts sarum fo rmosum); appear as dominants and eo-dominants, reach only up to - thorn-cushioncommunities (Astragalus microcephalus, 1000 - 1200 m. Due to local climatic peculiarities, the Onobrychis comuta, Acantholimon glumaceum); lower zones of the mountains are more humid than the - steppes (Stipa spp., Festuca valesiaca, Bromopsis riparia, upper zones: the mean annual precipitation in the sub­ Carex humilis). mountain area is 1700 mm (with a summer minimum), while the mean annual air precipitation above 2000 m is li/C. Hemi-xeric woodland zone. This zone, at 1600 - only 300-400 mm. 2300 (2400) m, includes: The mean annual air temperature in the foothill is not - Quercus macranthera woodlands; higher than 12-13 °C. According to a typical ZP A from the - low woodlands (Pyrus spp., Acer hyrcanum, Crataegus Talysh mountains a the following zonation types can be spp., Juniperus polycarpos); distinguished. - hemi-xeric shrublands ( Cotoneaster spp., So rhus graeca ); -steppes (Stipa tirsa, Festuca valesiaca, Koeleria cristata, IVA. Humid thermophilous Hyrkanic broad-leaved fo rest Nepeta grossheimii); zone. This zone occurs at altitudes< 1000 (1200) m. -thorn-cushioncommunities (Astragalusspp., Onobrychis NAI. Mixed broad-leavedforest belt. This belt is found up comuta, Acantholimon glumaceum); to 600 m and includes: - meadow steppes (Festuca ovina, Poa densa, Phleum -oak-parrotia,parr otia-hornbeam-oak,oak-hornbeam-azad phleoides, Carex humilis). forests (Quercus castaneifolia, Parrotia persica, Zelkova carpinifolia, Carpinus caucasica, Albizzia julibrissin,

Acta Phytogeogr. Suec. 85 - Main types of vegetation zonation on the mountains of the Caucasus - 15

Ficus hyrcana, Diospiros lotus, with shrubs and semi­ 1998) as well as in some other regions. Similar relations shrubs: !lex hyrcana, Ruscus hyrcanus, Danae racemosa, may be found with the Hyrkanic refugial type, but as and lianas: Smilax excelsa, Periploca graeca, Hedera mentioned above, evergreen elements are less distinctive pastuchowii); in this type. If we compare the Colchic and Hyrkanic - Quercus castaneifolia thermophilous forests. types with European mountain zonation types (Ozenda 1994), the convergentional features point rather to the IVA2. Oak fo rest belt. This belt occurs from 600 to 1000 southernAlps. (1200) m and includes: The similarity between the mountain zonations of - Quercus castaneifolia thermophylic forest sporadically West Caucasus and West Alps has been also mentioned with Parrotia persica, Zelkova carpinifolia, Acer velu­ earlier (Grebenschikov & Zimina 1974). The position of tinum, Gleditsia caspia; the South Caucasian type is more obvious - it should be - beech and beech-hornbeam forests (Fagus orientalis, united with the Front Asian (East Anatolian-West Ira­ Carpinus caucasica); nian) group of mountain zonation types. The issue of the - Quercus iberica-Carpinus caucasica forests. East Caucasian type requires additional investigation: this type, among other Caucasian types, holds a transient !VB. Mesicbeech fo rest zone, from1000-1600 (1800) m, position. Here, on the one hand we find xerophytic ele­ includes Fagus orientalis forests. ments in the lower belts, and on the other hand typical mesic beech forestsand semi-humid subalpine elfm woods !VC. Steppe and xeric dwarfsemi-shrub zone. From 1600 spread in the upper belt. to 2300 (2500)m we find: - steppes (Stipa tirsa, S. lessingiana, Festuca valesiaca, Koeleria cristata, Nepeta spp.); References - thorn-cushion communities (Astragalus aureus, A. lagurus, Onobrychis comuta); Alboff, N. 1896. Les forets de la Transcaucasie occidentale.­ - meadow steppes (Festuca valesiaca, Phleum phleoides, Bull. Herbarier Boissier 4: 61-78. Koeleria cristata, Carex humilis). Box, E.O. & Pujiwara, K. 1988. Evergreen broad-leaved forests of the Southeastern United States: preliminary description. -Bull. Inst. Environ. Sci.Technol. Yokohama Natl. Univ. Besides the above-described types in the northwestern 15(1): 71-93. and southwestern Caucasus,the nearby Crimea peninsula Cherepanov, S.K. 1995. Plantae vasculares USSR. - Nauka (Russia) and the upper part of the Chorokh river basin Leningrad, 509 pp. (Turkey), we come across local fragments of Crimea­ Dolukhanov,A. G. 1980. Colkhidianunde rwood.- Metsniereba, Mediterranean and Mediterranean types of zonation. Tbilisi, 262 pp. (In Russian.) Dolukhanov, A.G., Sakhokia, M.P. & Kharadze, A.L. 1942. On the zonation of high-mountain vegetation of the Caucasus. Conclusions - Trans. Tbilisi Bot. Inst. 8: 114-130. (In Russian.) Pujiwara,K. 1998. Competition between evergreen and summer­ Generally, mountain zonation typology in the Caucasus is green broadleaved forests along their ecotone. - Stud. 20: 41. defined first of all by the (1) transitional geographic Plant Ecol. Gagnidze, R. 1970. On the typology of altitudinal zonation in position of theregion between temperate deciduous broad­ the Caucasus. Bull. Acad. Sci. Georgia 59(3): 673-676. (In leaved forests and subtropical zones, (2) the location of Georgian.) the different phytogeographical regions (Mediterranean, Golubev, G.N., Tushinski, G.K., Grebenschikov, O.S. & Zimina, Minor Asian, Iranian) in the contact area, and (3) by the R.P. 1974. Geomorphology and altitudinal belts of natural evolutionary history of the original Caucasian flora. ecosystems. Baksan-Baksan river valley-Itkol. In: Prance­ The refugial history of thearea has greatly influenced Soviet Geographical Symposium 'Alps-Caucasus': Guide the Colchic type of zonation: it is the basic factor behind of the Caucasus (Symposium materials), pp. 78-85. Institute the biodiversity and phytogeographical individuality of of Geography of the Scientific Academy of the USSR. both the subregion and the whole region. Besides, the Nauka, Moscow-Tbilisi. (In Russian.) transitional location of the region may be considered as Grebenschikov, O.S. & Zimina, R.P. 1974. Natural ecosystems and vertical belts. In: Prance-Soviet Geographical Sympo­ one of the reasons for the maintenance of the Colchic sium 'Alps-Caucasus': Guide of the Caucasus (Symposium type. Analogies for a Colchic-type zonation can be found materials), pp. 8-12. Institute of Geography of the Scientific in areas characterized by transitional 'laurineous forests' Academy of the USSR, Moscow-Tbilisi. (In Russian.) (KlOtzli 1988), semi-evergreen forests (Box & Fujiwara Hildebrand-Vogel, R., Flores T.L., Godoy B.R. & Vogel, A. 1988)or 'warm evergreenFagusforests' (Fujiwara 1998), 1998. Synecology of Nothofagetea pumilionis- compari­ particular!y in the northernparts of the Southeastern United son between south Andean and European deciduous forests. States and the Southern Andes (Hildebrand-Vogel et al. - Stud. Plant Ecol. 20: 34.

Acta Phytogeogr. Suec. 85 16 N. Zazanashvili et al.

Klotzli, F. 1988. On the position of the evergreen broad-leaved Shiffers, E. V. 1953. Vegetation of the North Caucasus and its (non-ombrophilous) forests in the subtropicaland temperate naturalranges. - Publ. House of Acad. Sci. USSR, Moskva­ zones.- VerOff. Geobot. Inst. Eidg.Tech. Hochsch. Stift. Leningrad, 398 pp. (In Russian.) RiibelZlir. 98: 169-196. Sosnovsky, D.I. 1915. Sketch of the vegetation of the Upper Kolakovski, A.A., Maruashvili, L.l., Sokhadze, E.V., Ukleba, Svaneti.- Bull. Russ. Flora 1(3): 119-144. (In Russian.) D. B. & Urushadze, T.F. 1974. Geomorphology and vertical Stanukovich, K. V. 1955. General types of zonation for the belts of natural ecosystems. Sukhumi-Ritsa lake-Avadkhara. mountains of USSR.- Proc. USSR Geogr. Soc. 87(3): 232- In: France-Soviet Geographical Symposium 'Alps-Cauca­ 243. (In Russian.) sus': Guide of the Caucasus (Symposium materials), pp. Stanukovich, K.V. 1973. Mountain vegetation of the USSR.­ 152-162. (In Russian.) Donish, Dushanbe, 416 pp. (In Russian.) Kuznetsov, N. I. 1890. Geobotanicalinvestigation of the Cauca­ Wagner, M. 1848. Reise nach dem Ararat und dem Hochland sus northern slope. Proc. Imper. Russ. Geogr. Soc.17 , pp. Armeniens. Mit einem Anhange: Beitrage zur Natur­ 55-73. A.S. Suvorin' Typography, Sankt-Petersburg. (In geschichte des Hochlandes Armenien. - Verlag der Russian.) Gottiifchen Buchhandlung, Stuttgart, Tiibingen, XII+331 Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, pp. G.A.B. & Kent. J. 2000. Biodiversity hotspots for conserva­ Zazanashvili, N. 1999. On the Colchic vegetation. In: Klotzli, F. tion priorities.- Nature 403: 853-858. & Walther, G.-R. (eds.) Conference on recent shifts in Nakhutsrishvili, G. 1999. The vegetation of Georgia (Cauca­ vegetation boundaries of deciduous forests, especially due sus).- Centro Audiovisivi e Stampa, Camerino, 74 pp. to general global warming, pp. 181-197. - Birkhauser Ozenda, P. 1994. Vegetation du continentEuropeen.-Delachaux Verlag, Basel. et Niestle S.A., Lausanne, Paris, 271 pp. Zazanashvili, N., Sanadiradze, G. & Bukhnikashvili, A. 1999. Radde, G. 1899. Grundziige der Pflanzenverbreitung in den Caucasus. In: Mittermaier, R.A., Myers, N., Gil, P.G. & Kaukasuslandern. Die Vegetation der Erde, Vol. 3. - Mittermaier, C. G. (eds.) Hotspots: Earth'sbiologically rich­ Wilhem Engelmann Verlag, Leipzig, 500 pp. est andmost endangered terrestrialecoregions, pp. 269-273. - Toppan Printing Co., Japan.

Acta Phytogeogr. Suec. 85 Zonation and management of mountain forests in the Sierra de Manantlan, Mexico

Mi guel Olvera Vargas1*, Blanca Lorena Figueroa-Rangezt & Frans Bongers2

1 Instituto Manantldn de Ecologfa y Conservaci6n de la Biodiversidad, Centra Un iversitario de la Costa Sur, Universidad de Guadalajara, Apartado Postal108, C.P. 48900 Autldn, Jalisco, Mexico; 2Silviculture and Forest Ecology Group, Wa geningen Un iversity, P.O. Box 342, NL-6700 AH Wa geningen, The Netherlands; E-mail: [email protected]; * +5233811425; Corresponding author; Fax E-mail.· [email protected]

Abstract Introduction

The woody species composition and the environment of In tropical ecosystems tree species tend to be inter­ secondarymixed Que reus forest in the Sierrade Manantlan, mingled, but generally they form particular associations. Mexico are described, based on 60 circular permanent Sometimes these are easily recognizable in the field, yet plots of 500 m2 each, in which adult trees and saplings, as often the associations are complicated, and are difficult to well as environmental variables (altitude, aspect, slope, discernby simple fieldchecking. The CerroGrande pla­ physiography, litter and humus depth, grazing, erosion, teau, located in the southeastern portion of the Sierra de canopy openness, herbaceous and shrub cover) were re­ Mananth1n, shows a particular physiography with a large corded. Using classification by TWINSPAN we identi­ topographic and geomorphological variation. Apparently fied seven oakcommunity types, each dominated by one it is an isolated mountain without a clear connection with or two of the six oak species found. Canonical Corre­ the surrounding landscape, neither in physiography nor in spondence Analysis ordered the plots mainly by differ­ vegetation. Although, Quercus is clearly the most con­ ences in altitude and physiography. The six oak species spicuous genus on the plateau and is frequently observed were found in particular associations depending on the in a wide range of biogeographical conditions; species ecological zone where they were growing. The commu­ from this genus rarely tend to form monospecific stands nity dominated by Quercus castanea was found at the particularly in wet environments. In the wet habitats the lowest altitudes and on gentle slopes; the Q. laurina genus is commonly represented by Quercus candicans, community was found on abrupt slopes at intermediate Q. rugosa and Q. laurina, and generally eo-occurs with altitudes; whereas the community dominated by Q. other broad-leaved trees such as Oreopanax xalapensis, crassipes was found at the highest altitudes and on the Symplocos citrea, Styrax ramirezzi and Te rnstroemia flattest terrains. Based on species composition and envi­ lineata over a wide range of physical and biological ronmental variables, the study area can be divided in at conditions (Olvera-Vargas et al. 1997). In the driest habi­ least two different management zones, where manage­ tats and flattest terrains Quercus crassipes tends to form ment here is considered for silvicultural purposes only. monospecific associations although this species is also One zone withstands dominated by Q. crassipes could be frequently associated with Prunus serotina and Alnus managed using even-aged methods, e.g. a shelterwood jorullensis (Figueroa-Rangel 1995). cutting system. The second zone, mainly withstands domi­ In the present study we used both classification and nated by Q. candicans, Q. laurina and Q. rugosa, could be ordination techniques in order to recognize groups based managed using uneven-aged methods, e.g. a selection cut­ on species composition and stand structure, and to iden­ ting system. tify which environmental variables were related to differ­ ences in species composition among the groups. The aim Keywords: Classification; Communities; Ordination; was to discernmanagement zones based on the ecological Physiography; Quercus; Shelterwood cutting; Selection zonation that six oak species revealed in the Mountain cutting. Forests of the Sierra de Manantlan, Mexico.

Nomenclature:Vazquez-Garcia et al. (1995).

85: 17-22. 91-7210-085-0. Acta Phytogeogr. Suec. ISBN

Acta Phytogeogr. Suec. 85 18 M. Olvera Va rgas et al.

Methods Results and Discussion

Study area Vegetation classification

The study was undertaken in Cerro Grande in theSierra de According to TWINSPAN at the third levelof division we

Manantlan, located between the states of J alisco and distinguished seven species groups (Table 1 ). The canopy

Colima in Mexico, (19°24'32" - l9°31'02" N; 104°01'0 9"- of each species group (hereafter called community types) 103057'44" W). Cerro Grande is a large calcareous was dominated by one or two oak species. plateau (450 km2) of Upper Cretaceous limestones (Lazcano 1988), with a complex and highly variable relief 1. Quercus laurina community that generally faces West. Soils are classified as Andosols Quercus laurina dominated the canopy together with and a minor part as Cambisols and Luvisols (Anon. 1976). some individuals of Q. candicans and Oreopanax xala­ Slopes range from 10 to 60% and the largest part of the pensis. Codominant elements were Abies religiosa var. forested land stretches between altitudes of 2000 and emarginata, Clethra spp. and Symplocos citrea. Tern­ 2360 m. According to Garcfa (1987), the climate of the stroemia lineata was the most abundant and frequent region belongs to a temperate-subhumid type classified as species in the middle stratum, generally associated with Ca(w2)(w)(e)g. Cornus excelsa, !lex brandegeana, Styrax ramirezzi and Garrya laurifolia. Saplings of Ternstroemia lineata, G. Field data laurifolia and Styrax ramirezzi were found in abundance. Mean density of adult trees was estimated at 1005 indi­ The data set involved 60 circular permanent plots of viduals/ha and basal area at 35.2 m2/ha, mean height of all 500 m2 each established selectively over several oak adult species was 1 1.8m and mean diameter 16.2 cm. This (Quercus spp.) forest associations and topographic condi­ was one of the communities found in the wet habitats of tions. The presence of oaks (in either pure- or mixed-oak the plateau, mainly on slightly abrupt slopes at 2200 m. stands) and their most conspicuous physical characteris­ The canopy cover was estimated at 96.6% and the herb tics were initially recognized on aerial photographs (scale cover ( 10%) was sparsely distributed. 1 : 30 000), thematic maps and subsequently corroborated by field checking. Within each permanent plot, all adult 2. Quercus rugosa community trees � 5 cm DBH (diameter at breast height) were num­ Quercus rugosa was the most abundant species and bered and tagged to record species, DBH, and tree crown dominated the canopy. As in community 1, Ternstroemia class (dominant,codominant, intermediate and suppressed) lineata was also the most abundant and frequent species in following Kraft's tree classification (Smith et al. 1997). the middle stratum. Associated species in this community The number of saplings per species (individuals < 5 cm were Q. candicans, Q. castanea, Arbutus xalapensis, DBH and > 1.30 m height) was also registered. Environ­ Symplocos citrea, Xylosma flexuosum, Fraxinus uhdei, mental variables were considered according to Olvera et Comarostaphylis disco/or and Garrya laurifolia. There al. (1996): slope (%), altitude (m), aspect, litter depth were no saplings of the species present in the upper and (cm), thickness of the humus layer (cm), canopy openness middle canopy. Mean density of adult treeswas estimated 2 (% ) , percentage of herb and shrub cover (% ), physio­ at 766 individuals/ha and basal area at 37 m /ha; mean graphical units, erosion, and grazing. height of the adults was 12.3 m and mean DBH was 18.3 cm. Together withcommunity 1 it was found on the wet Data analysis side of the plateau at 2187 m mainly on slight slopes. Canopy cover was estimated at 92. 1%. We used TWINSPAN (Hill 1979) in order to group the plots based on their floristic composition. Subsequently 3. Mixed-oak community we perlormed two Canonical Correspondence Analyses Three oak species dominated the canopy of this com­ (CCA) on vegetation and environmental data (ter Braak munity: Que reus candicans, Q. castanea and Q. crassipes. 1990), one using density (number of individuals/ha) of In the middle stratum Alnus jorullensis, Cornus excelsa, each plot and another one using basal area (m2) per ha per Symplocos citrea and Arbutus xalapensis were found in plot. Aspect was codified according to Zerihun & abundance. The sapling community consisted of Q.

Lissanework (1989): N=O; E=2; S=4; W= 2.5; NE =1; castanea and S. citrea. 640 individuals/ha was the aver­ SE= 3; SW=3.3; NW =1.3. age density for adult species, while average numbers for basal area, height and diameterwere 31 m2/ha, 12.7 m and 20.8 cm respectively. The catena location was compara­ ble to community 2, however canopy cover was less (87 .2%) than for the above described communities.

Acta Phytogeogr. Suec. 85 - Zonation and management of mountain fo rests in the Sierra de Manantldn - 19

-Way table of plots and species for the oakforest in Cerro Grande, Mexico; results are based on a TWINSPAN run on the Table I. Two species abundance data. The hierarchic plot cluster structure is shown at the bottom of the table, the species clustering to the right.

Community type 2 3 4 5 6 7

Sample plot 03333444333445 00055554440071400 225551 556 22 1111 22211222330113 27689102345341 61423456784399058 015678 890 23 2376 45645789019012

Comus excelsa --33- 1 - 11- 41- 4 ------1 - 3 -----2 - --3 - 3 - ---�------00000 Symplocos citrea 12133325524423 ---2 ------374- 5 --13------00000 Abies re ligiosa var. emarginata ------1422------�------000010 Clethra spec. ------2 ---65------000010 !lex brandegeana - 61513 - 12 - 1 ------2 ------�------000010 Oreopanax xalapensis 22--1 - 1 - 2 - 3 - 4 ------1 ---1 ------000010 Quercus laurina - 477514724676------000010 Styrax ramirezzi ----1466652555 -----2 - 4646---1 - 1 ------000010 Garrya laurifolia 3252532- 1 - 1 --- 11----2 --51---1 -- - 13- ----�------000011 Te mstroemia lineata 66627695667876 75687986657775675 --1 ------00010 Zinowiewia concinna - 1 ------1 ------�------00010 Carpinus spec. ------�------1 ------000110 Comarostaphylis discolor ------4122-----5 ------000110 Fraxinus uhdei ------1 2 --- --1124-----1 ------000110 Litsea glaucescens ------33------000110 Lippia spec. ------2 ------000110 Xy losma flexuosum 1 ------12---1 - 2 - 1 ---1 ------000111 ------Acacia pennatula - 5-l 00100 Picramnia guerrerrensis ------1------00100 Quercus obtusata ------6 ------00100 Arbutus xalapensis ------3 1 ------1 - 2 - 155- 1 2 - 1 --2 3 - 1 ------00101 Quercus candicans --231253-----5 - 21----11- 51- 165- 63765- - 6 ------00101 Quercus rugosa ------15---1 -- 4 - 555555551- 2 ---1 -----4 12 4 --2 - 1 ------0011 Quercus castanea ------1 ------3 2 --1 ---131- 365545 175555 666 66 44------01 ------1 ------1 --- 10 Abies religiosa var. religiosa Pinus pseudostrobus ------1 ------55-- 110 Quercus crassipes - 1 ------1 ----- 554355 86 6678 99987799999878 110 Alnus jorullensis ------1 - 2 - 11 51 5525 554565641623 -- 111 Pinus leiophylla ---1 ------4 - 11------1 -----1 5563 55321--21----- 111 Prunus serotina --2 ------1 ----- 22- 6 56517566751--- 111

00000000000000 00000000000000000 000000 000 11 1111 11111111111111 00000000000000 00000000000000000 111111 111 00 1111 11111111111111 00000000000000 11111111111111111 000000 111 0000 11111111111111 00000011111111 00000000000111111 000001 00011111111111 00111 100000001 00000001111000001 01111 00001111111 0011111 0111111 00111 0111111

Number of species: 00000001000000 00000001100000000 000000 000 00 0000 00000000000000

4. Quercus castanea community 5. Quercus crassipes-Quercus castanea community Quereus castanea was the most abundant and frequent Quercus crassipes as well as Q. castanea dominated species and generally theunique element dominating the the canopy in this community, withAlnus jorullensis and canopy, although sometimes Q. obtusata codominated. Arbutus xalapensis in the middle stratum. Saplings were Acacia pennatula and Picramnia guerrerrensis were absent in these stands. Mean density of adult trees was present in the lower stratum. There were no saplings estimated in 1000 individuals/ha and basal area in 30 m2/ recorded in these stands. 433 was the number of indi­ ha; mean height was 11 m and mean DBH 14.7 cm. This viduals per ha, whereas 24 m2/ha was the average basal community was quite similar in many features to commu­ area; 15.4 m and 24.9 cm were the figures for mean nity 4, mainly in canopy cover, altitude and thickness of height and mean diameter. This community was found at the humus layer. It was also situated at the same catena 2033 m on smooth slopes, being in fact the inferior location as community 4. altitude for the whole set of the sampling plots. We observed that the dry gradient of the plateau startsin this 6. Quercus crassipes-Pinus leiophylla community community where the smallest canopy cover (84.2% ), as Quereus crassipes dominated the canopy, along with well as the thinnest litter depth and thickness of the some individuals of Pinus leiophylla. Quercus castanea humus layer, were recorded. codominated in the canopy, while Alnus jorullensis and Prunus serotina dominated the middle stratum.There were no saplings representing adult species. 785 individuals/ha

Acta Phytogeogr. Suec. 85 20 M. Olvera Vargas et al.

was the average density for adult species while average Table 2. Canonical coefficients (cc) of environmental numbers for basal area, height and DBH were 37 m2/ha, variables on axes 1 and 2 of the Canonical Correspond­ 12.8 m and 19.1 cm respectively. This community was ence Analysis for the 60 oak forest plots using density and located at 2300 m, one of the highest altitudes for oak area basal data. Only t-values (t) of the regression coeffi­ associations in the plateau, just at the border where pine cients are added; critical value (df� 18; a= 0.05) = 2.1.

(Pinus spp.) associations start to grow (pers. observ.). lt is Axis Density Basal Area important to remark that the canopy cover of this commu­ 1 2 1 2 nity was almost analogous to community2 (92.7% ), how­ Eigenvalues 0.6650 0.3925 0.4595 0.3596 cc cc cc cc ever, the herbaceous cover was comparatively high ( 40% ). Altitude (m a.s.l.) 0.48 4.33 --0.65 -4.40 --0. 16 -1.35 --0.16 --0.97 Aspect --0.30 -2.50 0.09 0.62 --0.09 --0.70 0.49 2.51 7. Quereus crassipes community ErosO(No erosion) 0.30 1.18 --0.14 --0.43 0.11 0.48 --0.62 -1.81 Que reus crassipes dominated the canopy. Sometimes, Eros1(Ligth erosion) 0. 14 0.59 --0.05 --0. 15 --0.34-1.44 --0.71 -2.12 GrazO (No grazing) 0.07 0.40 --0. 12 --0.50 0.27 1.95 0.25 1.29 this species was frequently associated with codominant Grazl (Light grazing) 0. 19 0.48 --0. 10 --0. 19 0.76 1.98 1.21 2.24 individuals of Pinus leiophylla, a species less abundant Graz2 (Moderate grazing) 0. 17 0.42 --0.31 --0.57 0.73 1.76 1.09 1.85 than in the previous community. Alnus jorullensis and Graz3 (Strong grazing) 0.20 0.65 --0.25 --0.60 0.40 1.17 0.30 0.63 Herbaceous cover (%) 0.30 2.60 0.44 2.80 0.29 2.37 --0.24 -1.41 Prunus serotina were present in the middle stratum. Sap­ Humus (cm) --0. 12 -1.08 --0.00 --0.01 0.02 0.24 --0 .12 --0.75 lings were only found of Quercus crassipes. This commu­ Litter depth (cm) --0.04 --0.33 --0.08 --0.54 --0.25 -1.94 --0.04 -0.24 Openness(%) 0.17 1.51 0.42 2.69 --0.10--0.72 --0.27 -1.39 nity is themost xeric in the study area, commonly located PhysO (Crest) --0. 14 --0.90 --0.04 --0.23 0.26 1.30 --0.31 -1 .13 on the flattest terrain and on gentle slopes. It was also the Phys 1 (Plateau) 0.25 2.30 0.17 1.19 0.38 3.68 --0.25 -1.72 Phys2 (Upper slope) 0.28 1.40 0.09 0.34 0.56 3.39 --0.29 -1 .12 community with only one oak species and, at the same Phys3 (Middle slope) 0.21 1.02 --0.20 --0.72 0.17 0.66 --0.95 -2.64 time, the densest of all the communities identified. Mean Phys4 (Lower slope) 0.23 1.14 --0.06 --0.22 0.49 2.09 --0.51 -1.54 density was estimated at 1271 individuals/ha and basal Shrub cover (%) --0.28 -3. 14 0.15 1.26 --0.37 -3.03 0. 13 0.90 Slope (%) --0.04 --0.35 --0.23 -1.29 --0.21 -1.46 --0.14 -0.71 area at 39 m2/ha; mean height was 1 1.7 m and mean DBH 16.8 cm. Canopy cover was estimated at 94.3%.

For basal area data, plots dominated by Q. crassipes Relationships between vegetation and were distinguished from the rest of the sites (Fig. 1 b) environment probably because of the differences in basal areaor due to their location in the driest environments (Olvera & Moreno There were clear differences if we use density or basal 1992) and on the flattest terrains of the plateau. These area data to demarcate species groups in relation to envi­ conditions were different from the other sites, which were ronmental variablesby Canonical CorrespondenceAnaly­ located in upper and middle slopes. Van Rompaey (1993), sis (CCA). Using density data, altitude was the main in a study of an African forest, found a relationship factor dividing groups of plots. For basal areadata, physi­ between moisture requirements of species and their estab­ ography was the main factor causing differentiation be­ lishment at their specific catenalocatio ns. Species charac­ tween communities (Table 2, Fig. 1 ). teristics of wet environments were generally found at Inthe density-based ordination (Fig. la) Axis 1 showed lower slopes of the catena, and species characteristics of the altitudinal variation. The plots dominated by Q. dry environments were located at the crest. In our study Q. crassipes (on the right) were located at higher altitudes crassipes was the species dominating the driest sites, than those dominated by other oak species (sites on the presumably because it was generally located on northern left). Sites dominated by Q. crassipes (from 2100 to 2300 aspects where humidity is lower due to orographic shade. m) containedonly holarctic genera like Alnus, Prunus and It is important to mention that Q. crassipes represented Pinus. In Chiapas (S. Mexico), Quintana-Ascencio & the biggest communities in extension as well as the most ' Gonzalez-Espinosa (1993) reported the above genera be­ abundant species on the plateau. tween 2300 and 2450 m. Kappelle ( 1995) found more On Axis 2, plots with Q. laurina were distinguished tropical genera in the Talamancan Mountain Forests in from the rest of the plots. This was due to Q. laurina forest Costa Rica, e.g. Nectandra, Geonoma, Clusia and Ocotea sites being found on southern oriented slopes, while the at 2300 m. Axis 2 separatedplots dominated by Q. castanea, other sites were found on northernoriented slopes (Fig. 1 b). with the highest herbaceous cover and the highest canopy Southern exposures in Cerro Grande are warmer than openness, from the rest of the plots (Fig. la). Herbaceous northern exposures, because they receive more solar ra­ cover was highly correlated with canopy openness, which diation during winter at these latitudes. Also the influence can beregarded as a result of physiographical attributes of of orographic shade causes higher precipitation in the the plots because it is related to light and water availability, south than in the north (Martfnez et al. 1992), which

which in turn is an allogenic consequence of climate, provides more moisture and better conditions for biomass weather, soil and topography (Smith & Huston 1989). growth and canopy closure.

Acta Phytogeogr. Suec. 85 - Zonation and management of mountain fo rests in the Sierra de Manantltin - 21

a 7: AXIS 2 X 6

5 X X

4

3 I

•

2 *PhysO *Phys3 e X .... X X

• Herb....

• *Physl • AXIS ! *Erosl *GrazO Aspect -· Slope -1,5 4 5 .. 5 2 . ' *ErosO 1, Humus Altitude *Phys4 Litter Shrub *Graz2 *Phys2 *Grazl +Q. crassipes + -2 • Q.crassipes, Pinus leiophylla

&Q. crassipes, Q. castanea

xQ. Castanea -3 :r;Q. candicans, Q. castanea, Q. crassipes

eQ. rugosa

+Q. laurina

4 A.XIS 2 b

3 + *Erosl + + +

+ + +

+ *Phys2 Litter + *Graz2 X *Grazl AXIS 1 HUJ!lUSf­ *Physl *ErosO ___ _ ---.\- _.,x..___--l�:::;;.__ '· -2 *Phys3 -1 2 • 3 4 • kerb *Graz3 penness •• • � _1 : •• -4 • • • • Jlfx• • *GrazO • 'Phy" X *PhysO • -2 1 ! +Q. crassipes • Q. crassipes, Pinus Jeiophylla

-3 a Q. castanea, Q. crassipes

xQ. Castanea

xQ. candicans, Q. castanea, Q. crassipes

eQ. rugosa

+Q. laurina

Fig. 1. Ordination diagram with axes 1 and 2 of a Canonical Correspondence Analysis of the 60 oak forest plots and theenvironmental data in Cerro Grande, Mexico; Using density data; Using basal area data. Altitude = Altitude (m. asl); Aspect = As pect codified a. b. according to Zerihun-Woldu & Lissanework (1989). ErosO =No erosion; Eros l = Light erosion; GrazO =No grazing; Grazl =Light grazing; Graz2 = Moderate grazing; Graz3 = Strong grazing; Herb = Percentage of Herbaceous cover (%); Humus = Thickness of the humus layer (cm); Litter = Litter depth (cm); Openness = Canopy Openness (%); PhysO = Crest, Physl = Plateau, Phys2 = Upper slope, Phys3 = Middle slope, Phys4 = Lower slope; Shrub = Shrub cover (%); Slope = Slope (%).

Acta Phytogeogr. Suec. 85 22 M. Olvera Vargas et al.

Conclusions References

In CerroGrande, the physiography is highly variable Anon. 1976. Carta edafol6gica escala 1:50,000. E13B21 y and in a small area differences in aspect, altitude, and soil E13B22. - CETENAL, Mexico DF. can be expected. This may cause a great variation in Figueroa-Rangel, B.L. 1995. Ecology of oak natural regenera­ micro-sites, and as a result, in the species composition of tion in mixed-oak forests in Cerro Grande, Mexico. - M.Sc. oak forests. The particular zonation has to be taken into Thesis, Wageningen Agricultural University, Wageningen. Garcia, E. 1987. Modificaciones al sistema de clasificaci6n account while defining management zones in the area. climatica de Koeppen. - Institutede Geograffa, Universidad Seven oak communities (four of them with Q. crassipes) Nacional Aut6noma de Mexico, Mexico DF. were distinguished. The ordination results suggest that Hill, M.O. 1979. TWINSPAN - A FORTRAN program for only two oak associations occur on the plateau. In this arranging multivariate data in ordered two-way table classi­ forested area management is mainly directed towards ficationof individuals and attributes. -Cornell University, silviculture. Therefore, for practical purposes, we believe Ithaca, NY. that the mixed-oak forests in Cerro Grande can be divided Kappelle, M. 1995. Ecology of mature and recovering Tala­ in two management zones. mancan Montane Quercus forests, Costa Rica. - Ph. D. The zone with stands dominated by Q. crassipescould Dissertation, University of Amsterdam. be managed using even-aged cutting systems like a Lazcano, S.C. 1988. Las cavemas de Cerro Grande, estados de Jalisco y Colima. -LaboratorioNatural Las Joyas, Univer­ shelterwood cutting system (Nyland 1996; Smith et al. sidad de Guadalajara, Guadalajara. 1997). This method involves tree cutting in three phases. Martfnez-Rivera, L.M., Sandoval, J .J. & Guevara, R.D. 1992. El In the first phase, the oldest trees or the ones with un­ clima en la Reserva de la Bi6sfera Sierra de Mananthin wanted characteristics (badly shaped or cracked trees) (Jalisco-Colima, Mexico) y en su area de influencia. - should be removed in order to reduce competition and to Agrociencia, Ser. Agua-Suelo-Clima 2: 107-119. avoid the establishment of undesirable species. In the Nyland, R.D. 1996. Silviculture: Concepts and applications.­ second phase, favourable conditions for the germination McGraw-Hill, New York, NY. and establishment of oak species (basically controlling Olvera-Vargas, M. & Moreno-G6mez, S. 1992. Estructura y the gap size) must be promoted. The thirdphase involves regeneraci6n de encinares en la Sierra de Man an thin, J alisco. the harvesting of the remainder trees, once the standing In: Arteaga, M.B. (ed.) Memoria del primer foro nacional sabre manejo integral forestal. - Universidad Aut6noma de trees have been improved and abundant regeneration has Chapingo, Chapingo, pp. 403-409. established. Olvera-Vargas, M., Figueroa-Rangel, B.L., Moreno-G6mez, S. & The second zone, mainly occupied by stands domi­ Solfs-Magallanes, A. 1997. Resultados preliminares de la nated by Q. candicans, Q. laurina and Q. rugosa, could be fenologfa de cuatroespecies de encino en Cerro Grande,Reserva managed using uneven-aged methods such as Selection de la Biosfera Sierra de Manantlan. - Biotam 9: 7-17. cutting systems (Nyland 1996; Smith et al. 1997). With Olvera-Vargas, M., Moreno-G6mez, S. & Figueroa-Rangel, this method only the mature oak trees or those trees with a B.L. 1996. Sitios permanentes de investigaci6n silvfcola: specifiedminimum diameter should be removed, although Manual para su establecimiento. - Libros del Institute Manantlan. Universidad de Guadalajara, Guadalajara. other commercial tree species (e.g. Symplocos citrea; I. brandegeana) could be considered for removal as well. Quintana-Ascencio, P.F. & Gonzalez-Espinosa, M. 1993. Afmidad fitogeograficapapel y sucesional de la flora lefiosa This method might require more technical skills during de Ios bosques de pino-encino de Ios Altos de Chiapas, harvesting as well as more ecological konwledge for its Mexico. -Acta Bot. Mex. 21: 43-57. application due to the presence of several species with Smith, D.M., Larson, B.C., Kelty, M.J. & Ashton, P.M.S. 1997. differential growth forms in a small area. This in turn The practice of silviculture: Applied forest ecology. -John necessarily implies the definition of several minimum­ Wiley & Sons, New York, NY. diameter cutting limits. As this method depends on the Smith, T. & Huston, M. 1989. A theory of the spatial and temporal successful recruitment of saplings from successive age dynamics of plant communities. - Vegetatio 83: 49-69. classes, it is very important to protect the stands from ter Braak, J.F. 1990. CANOCO - A FORTRAN program for grazing. canonical community ordination by [partial] [detrended] [canonical] correspondence analysis, principal components analysis and redundancy analysis (version 3.1.) - ITI­ TNO, Wageningen. van Rompaey, R. 1993. Forest gradients in West Africa. A spatial gradient analysis. - Ph.D Thesis,W ageningen Agri­ cultural University, Wageningen. Zerihun-Woldu, F.E. & Lissanework, N. 1989. Partitioning an elevation gradient of vegetation from southeastern Ethiopia by probabilistic methods. - Vegetatio 81: 189- 198.

Acta Phytogeogr. Suec. 85 The distribution of rare plant species on Mount Prado, a northern Apennine diversity hot spot

C. Ferrari*, G. Pezzi & A. Portanova

University of Bologna, Departmentof EvolutionaryBiology, via Irnerio42, 1-40126 Bologna, Italy. *Corresponding author; Fax +39051242576; [email protected]

Abstract Introduction

Mount Prado (2054 m a.s.l.) lies along the main ridge of The northern Apennineslie at ea. 44°N and the timberline the northern Apennines at 44o 14' N and 1 0°23' E of which is found at 1700- 1800 m a.s.l., which correspondslocally it is one of the highest peaks. Areas above the timberline to the altitudinal limit of beech (Fagus sylvatica) woods. along the ridge can be regarded as an 'archipelago' of As few peaks exceed 2000 m; their areas above timberline alpine islands on theboundary between the Central Euro­ are rare and scattered, and can be regarded as an 'archi­ pean and Mediterraneanregions in Italy. The timberline is pelago' of alpine islands. Moreover, the northern at ea. 1800 m a.s.l and correspondsto the local altitudinal Apennine ridge can be considered as the boundary limit of beech woods. The lithological substrate is a between the Central European and Mediterranean phyto­ turbiditic compact sandstone (upper Oligocene-lower geographic regions in Italy (Pignatti 1979). The sum­ Miocene ). Its northernslope is characterized by one of the mit flora contains 391 species. Phytogeographic analy­ widest and most typical Wtinnian glacial cirques of the ses (Foggi 1990; Tomaselli & Agostini 1994) showed northern Apennines. Its flora contains 208 species, in­ that its taxa are strictly connected with that of the Alps cluding mosses and lichens. There are 198 higher plant and, through these, with the alpine flora of other Cen­ species of the 391 taxa known from the northernApennine tral European mountains, mostly the Pyrenees, the summit flora. They form 22 plant communities (i.e. 73 % Carpathians and the Dinaric Alps. The few endemics of the plant communities described for the whole northern are closely related either to European orophytes or Apennine summit vegetation). Spatial data on plant com­ Alpic taxa. Ecological 'insularity' of the mountains munities and the distribution of rare species are com­ and peripheral occurrence in the distribution area of pared. The vegetation mosaic is dominated by two most taxa, make the summit flora of these mountains a Va ccinium-communities and the remaining ones cover vulnerable natural wealth. This is particularly true for small or very small areas. Plant communities with the the local rare species which are linked to very rare greatest frequency of rare species are mainly situated at habitats, such as Gnaphalium supinum to snow-beds, higher altitudes. These are some acidophytic grasslands,a Eriophorum angustifolium, Tricophorum alpinum and snow-bed community and the only rock-face community. ]uncusfilif ormisto acid fens, Rhododendron fe rrugineum As a whole, these communities are a focal point within the to sites covered with high snow. Nowadays, 30 plant vegetation considered. communities have been described for the alpine vegeta­ tion belt of the northernApennines (see Ferrari 1995, for Keywords: Apennines; Italy; Phytogeography; Plant a review). Va ccinium-heaths are the climatic climax (Pignatti 1994; Ferrari Piccoli 1997), but slope exposures diversity. & and landforms often generate fine vegetation mosaics. Within the chain of altitudinal islands of the northern Nomenclature: Pignatti (1982) for vascular plants except Apennines the northern slope of Mt. Prado is outstanding for the genus Festuca (Foggi & Rossi 1996; Foggi et al. for its vegetation diversity. It contains 22 plant communi­ 1999); Grabherr (1993) and Ellenberg (1988) for higher ties, some of thembeing restrictedto thismountain, such as syntaxa. For plant communities see references in Table 1. the snow-bed associations Salicetum herbaceae and Oligotricho-Gnaphalietum supini (Tomaselli 1991 ). Many areas above the northern Apennine timberline have expo­ sures and extensions similar to Mt. Prado glacial cirque but 85: 23-30. 91-72 10-085-0. do not have thesame species and plant communityrichn ess. Acta Phytogeogr. Suec. ISBN

Acta Phytogeogr. Suec. 85 24 C. errariF et al.

Itis right to assume that landform diversity is crucial for the taxon a Rarity Index (Gehu & Gehu 1980) was calculated: local plant wealth. All communities have a high degree of naturalness (as definedby Westhoff 1983). In the past, the RI ;;;;; 1-(n/ N) (1) grasslands at lower altitudes - just above the altitudinal limit of the beech woods - were grazed by sheep. Many where N is the number of quadrants of the grid of the Atlas years ago this impact has ceased and the whole area is of the Central European Flora (Ehrendorfer & Hamann protected as part of a Nature Reserve ('Parco Regionale 1965; Ferrari et al. 1993) which include the northern dell' Alto Appennino Reggiano'). Relationships between Apennine areas above the timberline (N;;;;28), and n is the the vegetation mosaic andsnow cover pattern on Mt. Prado number of quadrants where the taxon is found. Only taxa were studied (Ferrari & Rossi 1995), as well as the spatial with RI > 0.20 were considered as rare. Types of rarity patterns of the community. This pattern is strongly affected were defined according to Rabinowitz (198 1). Correla­ by land-form discontinuities in the glacial cirque which tions between plant community types and types of rarity cause steep ecological gradients. Community patchiness is were described by Analysis of Concentration(AOC; Feoli extensive and the patches mostly cover small or very small & Or16ci 1979). Community classification and types of areas (Ferrari & Pezzi 1999). Since Mt. Prado contains 73% rarity were clustered by a minimum variance criterion of theplant communitiesknown above thenorthern Apennine applied to a similaritymatrix based on Euclidean distance timberline, it may beconsidered a diversity hot spot (Wilson (Orl6ci19 78). Calculations were performed with the pack­ 1992) for the alpine vegetation of this chain, as well as a age MUL V A 5 (Wildi & Or16ci 1996). rewarding study area for investigating vegetation ecology Data on spatial patterns of plant communities were and monitoring the vegetation changes as bin-indicators of derived from a computerized vegetation map 1: 2000 climatic change. (Ferrari & Pezzi 1999). A GIS-analysis was done with ArcCAD© 11.4 on an IBM PC/AT. The map was quanti­ tatively analysed with ArcView©3.0a. We considered: (1) Study area total area of each plant community and (2) number of plant community patches (map polygons). Mt. Prado, or Prato (2054 m a.s.l.), is located along the main ridge of the northern Apennines at 44° 14' N and 10°23' E. Its northern slope is characterized by one of the Results widest and most typical Wtirmian glacial cirques of the northern Apennines (Losacco 1982). The timberline oc­ The 22 plant communities distinguished on Mt. Prado are curs at ea. 1750 m and the glacial cirque is situated above listed in Table 1. Their species pool includes 198 vascular the timber line. The cirque has steep and rocky walls, with plant species and 10 mosses and lichens. The checklist is prevailing N and N-W exposures. Its base is level and available from theauthors on request. crossed by three orders of moraine deposits which prob­ As compared to the species pool of the northern ably correspond to different stages of glacial retreat and Apennine alpine vegetation, the pool of the Mt. Prado which enclose small wet areas. Periglacial morphogenesis glacial cirque contains a very high number of circum­ is still active, mainly in spring and autumn, when the soil boreal and arctic-alpine taxa (Fig. 1). The chorological may be free from snow (Carton & Panizza 1988). The spectra of plant communities are shown in Table 2. lithological substrate is a turbiditic compact sandstone Orophytes dominate all the plant communities except (upper Oligocene-lower Miocene), locally named 'Maci­ acid fen (Cn) and stream community (Cc). High contents gno'. Annual precipitation ranges from 1000 to 3500 mm, of arctic-alpine species characterize the snow-beds (Sh, with an average of 2000 mm (Rossetti 1988). From No­ Og), most of the acidophytic grasslands (St, Ls, Gs), a vember to April precipitation mostly occurs as snow and Nardus-grassland (Ln), screes with big stones (Cd) and a snow melts from May to the beginning of July, according rock-face community (Dp). The rock-face community to differences in topography and exposure. Its vegetation (Dp) has the highest content of Apennine endemics and is dominated by Vaccinium-heaths, alpine grasslands and subendemics. scree communities. Table 3 shows the 43 taxa with a rarity index RI > 0.20 and their occurrence in the plant communities. 20 rare species occur only in one community. One species Data and Methods (Primula apennina) is an endemic and one (Senecio incanus) is a subendemic.The degree of rarity varies from The vegetation was sampled from 1992 to 1994. Plant very high values (RI ;;;;;0.97) to medium-low values (RI ;;;;; communities were identified by the Braun-Blanquet ap­ 0.23). High RI-values are prevailing: 21 species have RI � proach (Ferrari & Rossi 1995). Further surveys were made 0.80. As shown by Fig. 1 the content of rare taxa is for describing their full species compositions. For each particularly high in the endemics (70%) and relatively

Acta Phytogeogr. Suec. 85 - The distribution of rare plant sp ecies on Mount Prado - 25

Table 1. List of plant communities.

Plant community References

Vaccinium-heaths (Rhododendro-Vaccinietalia Br.-BI. in Br.-Bl. et Jenny 1926) Ve Empe tro-Vaccinietum gaultherioidis Br.-Bl.in Br.-Bl. et Jenny 1926 corr. Grabherr 1993 Ferrari & Piccoli ( 1997) juncetosum trifidi Vh Hyp erico richeri-Vaccinietum gaultherioidis Pirola et Corbetta 1971 nom. inv. Ferrari & Piccoli (1997) Vb Vaccinium myrtillus-Brachypodium genuense community Ferrari & Piccoli (1997)

Nardus-grasslands (Nardetalia Oberd. 1949 em. Prsg 1949) Gn Geo montani-Nardetum Liidi 1948 Tomaselli (1994) Vn Viola cavillieri-Narde tum Credaro et Pirola 1975 corr. Tomaselli 1994 Tomaselli (1994) Ln Luzula alpino-pilosa-Nardus stricta community Tomaselli (1994)

Acidophytic grasslands (Caricetalia curvulae Br.Bl. in Br.-Bl. et Jenny 1926) St Sileno excapae-Trifolietum alpini Tomaselli et Rossi 1994 Tomaselli & Rossi (1994) Ci Sileno excapae-Trifolietum alpini, Cetraria islandica variant Tomaselli et Rossi 1994 Ls Sileno excapae-Trifolietum alpini luzuletosum spicatae Tomaselli et Rossi 1994 Tomaselli & Rossi (1994) Gs Sileno excapae-Trifolietum alpini luzuletosum spicatae, Gnaphalium supinum var. Tomaselli et Rossi 1994 Tomaselli & Rossi (1994)

Mesophytic grasslands (syntaxonomy not yet defined) Fe Festuca spec. community Ferrari (unpubl.) Tf Trifolio-Festucetum puccinellii ass. nova prov. Rossi 1994 Rossi ( 1994) De Deschampsia caespitosa community Ferrari (unpubl.) Aa Aquilegio-Anemonetum narcissifloraeTomaselli 1994 Tomaselli (1994) Bg Brachypodium genuense community Barbero & Bonin (1980) La Trifolio-Festucetum puccinellii ass. nova prov., Luzula alpino-pilosa variant Rossi 1994 Rossi (1994)

Snowbed communities (Salicetalia herbaceae Br.Bl. in Br.-Bl. et Jenny 1926) Sh Salicetum herbaceae Riibel 1912 Tomaselli (1991) Og Oligotricho-Gnaphalietum supini Tomaselli 1991 Tomaselli (1991)

Big-stone scree vegetation (Androsacetalia alpinae Br.Bl. in Br.-Bl. et Jenny 1926) Cd Cryptogrammo-Dryopteridetum oreadis Rivas-Martinez in Rivas-Martinez et Costa 1970 Tomaselli (1994)

Rock face vegetation (Androsacetalia vandellii Br.-Bl in Meier er Br.Bl. 1934) Dp Drabo aizoidis-Primuletum apenninae Tomaselli 1994 Tomaselli (1994)

Acid fen vegetation (Caricetalia nigrae (W. Koch 1926) Nordh. 1936) Cn Caricetum nigrae Br.-Bl. 1915, Sphagnum subsecundum var. Ferrari (unpubl.)

Stream vegetation (Montio-Cardaminetalia Br.-BI. et Tx. ex Klika et Hadac 1944) Cc Chaerophyllo-Cardaminetum asarifoliae Gerdol et Tomaselli 1988 Gerdol & Tomaselli (1988)

high in the arctic-alpine chorotype (42%). The types of species with small population sizes (Sp) are correlated rarity mainly relate to peripheral geographic distribution with most of the acidophytic grasslands (Ls, St, Ci), to and small population size. Such types occur seldom some mesophytic grasslands (La, Aa, Bg), to aNardus­ together and are seldom related to specific habitats. community (Gn) and to the big-stone scree community The correlation between types of rarity and plant (Cd). Rare species with a peripheral geographic distribu­ communities as resulting from AOC is shown in Fig. 2. tion (Pd) are correlated with only one acidophytic grass­ The first two canonical variates account for 99 % of the land (Gs), to the snow-beds (Sh, Og), to a Nardus-com­ total variance. The Vaccinium-heaths are not corre­ munity (Vn) and to the mesophytic grassland communi­ lated with any particular type of rare species. Rare ties dominated by Festuca spec. (Fe) and Deschampsia

Table 2. Chorological spectra of plant communities on Mt. Prado . Legend as in Table 1 and Fig. 1.

Chorological Plantcommunity type

Ve Vh Vb On Vn Ln St Ci Ls Os Fe Tf De A a Bg La Sh Og Cd Dp Cn Cc 8.7 4.1 EA 12.0 12.5 12.8 31.0 21.4 15.0 9.6 22.2 13.3 13.6 4.5 23 .1 16.7 24.0 3.0 8.7 3.2 0.0 30.4 18.8 CB 22.0 20.8 25.6 20.7 14.3 10.0 9.6 16.7 11.7 11.4 13.6 8.7 11.5 10.0 12.0 14.3 12.1 13.0 16.1 9. 1 30.4 43 .8

AA 10.0 8.3 7.7 3.4 10.7 15.0 15.4 8.3 16.7 18.2 9. 1 8.7 0.0 10.0 4.0 20.4 21.2 17.4 16.1 18.2 13.0 12.5 OR 50.0 50.0 43.6 44.8 53.6 60.0 55.8 47.2 51.7 50.0 63.6 65.2 46.2 56.7 44.0 49.0 54.5 60.9 41.9 50.0 8.7 12.5 EN 2.0 2. 1 2.6 0.0 0.0 0.0 5.8 0.0 3.3 2.3 0.0 8.7 0.0 0.0 8.0 6.1 6. 1 0.0 9.7 13.6 4.3 0.0 ME 0.0 0.0 0.0 0.0 0.0 0.0 1.9 2.8 1.7 2.3 4.5 0.0 3.8 3.3 4.0 2.0 0.0 0.0 0.0 9.1 0.0 0.0 CS 4.0 6.3 7.7 0.0 0.0 0.0 1.9 2.8 1.7 2.3 4.5 0.0 15.4 3.3 4.0 4.1 3.0 0.0 12.9 0.0 13.0 12.5

Acta Phytogeogr. Suec. 85 26 C. errariF et al.

EA D Mt. Pmdospecies rare

11 Mt. Plado specles CB 11 NorthernApennine spec:1ee

AA CHOROTYPES

EAEurasian

CBcll'

ME Meditenanean Fig. 1. A comparison between the chorological spectra of Mt. ME Prado and of the entire alpine vegetation of the northern Ap­ CS ennines. The distributionof rare taxa within chorotypes found 0 20 60 so 100 120 140 180 180 on Mt. Prado is also shown.

caespitosa (De). Apennine endemics and subendemics terized by a high degree of patchiness. Communities are (Gr) characterize the rock-face community (Dp), while spatially subdivided into two to 32 patches. 50% of the rare species with high habitat specificity (Hs) characterize plant community are subdivided into fiveor more patches. stream (Cc) and acid fen communities (Cn) The digitized Apart from map mosaics there are only three plant com­ vegetation map (Fig. 3) covers an area of 463 982 m2. It munities withone map polygon. The dominant Vaccinium­ includes seven community mosaics. heath (V e) has the highest number of patches. As shown by Table 4, two Vaccinium-communities The greatest number of species is found in the dominate the landscape (Ve, Vh). They cover 58.8% of acidophytic grassland Ls. Apart from the dominant the total area. The remaining 41.2% is covered by 20 Vaccinium-heaths (Ve and Vh), the most species-rich communities and five map mosaics. Grassland communi­ communities are grasslands (St, La). The poorest is the ties prevail (including snow-beds, fens and mosaics with Deschampsia caespitosa grassland (De). Thespecies rich­ heaths). The vegetation spatial pattern is mainly charac- ness shows an increasing trend with community areas for most communities (Fig. 4). However, Nardus-grasslands, some mesophytic grasslands (Tf, Fe) the rock-face veg­ etation (Dp), the scree community (Cd) and the acid fen community (Cn) are poorer in species than the other communities with similar areas. Table 4 shows that the frequencyof rare species (RIN) 11 comparedto the species richness is high (i.e. R1N> 0.20) Sp • for some species-rich communities, as the grasslands St, Ls SI Ci La 1 0 Aa Gn Cd Bg 0 Ln Tf i La, Ls, Gs and the snow-bed community Sh. Itis notewor­ thy that high frequency values are found also in communi­ Ve Vh Vb ties with low numbers of species, such as the grasslands Fe, Tf, Ln and the rock-face community Dp. The lowest -2 -1 Gs Sh Og ce Pd value (RI = 0.05) is found in the stream community Cc. A Vn Fe De very low value (R = 0.08) characterizes the Va ccinium­ Hs • heath Vb and the acidophytic grassland Ci. The communi­ Dp -1 0 ties with the largest areas (Vaccinium-heaths Ve and Vb) Cc Cn • 0 have roughly average values of rare species frequencies, whilst the greatest frequencies occur in acidophytic grass­ -2 land communities (La, St, Ls ), some mesophyticgrasslands

Fig. 2. Reciprocal ordination of types of rarity and plant commu­ (Fe, Tf), the rock-face community (Dp) and the snow-bed nities according to the firsttwo canonical variates. Symbols as in community Sh. Table 3.

Acta Phytogeogr. Suec. 85 - The distribution of rare plant species on Mount Prado - 27

0 100 M

Fig. 3. Vegetationmap of the Mt. Prado glacial cirque (1 : 2000). Plant communities are listed inTable 1. Two communities (Aa and Cc) cannot be shown at the scale of the map.Grey patches: communities with the highest frequenciesof rare species and species density.

Acta Phytogeogr. Suec. 85 28 C. Ferrari et al.

Table 3. Rare species andtheir distribution in the plant comm.u- Table 4. Data on vegetation pattern and plant community corn- nities of Mt. Prado. Chorotypes andtypes of rarity are indicated. position. Data on Aa and Cc are not available (see Fig. 3). Chorotypes: = arctic-alpine; CB = circumboreal; CS = AA Community Area Patches Species Rare RJS cosmopolitan; EN = endemic; EA = Eurasian; OR = orophyte; (m2) (S) species (R) SC = subcosmopolitan; SN = subendemic. Rarity types: Gr = narrow geographic range; Hs =habitat specificity; Pd = peri- Ye 176469 32 58 11 0.19 pheral distribution; Sp = small population size. For other expla- Yh 96721 15 60 8 0.13 nations, see the Text. Yb 21747 7 48 4 0.08 Gn 8018 2 29 5 0.17 RI Taxon Plant communities Choro Type of Yn 12892 8 28 5 0.18 type rarity Ln 10340 20 20 3 0.15 St 19819 8 53 15 0.28 0.97 Hieracium glanduliferum St OR Pd/Sp Ci 447 1 37 3 0.08 0.97 Soldanella pusilia Og OR Pd/Hs Ls 15197 14 61 15 0.25 Senecio incanus St, Gs, Sh, La Pd Gs 14666 14 45 9 0.20 0.97 Ls, SN Fe 0.97 Salix herbacea St, Sh, La AA Pd/Hs/Sp 2832 3 22 6 0.27 Ye Pd/Sp 6406 4 23 6 0.26 0.97 Salix breviserrata AA Tf De 160 19 1 0.05 0.97 Pedicularis rostrato-spicata Ls OR Pd/Sp A a 30 5 0.17 0.97 Leucanthemopsis alpina St, Ls, Sh, Fe OR Pd Bg 251 2 26 3 0.12 0.90 Lychnis alpina Gs, Sh, La AA Pd/Sp La 3695 7 49 14 0.29 0.90 Luzula lutea Ye OR Pd Sh 6601 5 47 10 0.21 0.90 Gnaphalium supinum St, Ls, Gs, Sh, Og, Pd AA Og 4227 10 5 0.11 La,Yn, Ln, 44 Cd 15774 23 31 3 0.10 0.90 Carex fo etida Ln OR Hs/Sp Dp 6059 18 22 7 0.32 0.87 Polygonum viviparum Tf AA Hs/Sp Cn 5408 3 24 4 0.17 0.87 Luzula alpino-pilosa St, Ls, Gs, Sh, Og, AA Pd Cc 22 0.05 Cd, La, Gn, Yn, Aa 0.87 Luzula multiflora Bg CB Sp 0.83 Gentiana nivalis Ls AA Pd/Sp 0.83 Viola palustris Cn CB Hs/Sp 0.83 Soldanella alpina Yh, Ye, La OR Sp 0.83 Carex ornithopoda Tf, Aa EA Pd Conclusions 0.80 Silene acaulis ssp. bryoides St, Ls, Gs, AA Pd Sh, La, Dp 0.80 Primula apennina Dp EN Gr The northern slope of Mt. Prado is characterized by a rich 0.80 Carex canescens Cn CS Hs natural vegetation mosaic dominated by Vaccinium 0.77 Rhododendron fe rrugineum Yn OR Pd/Sp gaultherioides and Vaccinium myrtillus heaths. Itincludes 0.73 Euphrasia minima Yh, Ye, Yb, St, Ci, OR Pd 73% of the plant communities and 51% of the species Ls, Gs, Sh, Og, La, Gn, Yn, Fe presently described for the northernApennine alpine veg- Ln, 0.73 Euphrasia alpina St, Ci, Ls, Gs, Sh, OR Pd etation. As compared to the northern Apennine alpine Og, Tf, Gn, Yn flora, the Mt. Prado flora is very rich in arctic-alpine and 0.73 Erigeron uniflorus Ls AA Pd/Sp circumboreal species. Yh,Vb Pd/Sp 0.70 Lycopodium annotinum CB Species richness of the plant communities reaches 0.63 Armeria marginata La SN Gr 0.60 Saxifraga latina Dp EN Gr/Sp high values not only in the dominant heaths, but also in Ye, Ls, Dp Pd some acidophytic grasslands, for instance the Sileno 0.57 Sempervivum montanum Vh, OR Ye, Vb, La Sp 0.57 Huperzia selago Vh, se exscapae-Trifolietum alpini (Ls, St), the Trifolio- 0.50 Gentiana purpurea Ye, Gs, La, Fe OR Pd Festucetum puccinellii - Luzula alpino-pilosa variant (La) 0.43 Pinguicula vulgaris Cn, Cc EA Hs/Pd and the two snow-bed communities, Salicetum herbaceae 0.43 Leuchorchis albida Gn AA Sp Ye, Vb, St, Ls, Sp (Sh) and Oligotricho-Gnaphalietum supini (Og). These 0.40 Pulsatilla alpina Vh, OR La, Tf, Bg, Fe, Aa communities cover very small areas and have a high 0.40 Viola calcarata Yh, Ye, St, Ls, Gs, EN Gr species density. ssp. cavillierii Sh, La, Bg All communities of the Mt. Prado glacial cirque con- 0.40 Sempervivum arachnoideum St OR Sp tain species that are rarefor the northern Apennines, even 0.40 Festuca violacea Cd, Tf, Dp, Cn, De SN Pd ssp. puccinellii if their types and degrees of rarity are different. The very Ls, Gn 0.40 Botrychium lunaria se Sp rare species are prevailing. Rarity is mainly due to geo- 0.37 Saxifraga moschata Dp OR Sp graphic peripheral distribution, but small population size, 0.27 Festuca riccerii St EN Gr habitat specificity and narrow geographic range are (in 0.27 Astrantia minor Yh, Ve, St, OR Pd La, Fe, Aa descending frequency) the other types of rarity. 0.27 Aster bellidiastrum Ye, St, Ci, Gs, Cd, OR Sp The highest numbers of rare species can be found in La, Tf, Dp, Fe, Aa species-rich communities which cover small areas, i.e in 0.23 Aster alpinus St, Ls CB Sp communities with the highest species density. Such corn-

Acta Phytogeogr. Suec. 85 - The distribution of rare plant species on Mount Prado - 29

70

Ls • Vh 60 •• Ve St . 50 La ·· · • • §: Sh . V

10

0 0 2 3 4 5 6

Fig. 4. Relationship between species number and map area. Symbols as in Table 4.

munities are the Sileno exscapae-Trifolietum alpini (Ls, (MURST, ex -40%) and by the University of Bologna. We St), the Trifolio-Festucetum puccinellii, Luzula alpino­ wish to thank Prof. E. van der Maarel for critical reading pilosa variant (La), the snow-bed communities, Salicetum of the manuscript. herbaceae (Sh). They have a high content of rare species with small population sizes and/or peripheral geographic distribution and are linked to sites with a long duration of References snow cover (Ferrari & Rossi 1995). As far as we know their northern Apennine distribution is restricted to Mt. Barbero, M. & Bonin,. G. 1980. La vegetation de 1' Apennin Prado. To these communities we have to add the Festuca septentrional. Essai d'interpretation synthetique. - Ecol. grassland (which is characterized by rare species with Mediter. 5: 273-313. peripheral geographic distribution) and the rock-face Carton, A. & Panizza, M. 1988. ll paesaggio fisico dell' alto association Drabo aizoidis-Primuletum apenninae, which Appennino erniliano. Studio geomorfologico per 1' indivi­ ' is characterized by endemic and subendemic species. All duazione di un area da istituire a parco. - Gratis, Bologna. 182 pp. communities are situated mostly in the highest part of the Ehrendorfer, F. & Hamann, U. 1965. VorschHi.ge zu einer glacial cirque, on sites with mainly exposures (Fig. 3). NW floristischen Kartierung von Mitteleuropa. - Ber. Dtsch. The highest parts of the Mt. Prado glacial cirque show an Bot. Ges. 78: 35-50. unusual concentration of environmental resources for both Ellenberg, H. 1988. Vegetation ecology of Central Europe. - the rare communities and species within the alpine veg­ Cambridge University Press, Cambridge. 731 pp. etation of the northern Apennines. According to Walker Feoli, E. & Or16ci, L. 1979. Analysis of concentration and (1995), the summit sites of Mt. Prado altogether may be detection of underlying factors in structured tables. - considered a 'focal point' within a diversity hotspot ! Vegetatio 40: 49-54. Ferrari, C. 1995. La vegetazione dell' Appennino tosco-emiliano oltreil limite degli alberi. - Atti Convegni Lincei 115: 229- 253. Ferrari, C. & Pezzi, G. 1999. Spatial pattern analysis of the Mount Prado alpine vegetation (Northern Apennines, Italy). Acknowledgements A landscape approach. - J. Mediter. Ecol. 1: 77-84 Ferrari, C. & Piccoli, F. 1997. The ericaceous dwarfshrublands The research was funded by the Italian Ministry for Uni­ above the Northern Apennine timberline. - Phytocoenologia versity and for Scientific and Technological Research 27: 53-76.

Acta Phytogeogr. Suec. 85 30 C. errariF et al.

Ferrari, C. & Rossi, G. 1995. Relationships between plant com­ northern and central Apennines. - Fitosociologia 26: 5-17. munities andlate snow melting on Mount Prado (Northern Rabinowitz, D. 1981. Seven forms of rarity. In:Synge, H (ed.) Apennines, Italy). - Vegetatio 120: 49-58. The biological aspects of rareplant conservation. - Wiley, Ferrari, C., Bonafede, F. & Alessandrini,A. 1993. Rare plants of New York, pp. 205-217. the Emilia-Romagna region (Northern Italy): a data bank Rossetti, R. 1988. Condizioni termo-pluviometriche del versante and computer mapped atlas for conservation purposes. ­ padano della fascia appenninica tra la valle del torrente Bioi. Conserv. 64: 11-18. Scrivia e quella del torrente Reno. In: Carton, A. & Panizza, Foggi, B. 1990. Analisi fitogeografica del distretto appenninico M. (eds.) ll paesaggio fi sico dell' alto Appennino emiliano. tosco-emiliano. - Webbia 44: 169-196. - Grafis, Bologna, pp. 19-24. Foggi, B. & Rossi, G. 1996. A survey of the genus Festuca L. Rossi, G. 1994 ( 1991). Carta della vegetazione del Monte Prado (Poaceae) in Italy. The species of the summit flora in the (Parco Regionale dell' Alto Appennino Reggiano, Regione Tuscan-Emilian Apennines andApuanAlps. -Willdenowia Emilia-Romagna). Note illustrative.-Atti Ist. Bot. Lab. 26: 183-215. Crittogam. Univ. Pavia 7(10): 3-24. Foggi, B., Rossi, G. & Signorini, M.A. 1999. The Festuca Tomaselli, M. 1991. The snow-bed vegetation in the Northern violacea aggregate (Poaceae) in the Alps and Apennines Apennines. - Vegetatio 94: 177-189. (Central-Southern Europe).- Can. J. Bot. 77: 989-1013. Tomaselli, M. 1994. The vegetation of summitrock faces, talus

Gehu, J. M. & Gehu, J. 1980. Essai d' objectivation de I' evalu­ slopes and grasslands in the northern Apennines. - Fito­ sociologia 26: 35-50. ation biologique des milieux naturels. Exemples littoraux. In Gehu, J.M. (ed.) Seminaire de phytosociologie appliquee.­ Tomaselli, M. & Agostini, N. 1994. A comparative phyto­ Amicale Francophone de Phytosociologie, Metz, pp. 75-94. geographic analysis of the summit flora of the Tuscan­ Gerdol, R. & Tomaselli, M. 1988. Phytosociology and ecology Emilian Apennines and of the Apuan Alps (northern Apen­ of stream vegetation in the summit region of the northern nines). - Fitosociologia 26: 99- 109. Apennines. - Boil. M us. Storia N at. Lunigiana 6-7: 89-93. Tomaselli, M. & Rossi, G. 1994. Phytosociology and ecology of Grabherr, G. 1993 . Loiseleurio - Vaccinietea. In:Grabherr, G. & Caricion curvulae vegetation in the northern Apennines (N Mucina, L. (eds.) Die Pflanzengesellschaften Osterreichs. Italy). - Fitosociologia 26: 51-62. Walker, M.D. 1995. Patterns and causes of Arctic plant commu­ H. NatUrliche waldfreie Vegetation. - Fischer, Jena, pp. 447-467. nity Diversity. In: Chapin, F.S. & Korner, C. (eds) Arctic Losacco, U. 1982. Gli antichi ghiacciai dell' Appennino setten­ and alpine biodiversity: Patterns, causes and ecosystem trionale. Studio morfologico e paleogeografico. - Atti Soc. consequences. - Springer, Berlin, pp. 3-20. Nat. Modena 113: 1-224. Westhof f, V. 1983. Man's attitude towards vegetation. In: Orl6ci,L. 1978. Multivariate analysis in vegetation research. ­ Holzner,W., Werger, M.J.A. & Ikusima, I. (eds.) Man's Junk, The Hague. 451 pp. impact on vegetation. - Junk, Den Haag, pp.7-24. Pignatti, S. 1979. I piani di vegetazione in Italia. - Giorn. Bot. Wildi, 0. & Orl6ci, L. 1996. Numerical exploration of commu­ !tal. 113: 411-428. nity pattern. A guide to the use of MUL V A-5. 2nd ed. - Pignatti, S. 1982. Flora d' Italia. 3 vols. - Edagricole, Bologna. SPB Academic Publishing, The Hague. 71 pp. 2302 pp. Wilson, E.O. 1992. The diversity of life. -The Belknap Press Pignatti, S. 1994. The climax vegetation above timberline in the of Harvard University Press, Cambridge, MA. 406 pp.

Acta Phytogeogr. Suec. 85 Succession and zonation of vegetation in the volcanic mountains of the Hawaiian Islands

Dieter Mueller-Dombois

Department of Botany, University of Hawai'i at Maiinoa, #101-3190 MaiZe Wa y, Honolulu, HI 96822, USA; +18082541873; Fax E-mail [email protected]

Abstract Introduction

Vegetation succession in the Hawaiian Islands is quite This series of contributions deals with vegetation succes­ predictable in its montane rain forest biome up to an early sion and zonation in volcanic mountains. Succession is a stage of soil development. In lowland and seasonal envi­ time concept and zonation a space concept. One must be ronments, and in more advanced stages of soil develop­ careful not to confuse these two concepts, especially ment in rain forest environments, the influx of alien when bringing them together under one theme. However, species often produces changing patterns and processes. succession is most often inferred from the study of side­ An example is the explosive spread of the introduced by-side or space-for-time relationships of vegetation or Myrica fa ya (native to Macaronesia) into a geologically ecosystems, when their dates of origin following cata­ young montane seasonal zone, originally dominated by strophic disturbances are known and when they differ in the native Metrosideros polymorpha tree. In more ad­ an area with the same climatic regime. vanced stages of soil weathering, the dieback dynamics of Regarding these space-for-time relationships, one must M. polymorphaprovides entry of new alien dominants. A remember that time is a continuing dimension, which significant process of change is 'species packing' rather overrides the evolving and thereby varying processes of than 'species extinction', although the latter is certainly nature. This means that even in the absence of unique true for some of the rare species. In the early stages of changes, successional processes do not repeat themselves succession, vegetation zonation is primarily vertical, fol­ in exactly the same way. Yet approximate processes can lowing the outline of lava flows. As the rock surfaces repeat themselves in a 'chronosequence' in which con­ weather into soils, zonation becomes horizontal, thereby temporary differences (in space) are presented as time­ offering an enriched spectrum of habitats for species related events. This must be done with caution and good diversification. In spite of this habitat richness, the Ha­ ecological knowledge of the area. waiian flora did evolve only into ea. 1000 vascular plant Anarea with the sameclimatic regime can be defined as species. This is the principal reason for the vulnerability a vegetation zone or more precisely as a 'climatic vegeta­ of Hawai' i' s native ecosystems to alien species invasion. tion zone'. Other terms for the same zonation type are bioclimatic zone or biome. The distinction is relevant because one can also recognize 'edaphic vegetation zones', Keywords: Alien species; Dry forest; Ecological release; when a widespread substrate type can be recognized by Forest dieback; Invasion; Primary succession; Stability; the vegetation it occupies. The latter is also known as Tropical rain forest; Zonation. 'azonal vegetation'. Both types of vegetation zones, cli­ matic and edaphic, can occur in volcanic mountains, wherever they range from geologically young to old. Such situation applies to the Hawaiian Islands. I will discuss the theme of succession and zonation by summarizing results from a number of studies done in the Hawaiian Islands under four points: ( 1) Succession in rain forest and seasonal environments; (2) Shifting zonation from vertical to horizontal; (3) Alien species as causing changes in succession and zonation patterns; and (4) Acta Phytogeogr. Suec. 85: 31-40. ISBN 91-7210-085-0. Conclusions regarding ecosystem stability.

Acta Phytogeogr. Suec. 85 32 D. Mueller-Dombois

Succession in rrun forest and seasonal rock surfaces become invaded by ferns and seed plants. Most abundant among these are individuals of the sword environments fern Nephrolepis multiflora and the endemic 'ohi'a lehua tree, Metrosideros polymorpha (Myrtaceae). The lichen Since the mountainous Hawaiian Islands are situated in dominates the scene until the seedlings of M. polymorpha the tropical realm of the north-central Pacific and in the reach the young adult stage. This may take 50 -lOO yr. trade-wind zone, they have tropical rain forest environ­ Duringthat time, several pioneer shrub and forb populations ments on their windward NEand seasonally dry environ­ enter the lava fields, some only temporarily. Among the ments on their leeward SW sides. In terms of origin, the temporary shrubs is Dubautia scabra (Asteraceae ), an islands form a geo-chronological sequence. endemic mat-forming short-lived perennial. A few indi­ Located at the SE end of the island chain is theyoung­ viduals of this shrub may enter a new volcanic substrate in est island, Hawai'i, locally called the Big Island. Its one or several localities, then expand rapidly into small substrates range from current or zero to 1 million yr of age cohort populations. After developing a population peak in (Fig. 1). In NW direction the islands become progres­ a few years most individuals become senescent and begin sively older, with Kaua'i, the oldest large island, esti­ to die. A subpopulation or next generation may then mated as 5.1 million yr. The reason for this chrono­ appear in a neighbouring location on the new surface and sequence is a stationary 'hot spot' under the Pacific Plate repeat the growth cycle there. Its dying mats often become through which magma pours onto the plate, while the favourable places for early successional plants. A similar plate moves at a rate of ea. 10 crnlyr in NW direction. ephemeral behaviour is shown by the endemic shrub Primary volcanic succession has been studied on the Rumex giganteus (Polygonaceae). Other pioneer shrubs, Big Island by several researchers (Forbes 1912; Mac­ such as the endemic ericaceous Va ccinium reticulatum, Caughey 1917; Skottsberg 1941; Doty & Mueller-Dombois are longer-lived perennials. They also may swarm over 1966; Atkinson 1970; Eggler 1971; Smathers & Mueller­ only parts of the new surface. A tall fern, an 'almost -tree­ Dombois 1974; Vitousek et al. 1992; Drake & Mueller­ fern' , Sadleria cyatheoides, colonizes by widely scattered Dombois 1993; Stemmermann & Ihle 1993; Aplet & individuals or small groups. Much depends on the disper­ Vitousek 1994; Kitayama et al. 1995; Aplet et al. 1997; sal mechanisms and favourable microhabitats. In all cases, Clarksen 1997). On this island, the five contiguous and woody and herbaceous plant growth develops into only a partially intermingledvolcanic mountains, Kilauea, Manna sparse vegetation cover. That is why the lichen stage can Loa, Hualalei, Manna Kea, and Kohala have chrono­ persist for so many years. sequential surfaces, many with known dates of origin or A next major change occurs when the indigenous mat­ approximate ages. Manna Loa and Kilauea are the present­ forming fernDicr anopteris linearis is spreading itself by day active volcanoes. Both mountains support a number fast-growing stolons underneath the still open-grown and of side-by-side lava flows of different known young to juvenile Metrosideros forest. This fern causes the local older ages (from less than a decade to over a few thousand extinction of the Stereocaulon lichen cover. It also pre­ years) consistingof basaltic rock material of similar chemi­ vents any further establishment of Metrosideros seed­ cal composition but also of similar and different textures, lings. Although this fern typically becomes overwhelm­ such as pahoehoe, 'a'a, cinder, and volcanic ash. ingly dominant as undergrowth cover, a few other plants In some cases real time-series studies have been done usually are associated with it. They include the tall sedge or are under way, but since primary successional se­ Machaerina angustifolia and the clubmoss Lycopodium quences cannot be followed during a person's life time, cernuum. Among shrubs, Vaccinium calycinum may be the chronosequence (space-for-time) approach has been added or replace Va ccinium reticulatum. Another heath used to establish successional trends in the windward rain shrub Styphelia tameiameiae as well as the tall Sadleria forest and leeward dry forest or seasonal environments. fern may persist. Only the most general trends will be summarized here. When the Meterosideros forest develops a closer canopy tree ferns of Cibotium spp. begin to penetrate Primary succession on the windward side through the weakening Dicranopteris mat. The ex­ panding fronds of the tree ferns then aid in the decline In the windward rain forest environment, new lava flows of the light-demanding Dicranopteris. The tree ferns are invaded by cryptogams and higher plants as soon as overgrow the latter and exclude it by becoming the next the rock surfaces cool. Bluegreen algae and mosses ap­ dominant undergrowth layer. At that time also several pear immediately but do not become dominant. Instead a arborescent shrubs and small trees appear among the lichen, Stereocaulon vulcani, often forms a dominant tree ferns. They include members ofthe Aquifo liaceae, whitish-gray cover on the lava-rock surfaces already within Araliaceae, Gesneraceae, Lobeliaceae, Myrsinaceae, five years in the humid lowlands and more slowly in the Rubiaceae and Rutaceae, but no saplings of the myrta­ uplands. Simultaneously some of the cracks among the ceous canopy species Metrosideros polymorpha. Rarely

Acta Phytogeogr. Suec. 85 - Succession and zonation of vegetation in the volcanic mountains of the Hawai'ian Islands - 33

Kure Atoll em 29 M.\ Midway m 4 LEEWARD • 28 M"a .. Laysan11 ISLANDS •20Ma m .. - Gardner Pinnacles 6Bm• Necker Mm � 12Ma • 10Ma .oNihoa 2n m � 7Ma Nl'lhau 390m 4.9 Ma �

Fig. 1. Profile of the HawaiianArchipelago showing island ages, elevations, major moist-wind directions, andgeo-chronosequential 1998). alignment of the island chain from young at the SE end to old at the NW end (after Mueller-Dombois & Fosberg

does any member of the associated woody families join Primary succession on the leeward side the Metrosideros canopy. The mature Hawaiian rain forest thus develops a monodominant canopy in pri­ On the seasonally dry leeward side, succession takes very mary succession together with an almost monodominant much longer. On the northwest slope of Mauna Loa are tree fernsubcanopy. some lava flows that still remain edaphic deserts after ea.

In summary, rain forest development in primary vol­ 3000 years (Vitousek 1997). Older pahoehoe lavas show a canic succession can be visualized in four stages: 1. The sparse grass and shrub growth in the narrow cracks be­ early almost barren or desert-like bluegreen algal stage, tween the pavement-like basaltic rock slabs. Formally this which may last only 5 yr; 2. The Stereocaulon lichen grass was mostly pili grass, Heteropogon contortus. In the stage with sparse herbaceous fern,shrub and Metrosideros 1960s through 1970s it became largely replaced by seedling to sapling development, which may last 100 yr; Rhynchelytrum repens. Then feral goats were removed 3. The Dicranopteris fern stage exhibiting a dense almost from the area and an introduced Mrican savanna grass, complete ground cover under an open juvenile Metro­ Hyparrhenia rufa, became prevalent. This grass species sideros stand; this stage may last another 100 yr; 4. The turnoverwas not a part of primary succession in the past. It mature Metrosideros forest stage with Cibotium tree fern is a secondary successional response to goat release and undergrowth and a number of associated woody plants of alien tall-grass invasion (Mueller-Dombois & Spatz 1975; which few if any join the monodominant tree canopy. Mueller-Dombois 1981). Certain native shrubs however This mature forest stage may be achieved in 200 to 400 yr arecomponents of primarysuccession in the seasonally dry depending on substrate and altitude, i.e. faster on ash or zone on pahoehoe.They include here primarilyOsteomeles ash amended lava flows and more rapid in lowland versus anthyllidifo lia (Rosaceae), Styphelia tameiameiae (Epa­ upland rain fo rest environments. cridaceae), andDodonaea viscosa (Sapindaceae) . Whether If a stand-destroyingvolcanic eruption does not recur, the goats removed native tree species from the aging a fifth stage may be added. This is the canopy dieback pahoehoelava by bark strippingand predation of seedlings stage. Eventually, the mature Metrosideros forest will is still an unanswered question. Trees are found very reach a point in time when its canopy trees begin to rarely on pahoehoe, and then only on mounds with large

senesce. Ifin this stage, theforest is subjected to growth­ cracks. However, native trees grow on neighbouring 'a' a impeding climatic perturbations, its canopy can collapse lava of similar ages, where the rocks have not yet weath­ 'domino-style' over the entire volcanic substrate that gave ered to soil. This indicates that the climate is sufficient for riseto the forest (Akashi & Mueller-Dombois 1995). native tree growth with only 500- 700 mm mean annual

Acta Phytogeogr. Suec. 85 34 D. Mueller-Dombois

rainfall - I emphasize the term native, because the isola­ certain mesic rain forest areas.This treesuccession is in stark tion of the Hawaiian Islands has been a barrier to the contrast to the wetter rain forest in which Metrosideros is successful establishment of certain types of ecological perpetuating itself as the monodominant canopy species species, in particular, climax species. even on the oldest Hawaiian high island, Kaua'i, which has Here, in the seasonally dry environment, the different over 4 million yr old soils. These are deeply weathered and substrate texture and structure of the 'a' a lava provide for deficient in nutrients, particularly phosphorus (Crews et al. a soil moisture regime that allows tree development. In 1995). spite of the low annual rainfall, the occasional prolonged shower wets the 'a'a rock-rubble layer down to the solid Shifting zonation from vertical to rock core which underlies the rubble layer at ea. 1 m depth. horizontal Any water that penetrates below the surface to the solid core cannot escape by evaporation from dry air Vertical zonation moving over the lava flow. Theonly mechanism by which the available filmwater can be withdrawn is via roots of An important peculiarityof theHawaiian vegetation is the living plants. Here trees have the advantage over most wide altitudinal range of a number of native key species. herbaceous plants. The fr rst generation trees that establish Theyinclude among treesMetrosideros polymorpha, Aca­ on 'a'a lava are again Metrosideros polymorpha. cia koa, Sophora chrysophylla, and Myoporum sand­ Tree establishment appears to be a very rare event wicense. Among shrubs, they include Styphelia tameia­ however. But when it happens there seem to be a good meiae, Dodonaea viscosa, Vaccinium reticulatum, and number of individuals forming a cohort. Probably, new Chenopodium oahuense. Among these, Myoporum sand­ Metrosideros seeds are frequently available, while seed­ wicense and Chenopodium oahuense are today restricted ling establishment occurs only in unusually prolonged to coastal and montane/subalpine habitats. They are rarely wet periods, which are very rare in Hawai' i' s leeward found in between. environments. This wide altitudinal distribution of native key species Another tree species, Diospyros sandwicensis (Ebe­ represents a vertical zonation pattern, which is related to naceae), seems to become soon associated. Metrosideros two factors, 'ecological release', and 'edaphic zonation'. has tiny light-weight seeds that are wind dispersed, while Ecological release is the outcome of successful establish­ Diospyros has small hard-coated seeds that are more ment when the early island plant communities offered likely adapted for bird dispersal. These two tree species little or no competition to the advance of aggressive, are joined on older flowsby a good number of other native indigenous invaders. These species therefore spread over tree species so that in the Hawaiian dry-forest succession a wide altitudinal range from near sea level inward of the one can speak of a primary succession to a mixed-species strand zone to the high altitude environment at about 2500 forest. m. The shrubs Styphelia and Vaccinium climbed even Other more common successional dry-forest trees are much higher, where they form major components of the Psydrax odoratum (Rubiaceae), Erythrina sandwicensis alpine heath-scrub below and above 3000 m elevation, (Fabaceae), Myoporum sandwicense (Myoporaceae), while Sophora has replaced Metrosideros forming the Reynoldsiasandwicensis (Araliaceae), Nestegis sandwicensis treeline ecosystem on Manna Kea (Mueller-Dombois & ( Oleaceae ), andXylosma hawaiiense (Flacourtiaceae). How­ Krajina 1968), the older of the two highest (over 4000 m ever, there are usually ea. 20 native tree species in the few a.s.l.) mountains in Hawai'i. mature remnant stands left (Mueller-Dombois & Fosberg On the younger Manna Loa, zonation is typically vertical 1998). After several generations, the pioneering or azonal, following the course of lava flows. The stages Metrosideros becomes only a minor component in in rain forest succession, as indicated by lichen, fern, and Hawai'i's dry forests to eventually disappear in most. woody vegetation, are essentially occurringfrom sea level With soil weathering, trees tend to become more shrub­ to the inversion zone near 2000 m elevation. The only like, and shrubs, such as Dodonaea viscosa, assume preva­ altitudinal differentiation is the rate of primary succes­ lence. Also with soil formation two endemic legume trees sion, which is much faster at low elevation than near the can become successional to Metrosideros, forming the inversion zone. dominant trees in seasonally dry environments. These are The principal reason for the vertical zonation is that Sophora chrysophyllaand Acacia koa. Theirsuccessional the native pioneer flora is limited to a few species. As status has become clarified from studies of kipukas, is­ successful invaders on new volcanic surfaces, they pos­ lands of older vegetation surrounded by more recent lava sess a wide tolerance to the changing climate, in particu­ flows(Mueller -Dombois & Lamoureux 1967). Both range lar the decrease in temperature, along the altitudinal as successional trees from seasonally dry lowlandto upland gradient. enviromnents, andAcacia koaoverlaps withMetrosideros in

Acta Phytogeogr. Suec. 85 - Succession and zonation of vegetation in the volcanic mountains of the Hawai'ian Islands - 35

Horizontal zonation and by differentiating species in the undergrowth. The lowest unit (A1) has gone through a forest dieback on Following primary succession, when lava rocks weather volcanic shield remnants. These shield remnants are fre­ into soils, vegetation zonation becomes more strongly quently flooded and are slowly cut by streams into smaller controlled by climate and geomorphological develop­ topographic segments. On the Haleakala east slope, this is ment. The result is a shift to horizontal zonation as evi­ a zone of habitat diversification, which continues upslope denced by an altitudinal stratification of vegetation pat­ to about 1800 m elevation, where the frequently flooded terns. This shift in zonation from vertical to horizontal positions are more patchy. Above 1200 m, the shield is was also shown in the spatial analysis of the Metrosideros still very much intact. There the Metrosideros trees are rain forest dieback, which is a natural dynamic phenom­ taller and the substrate is more permeable and thus better enon following primary succession to mature rain forest. drained. This improved soil drainage patternoccurs within On the east slope of the geographically young Mauna the cloud-frequented zone up to the inversion zone near Loa, the dieback usually followed the outline of the lava 2000 m. Above the inversion layer, the climate becomes flows. In contrast, on the east slope of the geologically abruptly seasonal (with summer-drought). Here, Sophora older Mauna Kea, the spatial pattern of Metrosideros chrysophylla, together with other shrubs, forms a subalpine forest die back assumed a more patchy horizontal arrange­ scrub in which Metrosideros occurs only as scattered ment (Akashi & Mueller-Dombois 1995). Geomorpho­ remnant trees left over from a former pioneer stage. It still logical development on the east slope of Mauna Kea forms the treeline at 2200 m. shows an enlarging pattern of frequently flooded topo­ Thus, the altitudinal zones on the tradewind side of graphic flats and depressions among its downward flow­ Haleakala are controlled largely by two factors, the ing streams. This pattern, together with the dynamic re­ geomorphological aging process and the changing cli­ sponse of Metrosideros, displays an altitudinally confined mate upslope. These factors produce a horizontal zona­ stage in mountain evolution, in which the volcanic shield tion which is reflected in the undergrowth vegetation by a is being cut into smaller habitats by soil hydrological number of differentiating species, many of which are processes involving bog formation and stream ecosystem ferns.A native vine, Freycinetia arborea (Pandanaceae), evolution (Mueller-Dombois 1998). The same stage in can be used as a key indicator species, separating lowland geomorphological development, interacting with forest from montane rain forest at 1000 m. dieback, is indicated by the lowest segment on the topo­ graphic vegetation profile ofHaleakaHi, the shield-shaped and younger of two volcanic mountains on the island of Maui (Kitayama & Mueller-Dombois 1992, 1994). The topographic vegetation profile (Fig. 2) on the 410 Alien species causing changes in 000 yr old windward slope ofHaleakaHi Mountain on East succession and zonation Maui demonstrates the zonation shift from vertical to horizontal as follows: The leading canopy species, The majority of differentiating species, which character­ Metrosideros polymorpha, is distributed from the low­ ize the forest dieback zone floristically (zone A1 in Fig. land to the treeline at 2200 m. There aretwo overlapping 2), are introduced or alien species. Only two of the differ­ varieties, the pubescent leaved var. incana and the gla­ entiating species are indigenous. The indigenous fern brous leaved var. glaberrima. These varieties have been Odontosoria chinensis is strictly a heliophyte that became recognized as having a successional relationship, the pu­ established inresponse to the loss of Metrosideros canopy bescent being more prevalent in the early pioneer stages, here. The other is the native sedge, Machaerina maris­ the glabrous becoming more prevalent on older, more coides, which is an indicator of both wet soil and canopy developed soil substrates (Mueller-Dombois 1983; openness. In this case, the six alien species strengthen the Stemmermann 1983, 1986; Drake & Mueller-Dombois floristic differentiation of this lowland vegetation zone. 1993; Kitayama & Mueller-Dombois 1995a; Kitayama et The paperbark tree,Melaleuca quinquenervia (Myrtaceae), al. 1995, 1997). In terms of water relations, the pubescent native in New Caledonia and North Queensland, is now form can withstand soil drying to a greater degree than the rapidly spreading through the dieback zone, overlapping glabrous varieties. Thus, the altitudinally wide-ranging with the shrubby recovered growth of Metrosideros Metrosideros canopy tree dominates over a broad vertical (Mueller-Dombois 1988). M. quinquenervia was intro­ (altitudinal) zone. It also dominates over a long-term duced in 1920 with seeds from Florida (Wagner et al. substrate age gradient in the montane rain forest biome 1990), and seedlings were planted in parts of the dieback from the youngest to the oldest high island. zone in the 1930s to replace the forest cover that had Horizontal stratificationis beginning to show (on Fig. deteriorated through the dieback of the native Metrosideros 2) by a decreasing size and canopy structure downslope forest (Cuddihy & Stone 1990).

Acta Phytogeogr. Suec. 85 36 D. Mueller-Dombois

�

Winter ground frost (3"C) � '0 s i � Winter groundfrost (4"C) �ra ctuysophylla t lii - ­ E GeraniumCoProsmacuneetum montana � § Carexwahuensls e' en - Melro$kleros tree --­ Hne at2200m

� B3.Forest line comm. Pelea clusllfolla Asplenium ncrmale" il� Peperomla menilnlnaoea :§: e� Nertera granadensls" '0ID Pepecomlaeseena upeR �& Metroslderos polymO!Jlhav. lncana � a� :!! � �=�:r:rs �a flb::aGiaminltlsrav=mololhooker!" . �03 te• Adenophorus montanus Jhelypl8!fs IIIUldwlcensls Psychotrla hawallenals DI)'OP!erisacutkfene � Syzyglum sandwlcensls Peperomla hlrtipetlola � '0 �ore� �� · >- 'C .!! Sacc:lolepls lndlca•• 11bouchlna f Clklemla hlrtaherbecea"".. � Machaerina maflscoldes ssp. meyenll* 8 0 2.4 km I I 0

Fig. 2. Topographic vegetation profileof wet slope on Haleakala Mt., Island ofMaui (after Kitayama & Mueller-Dombois 1994, slightly modified). The profile demonstrates the shift from vertical zonation, as indicated by the two altitudinally wide-ranging dominant Metrosideros varieties, to horizontal zonation as indicated by the differentiating species, mostly in the undergrowth only. The lowland dieback zone is floristically differentiated primarily by alien species.

The best studied example of an alien species causing nally open Metrosideros forest is now in the process of changes in succession is Myrica fa ya (Myricaceae) becominga closed canopy forest dominated by two canopy (Vitousek et al. 1987; Vitousek & Walker 1989; Mueller­ species, the native Metrosideros and the alien Myrica. Dombois & Whiteaker 1990; Walker & Vitousek 1991; The aggressive behaviour of alien invaders, such as Tunison et al. 1993; Tunison 1998). The species was My rica fa y a, gave the impression that their distributional introduced around the turn of the century from the advances surpass that of the successfulindigenous invad­ Macaronesian Islands to the island of Hawai' i and others. ers. We tested this popular notion on the wet slope of One tree wasnoted in Hawai' i Volcanoes National Parkin Haleakala mountain by asking the question, "do alien the early 1960s (Doty & Mueller-Dombois 1966). By the plant species have wider ecological ranges than native late 1960s, the species began to spread explosively through­ species" (Kitayama & Mueller-Dombois 1995)? The an­ out the geologically young, open submontane seasonal swer turnedout to be negative. environment of the Park. This is one of the six climatic Although successional alien invaders may have an vegetation zones in this geologically young territory. The advantage by being a new ecological species not yet spread of Myricafaya was facilitated by two factors, alien available among the natives, their release in the new birds, particularly the Japanese white eye (Zosterops island environment is usually limited by their past adapta­ japonicus), and the presence of Metrosideros trees. The tion to the climatic and edaphic vegetation zones in their Japanese white eye feeds on the fruits of My ricafaya and home environment. For Myricafaya, this means the species is also attracted by the nectar in the Metrosideros flowers. will not spread from the submontane seasonal environ­ It uses the Metrosideros trees as perch trees and Myrica ment into the warm-tropical lowlands on either the sea­ oftenbecame establishednear the bases of theMetrosideros sonally dry leeward or the year-round moist windward trees. A third factor favoured the rapid establishment and sides. It will also not spread into the subalpine environ­ growth of Myrica. This is its abilityto fix nitrogen with an ment. All of these contain Metrosideros polymorpha in actinorrhizal symbiont (Frankia) living in nodules at­ geologically young areas. Myrica fa ya has also found its tached to Myrica' s feeder roots. Myrica trees grow many limitation in penetrating the geologically older, soil-nitro­ times faster than Metrosideros. In a decade, numerous gen enriched, wet montane rain forest environments. These Myrica trees were able to join the canopy of Metrosideros. are still dominated by Metrosideros forests. Myrica also Subsequently, the established Myrica trees themselves, has not yet invaded new lava flows which so far remain spread their seeds under and neartheir crowns. The origi- the domain of the native Metrosideros polymorpha.

Acta Phytogeogr. Suec. 85 - Succession and zonation of vegetation in the volcanic mountains of the Hawai'ian Islands - 37

In the meantime, a new insect, the two-spotted leaf forest on the lower wet slope of Haleakala. Here, the hopper (Sophonia rufofascia), has arrived from China, dieback was a natural process (Mueller-Dombois 1988), either naturally or by transfer on imported nursery but the invasion of Melaleuca was facilitated by human stock. This new insect has shown a preference for at­ transfer into the area. Similarly in the montane seasonal tacking Myrica faya leaves. The leaf hopper injects a environment of Hawai' i Volcanoes National Park, the poison that resulted in Myrica dieback in the dryer part presence of Myrica fa ya is human induced, but its explo­ of its invasion range and in leaf-margin yellowing and sive expansion is due to a combination of ecological leaf-size reduction in most of the other invasion areas. factors and mechanisms that together amount to a new We anticipate that the 'window of opportunity' has been natural dynamics. curtailed by this new insect pathogen and that the ag­ In terms of the MacArthur-Wilson theory, the distant gressiveness of Myrica is thereby reduced. This aspect is Hawaiian archipelago has become a near-source archi­ currently under investigation. pelago. Its original assemblage of species, in spite of almost 90% secondary enrichment through endemism, probably never attained a real equilibrium under isolation. Conclusions regarding ecosystem The assemblage of 1000 species, although distributed stability over the available habitat spectrum, did not accomplish a full use of available resources. Human-induced distur­ In their theory of island biogeography, MacArthur & bances, such as caused by feral goats and pigs, opened up Wilson ( 1967) presented species equilibrium models that a spectrum of newly available niches. These anthropo­ varied with island size and distance from their biological genic disturbances are added to the natural disturbances, source areas. These models imply, for example, that is­ such as those associated with volcanism,geomor phological land archipelagoes closer to a biological source area, have development, and climatic perturbations. Moreover, once a richer species equilibrium than similarly sized archi­ in a while a new plant species invades that brings its own pelagoes farther removed. The mechanism for the richer set of resources. This applies to My rica fa ya with its species equilibrium would be an accelerated rate of inva­ symbiotic N-fixing ability. This alien tree fitted well into sion balanced against a moderately increased rate of ex­ a geologically young habitat of extremely limited N­ tinction. availability. However, it would not have invaded this The Hawaiian Islands represent a large archipelago, habitat, were it not for Myrica's seed dispersal by intro­ far removed from any source area. Thus, their species duced birds. They in turn were attracted by the nectar in equilibrium, before human contact, was low, allowing for the flowers of that habitats' native key species, Metro­ ecological release of indigenous invaders and consider­ sideros polymorpha. Now this new alien, Myrica faya, able secondaryenrichment through speciation. The number fulfils a successional role. It may even exclude the native of indigenous colonizers of flowering plants are estimated Metrosiderosfrom parts of its habitat spectrum.But Myrica as 272 species (Fosberg 1948). About half of these gave will never bring Metrosideros to the brink of extinction as rise to 850 endemic species, resulting in a current estimate long as a wide dynamic habitat spectrum remains avail­ of ea. 1000 native Hawaiian species (Wagner et al. 1990). able. This includes the formation of new lava flows, The total adventive flora is estimated as ea. 4000 spe­ which may act as refugia for the !ow-N-tolerant Metro­ cies. Of these, 860 species have become naturalized with sideros tree in the future. Most likely, the introduced another 10% of them having assumed a distributional Myrica tree will remain as a naturalized species in the dominance that has affected changes in succession and area, in spite of being attacked by a newly introduced zonation. insect and in spite of continued control efforts by Park Compared to the species richness of similarly sized management. tropical land areas, Hawai'i's inventory of 1000 native Species packing will bring about a more complete species is far below its natural stocking capacity. Al­ assemblage of successional species, so that the simplified though the threat of extinction of rareendemics is real and auto-successional trends will change to successions in­ of great concern, what we witness far more than species volving new key species. Also, the globally increased extinction is the opposite, namely species packing. levels of C02 will most likely accelerate successions by The primary succession from barren lava flow to ma­ increasing therates of plantmetabolism (Mueller -Dombois ture Metrosideros-Cibotium rain forest is still functioning 1992). well in the volcanically active areas on Mauna Loa and The large ecosystems, the biomes, or climatic and Kilauea without much species packing. On older substrates, edaphic vegetation zones, will not change much in the the perpetuation of Metrosideros as the dominant canopy future, unless there is a real climate change as predicted species through auto-succession in form of chronological by some. Instead, however, the distributional dynamics in monocultures, is declining. An example is the influx of the Hawaiian vegetation (old and new) is such that zonal Melaleuca after landscape-level dieback of the native key indicator species will change (Mueller-Dombois

Acta Phytogeogr. Suec. 85 38 D. Mueller-Dombois

1992b ). Two alien species, Prosopis pallida and Leucaena Hawai'i. - Occas. Paper B.P. Bishop Mus. 5: 15-23 . leucocephala have been identified as key indicator species Fosberg, F.R. 1948. Derivation of the flora of the Hawaiian by Krajina (1963), for dry lowland zones. Egler (1939) Islands. In: Zimmermann, E.C. (ed.) Insects of Hawai'i. called the submontane rain forest zone on 0' ahu, the Vol. 1.-University of Hawai'i Press, Honolulu, pp. 107- Psidium guayava zone. Today, it is much more a mixed­ llO. Kitayama, K. & Mueller-Dombois, D. 1992. Vegetation of the species zone filled with a number of introduced fast­ wet windward slope of Haleakala, Maui, Hawai'i. - Pac. growing secondary tropical tree species, none of them Sci. 46: 197-220. native (Mueller-Dombois & Fosberg 1998). However, a Kitayama, K. & Mueller-Dombois, D. 1994. An altitudinal native Acacia koa zone is still evident and overlapping transect analysis of the windward vegetation on Haleakala, a with the former guava zone, and a Metrosideros zone is Hawaiian islandmountain: (2) vegetation zonation.- Phyto­ still evident throughoutthe high islands in the wet montane coenologia 24: 135-154. environments. Major conservation efforts are underway Kitayama, K. & Mueller-Dombois, D. 1995a. Vegetation changes to understand and maintain Hawai'i's native ecosystems along gradients of long-term soil development in the Ha­ through intensified ecological research and feedback to wai'ian montane rainforest zone. -Vegetatio 120: 1-20. Kitayama, K. & Mueller-Dombois, D. 1995b. Biological inva­ and fromnatural area management. sion on an oceanic island mountain: Do alien plant species have wider ecological ranges than native species? - J. Veg. Sci. 6: 667-67 4.

Kitayama, K., Mueller-Dombois, D. & Vitousek, P.M. 1995. Pri­ References mary succession of the Hawaiian montane rain forest on a chronosequence of eight lava flows. - J. Veg. Sci. 6: 21 1-222. Akashi, Y. & Mueller-Dombois, D. 1995. A landscape perspec­ Kitayama, K., Pattison, R., Cordell, S., Webb, D. & Mueller­ tive of the Hawaiian rainforest dieback. - J. Veg. Sci. 6: Dombois, D. 1997. Ecological and genetic implications of 449-464. foliar polymorphism in Metrosideros polymorpha Gaud. Aplet, G.H. & Vitousek, P.M. 1994. An age-elevation matrix (Myrtaceae) in a habitat matrix on Mauna Loa, Hawai'i.­ analysis of Hawaiian rainforest succession. - J. Ecol. 82: Ann. Bot. 80: 491-497. 137-147. Krajina, V.J. 1963. Biogeoclimatic zones of the Hawaiian Is­ Aplet, G.H., Hughes, R.F., & Vitousek, P.M. 1997. Ecosystem lands. -Hawaiian Bot. Soc. Newslet. 2: 93-98. development on Hawaiian lava flows: biomass and species MacArthur, R.H. & Wilson, B.O. 1967. The theory of island composition.-J. Veg. Sci. 9: 17-26. biogeography. -Princeton University Press, Princeton, NJ, Atkinson, I.A.E. 1970. Successional trends in the coastal and 203 pp. lowland forest of Mauna Loa and Kilauea Volcanoes, Ha­ MacCaughy, V. 1917. Vegetation of Hawaiian lava flows. ­ wai'i. - Pac. Sci. 24: 387-400. Bot. Gaz. 64: 386-420. Clarksen, B.D. 1997. Vegetation succession (1967-89) on five Mueller-Dombois, D. 1981. Vegetation dynamics in a coastal recent montane lava flows,Mauna Loa, Hawai'i. - N. Z. J. grassland of Hawai'i.- Vegetatio 46: 131-140. Ecol. 23 : 1-9. Mueller-Dombois, D. 1983. Canopy dieback and successional Crews, T.E., Kitayama, K., Fownes, J.H., Riley, R.H., Herbert, processes in Pacific forests. - Pac. Sci. 37: 317-325. D.H., Mueller-Dombois, D. & Vitousek, P.M. 1995. Changes Mueller-Dombois, D. 1988. Vegetation dynamics and slope in soil phosphorus fractions and ecosystem dynamics across management on the mountains of the Hawaiian Islands. - a long soil chronosequence in Hawai'i. - Ecology 76: Environ. Conserv. 15: 255-260. 1407-1424. Mueller-Dombois, D. 1992a. Potential effects of the increase in Cuddihy, L.W. & Stone, C.P. 1990. Alteration of native Ha­ carbon dioxide and climate change on the dynamics of waiian vegetation: effects of humans, their activities and vegetation. - Water Air Soil Pollut. 64: 61-79. introductions. - University of Hawai 'i Press, Honolulu, Mueller-Dombois, D. 1992b. Distributional dynamics in the 138 pp. Hawaiian vegetation. - Pac. Sci. 46: 221 -23 1. Doty, M.S. & Mueller-Dombois, D. 1966. Atlas for bioecology Mueller-Dombois, D. In press. Regenwalddynamik und Land­ studies in Hawai'i Volcanoes National Park. - Hawai'i schaftsentwicklung auf den Hawai'i Inseln. - Naturwiss. Botanical Science Paper No. 2, University of Hawai' i, Hono­ Rundsch. lulu, 507 pp. Mueller-Dombois, D. & Fosberg, F.R. 1998. Vegetation of the Drake, D.R. & Mueller-Dombois, D. 1993. Population develop­ tropical Pacific islands. - Ecological Studies Vol. 132. ment of rainforest trees on a chronosequence of Hawaiian Springer-Verlag, NY, 733 pp. lava flows. - Ecology 74: 1012-1019. Mueller-Dombois, D. & Krajina, V .J. 1968. Comparison of east­ Eggler, W.A. 1971. Quantitative studies ofvegetation on sixteen flank vegetations on Mauna Loa and Mauna Kea, Hawai'i. young lava flows on the island of Hawai'i. - Trop. Ecol. - Rec. Adv. Trop . Ecol. 2: 508-520. 12: 66-100. Mueller-Dombois, D. & Lamoureux, C.H. 1967. Soil-vegeta­ Egler, F .E. 1939. Vegetation zones of Oahu, Hawai'i. -Empire tion relationships in Hawaiian Kipukas. - Pac. Sci. 21: For. J. 18: 44-57. 286-299. Forbes, C.N. 1912. Preliminary observations concerning the Mueller-Dombois, D. & Spatz, G. 1975. The influence of feral plant invasion on some of the lava flows of Mauna Loa, goats on the lowland vegetation in Hawai'i Volcanoes Na-

Acta Phytogeogr. Suec. 85 - Succession and zonation of vegetation in the volcanic mountains of the Hawai'ian Islands - 39

tional Park. - Phytocoenologia 3: 1-29. Tunison, J.T., Castro, L.F. &Loh, R.K. 1998. Faya tree dieback: Mueller-Dombois, D. & Whiteaker, L.D. 1990. Plants associ­ distribution, demography, and associated ecological factors. ated with Myricafay a and two other pioneer trees on a recent - Cooperative National Park Resources Studies Unit, Uni­ volcanic surface in Hawai'i Volcanoes National Park. - versity of Hawai'i at Manoa Tech. Rept. 122, 15 pp. Phytocoenologia 19: 29-41. Vitousek, P.M. 1997. After the volcano. - Nat. Hist. 106: 48- Skottsberg, C. 1941. Plant succession on recent lava flows in the 53. Island of Hawai'i. - Goteb. K. Vetensk.Vitterhets. Samh. Vitousek, P.M. & Walker, R.L. 1989. Biological invasion of Handl. Foljden. Ser. B. 1-8. Myrica faya in Hawai'i: plant demography, nitrogen fixa­ Smathers, G.A. & Mueller-Dombois, D. 1974. Invasion and tion, and ecosystem effects. - Ecol. Monogr. 59: 247-265. recovery of vegetation after a volcanic eruption in Hawai'i. Vitousek, P.M., Aplet, G., Turner, D. & Lockwood, J.P. 1992. -National Park Service Scientific Monograph Series No. The Mauna Loa environmental matrix: foliar and soil nutri­ 5, 129 pp. ents. - Oecologia (Berl.) 89: 372-382. Stemmermann, L. 1983 . Ecological studies of Hawai'ian Vitousek, P.M., Walker, L.R., Whiteaker, L.D., Mueller­ Dombois, D. & Matson, P.A. 1987. Biological invasion of Metrosideros ina successional context. - Pac. Sci. 3 7: 361- 373. Myrica faya alters ecosystem development in Hawai'i. - 138: 802-804. Stemmennann, L. 1986. Ecological studies of 'ahi'a varieties Science (Metrosideros polymorpha, Myrtaceae), the dominants in Wagner, W.L., Herbst, D.R. & Sohmer, S.H. 1990. Manual of successional communities. - Ph.D. Dissertation, Univer­ the flowering plants of Hawai'i, 2 vols. - Univ. of Hawai'i sity of Hawai'i at Manoa, Honolulu. Press & Bishop Museum Press, Honolulu, 1853 pp. Tunison, J.T. 1998. Management and ecology of faya tree, Walker, L.R. & Vitousek, P .M. 1991. Interaction of an alien and Hawai'i Volcanoes National Park: Introduction; conclu­ a native tree during primary succession in Hawai'i Volca­ sions and recommendations. - Tech. Rept.,Cooperative noes National Park. - Ecology 72: 1449-1455. National Park Resources Studies Unit, University ofHawai 'i at Manoa, pp. 117-127.

Acta Phytogeogr. Suec. 85 Acta Phytogeogr. Suec. 85 Geographical determinants of the biological richness in the Macaronesian region

Jose Mar ia Ferndndez-Palacios1 * & Christian Andersson2

1Departamento de Ecologfa, Un iversidad de La Laguna, E-38206 Tenerife, Canary Islands, Spain; 2Baltic UniversityProgramme, Upp sala University, P.O. Box 2109, SE-750 02 Uppsala, Sweden; Present address: Departmentof Botany, Norwegian University of Science and Technology, Realfagbygget, N- 74 91 Trondheim, No rway; +349 22318311; *Corresponding author; Fax E-mail jmferpal@ ull. es

Abstract Introduction

Several geographicalfactors determinethe biological rich­ A survey of literature dealing with island biogeography ness of oceanic islands and archipelagos. To investigate and ecology (e.g. MacArthur & Wilson 1967; Carlquist the importance of these factors we used data on flowering 197 4; Gorman 1979; Williamson 1981; Mueller-Dombois plants, ferns, land birds, beetles and butterflies of the et al. 1981; Menard 1986; Whittaker 1998) led to the idea native biota of theMacaronesian region (the archipelagos that different geographical determinants can be claimed of the Azores, Madeira, the Canaries and Cape Verde). to play roles in controlling the biological richness of The fivetaxonomic groups vary in long-distance disper­ oceanic islands or archipelagos. Among the suggested sal ability. geographical determinants, we have chosen five major Three different analyses varying in level of approach features of islands: (1) Latitude; (2) Area; (3) Height were carried out: a Macaronesian or between-archipela­ a.s.l.; (4) Degree of isolation and (5) Geological age. gos approach, a Canarian or within-archipelago approach Island latitude influences the biological richness as it and finally a multiple island approach. It shows that at the controls the macroclimatic features of the island or is­ first level age and isolation are important factors in deter­ lands group, basically in terms of annual temperature mining the richness of the groups with low dispersal range and water availability. Thus, in the same way as ability, whereas area and height better predict the richness boreal or temperate forests are less rich than tropical ones, of taxonomic groups with good long-distance dispersal tropical islands (e.g. Hawai'i) are expected to have biotas ability. At the within-archipelago level, area and habitat richer than temperate (e.g. Azores) or boreal (e.g. Iceland) diversity usually were the most importantfactors determin­ islands (Mueller-Dombois 1992). Furthermore, the geo­ ing the biological richness, although some factors related graphical location of an island will determine, due to the with features of the nearest neighbouring island also play a general wind and marine current regimes of the zone, the role in explaining the richness of taxonomic groups with origin of the bulk of its biota and the pace of species good dispersal ability. Finally, two sets of islands that arrivalto the island. varied in area but not in habitat diversity and vice versa The role of the island area in controlling its species were selected and analysed to compare the 'area per se' number has been stressed by several authors (Arrhenius and the 'habitat diversity' hypotheses. The results do not 1921; Gleason 1922; Preston 1962).The importance of support the latter hypothesis. island area is included in two of the three main theories in island biogeography:

Keywords: Bird; Canaries; Insect; Island age; Species • The 'Random placement' or 'Null hypothesis theory' richness; Vascular plant. (Connor & McCoy 1979; Coleman et al. 1982) points out the importance of island areain relation to the likelihood

Abbreviations: ANI = area of nearest neighbouring is­ of colonization events of species from the continental land; DNI = distance to nearestneighbouring island; HD species pool.

= habitat diversity; HNI = height of nearest neighbouring • The 'Equilibrium theory' (MacArthur & Wilson 1963, island. 1967) highlights the roleof area for the size of the island populations involved and thus, for the likelihood of ex­ tinction events. Both theories, although arguing from different points of views, thus agree that richer biotas are found on larger Acta Phytogeogr. Suec. 85: 41-50. ISBN 91-7210-085-0. islands when compared to smaller ones.

Acta Phytogeogr. Suec. 85 42 J.M. Fernandez-Palacios & C. Andersson

The direct relation between height of the island and ries may be explained by an indiscriminate use of data to the number of different habitats that it can have is a major support them. Data have been obtained from real islands geographical factor explaining the richness of insular or archipelagos with different latitudinal location, origin biotas in the 'Habitat diversity hypothesis' (Williams (oceanic or continental), size, age or isolation, sometimes 1964). The higher the island, the more zonal ecosystems even from 'pseudo-islands' (as varied as mountain tops, (with an altitudinal distribution) exist on the island and habitat islands, ponds and inflorescences). Such data have thus, the larger are the differences in the environmental almost always involved different taxonomic groups and conditions to be exploited. Height repeats on the island the levels. Some authors (e.g. Connor & McCoy 1979) who latitudinal climatic variation, enabling the development, have tried to recompile available data to analyse them in although with some differences, of the latitudinal zonation the light of the different theories usually finished without of terrestrial vegetation. Moreover, altitude, together with clear conclusions. wind regime causes the island to develop well differenti­ The aim of our work has been to find more reliable ated slope types (windward and leeward) and can thus data and to investigate the relationships of geographical increase the habitat diversity and species number. and biological factors within oceanic archipelagos. The Isolation, expressed as the water gap existing between archipelagos within the Macaronesian region have a well­ the island and the nearest continent, is also claimed to be known flora and fauna and thus provide a good opportu­ a maj or determinant of the biological richness of an nity to compare how the selected geographic determi­ island by the 'Equilibrium theory'. The isolation will nants influence their species number. In particular we determine which species can reach an island, a large have studied the number of species in different taxonomic fraction of the continental species pool when the dis­ groups varying in dispersal ability in relation to a wide tance is short, or only species with good long distance range of latitude, area, height, isolation and age of the dispersal ability when it is large. Some classic works islands (cf. Table 1). (e.g. Firth & Davidson 1945; Carlquist 1965) state that there is a decrease in the number of species in a given taxonomic group in relation to the islands' distance to the Study area continent. However, isolation may also increase the bio­ logical richness of distant islands, as the non-existence The Macaronesian region (Fig. 1 and Table 1) is com­ of genetic interchanges between the continental and posed of four Atlantic archipelagos; The Azores, Madeira, island populations could give rise to speciation proc­ The Canaries and Cape Verde, located off the European esses that increase the island's biota. and African mainlands. The archipelagos embrace the lati­ Finally, geological age of an island has to be an tudes between 15° N (Cape Verde) and 40° N (Azores). important factor explaining its biological richness. Al­ Their degree of isolation, i.e. the distance from the nearest though this is not claimed in the classical theories deal­ continent, varies from 95 kmbetween theCanaries andthe ing with island biogeography. First, the likelihood of a Saharan coast to 1 450 km between the Azores and the species to arrive at an island is related to the island's age Portuguesian coast. The Canaries form the largest archi­ as it may be expected that a larger number of events have pelago (7480 km2) with the highest peak (3718 m on occurred on older islands than on younger. Second, Tenerife), but the Azores (235 1 m on Pica) and Cape younger islands usually acquire their biota from older Verde (2835 m on Fogo) are quite high too. ones. On the other hand and similar to the effect of isolation, the age of an island plays a role in the speciation processes, because the older the island the further has the speciation process on the island proceeded. Finally, the Table 1. Geographical (number of islands, latitude, area, height, erosion processes also influence the biota because they isolation and age) and biological (number of native vascular increase the actual area (not the projected area) of an plant species, land birds and Te nebrionidae beetles) features of island by creating new azonal habitats (such as ravines the Macaronesian archipelagos Azores (Az), Madeira (Ma), or cliffs) which tend to increase the richness of the biota. Canary Islands (Ca) and Cape Verde (CV). Other geographical determinants, such as distance to Az Ma Ca CV the nearest island, area of the nearest island, height of the No. of islands km2) 9 11 10 nearest island, distance to the nearest larger island, distance (> I 3 Latitude (0N) 37 -40 32-33 28-29 15-17 to the nearest older island, and so on, have been used in Area 2388 815 3580 (krn2) 7447 different biogeographical analyses (Johnson & Simberloff Height (m a.s.l.) 2351 1846 3718 2835 Isolation 1450 540 500 1974; Connor & Simberloff 1978; Nilsson et al. 1988). (km) 95 Age (mH!ion yr) 8.1 13.5 20.5 10.3 However, these factors are combinations of the main ones No. of vascular plants 300 750 1260 560 and may have a significanceonly for a single archipelago. No. of land birds 27 27 58 32 The existence of different island biogeography theo- No. of beetles 7 27 113 28

Acta Phytogeogr. Suec. 85 Geographical detenninants of the biological richness in the Ma caronesian region 43

AL 0 -- MC � 500 km EUROPE 0 50 100 km

PALMA LA UMUN>f£ r;!! TENERIFE tJW \} FUERTEVEN LA GOMERA \) AZORES 0 MADEIRA 0 EL HlERRO .... r::J GRAN CANARIA AFRICA

Fig. 2. Location of the CanaryIslands. AL= Alegranza, GR =La CANARY ISLANDS Graciosa, MC = Montana Clara, LO = Lobos.

AFRICA The ages of the archipelagos vary fromea. 20 million ... . �- yr for the Canaries (Coello et al. 1992) to ea. 8 million yr CAPE VERDE for the Azores (Ridley et al. 1974). The Cape Verde islands, traditionally considered as the oldest archipelago of theregion (due to findingsof rocks dated more than 100 Fig. 1. Location of the Macaronesian region. million yr old) are probably younger than the Canaries, nowadays considered to be between 8 and 10 million yr old (Bernard-Griffiths et al. 1975). Historic volcanism The volcanic origin of the archipelagos (from the European colonization onwards, i.e. the last 500 yr) has been present in the Azores (Sao Miguel, Terceira, The Macaronesian archipelagos share their volcanic ori­ Sao Jorge, Pico y Faial), Canaries (Lanzarote, Tenerife

gin. All the islands can be considered as oceanic, that is, and La Palma) and Cape Verde (Pogo) (Baez & Sanchez­ they have emerged after successive submarine eruptions Pinto 1983). This century eruptions include Capelinhos of basic magma (mainly basalts ), through different ocean (1957) on Faial, Chinyero (1909) on Tenerife, San Juan crust fractures. Some islands, notably Lanzarote and (1949) and Tenegufa (1971) on La Palma as well as Pico Fuerteventura (Canaries), are located at the transition do Pogo (1951, 1995) on Pogo. zone between the continental and oceanic crusts, because Macaronesia was first considered a biogeographical of their proximity to the Africancontinent; here the ascend­ regional unit due to the many common geographical and ing magma carried also fragments of sedimentary rocks biological characteristics. This has been accepted for more belonging to theAf rican continental margin. than a century (Sunding 1979). However, some authors Anyhow, the origin of all the archipelagos can be have recently highlighted the floristicheterogeneity within understood as a consequence of the North Atlantic inter­ the region (Lobin 1982; Nicolas et al. 1989) questioning nal geodynamics, with magma emitted mainly through the inclusion of Cape Verde and Azores in the same the Mid-Atlantic Ridge, but also through fracture zones biogeographical region. and transforming faults, since its opening almost 180 The Canary Islands (Fig. 2) arelocated in the eastern million yr ago. Recent oceanographic research of the part of the North Atlantic Ocean (28°N, 16°W), 95 km Atlantic floors has revealed that the ocean floor is moving from Punta Stafford on the Saharian coast. The archi­ away from the central ridge in both directions at a speed of pelago includes seven main islands - Tenerife, Fuerte­ 1-2 cm/yr. This means that the westernmostislands of the ventura, Gran Canaria, Lanzarote, La Palma, La Gomera Azores, Flares and Corvo, actually located on the Ameri­ and El Hierro- and four islets - La Graciosa, Alegranza, can tectonic plate, are being separated from the rest of the Lobos and Montana Clara. Although some earlier authors archipelago. (see Schmincke 1976) considered the easternmostislands With the exception of the Azores, the archipelagos are of the Canarian archipelago - Lanzarote, Fuerteventura included in the African tectonic plate, which implies that and the islets - as continental, nowadays the whole archi­ they have a weaker seismic and volcanic activity. On the pelago is regarded as oceanic (Banda et al. 1981). The other hand, the Azores, located at the NW edge of the central and western islands are separated from each other African plate and at both sides of the Mid Atlantic Ridge by ea. 3000 m deep water, whereas the eastern ones are show a high seismic and volcanic activity -during recent separated by relatively shallow waters <( 200 m); earlier years, some very destructive earthquakes have occurred they have formed one single island, the Eastern Canarian here (Garcfa-Talavera 1999). Ridge (Coello et al. 1992). No connection is known be-

Acta Phytogeogr. Suec. 85 44 J.M. Ferndndez-Palacios & C. Andersson

tween the archipelago and the African mainland, the distributed. On the other hand, isolation may be ignored water depth is about 1000 m. within an archipelago, as the distances between islands Within the Macaronesian region, the Canary islands withinan archipelago usually are shorter than thedistance were selected for an archipelago-level analysis due to the to the mainland. However, the possible role of the water­ high variability shown by the geographical detenninants gap between islands is analysed at the archipelago level. between the islands: (1) areas differing between 2034 km2 2. The second level is the 'archipelago approach', for Tenerife (the largest island ofMacaronesia) to 1.3 km2 which was followed for data from the Canaries. At this for Montana Clara; (2) heights varying from 3718 m level we analysed the influence of area, height and age and (Teide Peak on Tenerife, the highest point in the Atlantic also other factors such as habitat diversity, number of zonal and only surpassed by the Hawaiian peaks on the global ecosystems found on the island (sensu Humphries 1979), scale) and 122 m for Lobos; (3) isolations ranging from 95 area of the nearestisland (ANI),height of thenearest island kmfor Fuerteventurato 424 kmfor La Palma; (4) and ages (HNI) and distance to the nearest island (DNI) (Table 2). varying from ea. 20 million yr for the Eastern Canarian We also carried out three different correlation analy­ Ridge(Lanzarote and Fuerteventura) to several thousands ses. First, the geographical detenninants were subjected for the islets. to a correlation analysis to detect covariation within the Furthermore, the Canaries have by far the richest biota data set. Secondly, the richness of the different taxa were of the region, a biota closely related to the Mediterranean subjected to the same approach. Finally, single regres­ andSaharan ones. However, certain spectacular affinities sions of biological richness (dependent variables) to geo­ to South or East African and South American plant genera graphical factors (independent variables) were calculated (Sunding 1979) or Indo-Burman insect genera (B aez 1984) for the whole data set using four different regression are found. The proportion of endemic species varies be­ models: linear, logarithmic, exponential and power. tween the different taxonomic groups, from 100% for 3. A multiple comparison between single islands be­ reptiles to ea. 50% for vascular plants to low or non­ longing to the same or different archipelagos was also existent for groups with long-distance dispersal ability carried out. The aim was to analyse the influence of some such as land birds or winged insects. factors on islands similar to one another. It has been The fi rst data on island sizes and number of species for suggested (Abbott 1980; van der Werff 1983) that islands the Canary Islands were published by von Buch (1825). within an archipelago with the same number of habitats However, it was not until thebeginning of the 1960s that but different size, or vice versa, could be used to obtain Hemmingsen (1963) plotted the ftrst species-area curve information on the 'Area per se' versus 'Habitat diver­ for the archipelago. He included also Madeira and used sity' controversy (Simberloff 1974; Abbott 1980). data collected by Volsoe (1955) on breeding birds. First we compared islands within an archipelago -the Simberloff ( 1970) used the same data to calculate the eastern Canarian Islands and islets: Montana Clara (MC), taxonomic diversity of the Canary Islands' avifauna. Lobos, Alegranza, La Graciosa, Lanzaroteand Fuerteventura Connor & McCoy (1979) used data by Lems (1960) on (FV).These islands vary in area but not in habitat diversity, vascular plants and Simberloff s data ( 1970) on avifauna latitude and isolation. Theyhave only one habitat (subdesert to calculate correlation coefficients, slopes andintercepts scrub) but their areas range from 1 km2 (MC) to 1725 km2 for the species-area relationship using different regres­ (FV). A second comparison was carried out on islands with sion models. In recent years, different works dealing with similar areas but varying in height and thus in habitat general biogeographical aspects concerning the Canary diversity: Lanzarote (796 km2), Madeira (728 km2) and La Islands have appeared (Baez 1987; de Nicolas et al. 1989; Palma (729 km2), withhabitats: one, threeand ftve respec­ Becker 1992; Fernandez-Palacios & Andersson 1993). tively. For the selection of taxonomic groups we used two criteria: (1) the data should be reliable and update and (2) Methods the taxonomic groups should vary in long-distance dis­ persal ability. Five groups were found to fulfil these To analyse the geographical detenninants that control the requirements and were thus selected: ferns, flowering biological richness we used three different levels of scale: plants, land birds, beetles (Tenebrionidae) and butterflies. 1. The first level represents the 'Macaronesian ap­ Hansen & Sunding's (1985) checklist for both ferns proach' where we analysed the archipelagos as single and flowering plants of the Macaronesian flora was used entities and paid special attention to the consequences of for the Macaronesian analysis. The archipelago (Canarian) variation in latitude and isolation. Latitudinal variationhas analyses were instead carried out using the more recent little importance within an archipelago when the difference work by Kunkel (1993), that also includes data for the between the northernmostand the southernmost islands is islets. The data on flowering plants were also combined to only a few degrees. This is the case for the Macaronesian allow an analysis on as many islands as possible through­ archipelagos, because they are also mainly longitudinally out Macaronesia;in total data for 34 islandswere available.

Acta Phytogeogr. Suec. 85 Geographical detenninants of the biological richness in the Macaronesian region 45

Table 2. Geographical and biological features of the Canary Islands and Madeira. HD = habitat diversity; ANI = area of nearest neighbouring island; HNI = height of nearest neighbouring island; DNI = distance to nearest neighbouring island. Species richness, i.e. the number of species is indicated for ferns (fern), flowering plants (flow), land birds (bird), beetles (beet) and butterflies (butt); - =no data available.

HD ANI DNI fern flow bird beet butt Island area height isolation age HNI (m) (106 yr) (m a.s.l.) (km2) (km) (km2) (km) Tenerife 2034 3,718 292 11.9 6 370 1,487 30 41 804 49 47 24 Fuerteventura 1655 807 95 20.5 1 4 122 2 14 406 33 48 12 Gran Canaria 1560 1950 200 14.5 4 2034 3718 58 40 745 45 47 19 Lanzarote 807 67 1 132 15.5 1 27 266 1 12 382 32 34 9 La Palma 708 2423 424 1.5 5 370 1487 56 39 501 36 11 21 La Gomera 370 1487 340 12.0 3 2034 3718 30 33 518 37 24 22 El Hierro 269 1501 388 0.8 4 370 1487 63 24 402 30 15 14 La Graciosa 27.5 266 153 807 671 1 0 116 20 Alegranza 10.2 289 167 1.3 256 7 0 81 17 11 Lobos 4.4 122 122 1655 807 2 1 113 15 9 Montana Clara 1.3 256 162 27 266 2 0 62 10 3 Madeira 728 1846 540 5.0 3 69 520 40 73 711 27 37 15

The Canarianchecklist by Bacallado & Dominguez (1984) nant for beetles, whereas area only was found tobe a good on breeding land birds was completed for the Azores and predictor for land-birds richness. However, the analysis Cape Verde islands with Le Grand's works (1984, 1986) of the 34 islands showed that area and height, but not age, and for Madeira with Jones et al. (1987). Data of the were significant predictors of their richness. Therefore, Canarian islets, not included inthe checklist were obtained number of islands and latitude were the only factors not from Martin(pers. comm. ). We excluded sea birds fromthe found to be significant for any taxonomic group at the analyses. The works by Oromi ( 1982a, b) provided data on Macaronesian level. Canarian and Macaronesianbeetles ; these data were com­ As expected, the age of the archipelagos is correlated pleted for theCanarian islets (Oromi pers. comm.). Finally, with groups with large proportions of endemic species as the data from Baez (1984) on butterflieswere used only in time is a major requirement for speciation events. On the the Canarian analysis. other hand, age was not a significant factor for land-bird In general, the proportion of endemic species are low richness, a group with a low endemic rate in Macaronesia for groups having long-distance dispersal ability (ferns, andgenerally considered as good dispersors. Isolation could land birds and butterflies), and relatively high for plants: only explain beetles richness (high on the near-to-continent ea. 50% for the Canaries, beetles: ea. 50% for Madeira Canariesand Madeira and low on the far-to-continentCape and Cape Verde and up to 80% for the Canaries. Finally, Verde and Azores); this is reasonable if it is considered that it was assumed that the species lists include an error this is the group with the lowest long-dispersal ability source. However, the level of knowledge on the analysed amongst thestudied groups. We showed inan earlier study taxa in Macaronesia is great enough that one need not to (Fernandez-Palacios & Andersson 1993) that for the Ca­ consider it likely that the species number on each island is naries taxonomic groups which are assumed to have good better correlatedwith the number of scientific expeditions long-distance dispersal abilities as land-birds show a than to any geographical variable, as seems to occur for stochastic colonization patternin the archipelago,whereas instance in the Galapagos (Connor & Sirnberloff 1978). a deterministic pattern was found in groups with poor ability for long-distance dispersal. The analysis of the 34 islands, performed exclusively Results and Discussion for vascular plants, highlights the importance of both area and height of the island. Nevertheless, as seen above, both At the Macaronesian level, the only significant correla­ factors are correlated and it is not possible to distinguish tion identified between the geographical determinants which of them is ultimately responsible. Thus, specific was that between areaand height of the islands (Table 3a). analyses were performed, where variation in island area Table 4 shows the results of the Macaronesian approach, was not connected with variation in island height or vice where latitude, area, height, isolation and age of the four versa. archipelagos were correlated with number of vascular Finally, the fact that latitude is not found to be signifi­ plants, land-birds and beetle species. A further analysis of cant in any taxonomic group, may be attributed either to the 34 islands was carriedout, but only for vascular plants. its lack of explaining biological richness on islands, or to At this level, the age of the archipelago is a significant the latitudinal variation within Macaronesia {25°) as not predictor for vascular plants and beetles, but not for land­ being large enough to control the pattern of variation in birds. Moreover, isolation was also a significant determi- biological richness.

Acta Phytogeogr. Suec. 85 46 J.M. Ferndndez-Palacios & C. Andersson

Table 3a. Matrix of linear correlations between the independent variables throughout Macaronesia. b. Matrix of linear correlations

between the independent variables in the Canary islands. HD=h abitat diversity; ANI area= of nearest neighbouring island; HNI= height of nearestneighbouring island; DNI =distance to nearest neighbouring island. Only significant values are shown; n.s. = non-significant;

* = p < 0.05; ** = p < 0.01. a. b. Area n.s. Height 0.71 * Height n.s. 0.99** Isolation n.s. 0.67* Isolation n.s. n.s. n.s. Age n.s. n.s. -0.91 ** Age n.s. n.s. n.s. n.s. HD n.s. 0.96** 0.80** n.s. Latitude Area Height Isolation ANI n.s. n.s. n.s. n.s. n.s. HNI n.s. n.s. n.s. n.s. n.s. 0.8 1 ** DNI n.s. 0.68* 0.82** -0.78* 0.83 ** n.s. 0.68 * Area Height Isolation Age HD ANI HNI

When the multiple correlation analysis of the geo­ taxonomic richness vs. geographical features at the within graphical features in the Canaries was performed (Table archipelago level. As much as the 50% of all possible 3b), it showed that more than half of the combinations combinations are significantdescriptors of the richness of were correlated. Some of the correlations between geo­ any taxonomic group. From them, area has been found to graphical variables at the archipelago level, could be be the best descriptor for the richness of vascular plants, interpreted as generalfeatures for islands and archipelagos, land-birds and beetles. On the other hand, area was not a such as Area-Height, Height-HD or ANI-HNI. Othercorre­ significantfactor for ferns,where DNI was found to be the lations seem to be specific for the archipelago due to its best predictor, height was a better predictor for richness of shape. The correlations between DNI-HNI, DNI-HD and butterflies. Moreover, height was the single factor signifi­ Isolation-HDare likelyto be aneffect of the existence of an cant for all groups. Furthermore, habitat diversity ( corre­ eastern group of low islands close to each other and to lated with height) was found to be significant for all the African mainland. Finally, the significant correlation groups except beetles, whilst age (only significant for beetles) or isolation (only found to be significant for found between Age and HD may be attributed to thefact that old islands (as Lanzarote and Fuerteventura) which ferns) seems not to play an important role at this level. have been eroded for several millions of years and nowa­ Finally, it is worth noting that geographical features of the days arelow and flat, in fact display low habitat diversity. island nearest to the one analysed (such as distance to, and When the correlation between richness of the different area or height) were only significantfor groups with good taxonomic groups (Table 5) was calculated, we found dispersal ability. The population of the neighbouring is­ some interesting points: (1) land-bird richness is corre­ lands may thus act as a reservoir safeguarding local ex­ lated with the richness of all the other taxonomic groups; tinction. (2) vascular plants richness is a better predictor of fern The results obtained clearly differ from thosefound by richness than any of the geographical factors analysed; Connor & McCoy (1979), for the seven main Canary and (3) butterfly richness is only correlated with land-bird Islands. The improvement of the checklists over the last richness. In general, significant correlations were also decade and the inclusion of the islets in the analyses are obtained for a major part of the taxa group combinations. probably the reasons for thisdiscrepancy. Thus, comparisons Table 6 shows the results of the single regressions of carried out between taxa without and taxa with data for the islets may lead to unrealistic conclusions. A further prob­ lem in the interpretation of the data appears with the Table 4. Regressions of the main geographical factors versus taxonomicgroups (such as butterflies)where all thespecies richness (number of species of vascular plant, land birds and are present on the largest island, Tenerife, and the species Tenebrionidae beetles) on each Macaronesian archipelago, and number thereforedoes not reach the number in the species versus vascular plant richness of 34 islands throughout Maca­ pool asymptotically. ronesia (sensu Hansen & Sunding 1985) n = number of archi­ pelagos or islands included in the calculation; model = type of regression model; r = correlation coefficient. Only significant Table 5. Matrix of linear correlations between the richness of the values are shown: *=p < 0.05 and ** = p < 0.01. studied taxonomic groups in the Canary Islands. Only signifi­ cantvalues are shown; n.s. =non-significant; * = p<0.05; ** = Taxonomic group Factor n Model r Significance p

Acta Phytogeogr. Suec. 85 Geographical detenninantsof the biological richness in the Macaronesian region 47

Table 6. Best significant regression for each taxonomic group of Table 7. a. Power regressions of areavs. species number for the geographical factors vs. species richness in the Canary Islands. studied taxonomic groups on the Eastern Canarian Islands and

HD = habitat diversity; DNI = distance to nearest neigbouring islets, varying in area but not in number of habitats; n = number island; n = number of islands in calculation; model = type of of islands in the analysis; r = correlation coefficient and signifi­ regression model; r = correlation coefficient. Only significant cancelevel; n.s. =not significant; *=p< 0.01; ** = p<0.05. b. values are given: * = p < 0.05; ** = p < 0.01. Linear and power regression models of habitat diversity versus species number for the studied taxonomic groups on the islands Taxonomic group Factor n Model r Significance of Lanzarote, Madeira and La Palma, varying in number of Vascular plants Area 12 pwr 0.95 ** habitats but not in area. Height 12 log 0.92 ** HD 12 log 0.82 ** a. DNI 12 log 0.72 ** Taxonomic group Slope (z) -intercept Significance n Y Ferns DNI 8 pwr 0.82 * Isolation 8 pwr 0.81 * Flowering plants 6 0.97 0.27 56.11 ** Height 8 pwr 0.80 * Land birds 6 0.99 0.16 11.01 ** HD 8 pwr 0.77 * Beetles 5 0.98 0.34 3.95 ** Land birds Area 12 pwr 0.96 ** b. Height 12 log 0.88 ** Taxonomic group n r-linear Significance r-power Significance HD 12 lin 0.77 ** DNI 12 log 0.60 * Ferns 3 0.45 n.s. 0.79 n.s. Beetles Area 11 pwr 0.90 ** Flowering plants 3 0.36 n.s. 0.61 n.s. Age 8 pwr 0.81 * Land birds 3 0.44 n.s. 0.21 n.s. Height 11 pwr 0.63 * Beetles 3 -0.81 n.s. -0.70 n.s. Butterflies Height 8 pwr 0.88 ** Butterflies 3 1.00 ** 0.99 HD 8 pwr 0.85 ** HNI 8 pwr 0.78 * ANI 8 pwr 0.76 * DNI 8 pwr 0.76 * occurrence of the slopes in the expected range does not necessarily imply the acceptance of the theory. Moreover, a large proportion of power regression lines fitted this Of the regression models, the power model is the range and could sometimes be considered as a statistical better one with about 70% of all significant regressions. artefact (Loehle 1990). However, both the logarithmic (25%) and the linearmod­ On the other hand, the results of the second approach, els (5%) are also represented. Thus, it is difficult to where islands vary in HD but not in area and latitude, consider only one of them as universal in island biogeog­ show an erratic pattern, with trends varying between the raphy, as also has been pointed out by Connor & McCoy different taxonomic groups. With the exception of butter­ (1979) and Rydin & Borgegard (1988). flies, where the increment in habitat diversity is associ­ The results from the multiple island comparison level ated with a significant increment of the insular richness are shown in Table 7. It was not possible to compare the (for both the linear and the power model), the rest of the influence of a single geographical variable on the insular groups do not show an increasing trend, and even the biological richness. Thus, this analysis was exclusively beetles show a decreasing one. carried out to throw some light on the classical contro­ The results of the two multiple island comparisons versy on the role of area or habitat diversity in explaining seem to support both the Area per se and the Random the biological richness of a given island. placement hypotheses, opposite to what Mueller-Dombois Results of the first approach, where islands varying in & Fosberg (1998) found for the tropical Pacific islands. area but not in HD, isolation and latitude were compared The Area per se hypothesis explains the higher number of (the easternCanarian Ridge ), show an increase in richness species on larger islands as a result of lower extinction in all cases analysed (plants, land birds and beetles) al­ rates, due to larger populations involved, whereas the though there was no increase in the number of habitats. Random placement theory attributes the higher richness Results show that the slopes of the power regression lines, of larger areas to stochastic distributions. However, the z-values, for plants (0.27), fit well within the range of Habitat diversity theory where the species richness is values (0.18-0.35) (Diamond & May 1976) expected by largely due to the number of different habitats was not the Equilibrium theory (MacArthur & Wilson 1963, 1967) supported by our data. and was confirmed for different taxonomic groups in numerous oceanic archipelagos around the world (Dia­ mond & May 1976; Williamson 1981). However, both land birds (0.16) and beetles (0.34) z-values stay close to the lower and upper limits respectively, questioning its validity for the Canaries. Nevertheless, Connor et al. (1983), stated that the

Acta Phytogeogr. Suec. 85 48 J.M. Femdndez-Palacios & C. Andersson

Acknowledgements studies. - Am. Nat. 122: 789-796. de Nicolas, J. P., Fernandez-Palacios, J. M., Ferrer, F. & Nieto, We are grateful to Hakan Rydin and Graciela Rusch for E. 1989. Inter-island floristic similaritiesin the Macaronesian valuable comments on drafts of the manuscript. We also region. - Vegetatio 84: 117-125. Diamond, J.M. &May, R.M. 1976. Islandbiogeography and the thank Drs. Aurelio Martin and Pedro Ororniat the Depart­ design of natural reserves. In: May, R. (ed.) Theoretical ment of Zoology, University of La Laguna for kindly ecology. Principles and applications. - Saunders, Philadel­ permitting access to unpublished data. This study was phia, MA, pp. 228-252. made possible by a want from the Canarian Government Fernandez-Palacios, J.M. & Andersson, C. 1993. Species com­ to the first author. position and within archipelago eo-occurrence patterns in the Canary Islands. - Ecography 16: 31-36 Firth, R. & Davidson, J.W. 1945. Pacific Islands. Vol.1. Naval References Intelligence Division, London. Garcfa-Talavera, F. 1999. La Macaronesia. Consideraciones Abbott, I. 1980. Theories dealing with the ecology of land-birds geol6gicas, biogeograficasy paleoecol6gicas. In: Fernandez­ on islands. - Adv. Ecol. Res. 11: 329-371. Palacios, J.M., Bacallado, J.J. & Belmonte, J.A. (eds.) Arrhenius, 0. 1921. Species and area. J. Ecol. 9: 95-99. Ecologfa y cultura en Canarias. - Museo del Cosmos, Bacallado, J.J. & Domfnguez, F. 1984. Las aves de Canarias. In: Santa Cruz de Tenerife, pp. 39-63. Bacallado, J.J. (ed.) Fauna terrestre y marinadel archipielago Gleason, H. 1922. On the relation between species and area.- canario. - Edirca, Las Palmas de G.C., pp. 275-332. Ecology 3: 158-162. Baez, M. 1984. Los artr6podos. Fauna del archipielago canario. German, M. 1979. Island ecology. - Chapman& Hall, London, In: Bacallado, J.J. (ed.) Fauna terrestre y marina del 79 pp. archipielago canario.- Edirca, Las Palmas de Gran Canaria, Hansen,A. & S unding, P. 1985. Flora of Macaronesia.Checklist pp. 101-254. of vascular plants. 4th. rev. ed. - Sommerfeltia 1: 1-167. Hemmingsen, A. 1963. Birds on Hierro and the relation of Baez, M. 1987. Caracteres lies a l'insularite de la faune de number of species, of specific abundances andbody weights, I' Archipel des Canaries. - Bull. Soc. Zool. France 112: 143-152. to island area. - Vidensk. Medd. Dansk Naturhist. Foren. Baez, M. & Sanchez Pinto, L. 1983. Islas de fuego y agua. - 125: 207-236. Edirca, Las Palmas de Gran Canaria, 184 pp. Humphries, C.J. 1979. Endemism and evolution inMacaronesia. Banda, E., Dafiobeitia, J.J., Surinach, E. & Ansoge, J. 1981. In: Bramwell, D. (ed.) Plants and islands. -Academic Features of crystal structure under the Canary Islands. - Press, London, pp. 171-179. EarthPlanet. Sci. Let. 55: 11-24. Johnson, M. & Simberloff, D. 1974. Environmental determi­ Becker, P. 1992. Colonization of islands by carnivorous and nants of island species numbers in the British Isles. - J. herbivorous Heteroptera and Coleoptera: effects of island Biogeogr. 1: 149-154. area, plant species richness and extinction rates. - J. Jones, M., Lace, L., Hounsome, M. & Hamer, K. 1987. The Biogeogr. 19: 163-171. butterflies and birds of Madeira and La Gomera: taxon Bernard-Griffiths, J., Cantagrel, J.M., Matos Alves, C.,Mendes, cycles and human influence. - Bioi. J. Linn. Soc. 31: 95- F., Serralheiro, A. & Rocha Macedo, J. 1975. Donnees 111. radiometriques potassium-argon sur quelques formations Kunkel, G. 1993. Die Kanarischen Inseln und ihre Pflanzenwelt. magmatiques des iles de 1' archipel de Cap Vert. - C. R. - Gustav Fischer, Heidelberg, 185 pp. Acad. Se. Paris 280: 2429-2432. Le Grand, G. 1984. Status e distribu�ao da avifauna nidificante Carlquist, S. 1965. Island life. -Natural History Press, New no arquipelago dos A�ores. - PriOlo 2: 35-38. Le Grand, G. 1986. Bibliographia ornithologique de archipel York, NY, 45 1 pp. 1' Carlquist, S. 1974. Island biology. - Columbia University du Cap Vert. - Bol. Mus. Mun. Funchal 38: 120-129. Press, New York, NY, 660 pp. Lems, K. 1960. Floristic botany of the Canary Islands. - Coello, J., Cantagrel, J.M., Hernan, F., Fuster, J.M., Ibarrola, E., Sarracenia 5: 1-94. Ancochea, E., Casquet, C., Jamond, C., Dfaz de Teran, J.R. Lobin, W. 1982. Untersuchung tiber Flora, Vegetation und & Cendrero, A. 1992. Evolution of the eastern volcanic biogeographische Beziehungen der Kapverdischen Inseln. ridge of the Canary Islands based on new K-Ar data. -J. - Cour. Forsch.-Inst. Senckenberg 53: 1-112. Vole. Geol. Res. 53: 25 1-274. Loehle, C. 1990. Proper statistical treatment of species-area Coleman, B., Mares, M., Willig, M. & Ying-Hen, H. 1982. data. - Oikos 57: 143-144. Randomness, area and species richness. - Ecology 63: MacArthur, R.H. & Wilson, E.O. 1963. An equilibrium theory 1121-1133. of insular zoogeography. -Evolution 17: 373-387. Connor, E.F. & McCoy, E.D. 1979. The statistics and biology of MacArthur, R.H. & Wilson, B.O. 1967. The theory of island the species-arearelat ionship. - Am. Nat. 113: 791-833. biogeography. - Princeton University Press, Princeton, NJ, 203 pp. Connor, E.F. & Simberloff, D. 1978. Species number and compositional similarity of the Galapagos flora and fauna. Menard, H.W. 1986. Islands. Scientific American Library, - Ecol. Monogr. 48: 219-248. Washington, DC, 230 pp. Connor, E.F., McCoy, E.D. & Cosby, B.J. 1983. Model dis­ Mueller-Dombois, D. 1992. The Formation of Island Ecosys­ crimination and expected slope values in species-area tems. Geojournal 28: 293-296.

Acta Phytogeogr. Suec. 85 Geographical determinants of the biological richness in the Macaronesian region 49

Mueller-Dombois, D. & Fosberg, F.R. 1998. Vegetation of the Simberloff, D. 1970. Taxonomical diversity of islands biotas. ­ tropical Pacific Islands.-Springer, Berlin, 733 pp. Evolution 24: 23-47. Mueller-Dombois, D., Bridges, K.W. & Carson, H.L. 1981. Simberloff, D. 1974. Equilibrium theory of island biogeography Island ecosystems. Biological organization in selected Ha­ and ecology. - Annu. Rev. Ecol. Syst. 5: 161-182. waiian communities. Hutchinson Ross Publ., Philadelphia. Simberloff, D. 1978. Using island biogeographic distributions Nilsson, S.G., Bengtsson, J. & As, S. 1988. Habitat diversity or to determineif colonization is stochastic. - Am. Nat. 112: area per se? Species richness of woody plants, carabid 713-726.

beetles and land snails on is-lands. -J. Anim. Ecol. 59: 473- Sunding, P. 1979. Origin of theMacaronesian flora. In: Bramwell, 480. D. (ed.) Plants and islands, - Academic Press, London, pp. Preston, F.W. 1962. The canonical distribution of commonness 13-40. and rarity. - Ecology 43 : 185-215, 410-432. van der Werff, H. 1983. Species number, area and habitat Oromi, P. 1982a. Los tenebri6nidos de las Islas Canarias. - diversity in the Galapagos Islands. - Vegetatio 54: 167- Institute de Estudios Canaries, 50° aniversario ( 1932-1982), 175. pp. 267-292. Volsoe, H. 1955. The breeding birds of the Canary Islands. II.

Oromi, P. 1982 b. Distribuci6nde I os Tenebrionidae(Coleoptera) Origin and history of the Canarian avifauna. - Vidensk. en las Islas Athinticas. - Bolm. Soc. Port. 7: 215-231. Medd. Dansk Naturhist. Foren. 117: 117-178. Ridley, W.I., Watkins, N.D. & MacFarlane, D.J. 1974. The von Buch, L. 1825. Physicalische Beschreibung der Canarischen oceanic islands: Azores. In: Nairn,A.E. & Stehli, F.G. (eds.) Inseln. -Berlin, 201 pp. The oceansbasins andmargi ns. Vol. 2. The North Atlantic. Whittaker, R.J. 1998. Island Biogeography. Biology, evolution

-Plenum Press, New York, NY, pp. 445-483. and conservation. -Oxford University Press, Oxford, 285 Rydin, H. & Borgegard, S.O. 1988. Plant species richness on pp. islands over acentury of primary succession: Lake Hj illmaren. Williams, C.B. 1964. Patterns in the balance of nature and -Ecology 69: 916-927. related problems in quantitative ecology. -Academic Press, Schmincke, H. 1976. The geology of the Canary Islands. In: London. Kunkel, G. (ed.) Biogeography and ecology in the Canary Williamson, M. 1981. Island populations. - Oxford University Islands. -Junk, The Hague, pp. 67-184. Press, Oxford, 280 pp.

Acta Phytogeogr. Suec. 85 Acta Phytogeogr. Suec. 85 Succession and local species turnover on Mount St. Helens, Washington

Roger del Moral

Department of Botany, University of Washington, Box 355325, Seattle, WA 98195-5325, USA; Fax +206 6851728; E-mail [email protected]

Abstract Introduction

The 1980 eruption of Mount St. Helens provided a singu­ The lateral eruption of Mount St. Helens on 18 May 1980 lar chance to study community assembly. Permanent plots destroyed the northern half of its cone. This blast seared were used to monitor species richness, cover and even­ ridges while pyroclastic flows, pumice deposits and lahars ness. Structure stabilized quickly on mildly impacted created a new landscape (Foxworthy & Hill 1982). I sites, but secondary and primary successional sites con­ established sets of permanent plots starting in 1980 to tinued to develop. Annual DCA score increments changed monitor vegetation establishment and recovery. This pa­ least on tephra and most on primary sites. Vegetation per describes development on new surfaces (primary suc­ changed more rapidly soon after impact, more slowly cession), intensely damaged habitats (secondary succes­ recently. Net plot displacement after 18 seasons was 0.4 sion) and sites that recovered quickly. These patterns are half-change (HC) on tephra, 1.0 HC on secondary sites interpreted in terms of impact intensity and site isolation. and 1.1 HC on primary sites. Colonization and extinction Trends were described using detrended correspondence percentages in 0.25-m2 quadrats were estimated for three analysis, while I compared local colonization and extinc­ intervals. The extinction percentages declined from re­ tion percentages to determine the degree of local popula­ covered to primary sites, while the colonization percent­ tion dynamics. ages were similar in all. The proportion of extinction events to colonization events declined sharply. These results support the carousel model at an intermediate Study area spatial scale, even during succession. Degree of site isola­ tion, stress and impact intensity combined to form a Mount St. Helens is centred at46° 20' N, 122° 18' W, with mosaic of recovering vegetation that remains dynamic. an elevation of 2549 m. This study included 10 sites. At Each community had unique structure, developmental each, permanent plots that represent different combina­ sequences and local population dynamics . tions of impact intensity (Table 1) were monitored with permanent plots. Del Moral & Bliss (1993) provided a map and site descriptions. Keywords: Carousel model; Local colonization; Local Recovered sites were impacted by air-fall deposits of extinction; Primary succession; Secondary succession; coarse pumice (tephra) that buried vegetation on the south­ Volcano. ernslopes of the cone up to 20 cm. Coarse texture, shallow depth and erosion permitted significantsurvival and rapid Abbreviations: ANOVA = Analysis of variance; CV = recovery (del Moral 1983). Three sites (Tephra A, B and Coefficient of variation; DCA = Detrended Correspond­ C) had fullyrecovered by 1983. ence Analysis; HC = Half-change. Secondary succession sites received a variety of im­ pacts. Blast A is on the west side of the cone at the edge of Nomenclature: Titus et al. (1998). the directed blast that killed woody plants. Some dormant herbs survived. Rapid snow-melt formed lahars (cold slurries of mud and rocks) that scoured canyons and ridges before forming deposits at lower elevations (del Moral & Wood 1988). A lower site on an east slope ridge (Scour A) is within 100 m of surviving vegetation, while 85: 51-60. 91-72 10-085-0. a nearby upper site (Scour B) is more isolated. Lahars on Acta Phytogeogr. Suec. ISBN

Acta Phytogeogr. Suec. 85 52 R. del Moral

Table 1. Summary of characteristics of the ten study areas.

Site Location Impact type Elevation (m) Isolation Succession type

Tephra A South Flats Tephra 1370 None Recovered Tephra B South Flank Tephra 1525 None Recovered Tephra C South Flank Tephra 1600-1680 None Recovered Scour A East Ridge Light scour 1370 Low Secondary Blast A West Ridge Blast edge 1280-1340 Intermediate Secondary Scour B East Ridge Moderate scour 1510 High Secondary Scour C South Ridge Intense scour 1580-1710 Low Secondary Lahar South Flank Lahar 1320-1350 Low Primary Blast B Northwest Ridge Direct blast 1200-1310 High Primary Pumice North Flats Pumice deposit 1220-1280 High Primary

the south side scoured high elevation sites, permitting Detrended Correspondence Analysis (DCA) was per­ little survival (Scour C). formed with PC-ORD (McCune & Mefford 1997). This Three types of primary succession habitats were sam­ program corrects problems associated with input order pled. Lahars formed on meadows on the south slope. (cf. Tausch et al. 1995; Oksanen & Minchin 1997; Proximity to intact vegetation permitted rapid establish­ Podan.i 1997). Detrending was effected in 26 segments. ment. Blast B, on a northwestern ridge, was severely Species with fewer than five occurrences and plots with impacted by the lateral blast. All vegetation was killed only one species were removed. Rare species were and most soil was removed (del Moral 1993). The Pum­ down weighted. ice site, on the lower north slope (del Moral et al. 1995; del Moral & Wood 1993), was seared by the blast, and Turnover then covered by pumice. Colonization and extinction percentages were calculated from 0.25-m2 quadrats during three overlapping intervals. Methods In most cases, comparisons were from 1986 to 1991, 1989 to 1994 and 1991 to 1997. Pumice sampling started in Sampling 1989, so comparisons were 1989 to 1992, 1991 to 1995 and 1992 to 1997. Blast A was not sampled between 1988 and The 10 sites were sampled using permanently marked 1993. Comparisons were 1986 to 1994, 1987 to 1995 and plots to permitaccurate repeated sampling. Species cover 1994 to 1997. These analyses permit three independent was recorded at 1-m .intervals on four rad.i.i in 0.25 m2 turnover estimates. quadrats. Repeat non-destructive measures permit direct The extinction rate of a species was calculated by interpretations of colonization and extinction at this scale modification of the method of Froborg & Eriksson ( 1997). (Austin 1980). The design is 10 sites characterized by a Their method is sensitive to the usual preponderance of variable number of permanent 250 m2 plots that were empty quadrats and to different numbers of quadrats among sampled using 24 0.25 m2 quadrats each. the sites. Therefore, the index was expressed as a percent­ age of the number of sampled quadrats (see del Mora1 2000 Statistics for details): The number of quadrats occupied at Time0 that are empty at Time1, divided by the total number of quadrats Species richness (R), mean cover (C) and evenness (E) times 100% estimates the extinction percent. The coloniza­ were calculated for each plot. E = H/ln R, where His tion rate is the number of quadrats empty at T 0 that are the Shannon index. Changes in these measures were occupied at T 1 divided by the total number of quadrats tested for significance by ANOV A, followed by be­ times 100%. tween-mean comparisons with the Bonferroni statistic. Pooled plots may differ markedly due to different suc­ .cessional rates. The resultant high variance between samples of a given year often obscured trends. There­ fore, linear regressions between year and observations of individual plots on each variable were conducted (Anon. 1994).

Acta Phytogeogr. Suec. 85 Succession and local species turnoveron Mount St. Helens 53

Results Table 2. Community structure 18 seasons after the eruption. Plots summarized are those used in DCA and turnover studies. Community structure Richness is the meannumber of plantspecies in N 250-m2 plots. Cover is the mean percent cover in 24 0.25-m2 quadrats in each of N plots. Evenness is the mean from the N plots. Table 2 summarizes community structure at each site after 18 seasons. Data for these analyses may be viewed Study site N Richness Cover (%) Evenness at: http://www .biology. washington.edu/delmoraV. Tephra Recovered sites A had the highest cover and high richness. Strong domi­ Tephra A 3 21.3 57.3 0.419 nance by Agrostis pallens, Lupinus lepidus and Pinus Tephra B 4 22.8 52.7 0.611 Tephra C 4 contorta saplings produced low evenness. Tephra B was 17.5 30. 1 0.7 10 Secondary succession similar, but dominance was less pronounced. Tephra C is Scour A 4 19.5 49.7 0.569 more stressful, reducing richness and cover. Persistent Blast A 5 18.8 25.8 0.642 forbs such as Eriogonum pyrolifolium, Penstemon card­ Scour B 4 11.8 21.7 0.652 wellii and Phlox diffusa, rather than graminoids, domi­ Scour C 5 17.4 13.6 0.644 nated this higher elevation habitat. Primary succession Cover of secondary successional sites continued to Lahar 4 20.2 14.5 0.656 increase and to accrue species. Evenness is higher than Blast B 6 16.3 19.7 0.650 Pumice 5 15.8 3.8 0.833 in recovered plots. Scour A developed cover close to that of recovered plots, though mosses that were scarce on tephra contributed significantly. Some expected species, such as Calyptridium umbellatum,Hieracium albiflorum, Cover responded to weather variation in recovered and Achnantherum occidentale, were absent or rare, plots (del Moral & Bliss 1993; Fig. 2a). Tephra A cover while Agrostis pallens, Luetkea pectinata and Lupinus increased dramatically in 1981 due to Agrostis and Lupinus lepidus dominated. Many plants survived at Blast A and then fluctuated greatly. Tephra B still may be developing most expected species occurred. A. pallens, L. lepidus, since peaks in 1982, 1990, and 1997 were successively

Polygonum newberryi,Eriogonum pyrolifolium, Achillea higher. Tephra C appears to be stable. Cover in secondary millefo lium and Fragaria virginiana were common. plots continues to increase (Fig. 2b). Fluctuations in Scour

Agrostis, Luetkea, Penstemon and Eriogonum domi­ A were due to fluctuations in L. lepidus. Primary site cover nated scour B, but richness was low. Common subalpine continues to increase (Fig. 2c ). Blast B demonstrated large plants such as Luetkea and Calyptridium umbellatum fluctuations in L. lepidus cover. While lower elevation dominated scour C. Cover was low on these exposed plots may be at equilibrium, plots higher on this ridge have ridges, but richness was comparable to other secondary low cover. The Pumice site is stressful and establishment sites and the nearby Tephra C. was confined to favourable microsites (del Moral 1993). Primary successional plots had low cover and high Cover expansion was limited by low nutrients and drought evenness. The lahar plots had richness comparable to (del Moral & Bliss 1993). adj acent tephra, but lower cover. Abies lasiocarpa and No evenness trends were observed on recovered or

Pinus contorta dominated the plots near forest, while L. secondary sites other than Scour B, where evenness de­ lepidus, Penstemon and Calyptridium dominated more clined as dominance developed. Evenness on primary isolated ones. Blast B plots were on more protected sites sites declined as species established, except on Blast B. where large populations of L. lepidus had established, Here, strong L. lepidus dominance yielded low evenness. leading to high cover. Agrostis was the only other species As other species invaded, evenness similar to that of other with high cover. Pumice plots were the least developed, primary sites developed. showing very low cover and very high evenness. Agrostis ANOV A, followed by means-difference tests, were scabra and L. lepidus dominated these plots. conducted for richness, cover and evenness. Table 3 shows Fig. 1 summarizes the changes in richness. Richness the number of significantly different groups identifiedby increased quickly, then stabilized in recovered plots (Fig. Bonferroni comparisons, with a value of "1" indicating no la). Secondary plots recovered rapidly by 1983, then differences among groups. An asterisk indicates groups gradually accumulated species (Fig. 1b). The lowest rich­ thatare not sequentially related and not due to succession. ness occurred at Scour B, where colonization requires Richness at Tephra B was the only case where struc­ uphill migration. The primary sites lacked species in 1980 ture at a recovered site changed withtime. Secondarysites (Fig. le). Lahar plots were established in 1982, Blast B demonstrated significant increases in richness and cover, plots in 1984 and Pumice plots in 1989. The first plants but evenness declined significantly only on Scour B. Pri­ near the Pumice plots were noted in 1984 (Wood & del marysites demonstrated significantresponses inmost cases, Moral 1988). Once establishment started, richness in­ even though among-plot differences obscured patterns. creased steadily. Lahar and Pumice plots each had several clusters of means.

Acta Phytogeogr. Suec. 85 54 R. del Moral

_ 70 ,. • . . 23 . .

... -; 22 0 60 -a 21 � ...... i ...... 20 ...... · } 50 .. 19 =• . E:I ..5. i 18 • lit , · '·• ·-· · ...•. � 40 a ..... ,. c 17 ,• • .c= Q. u 16 ..... _ .,.•' t ii: � 15 u • 30 ,· = .. , I 14 '• lCl) 13 __._ Tephra A Tephra B 20 12 · -· -·· Tephra C 11 L-�----�--�--�--�--�----�------�--�--- 10 1� � 1� 1� 1� 1� 1� 1� 1� 11980 1 9 1994 1996 1998 1� 1982 1984 1986 1988 �ear990 92 Year

...... 20 ... 50 . . .

1 8 ... · . .... -; 16 40 ! . 14 ...... · 8. · · .t. .... 12 _ .....· � 30 . • · . •. . . ! ··- ...... ,..... - , ··- ·· - ·- lii . E:I ...-· .... 10 � _ ...... ·· _ .. u . - ...-...... 5.., .....·-··-· ..c: Ill :. - .. · · • · 20 c: :• - :11 .c · u 6 I ii: ! - 4 • Blast A um 10 I ScourA I · !. - Scour S 11) ...... •· -· 11111 Scour C 0 0

1 62 1 1980 9 1984 86 1994 1� 1� 1� 1� 9 1988 1990 1992 1996 1998 1� 1� 1� 1� 1� � Year Year

22 .A .. 25 ...... ,. .. . 20 .... . A 18 20 i . i5...... i 16 ! 14 t> 0 e:I u 15 12 .. ..5. t:D ...... 10 i c . . . .c 10 . .!! 8 IL D: �

: 6

iCl. ··· • 5 . . - 4 , . .. Cl) • .t.· ···A ,•· - ··- ...- ·' 2

0

1980 1982 1984 1986 1988 1990 1992 1994 1998 1996 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 Year Year

Fig. 1. Changes inrichness (mean number of species per 250m2 Fig. 2. Changes in cover percentage (mean per 250m2 quadrat) quadrat) in permanent plots; a. recovered sites; b. secondary in permanentplots; a. recovered sites; b. secondarysuccessional; successional; c. primary successional sites. c. primary successional sites.

Acta Phytogeogr. Suec. 85 Succession and local species turnoveron Mount St. Helens 55

Table 3. Number of groups resulting fromANOV A fo llowed by Table 4. Regression index at each site. Richness, cover and Bonferroni comparisons of the means of richness, cover, and evenness as described in Table 2. evenness (see Table 2 for definition). Each group consists of Site Cover Richness # # Evenness # means not significantly different (P < 0.05) for the structural P< 0.01 P< 0.01 P< 0.01 variable. An * indicates significant differences that were not Recovered sites sequential in time, indicating only response to climatic varia­ Tephra A (n =3) 2.33 3 0.67 0 0 tion. Tephra B (n =4) 2.25 3 2.33 I 0 0 Site Richness Cover Evenness Tephra C (n=4) 0.5 0 0 2.75 3 Secondary sites Recovered sites 4 5.0 4 0.5 Tephra A 3* 4* 2* Scour A (n=4) 3.75 Blast A (n=5) 1.4 2 4. 1 5 0.8 Tephra B 3 2* 1 Tephra C 1 1 1 Scour B (n=4) 4.75 4 5.0 4 3.75 4 Scour C (n=5) 3.0 3 5.0 5 2.4 3 Secondary sites Primary sites Scour A 2 5 2* Lahar(n =4) 5.0 Blast A 2 3 1 4 5.0 4 4.75 4 Blast B 5.0 6 3.0 6 1.67 3 Scour B 3 3 3 (n=6) Pumice (n=5) 5 2.0 3 0.6 Scour C 2 1 1 2.2 2 Primary sites Lahar 6 5 5 Blast B 5 1 angustifolium, Hypochaeris radicata, Saxifragaferruginea, Pumice 3 3 4 ]uncus mertensianus, Carex microptera and rock crevice ferns (e.g. Cryptogramma cascadensis). They also included more widely distributed species such as Agrostis pallens Differences between successive years may fall within and Lupinus lepidus. Bonferroni limits because the combined plots were not Sites with very high Y-axis scores were dominated by replicates. Therefore, linear regressions between year and Saxifraga fe rruginea, Polygonum newberryi and Carex each parameter were calculatedfor each plot. A weighted mertensii. All Scour C#4 plots had extreme values. Plots average summarized theseregre ssions (Table 4) . Theindex missing from the middle were Pumice, Blast B and early was derivedas follows: A linear regression at P < 0.01 ;;;;; 1; lahars, which dominated the low values. The species con­ centrated withlow Y -scores included L. lepidus, A. pallens, p < 0.005 ;;;;;2; < 0.001 ;;;;;3; p < 0.0005 ;;;;;4; p < 0.0001 ;;;;;5. The weights were added and divided by the number of A. scabra, H. radicata and Lomatium martindalei. regressions to yield the index. This arbitrary index pro­ Table 5 shows the mean annual Euclidean distance vides a descriptive comparison among the sites and is not change along DCA X and Y for each plot and the CV of intended to test statistical differences. these changes. The mean increment measures the magni­ The number and strength of time-related relationships tude of annual change, while the CV measures the steadi­ increased from recovered to primary sites. Cover in­ ness of change. Changes are not always unidirectional, creased significantly in two of 11 recovered plots, all as a sample can fluctuate about a mean. secondary plots, and all but two primary plots. Pumice The firsttwo increments where cover was more than showed fewer cover trends because annual fluctuations 0.3% were compared with the last two increments to were large and the recovery small. determine rates of change (Table 6). This table also shows the net change from first to last samples to indi­ Vegetation change cate overall compositional change. Mean annual change of Tephra plots was low, de­ Directional vegetation change was explored by DCA. spite initially large changes. These sites vary around All data were analysed together to permit direct com­ 0.15 HC/yr; CV was ea. 60%. Early samples changed parisons of species change. DCA used 58 species in 613 more than late samples. The directional change ranged samples. The first three eigenvalues were 0.492, 0.391 from about 0.2 to 0.5 HC in X and from 0.2 to 0.3 HC in and 0.289. The X-axis spanned 3.95 HC, while the Y­ Y. Euclidean change ranged from 0.3 to 0.5 HC. Changes axis spanned 3.32 HC. Along X, high values were asso­ in Tephra A plots were associated with the increasing ciated with Lupinus latifolius, Luetkea pectinata and dominance of Agrostis pallens, decline of Eriogonum Luzula parviflora. These were species common in late pyrolifolium and fluctuations in Lupinus lepidus and snowmelt habitats. The middle of X included samples Lomatium martindalei. Changes in Tephra B involved from nearly all sites. Between X-values 191 and 199, less dramatic increases in Agrostis, development of only the Blast B and Pumice samples did not occur. Danthonia intermedia and a decline in Lomatium and Primary succession plots dominated the low end of X. Lupin us. Tephra C changes were associated with devel­ Plots with low DCA scores were characterized by sparse opment of grasses, Juncus parviflora and Penstemon cover and dominated by Anaphalis margaritacea, Ep ilobium cardwellii, while Lupinus and Calyptridium fluctuated.

Acta Phytogeogr. Suec. 85 56 R. del Moral

Table 5. Floristic changes (Euclidean distance) in DCA space; excludes samples with cover less than 0.4%. Units are HC (X 100). 8.n is the mean annual DCA increment for an individual plot. CV is the coefficient of variation of the increments.

Study Site Mean D1 Mean D2 Mean D3 Mean D4 Mean D5 Mean D6 Grand Mean

Tephra A 16.0 12.6 12.3 13.6 CV 66.9 60.0 47.6 58.2 Tephra B 13.0 14.5 18.0 14.5 15.0 CV 55.6 72.1 64.3 33.3 56.3 Tephra C 14.6 10.8 11.1 10.3 15.1 CV 46.5 55.3 57.6 93.5 63.2 Scour A 11.6 18.5 39.5 17.8 21.9 CV 99.0 82.6 104.4 89.0 73.7 Blast A 43.2 28.4 41.5 17.9 42.8 34.9 CV 78.0 70.8.0 100.2 100.9 115.0 92.8 Scour B 27.2 28.4 17.8 19.6 23.3 CV 84.9 76.9 92.9 71.6 81.6 Scour C 14.8 43.1 29.3 28.2 15.6 26.2 CV 69.5 51.3 125.5 84.7 74.8 81.2 Lahar 31.9 32.7 21.5 40.7 31.7 CV 112.9 56.2 65.6 82.7 79.4 Blast B 19.3 28.0 42.8 44.6 43.4 51.4 38.3 CV 58.1 146.6 276.2 117.9 120.3 166.2 147.6 Pumice 33.7 28.8 38.1 33.6 27.7 32.4 CV 54. 1 68.2 40.9 60.5 52.2 55.2

Secondary sites showed substantial mean increment The primary sites had lower cover than other plots, yet change and all had CVs larger than CV s on tephra. (The their mean annual change was higher than all but Blast A. increment between 1987 and 1994 in Blast A was ex­ The contrast between initial and recent changes varied. cluded from the calculation.) Early recovery substan­ Initial changes were high, similar to secondary plots. tially exceeded recent changes. Directional change was Recent increments were much higher than on recovered large, ranging from 0.5 to 1 HC in X and 0.5 to 0.7 HC in sites, and, except for the lahar, higher than secondary Y. Euclidean change ranged from 0.8 to 1.2 HC. plots. Directional X change varied from 0.4 to 0.9 HC, Scour A changes with the increase inAgrostis, Lupinus while that of Y was from 0.4 to 1.3 HC. Euclidean change lepidus, Luetkea pectinata, Penstemon and mosses. Sev­ ranged from 0.9 to 1.5 HC. The Lahar changed in re­ eral species of Carex tended to increase.Blast A changed sponse to the development of Abies, Pinus, June us, with the development of Agrostis, Achillea millefolium, Luetkea, Penstemon, Calyptridium, mosses and several Castilleja miniata, Lomatium, Penstemon and Racomi­ grasses. Blast B fluctuated substantially due to large trium canescens, the decline of Lupinus latifolius and changes in L. lepidus. Directional change occurred in variation in L. lepidus. Scour B also changed with the response to the development of Agrostis, Anaphalis, development of Agrostis, Eriogonum, Penstemon and Penstemon and mosses. The Pumice plots have changed mosses. Scour C changed primarily with the development in response to Agrostis scabra, Penstemon, Calyptridium of Carex mertensii, Eriogonum, Penstemon andPolygonum. and mosses.

Table 6. Floristic changes (two-dimensional Euclidean distance) in DCA space comparing fr rst two annual increments with last two increments. A Earlyis the mean difference between first two increments; A Late is mean of the last two increments. DXY is overall mean change in plots during study. Mean M, AY and MY are the mean of the extreme values of samples in the ordination space. [Units are HC x 100.]

Study Site �Early �Late Early/Late Mean Mean Mean Ratio AX �y MY Tephra A 17.8 9.5 1.87 45.3 31.8 48.5 Tephra B 15.9 9.9 1.61 28.3 14.5 32.1 Tephra C 12.9 11.3 1.14 20.5 20.0 29.7 Scour A 52.1 11.4 4.56 63.8 61.0 97.1 Blast A 54.3 19.9 2.73 98.2 56.4 118.8 Scour B 35.6 15.9 2.24 62.0 63.0 94.9 Scour C 47.5 15.0 3.17 54 60.8 85.3 Lahar 61.5 14.6 4.21 43.0 123.0 142. 1 Blast B 52.9 25.3 2.09 88.2 46.0 101.3 Pumice 41.9 28.9 1.45 63.8 48.2 88.2

Acta Phytogeogr. Suec. 85 Succession and local species turnoveron Mount St. Helens 57

Turnover Secondary successional sites suffered different im­ pacts, but each lost species and was isolated from poten­ Table 7 summarizes extinction and colonization percent­ tial colonists to different degrees. Blast A lost its forest ages for three several-yearintervals in each habitat. By 1997, cover, some soil and species with exposed parts. The richness was similar among most plots (Fig. 1), but substratewas unstable, and the site was exposed to drought, successional plots had lower cover than did recovered high temperatures and wind. Therefore, recovery has been ones. There were always more colonization than extinc­ slower than at other sites. Richness continued to increase. tion events, reflected by E/C ratios less than 1.0 in each Cover remained incomplete. As competitive dominance case. by species such asAgrostis, Lupinus lepidus and Penstemon Extinction percentages were high on tephra (4.6 to continues to be asserted, evenness should decline. 12.3% of total quadrats), moderate on secondary sites (1.2 The scours offer direct contrasts of isolation and im­ to 4.9% of quadrats) and low on primary sites (0.6 to pact effects. Scour A was close to intact understorey 5.2%). Colonization percentages demonstrated no pat­ vegetation. Scouring was incomplete and burial shallow tern.Tephra sites varied from 3.9 to 17.0 % of the quadrats, (del Moral 1983). This site developed rapidly and now secondary sites varied from 5.8 to 15.3% of the quadrat, resembles tephra sites. Scour B lost most vegetation (del and primary sites variedfrom 3.9 to 15.9% of thequadrats . Moral 198l ). Isolation, exposure and the modest dispersal The disparitybetween these two processes is reflected by ability of most colonists (Wood & del Moral 1987), have the ratio of extinctions to colonizations. These ratios limited species richness. Cover continues to increase, but approach 0.9 on recovered sites, varyfrom 0.26 to 0.45 on was half that of Scour A. Scant soil, strong winds and secondary sites and from 0.19 to 0.26 on primary sites. shorter growing season reduced the development rate. Scour C was intensely impacted. Itis at high elevation and drought stressed. Because it is near intact vegetation, it Discussion has 2/3 more species than Scour B, but only 2/3 as much cover. Changes on secondary sites were time-correlated, Communitystructure despite substantial variation in response to weather pat­ terns. The number of significant cover groups was corre­ The degree of development reached by 1997 resulted lated to the initial impact intensity. from combinations of factors. Tephra plots recovered Each primary successional site has developed under quickly from disturbance. While many plants died, there unique circumstances. The lahars consist of reworked was no evidence that any vascular plant species was volcanic material with more nutrients than new substrates eliminated. Cover doubled in Tephra A within two years, (del Moral & Clampitt 1985). They were deposited next to possibly the result of a pulse of nutrients released by intact vegetation. Rapid dispersal produced richness com­ decomposing plants (del Moral 1983) and reduced evapo­ parable to that of recovered sites, though the species ration. Tephra sites differed in elevation and exposure, composition is dramatically different (del Moral 1998). resulting in different rates of productivity. Cover fluctu­ Though not recovered, cover was similar to more extreme ated in response to the amount and distributionof summer secondary sites. Evenness was typical of the open com­ rain (Pfitsch & Bliss 1988), while evenness increased due munities studied. Blast B was exposed and continued to to decreasing grass dominance. Changes after 1982 were receive propagules from up-valley winds. Epilobium spp., associated with secular fluctuations in summer precipita­ Asteraceae, ferns and mosses were common. Species tion, so there were no meaningful correlations between richness was greater than that of Scour B and cover was structure and year. high due to Lupinus lepidus. This short-lived, nitrogen-

Table 7. Mean extinction and colonization percentages. The extinction to colonization ratio was calculated fromthe mean percentages.

I, II and m represent three comparison periods. Recovered Secondary sites Primary sites Property Tephra A Tephra B Tephra C Scour A Blast A Scour B Scour C Lahar Blast B Pumice 5.3 4.7 1.8 2. Extinction Percent I 4.6 2.6 5 2.0 0.9 1 .8 0.7 12.3 6.3 4.8 2.8 4.9 3.3 1.2 Extinction Percent 11 0.9 0.6 1.5 1 6.5 4.7 Extinction Percent Ill 9. 6.8 3.7 4.7 2.5 5.2 4.7 1.8 Mean Extinction % 8.9 5.9 5.3 3.0 3.8 3.5 1.9 2.3 2.3 1.3 17.0 6.5 13.2 12.0 5.8 Colonization Percent I 6.2 5.8 6.2 13.2 3.9 Colonization Percent 11 6.1 3.9 7.1 7.6 15.3 8.3 6.8 12.4 15.9 6.1 6.2 6.5 Colonization Percent Ill 10.3 9.3 6.8 8.7 9.5 7.9 6.5 6.7 Mean Colonization % 9.8 6.9 7.5 6.7 11.7 9.7 7.4 8.8 11.9 5.6 Extinction/Colonization 0.91 0.86 0.71 0.45 0.32 0.36 0.26 0.26 0.19 0.23

Acta Phytogeogr. Suec. 85 58 R. del Moral

fixing species fluctuates significantly over a 4- to 5-yr highest among secondary sites. The other scours experi­ period (Bishop & Schemske 1998). As a result of its enced moderate annual change, were highly variable and facilitative effect, cover was enhanced on some plots on changed substantially at fi rst, less so more recently. The this ridge. Blast B retains older substrates and protected net change of secondary sites was three times that of the microsites that enhance the recovery rate. Pumice consists tephra sites. of newly formed material, is isolated, nutrient poor and All primary sites changed significantly between years. lacks significant ameliorative features (Wood & del Moral Lahar variation was comparable to that of secondarysites, 1988; del Moral 1993). Though species richness was but low cover on Pumice led to low annual variation. In moderate, cover was the least of any in this study. There contrast, Blast B had very high annual variation, due to L. were few species interactions and evenness was very lepidus variation. Net change on Pumice was low, but high. All primary sites showed rapid richness increases, recent changes were the highest of any site. The rate of which suggested that richness accumulates more quickly development on Blast B was accelerating in more ex­ than other aspects of structure. Lahar and pumice sites posed plots. have demonstratedsignificant cover and evennesschange s. DCA permitteda direct comparison among the sites. It Blast B lacks significant cover and evenness groups, an demonstrated that each had a unique response to the artifact of large fluctuations in L. lepidus. disturbance regime. The details of the numerical response Regressions of the structuralvalues of individual plots are affected by richness and cover, but the analysis is through time eliminated the pseudoreplication problem. robust, and readily revealed patterns. These results sup­

Significant (P < 0.05) richness increases occurred in all port the finding of Myster & Pickett ( 1994) that early but two Tephra sites and one Blast A plot. The strength of change is usually more rapid than later change. these relationships varied with impact intensity. Cover relationships were moderate on secondary sites, which Turnover started with low cover but have gradually accumulated biomass. Low cover on Pumice and extreme L. lepidus The concept of local species turnover is embedded in the variation on Blast B obscured cover relationships on lottery model (Lavorel & Lebreton 1992; Laurie & Cowl­ primary sites. ing 1995), the carousel model (van der Maarel & Sykes Evenness is sensitive to changes in both cover and 1993), mass effect (Shmida & Ellner 1984), the hierarchi­ richness. Declining evenness indicates the development cal continuum concept (Hoagland & Collins 1997) and of a dominance hierarchy. This patternis best revealed on the core and satellite hypothesis (Hanski 1982; Collins et scours and the lahar where initial cover was very low and al. 1993). The turnover rate at small scales in stable different species developed differentially. Evenness re­ meadows is high (Herben et al. 1997). Tsuyuzaki (1991) mains high on most pumice plots. Evenness of Blast B found turnover early in volcanic succession to be domi­ plots varied with L. lepidus fluctuations. nated by colonization, with few extinction events in large quadrats. Froborg & Eriksson (1997) showed that local Detrended Correspondence Analysis colonization and extinction in a stable forest understory at a scale of 100 m2 were high, and that some of this effect DCA permits direct comparison of the rates and direction was correlatedto seed weight and dispersal types. V an der of vegetation change in permanent plots through time. Maarel et al. ( 1995) suggested thatthere was little evidence This study confrrmed that responses to disturbance are for niche structure on a small scale, indirectly supporting a related to impact intensity and isolation. Equally isolated carouselmodel (but see Kikvidze 1993). V an der Maarel & sites developed differently in response to local factors, Sykes (1997) quantified the rate of local mobility among such as soil fertility. species and found it much larger than generally recognized. Annual change (.Ll), net change and fluctuation (CV) The present study indicates that species do have high were small in recovered plots. These measures form a local mobility. It differs from previous studies of local baseline of normal variation. Annual changes on second­ turnover in two ways: it explores an intermediate scale ary sites were large and moderately directional. The re­ and it investigates the process against a successional cent rate of change of Scour A was similar to that of background. Species present in 1986 may have survived tephra, its annual change was the lowest of successional the 1980 impacts, arrived by long-distance dispersal or sites and its CV was the least of the moderately vegetated grown into the quadrat. Once present, a plant may be lost sites. Two-dimensional movement that is about three by senescence (e.g. Lupinus lepidus,Hypochaeris radicata times that of tephra reveals its successional status. Blast A or Hieracium gracile), disturbance (e.g. elk or gopher), changed substantially between years. The annual incre­ dieback or competitive displacement. ment was high because, in addition to the increase of Extinction percentages declined from recovered to species adapted to open conditions, there was a marked primary sites as expected. The decline occurred because decline of Lupin us latifolius. Thus, the net change was the there were fewer quadrats with species at risk in primary

Acta Phytogeogr. Suec. 85 Succession and local species turnover on Mount St. Helens 59

sites than in stable ones. Colonization events were more Acknowledgements numerous than extinction events and colonization per­ centages were high. Empty quadrats are colonized at National Science Foundation Grants BSR-89-06544 and similar rates on all sites. No trends were evident. How­ DEB-94-06987 provided funds for this study. The Mount ever, the ratio of extinction percentage to colonization St. Helens National Volcanic Monument staff facilitated percentage provides a basis for comparison. In recovered this study. I am particularly grateful to Helen de la Hunt, sites, extinction events are nearly as frequent as coloniza­ Katrina Dlugosch, Roger Fuller, Richard Robham, Jon tion events. In secondary sites, they are only one-third as Titus and David Wood, who all made significant contri­ common and on primary sites they are one-fourth as butions to data collection. The manuscript was improved common. All habitats are dynamic at this se-ale. by valuable input to the manuscript from Beth Brosseau, Small-scale extinction and colonization appear to be based Roger Fuller, Chad Jones and Dennis Riege, as well as by on life histories, dispersal ability and proximity to seed two anonymous reviewers. Perceptive papers by Prof. sources. Some deterministicextinction may have occurred, Eddy van der Maarel stimulated this study. but rates were high in open quadrats. Colonization per­ centages were no lower in recovered sites than the others, suggesting that the presence of less open vegetation does References not inhibit invasion. Stochastic extinction and coloniza­ tion appear to dominate local population dynamics in Analytical Software. 1994. Statistix 4.1 for Windows - User's most cases. Manual. - Analytical Software, Tallahassee, FL, 329 pp. This study examined succession in permanent plots Austin, M.P. 1980. Permanent quadrats: An interface for theory from three perspectives and provided insights into mecha­ and practice. -Vegetatio 46: 1-10. Bishop, J.G. & Schemske, D.W. 1998. Variation in flowering nisms of community assembly. Each site has a different phenology and its consequences for lupines colonizing Mount combination of impact intensity and type (Walker 1999), St. Helens. -Ecology 79: 534-546. biological legacies (Franklin et al. 1985) and isolation Collins, S.L., Glenn, S.M. & Roberts, D. W. 1993. The hierarchi­ (del Moral 1993). Isolation creates a selective filter that cal continuum concept. - J. Veg. Sci. 4: 149-156. limits colonization rates. Isolated sites have a species del Moral, R. 1981. Life returns to Mount St. Helens. -Nat. composition distinct from similar sites near sources of Hist. 90: 36-49. colonists. Site stress affects cover strongly and has some del Moral, R. 1983. Initial recovery of subalpine vegetation on impact on species composition. The principal site factor is Mount St. Helens.-Am. Midi. Nat. 109: 72-80. substrate age. New surfaces (pumice, pyroclastic materi­ del Moral, R. 1993. Mechanisms of primary succession on als) had the least cover. Old devastated surfaces (blasted volcanoes: a view from Mount St. Helens. In: Miles, J. & ridges, scours) may develop more quickly, but the details Walton, D .H. ( eds.) Primary succession on land.- Blackwell Scientific Publications, London, pp. 79-100. depend on stress factors. Reworked old substrates (lahars) del Moral, R. 1998. Early succession on lahars spawned by can support many species and moderate cover, but isola­ Mount St. Helens. - Am. J. Bot. 85: 820-828. tion effects have been noted on some lahars (del Moral del Moral, R. In press. Local species turnover on Mount St. 1998). Helens, Washington. In White, P.S., Mucina, L. & Leps, J. The developing meadow vegetation of Mount St. Hel­ (eds.) Vegetation science in retrospect and perspective. ­ ens demonstrates high rates of local colonization and Opulus Press, Uppsala. extinction. Successional status affects these rates, but the del Moral, R. & Bliss, L.C. 1993. Mechanisms of primary composition of each site is dynamic. These results, ob­ succession: insights resulting from the eruption of Mount St. tained in recovered and successional sites with diverse Helens. - Adv. Ecol. Res. 24: 1-66. traits, support the carousel model. The species in these del Moral, R. & Clampitt, C.A. 1985. Growth of native plant species on recent volcanic substrates from Mount St. Hel­ habitats do not have narrow, discrete niches. Therefore, ens. - . Midi. Nat. 114: 374-383. they may coexist or they may replace one another in a Am del Moral, R., Titus, J.H. & Cook, AM. 1995. Early primary non-deterministic way. succession on Mount St. Helens, Washington, USA. - J. Veg. Sci. 6: 107-120. del Moral, R. & Wood, D.M. 1988. Dynamics of herbaceous vegetation recovery on Mount St. Helens, Washington, USA, after a volcanic eruption. - Vegetatio 74: 11-27. del Moral, R. & Wood, D.M. 1993. Earlyprimary succession on the volcano Mount St. Helens. - J. Veg. Sci. 4: 223-234. Foxworthy, B .L. & Hill, M. 1982. Volcanic eruptions of 1980 at Mount St. Helens: the first 100 days. - U.S. Geological Survey Professional Paper 1249, U.S. Government Printing Office, Washington, DC.

Acta Phytogeogr. Suec. 85 60 R. del Moral

Franklin, J.F., MacMahon, J.A., Swanson, F.J. & Sedell, J.R. lowing the 1980 eruption. - Can. J. Bot. 66: 989-997. 1985. Ecosystem responses to catastrophic disturbances: Podani, J. 1997. On the sensitivity of ordination and classifica­ Lessons from Mount St. Helens. - Nat. Geogr. Res. 1(2): tion methods to variations in the input order of data. - J. 198-216. Veg. Sci. 8: 153-156. Froborg, H. & Eriksson, 0. 1997. Local colonization and extinc­ Shmida, A. & Ellner, S. 1984. Coexistence of plant species with tion of field layer plants in a deciduous forest and their similar niches. - Vegetatio 58: 29-55. dependence upon lifehistory features. - J. Veg. Sci. 8: 395- Tausch, R.J., Charles, D.A., Weixelman, D.A. & Zamudio, D.C. 400. 1995. Patterns of ordination and classification instability Hanski, I. 19 82. Dynamics of regional distribution: the core and resulting from changes in input data order. -J. Veg. Sci. 6: satellite species hypothesis. - Oikos 38: 210-221. 897-902. Herben, T., Krahulec, F., Hadincova, V., Pechackova, S. & Titus, J.H., Moore, S., Arnot, M. & Titus, P.J. 1998. Inventory of Movarova, M. 1997. Fine-scale spatia-temporal patterns in the vascular flora of the blast zone, Mount St. Helens, a mountain grassland: do species replace each other in a Washington. - Madroiio 45: 146-161. regular fashion? - J. Veg. Sci. 8: 217-224. Tsuyuzaki, S. 1991. Species turnover and diversity during early Hoagland, B.W. & Collins, S.L. 1997. Gradient models, gradi­ stages of vegetation recovery on the volcano Usu, northern ent analysis, and hierarchical structure in plant communi­ Japan. -J. Veg. Sci. 2: 301-306. ties. - Oikos 78: 23-30. van der Maarel, E. & Sykes, M.T. 1993. Small-scale plant Kik:vidze, Z. 1993. Plant species associations in alpine-subnival species turnover ain limestone grassland: the carouselmodel vegetation patches in the Central Caucasus. - J. Veg. Sci. and some comments on the niche concept. - J. Veg. Sci. 4: 4: 297-302. 179-188. Laurie, H. & Cowling, R.M. 1995. Lottery coexistence models van der Maarel, E. & Sykes, M.T. 1997. Rates of small-scale extended to plants with disjoint generations. - J. Veg. Sci. species mobility in alvar limestone grassland. - J. Veg. Sci. 6: 161-168. 8: 199-208. Lavorel, S. & Lebreton, J.D. 1992. Evidence for lottery recruit­ van der Maarel, E., Noest, V. & Palmer, M.W. 1995. Variation ment in Mediterranean old fields. - J. Veg. Sci. 3: 91-100. in species richness on small grassland quadrats: niche McCune, B. & Mefford, M.J. 1997. PC-ORD, multivariate structure or small-scale plant mobility?- J. Veg. Sci. 6: analysis of ecological data, version 3.0. - Mj M Software 741-752. Design, Gleneden Beach, OR, 126 pp. Walker, L.R. 1999. Ecosystems of disturbed ground. - Elsevier, Myster, R.W. & Pickett, S.T.A. 1994. A comparison of succes­ Amsterdam, 868 pp. sion over 18 years in 10 contrasting old fields. - Ecology Wood, D.M. & del Moral, R. 1987. Mechanisms of early pri­ 75: 387-392. mary succession in subalpine habitats on Mount St. Helens. Oksanen, J. & Minchin, P .R. 1997. Instability of ordination -Ecology: 68: 780-790. results under changes in input order data: explanations and Wood, D.M. & del Moral, R. 1988. Colonizing plants on the remedies .-J. Veg. Sci. 8: 447-454. Pumice Plains, Mount St. Helens, Washington. - Am. J. pfitsch, W.A. & Bliss, L.C. 1988. Recovery of net primary Bot. 75: 1228-1237. production in subalpine meadows of Mount St. Helens fol-

Acta Phytogeogr. Suec. 85 Primary succession on lava flows on Mt. Etna

Emilia Poli Marchese* & Maria Grillo

Istituto di Biologia ed Ecologia vegetate, Universitil di Catania, Via Etnea 440, /-95128 Catania, Italy; *Corresponding author; Fax +39095553273; E-mail [email protected]

Abstract Introduction

Vegetation development during succession on six of the Colonization and establishment of plantson new surfaces more recent lava flows of Mt. Etna is exemplified by is one of the most interesting ecological phenomena. analysing the structure and species composition of the Successional stages of vegetation may relate primarily to vegetation. The following dynamic structural stages are habitat differences. They cause changes in vegetation identified: 1. Lichen and moss colonies on rocks, repre­ characteristics such as structure, density, cover, floristic sented by the Rhizocarpetea geographici and Grimmio­ composition, species richness and diversity. Many au­ Rhacomitrietea heterostichi communities respectively. 2. thors have dealt withthe topic of primary succession (e.g. Herbaceous communities, mostly of theclass Tuberarietea Beardsley & Cannon 1930; Smathers & Mueller-Dombois guttatae and in some sites of the Stellarietea mediae , on 1974; Whittaker et al. 1989; van der Maarel et al. 1985; shallow soil, dominated by dwarf therophyte species or by Mueller-Dombois 1992). Colonization andvegetation suc­ perennials. 3. Shrub communities established in cracks in cession on barren volcanic substrate have attracted a great the rocks and mostly dominated by Spartiumjunceum, deal of scientific interest. In Europe studies on these Pistacia terebinthus at lower elevations and by Genista successions have been conducted by various scientists. aetnensis at higher elevations. 4. Woody communities, Particularly worth mentioning are the studies by localized in deeper cracks and characterized by Quercus Fridriksson( 1966, 1975, 1992); Fridriksson & Magnusson ilex and thermophilous deciduous oaks. 1992; Bjarnason(19 91) and other botanists in Iceland, by Both the shrub and the woody stages, in which groups Raus (1986, 1988) of the volcanic islands in the Aegean; of species which dominate the herbaceous communities by Gonzales et al. (1990) and other botanists of the are often represented as well, gradually tend to form Canary Islands. In Italy studies have been devoted mainly communities of the evergreen Mediterranean vegetation to Mt. Vesuvius (Agostini 1975; Mazzoleni et al. 1988, of the class Quercetea ilicis. 1989; Mazzoleni & Ricciardi 1993) and Mt. Etna (Poli All the stages identified here belong to the dynamic 1965,. 1970a, b; Poli 1971; Poli & Grillo 1975; Di Bene­ series of this vegetation. They can be related to the age of detto 1983; Poli Marchese et al. 1995). In these studies the lava and to other factors depending on the lava surface some aspects of the primary succession related to theage morphology, soil conditions, climate and microclimate, and characteristics of the substrates were shown. as well as biological characteristics of the participating Here the main stages of plant colonization on six lava species, notably dispersal. flows located on the southern slopes of Mt. Etna will be treated. The aim is to point out the structure, floristic composition and dynamics of the vegetation on different Keywords: Dynamical stage; Quercetea ilicis; Vegeta­ lava flows located in the same area. tion dynamics; Volcano. Study area

Nomenclature: Pignatti (1982) for phanerogams; Nimis Mt. Etna, situated in the northeast of the island of Sicily, (1992) for lichens. Italy, is a relatively recent volcano in that it was formed at the beginning of the Quaternary period. The volcanic structure rests on a layer of Pleistocene clays, reaches a height of 3350 m a.s.l. and has a basal area of 1500 kmz. The morphology of Mt. Etna is characterized by the presence of more than 200 secondary craters or 'conetti'

Acta Phytogeogr. Suec. 85: 61-70. ISBN 91-7210-085-0. (little cones), formed during lateral eruptions, by a large

Acta Phytogeogr. Suec. 85 62 E. Poli Marchese & M. Grillo

depression - an ancient caldera - on its eastern slope called 'valle del Bove' and by numerous lava flows. The composition of the volcanic substrate is the result of lava flows of different ages, from prehistoric times to today. The study area is located on the southernslopes of Mt. Etna at the lowest altitudinal belt between 700 and 1 100 m elevation, in the zone most affected by human influence. Thearea is characterized by bothold and recent lava flows (Fig. 1). Thisstudy was conducted on six more recent lava flows dating backto either 812 or 1169, 1536, 1780, 1886, 1910 and 1983 respectively (see Romano et al. 1979). The surface morphology of these lava flows varies according to the lava type. The lava may be composed of rough-textured porous clinker-like rubble ('aa' type) of solid smooth-surface ('pahoehoe' type ) or of loose fine material such as lapilli and volcanic ash, called 'ejecta' (Romano et al. 1979). The ecological factors on each lava flow vary according to the surface morphology and microclimate. This gives rise to a mosaic type arrange­ ment of vegetation (Poli 1970a, 1971). As regards the climate, data were collected from the Nicolosi meteorological station, located at 698 m a.s.l., where the mean annual rainfall is 1110 mm. The mean Fig. 1. Location of the six lava flowsfrom the southern slope of annual temperature reaches 14.3 °C. The climate, with a Mt. Etna. dry period during summer, is of the Mediterranean type. According to Bagnouls & Gaussen (1957) it can be con­ sidered meso-Mediterranean.

Methods

The vegetation was analysed according to the Braun­ Results Blanquet (1964) method in each of the six lava flows in randomly selected areas located on different microhabitats Although the area studied is situated in a zone that is and substrate types and at different elevations. The areas continually subjected to man's impact, various stages of were selected randomly, with a minimum distance of 50 the primary successions remain. These stages are distrib­ m from each other. Thestudy was carried out at altitudes uted in various ways on the different lava flows. The colonization process occurring on thelava flows between 700 and 1100 m, that is in the belt with the Quercion ilicis as type of climax vegetation. The marginal studied proceeds as follows. areas of the various lava flows were not considered be­ cause there theplant cover is greatly influenced by that of 1. Cryptogam stages the adjacent areas. Some plant communities were identified as real dy­ The first macroscopic stage of plant colonization is repre­ namic stages by means of releves. For each plant commu­ sented by the lichen and moss colonies which cover the nity in the table, the left hand column refers to the life­ roughness and little hollows in the rock. On the younger form type: T= Therophytes; G= Geophytes; H= Hemi­ lava flows, in particular those from 1910 and 1780, the cryptophytes; Ch= Chamaephytes; NP= Nano-Phanero­ cryptogam vegetation cover is as high as 80-90 %. The phytes; P= Phanerophytes. older lava, from 812 or 1169 (the age of this lava flowis Through comparison of the plant communities for doubtful in literature), is well colonized by various stages each habitat ('Habitat comparison' sensu Braun-Blanquet of phanerogamic vegetation, although the cryptogam 1964) it is possible to show the principal dynamic stages. vegetation is still well-represented. The most recent lava of 1983 lacks a cryptogam stage or any other coloniza­ tion stage. Only some rocky micro-surfaces harbour rare and small fragments of colonies of algae, mosses, and lichens in the process of organization of the thallus.

Acta Phytogeogr. Suec. 85 - Primary succession on lava flows on Mt. Etna - 63

Table 1. Parmelia pulla-P. conspersa community. Table 2. Sedum stellatum community.

Releve no. 1 2 3 4 Year of lava flow 1886 Altitude (m a.s.l.) 810 Year of lava flow 812/ 812/ 812/ 1910 Area (m2) 4 1169* 1169* 1169* Cover of vegetation (%) 50 Height of vegetation (cm) 15 Altitude (m a.s.l.) 780 720 780 1060 Area (cm2) 50 40 70 50 T Sedum stellatum 3.2 Species of the Tuberarietea guttatae Parmelia pulla 2 2 2 T Trifolium campestre 1.1 T Aira cupaniana +.2 Parmelia conspersa + I + T Rumex bucephalophorus T Trifolium arvense + Species of the Parmelienalia conspersae + Species of the Stellarietea mediae Lecidiafu scoatra + + + + T Bromus sterilis T Anthemis arvensis Lecanora bolcana 1 + + +.2 + T Lupinus angustifo lius T Sonchus oleraceus Amandinea punctata + + + T Vulpia ciliata + Species of the Apicilietalia gibbosae Other species Lecanora rupicola + + + T Centranthuscalcitrapa 1.1 T Hypochoeris glabra +.2 Aspicilia intermutans + + + Hypochoeris radicata H Linaria heterophylla Caloplaca crenularia H + + + + + T Misopates orontium + T Petrorhagia velutina Lecidella carpathica + + + + H Rumex multifidus T Senecio vulgaris Species of the Rhizocarpetea geographici T Silene conica + + Candelariella vitellina + Musci spec. pi. Aspicilia spec. + Rhizocarpon tinei + + + Parmelia saxatilis + Tephromela atra + Table 3. Sedum caeruleum community. Other species Year of lava flow 812/1169 Altitude (m a.s.l.) 760 Stereocaulon vesuvianum + + 2 Area (m2) 1 Cladonia spec. + + + Cover of vegetation (%) 50 Height of vegetation (cm) 8 Cladonia rangiformis + Lecidea spec. + T Sedum stellatum 3.2 Aspicilia spec. + Species ofthe Tuberarietea guttatae T Rumex bucephalophorus 3.2 T Trifolium arvense *The age of this lava flow is doubtful in literature. + T Oglifa gallica + Species of the Stellarietea mediae + The lichen vegetation is often dominated by the T Bromus tectorum T Echium plantagineum + Other species Stereocaulon vesuvianum community belonging to the T Sedum rubens + T Sedum stellatum + Poa bulbosa class Rhizocarpetea geographici. At times this commu­ H Agropyron panormitanum + H + Parmelia conspersa nity covers vast lava surfaces. Where the lichen vegeta­ + tion manages to develop, other communities of the class Table 4. Sedum tenuifolium community. Rhizocarpetea geographici become established. One of Year of lava flow 1780 Altitude (m a.s.l.) 1100 the most widespread and well organized of these is the Area (m2) I Cover of vegetation (%) Height of vegetation (cm) Parmeliapulla-P. conspersa community (Table 1). This 50 10 vegetation, which has already been identified on other Ch Sedum tenuifolium 3.2 Species of the Tuberarietea guttatae lava flows on Mt. Etna (Poli Marchese et al. 1995), T Briza maxima 1.1 T Airacupaniana + has a xerophilous character and is typical of acid Species of the Stellarietea mediae T + Bromus tectorum substrates. Bromus madritensis T + Other species Insmall cavities inthe rock where there is a thin layer H Agropyron panormitanum H Chondrilla juncea + + of soil, mosses become established. These are represented by various species including those belonging to the genus Rhacomitrium (R. heterosticum, R. canescens). The moss stellatum community (Table 2) and the Sedum caeruleum communities belong to the classes Grimmio-Rhaco­ community (Table 3) are the most widespread. mitrietea heterostichi and Ceratodonto-Polytrichetea In more restricted areas communities dominated by pilife ri (Poli Marchese et al. 1995). the perennial Sedum tenuifolium (Table 4) occur, in which moss and lichen colonies are still present. Other commu­ 2. Dwarftherophyte vegetation nities are characterized by other therophytic species such as Airacupa niana (Table 5) or other grass species (Vulpia Some communities made up of dwarf herbaceous plants, ciliata, Catapodium rigidum, Lamarckia aurea) and yet almost exclusively annuals, became established on very others by Leguminosae or by still other species. These are small surfaces, where there is a thin layer of soil on top of all pioneer stages of a xerophilous character belonging to moss colonies (cf. Poli 1970b ). They are the firstphanero­ the class Tuberarietea guttatae and identified in the Etna gamic stage of the primary succession and are present on territoryon various lava surfaces (Poll 1970b; Poli & Grillo all the lavas except the youngest, that of 1983. 1975; Di Benedetto 1983; Poli Marchese et al. 1995). This vegetation, which is very poor, is widespread on In the areas most influenced by man the vegetation the various lava flows and has a variable floristic compo­ includes ruderal species characteristic of the class Stella­ sition. Some communitiesare dominated by annual Crassu­ rietea mediae sensu Rivas-Martinez (1975). In some areas laceae, above all those of the genus Sedum. The Sedum these species make up the only significant floristic corn-

Suec. Acta Phytogeogr. 85 64 E. Poli Marchese & M. Grillo

Table 5. Aira cupaniana community. Table 6. Trifolium angustifo lium community.

Releve no. 1 2 3 Year of lava flow 812/1 169 Altitude (m a.s.l.) 800 Year of lava flow 1780 812/ I910 Area (m2) 1 1169 Cover of vegetation (%) 80 Height of vegetation (cm) 15 Altitude (m a.s.l.) liOO 790 1070 Area (cm2) I T Trifolium angustifolium 3.2 I 1 Cover of vegetation (%) 35 50 10 Species of the Brometalia rubenti-tectori andStellarietea mediae Height of vegetation (cm) lO 6 10 T Bromus tectorum + T Vulpia ciliata + T A vena barbata + T Stipa capensis + T Aira cupaniana l.l 1.1 1.1 Other species Species of the Tuberarietea guttatae Ch Sedum tenuifolium 3.2 T Trifolium arvense 2.2 1.2 T Briza maxima + + Table 7. Micromeria graeca community. T Oglifa gallica + + T Rumex bucephalophorus 2.2 Year of lava flow 812/ll69 Altitude (m a.s.l.) 800 T Trifolium campestre + Area (m2) 4 T Plantago bellardi + Cover of vegetation (%) 60 Height of vegetation (cm) 20 Other species Ch Micromeria graeca 3.2 H Poa bulbosa 1.2 + Ch Sedum tenuifolium + + Species of the Tuberarietalia and Tuberarietea guttatae T Petrorhagia velutina + + T Trifolium arvense 1.1 T Trifolum stellatum 1.1 H Agropyron panormitanum + + T Aira cupaniana +.2 T Briza maxima + T Sedum rubens + + T Plantago bellardi + T Rumex bucephalorophus + T Bromus madritensis + + T Trifolium campestre + Y Oglifa arvensis + Species of the Brometalia rubenti-tectori and Stellarietea mediae T Parentucellia viscosa + T Trifolium stellatum 1.1 T Vulpia ciliata 1.1 T Myosotis incrassata + T Avena barbata + T Echium plantagineum + Candelariella vitellina + + T Trifolium cherleri + T Trifolium glomeratum + Caloplaca crenularia + Other species Aspicilia spec. + T Biscutella didyma + H Lactuca viminea + T Oglifa arvensis + T Petrorhagia velutina + T Sedum caeruleum + T Sedum rubens + Ch Sedum tenuifolium + ponent (Table 6). This stage, which is the firstphanerog amic stage on young volcanic substrates, represents in some Table 8. Asphodeline lutea community. areas a successionally intermediate stage between the first Year of lava flow 1780 Altitude (m a.s.l.) 900 cryptogam stages and the following phanerogams. Area (m2) 15 Cover of vegetation (%) 80 Height of vegetation (cm) 40

Asphodeline lutea 3.2 3. Other herbaceous stages Species of the Brometalia rubenti-tectori and Stellarietea mediae T Bromus tctorum 2.2 T Vulpia ciliata 2.2 On the lava surfaces there are some herbaceous communi­ T Avena barbata 1.2 T Bromus madritensis + T Echium plantagineum + T Erodium cicutarium + ties with a variable floristiccomposition . They colonize Species of the Tuberarietalia and Tuberarietea guttatae substrates where a certain amount of soil has accumu­ T Trifolium arvense 2.2 T Aira cupaniana 1.2 T Crupina crupinastrum + T Ornithopus compressus + lated. This vegetation, often with steppe characteristics, T Plantago bellardi + T Trifolium campestre + discontinuously covers the surfaces on which it becomes T Trifolum stellatum 1.1 established. The areas most affected by rich in Species ofthe Poetea bulbosae man are H Poa bulbosa 1.2 T Parentucellia latifolia + nitrophilousspecies . Examples include communities domi­ Other species nated by Micromeria graeca (Table 7) and Asphodeline T Lagurus ovatus 1.2 T Biscutella lyrata 1.2 H Isatis tinctoria 1.1 H Achillea ligustica + lutea (Table 8) respectively. A community dominated by T Arenaria serpyllifo lia + G Asphodelus microcarpus + Ferula communis (Table 9) is widespread on the oldest H Carlina corymbosa + T Centaurea cyanus + H Ferula communis + T Hypochoeris glabra + lava, on areas frequently subjected to pasture activity. H Linaria purpurea + T Medicago spec. + These communities belong to the class Tuberarietea T Silene conica + T Silena gallica + H Silene vulgaris ssp. vulgaris + T Trifolium molinerii + guttatae and often contain ruderal elements of the order T Daspyrum villosum + T Vicia sativa* + Vicia villosa ssp. varia + * ssp. angustifolia Brometalia rubenti-tectori and of the class Stellarietea T mediae. In some communities the species of this class predominate. The presence in the same vegetation of elements be­ Senecio ambiguus, Scrophularia canina, Helichrysumitali­ longing to the two classes cited can be attributed - in cum and, in some areas, also Euphorbia characias, a dwarf agreement with Izco ( 1977) and other authors - to a phanerophyte. transition between one form of vegetation and another These species often colonize, also alone, incoherent caused by the influence of man. and changeable substrates. Below and around the suf­ Between the lava blocks, on deeper soil, discontinu­ frutescent species various species become established. ous vegetation dominated by perennial herbs is wide­ Sometimes these also include seedlings of tree species spread. Here, an important role is played by Rumex (Table 13). This confirms the role they play in the lava scutatus, Centranthus ruber and other species such as colonization process.

Acta Phytogeogr. Suec. 85 - Primarysuccession on lava flowson Mt. Etna - 65

Table 9. Ferula communis community. Table 11. Rumex scutatus-Centranthus ruber community.

Year of lava flow 812/1 169 Altitude (m a.s.l.) 800 Releve no. 1 2 3 Area (m2) 10 Year of lava flow 1910 1780 1886 Cover of vegetation (%) 40 Height of vegetation (cm) 30 Altitude (m a.s.l.) 1070 1100 830 Area (m2) 50 50 30 2. 1 H Ferula communis Cover of vegetation (%) 60 40 40 Species of the Tuberarietea guttatae Height of vegetation (cm) 30 25 25 + T Briza maxima + T Crupina crupinastrum + H Rumex scutatus 2.2 1.2 1.2 T Trifolium arvense + T Trifolum campestre 1.1 1.1 1.1 Trifolium stellatum + T Aira cupaniana T T Galium parisiense Species of the Stellarietea mediae + T Crupina crupinastrum + T Avena barbata + T Bromus madritensis + H Jasione echinata + T Vulpia ciliata + Species of theStellarietea mediae Other species T Bromus madritensis + 1.2 Ch Sedum tenuifo lium 1.1 Ch Micromeria graeca +.2 T Bromus tectorum 1.1 G Asphodelus microcarpus + H Carlina corymbosa + T Bromus sterilis + Ch Centranthus ruber + T Cy nosurus echinatus + Vulpia ciliata + NP Euphorbia characias + Lagurus ovatus + T T Echium plantagineum + T Orobanche rapum-genistae + T Trifolium molinerii + T Other species T Andryala integrifolia + + T Sedum rubens + + H Daucus carota + + Table 10. Rumex scutatus community. H Carlina corymbosa + + G Umbilicus rupestris + Re1eve no. 1 2 H Lactuca viminea + Year of lava flow 1910 1780 H Linaria heterophylla + Altitude (m a.s.l.) 1070 990 H Poa bulbosa + Area (m2) 50 40 H Isatis tinctoria + + Cover of vegetation (%) 60 25 H Agropyron panormitanum + 2 15 + Height of vegetation (cm) 0 T Geranium robertianum Ch Helichrysumitalicum +.2 4.3 2.2 H Rumex scutatus Ch Senecio ambiguus + Species of the Tuberarietea guttatae Ch Sedum tenuifolium + T Briza maxima 1.1 + Cynosurus echinatus + T + Species of the Stellarietea mediae H Hypochoeris radicata Vulpia ciliata +.2 H Scrophularia canina + T T Bromus tectorum + Stereocaulon vesuvianum + + Candelariella vitellina + Other species Cladonia rangiformis + T Andryalaintegri folia + H /satis tinctoria + + H Hypochoeris radicata + Table 12. Centranthusruber community. T Sedum rubens + H Crepis leontodontoides + Year of lava flow 1780 990 T Bromusgussonei + Altitude (m a.s.l.) + Area (m2) 40 T Dasypyrum villosum H Linaria heterophylla + Cover of vegetation (%) 60 Height of vegetation (cm) 30 Ch Sedum tenuifolium + Ch Centranthus ruber 3.3 T Cynosurus echinatus + G Umbilicus rupestris Species of the Tuberarietea guttatae + + Candelariella vitellina + T Briza maxima T Galium parisiense + H Jasione echinata + Species of the Stellarietea mediae T Bromus tectorum 2.2 H Reichardia picroides + T A vena barbata + Other species Ch Sedum tenuifolium 1.2 T Cynosurus echinatus + Some examples of the vegetation found in these habi­ H Anthoxanthum odoratum 1.1 T Oglifa arvensis 1.1 +.2 +.2 tats are shown in Tables 10 - 14. From these it can be seen T Geranium robertianum Ch Micromeria graeca H Linaria heterophylla + H Isatis tinctoria + that the vegetation includes species both from the Sedum rubens + G Umbilicus rupestris + T xerophilous communities(Tuberarietea guttatae) and from H Agropyron panormitanum + H Crepis leontodontoides + T Petrorhagia velutina + T Arenaria serpyllifolia + the ruderal communities (Stellarietea mediae) the latter T Lagurus ovatus + being favoured by man's activity in the area. Only rarely are they absent (see Tables 13 and 14 ) . Although these communities are dominated by species such as Rumex scutatus, Centranthus ruber and more 4. Shrub-dominated stages sporadically Scrophularia canina, of the class Thlaspietea rotundifoliae,typical of incoherent and changeablesubstra tes, In cracks in the rock where the soil is deeper isolated they cannot be said to belong to this class which is well shrubs become established. With time, asthe colonization represented in other phytogeographic regions. proceeds, the vegetation cover gradually increases and the In the areas most affected by pasture activity, species shrubs begin to become organized intocommunit ies. These belonging to the class Poetea bulbosae are widespread communities, which have a variablefloristic composition, (see Poli Marchese et al. 1995). constitute a well defined stage of the primary succession.

Acta Phytogeogr. Suec. 85 66 E. Poli Marchese & M. Grillo

Table 13. Senecio ambiguus community. Table 16. Spartiumjunceum-Quercus pubescens community.

Year of lava flow 1886 Altitude (m a.s.l.) 800 Year of lava flow 812/1 169 Altitude (m a.s.l.) 780 Area (m2) 30 Area (m2) 30 Cover of vegetation (%) 40 Height of vegetation (cm) 35 Cover of shrub vegetation (%) 60 Height (m) 1.30 Cover of herb vegetation (%) 50 Height (cm) 30 Ch Senecio ambiguus 3.2 Sp artium junceum 3.2 Quercus pubescens s.l. Species of the Tuberarietea guttatae p p u T Briza maxima + Airacupaniana + Species of the Quercetea ilicis T H Trifolium arvense NP Euphorbia characias 2.2 + p Pistacia terebinthus + Other species Species of the Brometalia rubenti-tectori and Stellarietea mediae + H + 1.2 T H Reichardia picroides Daucus carota T A vena barbata Bromus sterilis + H + + T T Linaria heterophylla Ch Sedum tenuifolium Bromus madritensis + Geranium molle + + H Umbilicus rupestris T Ch Helichrysum italicum + T Vulpia ciliata + Urospermumpicroides + H + + Poa bulbosa H Isatis tinctoria Species of the Tuberarietea guttatae T Bellardia trixago + Quercus pubescens s.l. + Trifolium stellatum + + p T T Cerastium H Ceterachoffi cinarum + Aspicila spec. + T T semidecandrum Rumex bucephalophorus + Trifolium campestre + Other species T Galium aparine 2.2 H Isatis tinctoria 2.2 14. Table Scrophularia canina community. T Geranium robertianum 1.1 Ch Centranthus ruber 1.1 Vicia sativa ssp. angustifolia + Ferula communis + Year of lava flow 81211 169 Altitude (m a.s.l.) 780 T H Cynosurus echinatus Vicia villosa varia + Area (m2) 10 T + T ssp. Leopoldia comosa Rubus ulmifolius + cover of vegetation (%) 40 Height of vegetation (cm) 35 G + NP Silene vulgaris . H + T Lagurus ovatus + 2.2 Carduus pycnocephalus + Ch Scrophularia canina H + T Hypochoeris glabra Linaria reflexa + Species of the Tuberarietea guttatae T + T Oglifa arvensis Briza maxima + Rumex bucephalophorus + Rumex scutatus + Veronica cymbalaria + T T H T + Trifolium campestre + H Trifolium arvense T Other species Table 17. Genista aetnensis community; releves 1 and 2: Ch Senecio ambiguus 1.2 Ch Helichrysum italicum 1.1 +.2 stage; rels. 3 and 4: evolved stage. Ch Teucrium italicum T Echium plantagineum + initial H Carlina corymbosa + Ch Centranthus ruber + Releve no. 1 2 3 4 NP Euphorbia characias + H Ferula communis + Year of lava flow 1910 1886 1886 1780 H Isatis tinctoria NP Rubus ulmifolius + + Altitude (m a.s.L) 1070 830 890 H Rumex scutatus + T Sedum caeruleum + lUO Area (m2) 50 20 20 20 Cover of shrub vegetation (%) 80 50 50 60 Table 15. Spartiumjunceum community. Height (m) 2.2 2.0 2.0 3.0 Cover of herb vegetation (%) 40 50 30 30 Year of lava flow 1886 Altitude (m a.s.l.) 890 Height (cm) 50 50 30 50 Area (m2) 30 4.2 3.2 3.1 3.2 Cover of shrub vegetation (%) 40 Height (m) 1.3 p Genista aetnensis 1.2 Cover of herb vegetation (%) 30 Height (cm) 30 p Quercus pubescens s.L (shrub layer) 1.1 Quercus pubescens s.L (herb layer) + P Spartium junceum 2.2 p Species of the Tuberarietea guttatae Species of the Tuberarietea guttatae + 1.2 T Briza maxima + Briza maxima 1.2 T Aira cupaniana +.2 + 1.2 T T Aira cupaniana H Trifolium arvense + T Jasione echinata + Crupina crupinastrum T l.l Species of the Stellarietea mediae T Galium parisiense + T Bromus madritensis 1.2 H Reichardia picroides + H Trifolium arvense + T Vulpia ciliata + Species of the Stellarietea mediae Other species T Bromus madritensis + 1.2 1.2 1.1 T + +.2 Ch Centranthus ruber + Ch Sedum tenuifolium A vena barbata H Rumex scutatus +.2 Ch Helichrysum italicum +.2 H Reichardia picroides + H +.2 + Vulpia ciliata Scrophularia canina H Isatius tinctoria T + G + + T Bromus sterilis Umbilicus rupestris H Linaria heterophylla + T + Lagurus ovatus + Cynosurus echinatus T Ch Senecio ambiguus + H Daucus carota + Other species T + + T Geranium robertianum 1.2 2.3 Sedum rubens H Lactuca viminea + + Musci spp. + T Rumex scutatus + 1.2 1.1 Parmelia pulla + + H Isatis tinctoria + + + l.1 Ch Centranthus ruber 1.1 + 1.2 Sedum tenuifolium 1.1 + Ch Ll Umbilicus rupestris + + + At lower elevations the communities are .most fre­ G + H Carlina corymbosa + + quently dominated by Sp artium junceum, at higher eleva­ Senecio ambiguus Ch L2 + + + tions by Genista aetnensis.This species has an important T Lagurus ovatus Musci spp. + + role in the colonization of the Etna lavas, where it charac­ terizes various types of vegetation as it has a very variable floristic composition. In the shrub vegetation we can identify both initial rietea guttatae and Stellarietea mediae, respectively) are stages which are more similar to the herbaceous stages present. This shows the dynamic relationships that exist and more mature stages which are more like woodland between thesetypes of vegetation and the shrubvegetation. vegetation. Two examples of the former are giv�n in On the older lava substrates, where the soil has devel­ tables 15 and 17, where some herbaceous species belong­ oped more, the shrub communities include other woody ing to both xerophilous and ruderal vegetation (Tubera- species (Tables 16 and 17: rels. 3 and 4) and frequently,

Acta Phytogeogr. Suec. 85 - Primary succession on lava flowson Mt. Etna - 67

Table 18. Quercus ilex-Quercus pubescens scrubland. Table 20. Quercus ilex community.

Year of lava flow 812/1 169 Altitude (m a.s.l.) 810 Year of lava flow 812/ 1169 Altitude (m a.s.L) 790 Area (m2) 30 Area (m2) 40 Cover of shrub vegetation (%) 80 Height (m) 3.0 Cover of tree vegetation (%) 95 Height (m) 8.0 Cover of herb vegetation (%) 30 Height (cm) 30 Cover of shrub vegetation (%) 10 Height (m) 1.5 Cover of herb vegetation (%) 70 Height (cm) 25 P Quercus ilex (s) 2.2 P Quercus ilex (h) 1.2 P Quercus pubescens s.I. (s) 1.1 P Quercus pubescens (h) + Quercus ilex (t) 4.5 1.1 p p Quercus ilex (s) Species of the Quercetea ilicis p Quercus ilex (h) + G Asparagus acutifolius 1.2 NP Euphorbia characias 2.2 Species of the Quercetea ilicis P Pistacia terebinthus (s) l.l P Pistacia terebinthus (h) + NP Euphorbia characias 1.1 G Asparagus acutifolius + Other species Other species T Galium aparine 2.2 Ch Micromeria graeca 1.2 H Ferula communis 2.2 T Bromus madritensis 2.2 T Bromus madritensis + G Asphodelus microcarpus + T Trifolium molinerii 2.2 T Biscutella lyrata +2 H Isatis tinctoria + T Trifolioum arvense + T Galium aparine +.2 H Thapsia garganica + G Allium subhirsutum + T + Rubus ulmifolius + T Briza maxima NP + T + T T + Hypochoeris achyrophorusa Cynosurus echinatus Vulpia ciliata + Cynosurus echinatus T Stellaria media + H Daucus carota + H Crepis leontodontoides + T Scandix pecten-veneris + H Festuca circummediterranea + T Geranium rotundifolium + H Hypochoeris radicata + H Isatis tinctoria + G G + Table 19. Quercus pubescens community. Cyclamen hederifolium + Leopoldia comosa T Silene gallica + T Silene vulgaris* + + Year of lava flow 1780 Altitude (m a.s.l.) 900 H Ranunculus spec. T Vicia pseudocracca + T + Area (m2) 30 Silene spec. *ssp. vulgaris Cover of tree vegetation (%) 10 Height (m) 4.0 Cover of shrub vegetation (%) 50 Height (m) 1,0 Cover of herb vegetation (%) 40 Height (cm) 40

P Quercus pubescens (t) 3.2 P Quercus pubescens (s) 1.2 Tuberarietalia Tuberarietea guttatae Table 21. Quercus ilex-Quercus pubescens woodland. Species of the and T Trifolium stellatum + T Cerastium + semidecandrum T + T + Releve no. 2 3 4 Rumex bucephalorophus Trifolium campestre I T Briza maxima + T Airacupaniana + Year of lava flow 1536 1536 1780 1780 T Ornithopuscompressus + T Crupina crupinastrum + Altitude (m a.s.L) 1070 1070 920 900 Area (m2) 30 30 30 40 Other species Cover of tree vegetation (%) 60 90 60 60 T Avena barbata 1.2 Vicia sativa* 1.2 T Height (m) 7.0 5.0 6.0 4.0 H Poa bulbosa 1.2 T Geranium robertianum 1.1 Cover of shrub vegetation (%) 10 10 15 10 T Bromus sterilis + T Galium aparine + Height (m) 1.0 1.0 1.0 1.0 H Ferula communis + Cynosurus echinatus + T Cover of herb vegetation (%) 30 20 20 60 T Vicia villosa ssp. varia + H Crepis leontodontoides + 30 40 25 40 G + Height (cm) Umbilicus rupestris H Daucus carota + G Ranunculus spec. + T Anthoxanthum odoratum + 2.1 4.2 3.2 1.2 p Quercus ilex (t) T Centaurea cyanus + G Cy clamen hederifolium + 1.1 + + p Quercus ilex (s) Genisga aetnensis (e) + T Lathyruspurpurea + p + p Quercus ilex (h) Parentucellia viscosa + T Petrorhagia velutina + 2.1 1.2 1.1 3.2 T p Quercus pubescens (t) H Smyrnium rotundifolium + * ssp. angustofo lia + 1.2 p Quercus pubescens (s) + + + + p Quercus pubescens (h) Species of the Quercetea ilicis Euphorbia characias 1.1 + NP l.l together with species of the herbaceous vegetation, spe- + p Lonicera etrusca + cies belonging to the woodland vegetation of the class p Pistacia terebinthus (s) l.l 1.2 p Rhamnus alaternus (s) Quercetea ilicis (see Table 16) present. In some + are areas p Calitome villosa + shrublands begin form more evolved vegetation types p Rubia peregrina to G Asparagus acutifolius + dominated by arboreal species characterizing the woods of Celtis tournefo rtii 1.1 p (s) the region. Anexample is thevegetation shown in Table 18, Species of the Tuberarietea guttatae T Trifolium stellatum + +.2 1.1 which is located on theoldest lava flow. Here, together with T Briza maxima + + the dominant elements of the woodland vegetation, other T Trifolium arVense + T 1.2 species of the class Quercetea ilicis are present. Vulpia myuros There are more mature shrublands, which at times Other species appear as brushwoods of the Quercetea ilicis class on the T Galium aparine + + + + H Crepis leontodontoides 1.1 + + + oldest lava flows. G Umbilicus rupestris + 1.1 + + H Ferula communis + + + T 1.1 + Geranium robertiarium + 5. Tree-dominated stages H Poa bulbosa 1.2 + + H +.2 Ceterach of ficinarum + + + + H Linaria heterophylla + Where the soil has reached a certain thickness, as in T Cynosurus echinatus + + 1.2 + cavities and deep cracks in the lava flows, vegetation T Anthoxanthum odoratum H Asplenium trichomanes · + + types occur whose composition and structure closely re- G Leopoldia comosa + + semble those of woodlands. Often only fragments of T Sedum rubens + + T Vicia sativa ssp. angustifolia + 1.1 woodlands are found here. These types are usually domi- H Thapsia garganica 1.1 + nated by Que reus ilex and by Quercus pubescens s.l. or by Ch Micromeria graeca +.2 +.2

Acta Phytogeogr. Suec. 85 68 E. Poli Marchese & M. Grillo

both species (see Tables 19 - 21). They represent a succes­ As put forward by other authors, there are also other sionally intermediate stage between the shrub vegetation factors to be considered. For example Walker & Chapin and the more mature woodlands belonging to the class (1987) and del Moral & Wood (1993) showed how the Quercetea ilicis. The Quercus pubescens s.l. community primary succession can be influenced by the amelioration (Table 19) whichis rich in species of theclass Tuberarietea of a barren substrate before the first seed lands or by other guttatae and does not contain elements of the class unpredictable events (e.g. seasonal changes of climatic Quercetea ilicis, is the less mature stage. conditions). Some improvements depend on erosion fac­ The floristic composition gradually changes during the tors, as Timmins (1983), Kadomura et al. (1983) and del succession up to the last dynamic stage. On sites disturbedby Moral & Wood (1993) have pointed out. Other soil factors man's activity the vegetation cannotreach the climax stage. (e.g. nutrients in the substrate, soil moisture) are highly variable between microsites (Fig. 3). Theplant colonization process is also regulated by Discussion biological factors connected with the colonizing species concerned, such as dispersal type. In the stages we have The establishment of plant cover on the lava flowsstudied identified, for example, the herbaceous species with the has reached different stages. Theyare not uniformly present greatest cover are almost all dispersed by the wind. on allthe lava flows.For example the most evolved stages Other factors connected with the characteristics of the characterized by woody vegetation are only present on the species are: germination capacity, growth and longevity. oldest lava flowsand not on the more recent ones of 1910 This was emphasized by Chevennement (1990), who and 1983. On the 1983 flow, not even 20 years old, none referred to Connell & Slatyer's (1977) 'tolerance' plant of the stages is present even though in some rocky succession model. microhabitats fragments of algae, lichens or bryophytes In our study area we consider the same authors' 'fa­ are beginning to appear and, in very rare sites; near the cilitation' model to be applicable. This model, which is limits of the lava flows, specimens of Rumex scutatus also supported by Lebrun (1959), Leonard (1959) and occur. The various stages identified belong to the same many others, is based on the principle that the species dynamic series: the series of Mediterranean evergreen present in each stage of the succession modify the envi­ vegetation of the class Quercetea ilicis (Fig. 2). ronment where they appear making it more favourable for Thedynamic stages present on each lava flowcannot other species, which become established and then com­ only be related to the age of the lava as other factors come pete with and eventually eliminate the firstones. With this into play. Of particularimportance are the factors depend­ principle in mind we have indicated the main successional ing on the lava substratesuch as surface morphology and pathways (Fig. 2). the depth of the soil as well as the climatic and microclimatic conditions.

Potential natural vegetation

ercetea ilicis communities)

Bushwood dominated by:

Quercusilex thermophilous deciduous oaks (Q. roburgroup)

Shrub stages dominated by:

shrubs: trees with shrub habitus:

Sp attiumjunceum Quercu. ile;x; Calicotome villosa termophi/ou. deciduous oaks Pistacia terebimhus (Q. robur group) Genistaaetnensis Rhamnu. alatemu., Celtis tournefortli

Herbaceous stages ( Tuberarietea guttatae, r--- Brometalia rubenti-tector�) I L...... -...--_____J Dwarftherophytic vegetation (Tuberarietea gutta/ae)

C1}11togam stages (Mosses Lichens) and Fig. 2. Pathways of primary suc­ cession on lava flows on the southern slopes of Mt. Etna. Crags and hollows Thin layers of soil Lava blocks Deepcrevices Ejecta

Acta Phytogeogr. Suec. 85I - Primary succession on lava flows on Mt. Etna - 69

Time Surfiu:e morphology of the lava Climate Biological factors (characteristics of the species) (age of lava) ! Microclimate

I Kind of site Dispersal Growth Longevity Number ofsites Gennination Reproductivesu ccess Soilcondit ions I � (depth, moisture, nutrients) mhobit• ooOOitiDM l J Amelioration of barren substrate -1--�

Unpredictable events

Vegetation stages

Stnu:ture div..-sity coverof vegetation

L------Vegetation development Fig. 3. Main factors influencing the primary succession on lava flows on Mt. Etna and their dynamic relations.

This study, which follows others of other lava flows References on the lower slopes of Mt. Etna (see Poli & Grillo 1975; Di Benedetto 1983; Poli Marchese et al. 1995), provides Agostini, R. 1975. Vegetazione pioniera del Monte Vesuvio: aspetti fitosociologici ed evolutivi. -Arch. Bot. Biogeogr. further information about the plant colonization process !tal. 51: 11-34. in the Mt. Etna territory.The stages of the primary succes­ Bagnouls, F. & Gaussen, H. 1957. Les climats biologiques et sion on the lava flows of the southern lower slopesof the leur classification. -Ann. Geogr. 355: 193-220. volcano have a certain similarity in accordance with the Beardsley, G.F. & Cannon W.A. 1930. Notes on the effects of a age, physical structure of the relevant substrates and the mud flow at Mount Shasta. -Ecology 11: 326-336. other main colonization factors. On the more humid east­ Bjarnason, A.H. 199l.Vegetation on lava fields in the Hekla ern slopes the primary succession proceeds more rapidly. area, Iceland. - Acta Phytogeogr. Suec. 77: 5-114. Here, on the 1651 lava flow the dynamic stages are more Braun-Blanquet, J. 1964. Pfl anzensoziologie. Grundziige der Vegetationskunde. 3rd. - Springer, Wien. similar to those which on the southern face are found on Ed. the lava of either 812 or 1169 than to the almost coeval Chevennement, R. 1990. La colonisation vegetale d'un champ de lava de la Reunion. - Soc. Biogeogr. 66 (2): 47-61. lava of 1669. Moreover, the two lava flows, those of 1651 C.R. Connell, J.H. & Slatyer, R.O. 1977. Mechanisms of succession and of either 812 or 1169, have a more or less similar in naturalcommunities and their role in community stability physiognomy as concerns the plant cover considered as a and organization. -Am. Nat. 111: 1119-1144. whole. In this case it is the macroclimate that plays an Del Moral, R. & Wood, D.M. 1993. Earlyprimary succession on important role. the volcano Mount St. Helens. - J. Veg. Sci. 4: 223-234. It is, in fact, the conditions of the typical Mediterra­ Di Benedetto, L. 1983. Evoluzione della vegetazione sulla colata nean climate that make the primary succession processes lavica etneadel 1669. - Not. Soc. !tal. Fitosociol. 18: 19- slower on Mt. Etna. Moreover, the combination of the 35. summer dryness with the unfavourable conditions pro­ Fridriksson, S. 1966. The possible oceanic dispersal of seed and other plantparts to Surtsey. - Surtsey Res. Progr. ep. 2: 59- vided by the volcanic substrate promotes microclimatic R 62. overheating of the black rocks and makes the primary Fridriksson, S. 1975. Surtsey. Evolution of life on a volcanic succession even more difficult (see Richter 1984; Raus island. - Butterworths, London. 1986). Fridriksson, S. 1992. Vascularplants on Surtsey 1981-1990. - Surtsey Res. Progr. Rep. 10: 17-30. Fridriksson, S. & Magnusson, B. 1992. Development of the ecosystem on Surtsey with references to Anak Krakatau. - GeoJournal 28: 287-291. Gonzales, J.M., Sanchez, L., Beltran, E. & Losada, A. 1990. Contribucion al estudio floristico de las coladas historicas

de las Islas Canarias. I Chinyero (Tenerife). - An. lard. Bot. Madrid 46: 437-444.

Acta Phytogeogr. Suec. 85 70 E. Poli Marchese & M. Grillo

Kadomura, H., Imkagawa, T. & Yamamoto, K. 1983. Eruption­ colata lavica etnea del 1381. - Atti 1st. Bot. Lab. Critt. induced rapid erosion andmass movements on U su Volcano, Univ. Pavia 1975: 127-185.

Hokk:aido.-Z. Geomorph. 46: 123-142. Poli Marchese,E., Grillo, M. & Lo Giudice, R. 1995. Aspetti del Izco, J. 1977. Revision sintetica de Ios pastizales del suborden dinamismo della vegetazione sulla colata lavica del Brometalia rubenti-tectori. - Coli. Phytosociol. 6: 37-54. 1651 del versante orientate dell 'Etna. - Coli. Phytosociol. Lebrun, J. 1959. Sur les processus de la colonisation vegetale 24: 241- 264. des champs de lava de Virunga (Congo Beige). - Bull. Raus, T. 1986. Floren- und Vegetationsdynamik auf der Acad. Roy. Belg. Sci. 45: 759-776. Vulkaninsel Nea Kaimeni (Santorin-Archipel, Kykladen,

Leonard, A. 1959. Contribution a I' etude de la colonisation des Griechenland). - Abh. Landesmus. Naturk. Munster 48: laves du volcan Nyamuragira par les vegetaux. - Vegetatio 373-394. 7: 250-258. Raus, T. 1988. Vascular plant colonization and vegetation de­ Mazzoleni, S. & Ricciardi, M. 1993. Primary succession on the velopment on sea- born volcanic islands in the Aegean cone of Vesuvio. - Bull. Brit. Ecol. Soc. 12: 101-112. (Greece). - Vegetatio 77: 139-147. Mazzoleni, S., Aprile, G.G. & Ricciardi, M. 1988. Prime Richter, M. 1984. Vegetationsdynamik auf Stromboli. (Zur acquisizioni sulla vegetazione pioniera dell' area vesuviana. Geookologie trocken-mediterranerStandorte ) . -Aechener -Giom. Bot. !tal. 122 Suppl. 1: 71. Geogr. Arb. 16: 41-110. Mazzoleni, S., Ricciardi, M. & Aprile, G.G. 1989. Aspetti Rivas-Martinez,S. 1975. Mapa de vegetacion de la provincia de pionieri della vegetazione del Vesuvio. -Ann. Bot. (Roma) Avila. -Anal. Inst. Bot. Cavanilles 32: 1493-1556. 47, Suppl. 6: 97-110. Romano, R., Sturiale, C. & Lentini, F. 1979. Geological map of Mueller-Dombois, D. 1992. Distributional dynamics in the Ha­ Mt. Etna, scale 1:50. 000. - International Institute of Vul­ waiian Vegetation. -Pac. Sci. 46: 221-231. canology, Catania. Nimis, P.L. 1992. The lichens of Italy. An annotated catalogue. Smathers, G.A. & Mueller-Dombois, D. 1974. Invasion and -Bolletino Museo Regionale Scienze Naturali, Torino. recovery of vegetation aft er a volcanic eruption in Hawaii. Pignatti, S. 1982. Flora d' Italia. Vols. 1-3. Edagricole, Bolo­ -National Park Service, Honolulu. gna. Timmins, S. 1983. Mt. Tarawera: I. Vegetation types and suc­

Poli, E. 1965. Lavegetazione altomontana dell'Etna. Gianasso, cessional trends. -N.Z.J. Ecol. 6: 99-105. Sondrio. van der Maarel, E., Boot, R., van Dorp, D. & Rijntjes, J. 1985.

Poli, E. 1970a. Vegetationsgrenzen inVulka ngebieten. -Arch. Vegetation succession on the dunes nearOostvoorne. The Bot. Biogeogr. !tal., 46: 1-24. Netherlands. A comparison of the vegetation in 1959 and Poli, E. 1970b. Vegetazione nano-terofiticasu lave dell'Etna. ­ 1980. - Vegetatio 58: 137-187. Arch. Bot. Biogeogr. !tal. 46: 89-100. Walker, L.R. & Chapin, F.S. 1987. Interactions among proc­ Poli, E. 1971. Aspetti della vita vegetale in ambienti vulcanici. esses controllingsuccessional chage. - Oikos 50: 131-135. -Ann. Bot. (Roma) 30: 47-73. Whittaker, R.J., Bush, M.B. & Richards, K. 1989: Plant Poli, E. & Grillo, M. 1975. La colonizzazione vegetate della recolonization and vegetation succession on the Krakatau Islands, Indonesia. - Ecol. Monogr. 59: 59-123.

Acta Phytogeogr. Suec. 85 Vegetation dynamics on Paricutin, a recent Mexican volcano

Alejandro Ve ldzquez,1•2*, Joaqufn Gimenez de Azcdrate2, Ma rtha Escamilla We inmann1,Gerardo Bocco3 & Eddy van der Maarel4

1Laboratoriode Biogeograffa y Sinecologfa, Facultad de Ciencias, UNA M, /nsurgentes sur sin, C.P. 04510, Mexico D.F., Mexico; 2Departamento de Geografta Ffsica, Instituto de Geografta, UNAM, Insurgentes sur sin, C.P. 04510, Mexico D.F.,Me xico; 3Laboratorio de Geoecologfa, 1nstituto de Ecologfa, UNA M, Insurgentes sur sin, C.P. 04510, Mexico D. F.,Mex ico; 4Department of Plant Ecology, Upp sala University, Villaviigen 14, SE-752 36 Uppsala, Sweden; and Departmentof PlantBiology, University of Groningen, P.O.Box 14, 9750AA Haren, The Netherlands; * 14, 14370, Corresponding Author; Cruz Azul No. Colonia liizaro Cardenas, C.P. Mexico D.F.,Mexico; +52 56798171; Fax E. mail [email protected]. unam.mx

Abstract Introduction

Vegetation goes through continuous processes of devel­ It is now generally recognized that vegetation structure opment across all time and space scales. This is most and composition are neither steady nor stable (e.g. Spies evident on re-vegetation processes after volcanic events. & Turner19 99). On the contrary, it is widely accepted that The present paper aims at describing major trends and vegetation changes through different ecological processes vegetation patterns in recent Mexican volcanic forma­ of development so that a large amount of variation is tions, notably the Pari cutin volcanoand surrounding areas found across time and space (e.g. vander Maarel 1988; in Central Mexico. 185 releves were made. Cluster analy­ Velazquez 1994; Bakker et al. 1996; Hong 1998). Vegeta­ sis (TWINSP AN) was performed in order to measure tion, as the expression of plant species aggregation, forms floristic affinity among releves. Detrended Correspond­ whimsical land cover patterns as a result of small, medium ence Analysis (DCA) was performed to characterize the and large landscape changes (e.g. Naveh & Lieberman dynamical relations between the community types. Nine 1993). These dynamic processes take place regardless of different plant community types were distinguished. DCA the time and space resolution and are important to be showed two major trends in vegetation composition: (1) a understood in order to provide coherent interpretations unidirectional progression from colonization stages to for land use, conservation andrestoration purposes (Bakker mature forest and (2) a gradient of disturbance. et al. 1998; Spies & Turner 1999). Mexico's biodiversity is partly explained by the dy­ namic character of the territory (Ramamoorthy et al. Keywords: Colonization; Conifer forest; Detrended Cor­ 1987; Flores-Villela & Gerez 1994). To illustrate this, respondence Analysis; TWINSPAN origin of flora, land uplifting, land use (colonial and pre­ colonial practices) and volcanism are considered among the most relevant to architecture current Mexican land­ Abbreviation: DCA = Detrended Correspondence Analysis. scape patterns (Lugo 1984; Ferrusquia 1987). Changes in climate, terrain, soil, water availability, land use and, consequently, present Mexican vegetation patterns at vol­ Nomenclature: Rzedowski (1988). canic landscapes are still poorly understood. Addition of ash to the atmosphere, for example, and its later deposi­ tion promotes direct modifications in geomorphology and soils. Vegetation, in turn, responds through colonization depending upon the extent and nature of the changes that took place (e.g. Griggs 1933; Eggler 1959; Vargas 1985; Beltnin 1994). Thispaper aims at describingma jor patterns and trends in the vegetation on the slopes of the recent Paricutin

Acta Phytogeogr. Suec. 85: 71-78. 91-7210-085-0. volcano and surrounding areas in Central Mexico. ISBN

Acta Phytogeogr. Suec. 85 72 A. Veldzquez et al.

Study area and environmental setting

The Paricutin volcano is located in the province of Michoacan (19° 30' N; 102° 15' W) in Central Mexico

(Fig. 1). It is surrounded by the Tancitaro formation, which forms part of the Mexican Neovolcanic Belt. This belt harbours hundreds of volcanoes of Plio-Quatemary origin (Segerstrom 1950; Demant 1978). The elevation of the study area ranges from 2300 to 2800 m. The volcano uprising took place between February 20, 1943 and March 4, 1952 (Eggler 1948). 1ts eruption led to the construction of a cineritic cone of 410 m height, and a lava flowcovering an area of ea. 25 km2 with a thickness ranging from 8 to 242 m (Rees 1979). Consequently the surrounding landscape changed into four major terrain units (Eggler 1959): (1) volcanic cone (comprised of fm e graveland ash deposits),(2) malpais (blocky lava flow), (3) accumulation plains (found within and on the edges of lava flows), and (4) neighbouring areas where the local Fig. 1. Location of the study area. The Paricutin volcano is found geomorphology remained but vegetation processes were on the north footslope of the Tancitarofor mation, in Michoacan slightly modified. province, �exico. In the area, Lithosol, Regosol and Andosol soil types prevail in descending order (INEGI 1985; Bocco et al. 1998). Two major climates are typical of the area (Garcia 1973): (A) C (w2) (b) at the lower parts between 2200 and 2500, and C (w2) (w) (b') towards the summit of the Paricutin volcano. The first is considered the most humid foxes) are the major groups inhabiting this newly formed of the quasi-warmtemperate types, whereas the second is habitats (Sosa 1996; Lobato 1998). regarded as the most humid of the quasi-cool temperate Thearea is under the jurisdiction of three indigenous types. The area lies within biome II, the Tropical zone, of Indian communities ('Nuevo San Juan Parangaricutiro', Waiter (1985). According to the global classification sys­ 'Angaguan' and 'Calzontzin'). These communities are tem of Rivas-Martinez (1997) the area is included in the extremely interested in finding new alternative uses of bioclima tropical pluviestational. Herein, theMesotropical their natural resources. The information of the present and Supratropical bioclimatic belts are recognized, both paper is included in a database, which served to develop belonging to the humid ombrotypes. an ecologically sound plan management for Nuevo San Information from eight previous botanical surveys Juan Parangaricutiro indigenous community. (quoted by Gimenez et al. 1997) shows that the number of plant species present on the recent volcano has increased significantly. The dominantvegetation in the region has a Methods temperate affinity(Rzedowski 1988). Conifer and broad­ leaved forests prevail at places where the best developed Vegetation sampling soils occur. Shrubs and grassland communities, in con­ trast, dominate on young soils or rocky substrata. A thor­ First the study area was divided into the major terrain ough descriptionof all azonal plant community types was units as mentioned above. Within every unit, 'walking given by Gimenez et al. (1997). transects' were established and the major physiognomic Although most of the fauna is of Nearctic affinity, a plant community types recorded. In every physiognomic fairly large number of species from the tropics eo-occur in type all plant species were inventoried and releves made the area. The region is located at the most narrow transi­ with cover estimations, following the Braun-Blanquet ap­ tional zone between the Nearctic and Holarctic bio­ proach (Mueller-Dombois & Ellenberg 1974; Westhoff & geographic realms (Bocco et al. 1998). As a result, a van der Maarel 1978). The number of releves per complex fauna! composition is expected, but no system­ physiognomic type varied accordingly their species rich­ atic surveys have been conducted yet. Personal observa­ ness. In total 185 releves were made. tions on vertebrates suggest that reptiles (spiny lizards and rattlesnakes), birds (mainly sparrows, crows androb­

ins) and mammals (mostly shrews, rodents, rabbits and

Acta Phytogeogr. Suec. 85 - Vegetation dynamics on Paricutin, a recent Mexican volcano - 73

Vegetation data analysis The maj or differences are shown in Table 2. In this table three sets of species with different ecological amplitudes Cluster analysis was performed in order to quantify the are shown. First, those of narrowdistribution - most of floristic affinity among the releves. Two way indicator them restricted to one community - and high frequency species analysis (TWINSP AN; Hill 1979) was used to and cover values: the exclusive species. Second, those of differentiate plant community types. A minimum of four broader distribution - occurring in three or more com­ releves was required to establish a plant community type. munities, but with larger importance values in one ore The data were arranged in traditional phytosociological two communities: preferential species. Third, species tables, from which characteristic species (exclusive and with a low frequency in only one or two types, the preferential) of all different plant community types were incidental species. Species of a fourth group with a wide detected (sensu Braun-Blanquet 195 1). distribution and generally large importance values (> Allcharacteristic species were comprisedin a fidelity than class II for presence degree and 2 for importance table (sensu Szafer & Pawlowski, quoted inBraun-Blanquet value), the constant species, were hardly present in this 1951; see also Westhoff & van der Maarel 1978). The material. hierarchical order of characteristic species and plant com­ Each of the nine community types is characterized by munity types was used to base upon a final table array. several exclusive and preferential species and by a certain Total species richness per plant community type was plot­ total number of species (Fig. 3). Two trends can be recog­ ted to detect trends of vegetation diversity. nized among plant communities. Comm. VIII (Gnapha­ In order to better understand the dynamic relations lium americanum-Pinus montezumae) and V (Eupato­ between the plant community types, Detrended Corre­ rium glabratum-Baccharis heterophylla) are species-poor spondence Analysis ( ter Braak 1986) was performed. communities, whereas Comm. VI (Quercus laurina­ Finally, the results were discussed in the light of different Senecio angulifolius) and IX (Trisetum virletii-Abies vegetation-dynamical approaches. religiosa) are the richest communities. Comm. IX and VI are also characterized by having many species sharing relatively similar importance values, whereas Comm. VIII Results has few species with large dominance values (Table 2). Thus one trend is that species richness increases from Nine different plant community types were distinguished pioneer to well-established forest communities. The two (Fig. 2). These are differentiated from each otherby their species-poor Pinus forest Comm. VII (Piptochaetium species composition (preferential and exclusive species), virescens-Pinus leiophylla) and VIII (Gnaphalium physiognomy, soils and degree of disturbance. A synoptic americanum-Pinus montezumae) are intensively managed table was prepared to summarize ( dis )similarities among for forestry purposes. Thus, a second trend is that species plant community types on the basis of the attributes de­ richness, which is potentially high in forests, is reduced scribed above (Table 1). On the whole, large floristic by heavy human disturbance. differences were detected among plant community types.

I A-=0.998

I A-=0.320 A-=0. 173 I I A-=0. 167 1 1 A.=0�380 A-=0.060 I I A-=0.048 A-=0.098

III Open short shru � community

IV Dwarf woodland

I Fernherbby community II Open grassy community V Scattered shrub

VII Open pine and oak forest

VIII Dense pine fo res

Fig. 2. Dendrogram showing the affinity VI Open oak fo rest among the plant communities distinguished Open fir fo rest (TWINSPAN clustering sensu Hill 1979). IX

Acta Phytogeogr. Suec. 85 74 A. Ve ldzquez et al.

Table 1. Synoptic description of all plant community types present in the study area. The roman numbers were used as plant community codes throughout the paper.

Community Characteristic species Physiognomy Land form Dominant soil type Current disturbance

Elaphoglossum pringlei Fern-herb comm. Lava block cracks Leptosoil, lithic None Phlebodium araneosum on malpais II Aegopogon cenchroides Open grass comm. Lava fords Leptosoil, lithic Rock creeping Gnaphalium semiamplexicaule Ill Gaultheria lancifo lia Open short shrub comm. Vo lcanic cone Leptosoil, lithic Ash and gravel Gnaphalium canescens with straight slopes creeping IV Buddleia cordata Dwarf woodland Concane depression Leptosoil, humic Deposition of material Coriaria ruscifolia in lava flow V Baccharis heterophylla Scattered shrub comm. Alluvial plains with Regosol Extensive grazing Eupatorium glabratum ash deposits VI Festuca amplissima Open oak forest Undulating rolling Andosol, humic Extensive logging Quercus laurina footslopes Pinus leio h a Undulating rolling Andosol, humic Extensive logging VII p yll Open pine oak forest Piptochaetium virescens quasi-flat footslopes VIII Gnaphalium americanum Dense pine forest Undulating Andosol, apic Intensive forest Pinus montezumae footslopes management Open forest Steep irregular Andosol, humic Extensive grazing IX Abies religiosa fir Trisetum virletti footslopes and logging

The Detrended Correspondence Analysis shows two Plant communities were less clearly separated along

major gradients along which vegetation trends were in­ DCA-axis 2 (/...,= 0.326) with the following sequence of

ferred (Fig. 4). Plant communities following a unidirec­ communities: VI, Vll, V and Vill, reflecting this gradient. tional progression from colonization stages to well-estab­ This sequence suggests a gradient of disturbance, since lished forest were placed along DCA-axis 1 (/...,=0.823). comm. VI (Quercus laurina-Senecio angulifolium) at the This axis is interpreted as a gradient of soil development. lower end of the axis is a forest with minor disturbance,

To illustrate this further, Comm. Ill (Gnaphalium while Comm. Vill( Gnaphalium americanum-Pinus monte­ canescens-Gaultheria lancifolia) at the left end of axis 1, zumae) is the locally most disturbed type. Comm. V grows on blockylava flowswhere soils (sensu stricto) are (Eupatorium glabratum-Baccharisheter ophylla) represents scanty, while Comm. IX, the fir forest, which prevails on an open shrubland which could further develop into forest Andosol-Humic soil types, is found at the right end of this which is comparable to Comm. Vill. axis. (Axis ll ) 4 D

• V Nmnber u .IX of species 3 a a2 140 2 Ill tt D D\3 8 4 b §la I ·Il Dfll 7 a a D 15 s 6 a a a a fls 17 120 n •I • 12 IV D a14 e •VII 100

80 o +---�--�--�--,---�---r�v�·���--�(A�x=is�I) 0 2 4 5 Soil 60 maturity

Fig. 4. DCA ordination diagram of representative sampling units 40 and exclusive species. Roman numbers refer to the plant com­ munity types. Exclusive species: 1. Aegopogon cenchroides; Abies religiosa; 3. Alchemilla pringlei; 4. CVIII CV Cl CII CIII CVII CIV CVI CIX 2. Baccharis heterophylla; 5. Buddleia cordata; 6. Coriaria ruscifolia; 7. Elaphoglossum pringlei; 8. Eupatorium glabratum;9. Gnaphalium canescens; 10. Gnaphalium semiamplexicaule; 11. Gaultheria lancifolia; Pinus Fig. 3. Comparison among plant communities on the basis of 12. leiophylla; 13 Pin montezumae; 14. Pipthocaetium virescens; 15. Phlebodium species richness. Dense pine forest (Vill) is. the poorest, . us fir araneosum; 16. Quercus laurina; 17. Senecio angulifolium. forest (IX) the richest community.

Acta Phytogeogr. Suec. 85 - Vegetation dynamics on Paricutin, a recent Mexican volcano - 75

Table 2. Exclusive, preferential and incidental species in the plant communities of the study area. Constant species are listed at the bottom. Frequency categories: II = 21-40%; Ill = 41 -60%; IV = 61-80%; V= 81-100%. Cover scale: 1 = < 1 %; 2 = 1-5%; 3 = 6-10%; 4 = 11- 20%; 5 = 21 - 40%; 6 = 41 - 60%; 7 = 61 -80%.

Plant community code I 11 IV IU V VII IX VIII VI 9 11 15 23 5 10 17 4 6 No. of sampling units Elaphoglossum pringlei V2 III 2 Phlebodium araneosum IV 1 IV 2 Cheilanthes kaulfusii III 2 11 2 11 2 Dryopteris rossii V2 V2 IV 2 V2 IV 2 V2 Gnaphalium semiamplexicaule lVI Pellaea temifo lia IV 2 IV 2 IV 2 Aegopogon cenchroides V5 V3 V4 Buddleia cordata V6 Coriaria ruscifolia V7 IIl III4 Gaultheria lancifolia III 2 V4 Gnaphalium canescens IV 2 Andropogon virginicus III 2 Senecio bellidifolius III 2 Hypericum silenoides var. silenoides III 2 Baccharis heterophylla V3 Eupatorium glabratum III 2 V2 III 3 Ill Pinus pseudostrobus V2 IV 5 IV 5 Pinus leiophylla 11 3 V4 11 2 2 Pipthocaetium virescens IV 3 11 111 3 Baccharis conferta IV 4 Crataeguspubescens III 2 Abies religiosa (var. emarginata) IV 5 Trisetum virletii III 2 Fuchsia microphylla Ill ! Didymaea alsinoides Ill l Galium mexicanum Ill l Gnaphalium americanum IV 1 Ta getes filifolia II 2 IV 4 Jaegeria hirta III 4 Phacelia platycarpa III 4 Senecio salignus III 2 Heteroteca inuloides III 2 Quercus laurina III 3 IV 3 IV 2 Festuca amplissima II 2 III 5 Te mstroemiapringlei II 2 III 3 Cheilanthes fa rinosa V2 IV 2 111 2 11 2 Polypodium madrense V2 III 2 V3 II2 IV IV 2 Baccharis glutinosa I Bidens odorata III 2 II 2 II 2 Pinus montezumae IV 3 II 2 112 III 6 V4 Eupatorium patzcuarense II 2 III 2 III 2 11 2 III 3 IV 2 Asplenium preamorsum IV 2 IV 2 12 112 Adiantum raddianum lVI IV 2 Cheilanthes angustifo lia 112 Clethra mexicana II 2 I3 Fuchsia fulgens II 3 Stevia ovata var. ovata 11 5 Hedyotis cervantesii III 2 Rhynchelytrum roseum III 2 Lycopodium clavatum III 2 Ta getes coronopifolia III 2 II 4 II3 Quercus rugosa IV! II 2 Senecio stochaediformis Ill ! III Bidens aequisquama II 2 I III III 2 Salvia mexicana I IV 2 III Smilax moranensis Il l l Ill l I2 Lopezia racemosa Ill l II 2 11 2 IV Alchemilla pringlei Il l I Alnus jorullensis 11 2 11 2 Monnina ciliolata Ill l Senecio angulifolium IV 2 IV 3 Adiantum andicola 11 2 II 2 Geranium seemanii III 2 Ill ! Satureja macrostema 11 4 Stevia nelsonii 11 4 Eupatorium mairetianus II 4 Oenothera pubescens Ill l

Addendum. The following constant species occurred in only one type with frequency and cover 2: Comm. I: Asplenium monanthes; Comm. 11 11: Bromus carinatus; Comm. Buddleia parviflora, Cheilanthes bonariensis, Cheilanthes hirsuta, Crusea longifolia, Cystopteris fragilis, IV: Heliopsis annua, Peperomia quadrifolia, Phlebodium pseudoaureum, Senecio cinerarioides, Stellaria cuspidata; Comm. VI: Cestrum anagyris; Comm. Bacopa procumbens, Cyperus flavus, Muhlenbergia minutissima; Comm. VI: Arbutus glandulosa, Arctostaphylos disco/or, Quercus Ill: crassipes; Comm. Arracacea atropurpurea. IX:

Acta Phytogeogr. Suec. 85 76 A. Ve ldzquez et al.

Discussion Table 3. Survey of case studies of vegetation succession on Americanvolcanoes with a selection of references; see also the

Throughout the 20th century many approaches have been reference lists in other contributions in this Special Volume. suggested to understandshort- (Franklin& Hemstrom19 81; Volcano References van der Maarel & Sykes 1997), medium- (Gleason 1926) Katmai, Alaska Griggs ( l933) and long-term (Clements 1936; Rivas Martfnez 1994) veg­ Moon National Monument, Idaho etation dynamics (Glenn-Lewin & van der Maarel 1992). Eggler (1941) Great Smoky Mountains Whittaker (1956) These studies have made clear that experimental and moni­ Jorullo, Mexico Eggler (1959) toring approaches arenot always effective for gaining a full Arena!, Costa Rica Vargas (1985) understanding of how vegetation develops after the occur­ Nevada de Ruiz, Colombia Salamanca (1991) Mt. St Helens, Washington Franklin et al. (1985) rence of disturbing causes of long term influence. del Moral & Wood (1993) V olcanism, as a natural phenomenon, functions as Tsuyuzaki & del Moral (1995) trigger of primary vegetation succession. Early vegeta­ del Moral et al. (1995) tion recovery patternshave been documented at several localities. Among these, there are many well-docu­ mented cases (Table 3). Most previous studies, provide The dynamical status of the different communities detailed data on floristic colonization, including num­ can be summarized as follows. bers of species, whereas plant community development • Comm. I (Phlebodium araneosum-Elaphoglossum is also relevant to comprehend vegetation establish­ pringlei), may be considered stable and permanent; it is ment (Bakker et al. 1998). The development of plant restricted to cracks within lava blocks and adapted to communities rather than the population demography of strong fluctuations, especially regarding water availabil­ a few species offers a wider perspective on succession ity. Its establishment follows a stochastic pattern, so that its and its environmental characteristics (e.g. Velazquez trend towards other communities is not predictable.

1994; Bakker et al. 1996). Considerable research effort • Cornm. ll (Gnaphalium semiamplexicaule-Aegopogon has been made to document floristic colonization on cenchroides) and comm. Ill (Gnaphalium canescens­ recent volcanoes in Mexico (Eggler 1948, 1959, 1963; Gaulteria lancifolia), in contrast, are regardedas change­ Rzedowski 1954; Segerstrom & Gutierrez 194 7; Beaman able, chamaephytic species aggregates. These are favoured 1960; Espinosa 1962; Rej manek et al. 1982; Cano­ by shallow, gravely volcanic, suitable for improvement of Santana & Meave 1996; Gimenez et al. 1997). Yet, plant better conditions for progressive development of shrub community development is still not well-documented, communities. neither regarding primary nor secondary succession • Comm. IV (Buddleia cordata-Coriaria ruscifolia) can (Rzedowski 1988). be considered as the follow-up community normally re­ In our study area, fine-scale geomorphologic features placing Comm. II and Ill, once shallow soil layers have -produced during the period of volcanic activities and ash developed on ashes, gravels and rocks. deposition- promoted different substrataupon which plant • Comm. V (Eupatorium glabratum-Baccharis hetero­ species have started primary colonization. As a result, a phylla) distributes as a strip-like pattern, found on plains variety of vegetation patterns is found across the recent where it forms scattered mosaics of dense shrubs. This volcanic landscape of Paricutin. This can be illustrated community contributes with a larger amount of organic by the presence of different life forms such as: phane­ material than the other communities, Two trends of fur­ rophytes, chamaephytes, hemicryptophytes and thero­ ther development are possible. phytes (sensu Raunkiaer 1934) and the occurrence of • The first trend is the establishment of resilient tree widely different physiognomic types, ranging from short species in Comm. VI (Festuca amplissima-Quercus grass-dominated to tall-arboreal structures. Similar veg­ laurina), VII (Piptochaetium virescens-Pinus leiophylla) etation mosaics have been reported elsewhere by V argas on soils of the Andosol aplic soil type.

(1985), Beltran (1994), Franklin et al. (1985) and Grishin • The second trend is the development of Cornm. IX et al. (1996). The rate of this colonization process de­ (Trisetum virletii-Abies religiosa) on the Andosol humic pends mostly on local conditions, such as water avail­ soil type. ability, substrate thickness and the resilience of the colo­ • Comm. VIII ( Gnaphalium americanum-Pinus monte­ nizing species. Invasive species settled first (as sug­ zumae) may develop from any of the Comm. V, VI or VII gested by Armesto et al. 1991) . Continuous processes of but its current establishment is definitely influenced by exclusion and inclusion of species take place until the man. The environmental conditions are slightly cooler local conditions have changed into more suitable envi­ and soils are more acid than its preceding types. Most of ronments for other species (Connell & Slatyer 1977), the wood needed to fulfil the demands of the Comunidad better adapted to the new conditions (Valiente- Banuet & Indfgena de Nuevo San Juan Parangaricutiro is obtained De Luna 1990). from stands of this community.

Acta Phytogeogr. Suec. 85 - Vegetation dynamics on Paricutin, a recent Mexican volcano - 77

Concluding remarks Connell, J. & Slatyer, 0. 1977. Mechanisms of succession in natural communities and their role in community stability

The results may be interpreted in terms of changeable and organization. - Am. Nat. 111: 1119-1144. species assemblages that are replaced by new clusters of del Moral, R. & Wood, D.M. 1993. Early primary succession on the volcano Mount St. Helens .-J. Veg. Sci. 4: 223-234. species once environmental conditions are suitable for del Moral, R., Titus, J.H. & Cook, A.M. 1995. Early primary this continuous process (e.g. Whittaker 1973; van der succession on Mount St. Helens, Washington, USA. - J. Maarel 1988; van der Maarel & Sykes 1997). Veg. Sci. 6: 107-120. Vegetation dynamics takes place at all temporal and Demant, A. 1978. Caracteristicas del eje neovolca.nico spatial scales. Studies with a fine resolution (population transmexicano y sus problemas de interpretaci6n. - Rev. level) should, therefore, be consistent with studies . con­ Inst. Geol. Univ. Nac. Aut6n. Mex. 2: 172-188. ducted with a coarser resolution (landscape level). Eggler, A. 1941. Primary succession on volcanic deposits in southern Idaho. -Ecol. Monogr. 3: 277-298. Eggler, A. 1948. Plant communities in the vicinity of the vol­ Acknowledgements cano El Paricutin, Mexico, after two and a half years of eruption. - Ecology 29: 415-436. Eggler, A. 1959. Manner of invasion of volcanic deposits by The authors would like to thank on the one hand the plants withfurther evidence fromParicutin and Jorullo. ­ academic support provided by C. Medina and F. Guevara Ecol. Monogr. 23: 267-284. in species identificationand on the other to J. Amigo and G. Eggler, A. 1963. Plant life of Paricutin volcano, Mexico eigth

Reil, for the comments on an early version. The fm ancial . years after activity ceased. - Am Midi. Nat. 69: 38-68. support from the International Foundation for Science (D/ Eggler, W.A. 1971. Quantitative studies of vegetation on sixteen 2540-1), the DGAPA-UNAM (IN 101196) andUniversity young lava flows on the island of Hawaii. - Trop . Ecol. 12: of Amsterdam is gratefully acknowledged. 66-100. Espinosa, J. 1962. V egetaci6n de una corriente de lava de formaci6n reciente, localizadaen el dedi ve meridional de la References sierra Chichinautzin. - Bol. Soc. Bot. Mex. 27: 67-126. Ferrusquia, V.I. 1987. Geology of Mexico. In: Ramamoorthy,

Anon. 1985. Sintesis geognifica del Edo. de Michoacan, Mex. T.P., Bye, R., Lot, A. & Fa, J. (eds.) Biological diversity of -INEGI, Secretaria de Programaci6n y Presupesto, Mexico, Mexico. - Oxford University Press, Oxford, pp. 3-109. pp. 185. Flores-Villela, 0. & Gerez, P. 1994. Biodiversidad y conservaci6n

Armesto J., Pickett, S. & Donnell, Me. 1991. Spatial heteroge­ en Mexico: vertebrados, vegetaci6n y uso del suelo. - neity during succession: a cyclic model of invasion and Facultad de Ciencias, UNAM, Mexico, DF, 439 pp. exclusion. In: Kolsa, J. & Pickett, S.T. (eds.) Ecological Franklin, J.F. & Hemstrom, M.A. 1981. Aspects of succession heterogeneity.- Springer-Verlag, New York, NY, 216 pp. in the coniferous forest of the Pacific Northwest. In: West, Atkinson, I.A.E. 1970. Successional trends in the coastal and D.C.,Shugart, H.H. & Botkin, D.B. (eds.). Forest succes­ lowland forest of Mauna Loa and Kilauea Volcanoes, Ha­ sion: concepts and applications.- Springer-Verlag, New

waii. -Pac. Sci. 24: 387-400. York, NY, pp. 219-229. Bakker, J.P., Willems, J.H. & Zobel, M. 1996. Long-term veg­ Franklin, J.F., MacMahon, J.A., Swanson, F.J. & Sedell, J.R. etation dynamics.-J. Veg. Sci. 7: 145-228. 1985. Ecosystem responses to the eruption of Mo1lnt St. Bakker, J.P., van Andel, J. & van der Maarel, E. 1998. Plant Helens.-Natl. Geogr. Res. Spring 1985: 198-216. species diversity and restoration ecology: Introduction. - Garcia, E. 1973. Modificaciones al sistema de clasificaci6n Appl. Veg. Sci. 1: 5-8. climatica de Koppen (para adaptarlo a las condiciones de la Beltnin, M.E. 1994. Colonizaci6n vegetal y condiciones Republica Mexicana). - Institute de Geografia UNAM, climaticas en Ios volcanes hist6ricos de Tenerife (I. Canarias). Mexico, DF, pp. 85. Gimenez, J ., Escamilla, M. & Velazquez, A. 1997. Fitosociologia -An. Inst. Super. Agron. 44:41 6-430. Beaman, H. 1960. Vascularplants on the cinder cone of Pari cutin y sucesi6n en el volcan Paricutin (Michoacan, Mexico). ­ volcano in 1958. - Rhodora 62: 175-186. Caldasia 19: 487-505. Bocco, G., Torres, A,.Vehizquez, A. & Siebe, C. 1998. Manag­ Gleason, H. 1926. The individualistic concept of plant associa­ ing natural resources in developing countries. - Conserv. tion. - Bull. Torrey Bot. Club 53: 7-26. Voices 16: 71-84. Glenn-Lewin, D.C. & van der Maarel, E. 1992. Patterns and Braun-Blanquet, J. 1951. Pflanzensoziologie, Grundziige der processes of vegetation dynamics. In: Glenn-Lewin, D.C., Vegetationskunde. 2nd ed. - Springer, Vienna, 865 pp. Peet, R.K. & Veblen, T.T. (eds.) Plant succession - theory Cano Santana,Z. & Meave, J. 1996. Sucesi6n primariaen derrames and prediction. - Chapman & Hall, London, pp. 11-59. volcarucos: El caso del Xitle. - Ciencias 41: 58-68. Griggs, F. 1933. The colonization of the Katmai ash, a new and

Clarksen, B.D. 1997. Vegetation succession (1967-89) on five inorganic soil. - Am. J. Bot. 20: 92-113. recent montane lava flows, Mauna Loa, Hawaii. - N. Z. J. Grishin, S.Y., del Moral, R., Kreston, P.V. & Verkholat, V.P. Ecol. 23: 1-9. 1996. Succession following the catastrophic eruption of

Clements, F.E. 1936. Nature andstructure of the climax. - J. Ksudach volcano (Kamchatka, 1907). - Vegetatio 127: Ecol. 24: 252-284. 129-153.

Acta Phytogeogr. Suec. 85 78 A. Ve ldzquez et al.

Hill, M.O. 1979. TWINSPAN- a FORTRAN program for Segerstrom, K. & Gutierrez, C. 1947. Activity of Paricutin Detrended Correspondence Analysis and Reciprocal Aver­ Volcano from May 4 to September 18, 1946. - Trans. Am. aging. - Cornell University, Ithaca, NY, pp. 76. Geophys. Union 28: 559-566. Hong, S.K. 1998. Changes in landscape patterns and vegetation Segerstrom, K. 1950. Erosion studies at Paricutin volcano, State process in the Far-Eastern cultural landscapes: Human ac­ of Michoacan,Mexico .-U.S. Geol. Surv. Bull. 965A: 1- tivity on pine-dominatedsecondary vegetation in Korea and 164. Japan. - Phytocoenologia 28: 45-66. Segerstrom, K. 1960. Ero sion and related phenomena at Paricutin Lobato, J. 1998. Los marniferos de la comunidad de San Juan volcano in 1957. - U.S. Geol. Surv. Bull. 1104A: 1-18. Nuevo, Parangaricutiro, Michoacan, Mexico. - Tesis de Sokal, R.R. & Rohlf, F.J. 1981. - Biometry. Freeman, San Licenciatura, Universidad Michoacana de San Nicohl.s Francisco, CA, 412 pp. Hidalgo, pp. 87. So sa, N. 1996. El efecto de la fragmentaci6n en comunidades de Lugo, H. 1984. Geomorfologia del sur de la cuenca de Mexico. aves de la Comunidad de Nuevo San Juan Parangaricutiro, -Instituto de Geografia, Serie Varia, Vols. 1-8, Mexico. Michoacan, Mexico. - Tesis de Licenciatura, Universidad Mueller-Dombois, D. & Ellenberg, H. 1974. Aims and methods Michoacana de San Nicohl.s Hidalgo, pp. 112. of vegetation ecology. - John Wiley, New York, NY, 547 Spies, T.A. & Turner, M.G. 1999. Dynamic forest mosaics. In: pp. Hunter Jr., L.M. (ed.) Maintaining biodiversity in forest Naveh, Z. & Lieberman, A.S. 1993. Landscape ecology, theory ecosystems. - Cambridge University Press, Cambridge,

and application. - Springer-Verlag, New York, NY. pp. 95-160. Ramamoorthy, T.P., Bye, R., Lot, A. & Fa, J. (eds.) 1987. ter Braak, C.J.F. 1986. Canonical correspondence analysis: a Biological diversity of Mexico. - Oxford University Press, new eigenvector technique for multivariate direct gradient Oxford, 792 pp. analysis. - Ecology 67: 1167-1179. Raunkiaer, C. 1934. The life forms of plants and statistical plant Tsuyuzaki, S. & del Moral, R. 1995. Species attributes in early geography. - Clarendon Press, Oxford, 632 pp. primary succession on volcanoes. - J. Veg. Sci. 5: 617-622. Rees, D. 1979. Effects of the eruption of Pari cutin Volcano on Valiente-Banuet, A. & De Luna, E. 1990. Una lista floristica landforms, vegetation, and human occupancy. In: Rees, D. actualizada para la reserva del Pedregal de San Angel, (ed.) Volcanic activity and human ecology. - Academic Mexico, DF. -Acta Bot. Mex. 9: 13-30. Press, New York, NY, pp. 250-292. Vehl.zquez, A. 1994. Multivariate analysis of the vegetation of Rej manek, M., Haagerova, R. & Haager, J. 1982. progress of the volcanoes Thiloc and Pelado, Mexico. - J. Veg. Sci. 5: plant succession on the Paricutin volcano: 25 years after 263-270. activity ceased. - Am. Midi. Nat. 108: 194-198. van der Maarel, E. 1988. Vegetation dynamics: patterns in time Rivas Martinez, S. 1994. Dynamic-zonal phytosociology as and space. - Vegetatio 13: 7-19. landscape science. - Phytocoenologia 24: 23-25. van der Maarel, E. & Sykes, M.T. 1997. Rates of small-scale Rivas-Martinez, S. 1997. Syntaxonomical synopsis of the North species mobility in alvar limestone grassland. - J. Veg. Sci. American natural potential vegetation communities, I. - 8: 199-208. Itin. Geobot. 10: 5-148. V argas,G. 1985. La sucesi6n vegetal primaria en una region de Rzedowski, J. 1954. Vegetaci6n del Pedregal de San Angel vulcanismo reciente en el volcan Arenal y sus alrededores, (Distrito Federal, Mexico). - An. Esc. Nac. Cienc. Biol. Costa Rica. -Rev. Bioi. Trop. 33: 171-183. Mex. 8: 54- 129. Walter, H. 1985. Vegetation ofthe earth and ecological systems Rzedowski, J. 1988. Vegetaci6n de Mexico. Editorial Limusa, of the geobiosphere. - Springer, Berlin, 128 pp. Mexico, DF. Westhoff, V. & van der Maarel, E. 1978. The Braun-Blanquet

Salamanca, S. 1991. The vegetation of the paramo and its approach. In: Whittaker, R.H. (ed.) Classification of plant dynamics in the volcanic massif Ruiz-Tolima (Cordillera communities. -Junk, The Hague, pp. 287-408. Central, Colombia). -Dissertation, University of Amster­ Whittaker, R. 1956. Vegetation of the Great Smoky Mountains. dam. - Ecol. Monogr. 26: 1-80.

Acta Phytogeogr. Suec. 85 Hookeria lucens, a new record fo r the moss flora of Iceland - with some remarks on recent lava flows

Agustll. Bj arnason

Laugateigi 39, 105 Reykjavik, Iceland; E-mail [email protected]

Abstract Introduction

The occurrenceof the moss Hookeria lucens (Hedw.) Sm. The geology of Iceland is quite different from the geology in the lava field Eldborgarhraun is described. This lava of the continents. Iceland is situated on the North-Atlantic flow was formed during a single eruption 5000 to 9000 yr ocean ridge, in fact the largest part of an ocean ridge ago. Some information is presented on this and five other above sea level. Iceland is therefore part of the basaltic lava flows in Hnappadalur, western Iceland. There is vulcanism of the ocean floor. Its existence as an island is geothermal activity throughout this area. The moss was due to the combined magma-production of the ridge itself found in a depression where the ground water is heated by and the production of a very effective hot spot. The the magma. The vegetation of the depression is different bedrock of Iceland consists therefore almost entirely of from other places in the lava field. There is shrub layer layers of basaltic lavas. with Betula pubescens, Salix lanata and Sorbus aucuparia Iceland is also geologically quite young, the oldest and a rich fern- and moss layer. On shaded rock ledges a lavas being less than 20 million yr old. The oldest rocks moss/liverwort vegetation is found including Hookeria are found in the westernmost and the easternmostparts of lucens. the island, the youngest in the active volcanic zones in the The newly found locality of Hookeria lucens means a centre. The bedrock is divided into three formations: (1) northern extension of the distribution limit of the species, the tertiary formation; (2) the plio-pleistocene formation, which occurs in parts of NW Europe and N America with 3-3.5 million yr old; (3) the postglacial lavas from the last an oceanic climate. 10 000 yrwhich have not been eroded by glaciers and kept their original surface. Keywords: Eldborgarhraun; Geothermal activity; Oce­ Recent vulcanism is quite active, with one eruption anic climate; Volcanic zone. every fifth year on average. The active volcanic zones are restricted to an area of ea. 25 000 km2, one fourth of the Nomenclature: Bjarnason (1983) for vascular species, island. These volcanic zones are divided into rift zones J6hansson (1998) for mosses and liverworts. where the drift of the North-Atlantic ridge crosses the country, and flank zones without any drifting. The com­ position of the igneous rocks of the volcanic zones is different. The rocks of therift zonesbelong to the tholeiitic series; the rocks of the flank zones to the alkalic series. The North-Atlantic ridge reaches Iceland in the southwest at the Reykjanes peninsula and continues along the Reykjanes rift-zone up to Langjokull in the middle of the country, the direction of the riftingrunning NE-SW. From Langjokull the rifting jumps to the east, to the northern part of the Vatnajokull area and continues to the north along the NE rift zone. From the northern part of Vatnajokull there is a propagating rift zone to the south­ east. This volcanic zone continues as a flank zone towards the Vestmanna islands. The other flank zone forms the Snrefellsnes peninsula. The volcanic zones are divided into individual vol­

Acta Phytogeogr. Suec. 85: 79-84. ISBN 91-7210-085-0. canic systems of which 29 have been active in post glacial

Acta Phytogeogr. Suec. 85 80 A.H. Bjarnason

time. They tend to be long (ea. 100 km)and narrow(ea. 10 Table I. Recent lava flows in Hnappadalur. (From J6hannesson km)with a central volcano (a stratovolcano or a caldera) in 1978.) the middle and a swarm of fissures, faults and crater rows at Lava flow Size (km2) both sides. A few systems have not developed a distin­ Eldborgarhraun 33.4 guished centre. Each volcanic system seems to be active Raudhalsahraun 12.7 separately, without connections to other systems (Fig. 1). Barnaborgarhraun 9.5 Gullborgarhraun 14.6 Rauaamelshraun 2.7 Halsahraun 2.3 Lava fields in Hnappadalur

The origin of the flankzone of the Snrefellsnes peninsula It is worth pointing out that in the middle of the lava deserves a separate description. The older part of the field there is a black scoria lava field - aa - with a rough bedrock here consists of an old tertiary rift zone with and crumbly surface, made of loosely stacked scoria tholeiitic rocks which became extinct ea. 6 million yr ago. lumps, stretching north, west and south from the Eldborg (Vulcanism re-appeared ea. 2 million yr ago as a flank lava ring. This part of the lava area is shown in Fig 3. It zone with alcalic rocks.) The direction of the fissures is differs from the main lava fieldwhich is a ropy lava field about E-W and no rifting has taken place. Volcanic activ­ - pahoehoe - mostly covered with vegetation where the ity has been scarce. There are three volcanic systems in surface is fmn , even and plate like with a conspicuous this zone. Theeasternmost volcanic system is the Lj 6sufjoll rope like structure. volcanic system. It reaches from the Rauoakula crater in Because of this difference in the lava fields some Berserkj ahraun lava field in the west to the Gnibr6k geologists (Askelsson 1955; Einarsson 1970) stated that

craters in the upper Borgarfjorour region in the east. It is the aa lava field is much younger or has been produced ea. 90 km long and ea. 20 km wide. (There is no central after the settlement of the country but the ropy lava field is volcano but the abundance of acid rock here suggests that much older or from the time before the settlement. They this is the probable centreof theLj 6sufjoll area.) More than maintain in other words that the Eldborgarhraun is there­ 20 postglacial lava fields have been formed in this system fore in fact two different lava fields. But according to of which six are in the Hnappadalur area (see Fig. 2 and J6hannesson (1978) the lava field Eldborgarhraun was Table l). formed during a single eruption 5 000-9000 yr ago, like Of these six lava fields in Hnappadalur thelava field all other lava fields in Hnappadalur - except the lava Eldborgarhraun is the largest and best known. Its name fields Rauohalsahraun and Rauoamelshraun, which are comes from a very beautiful crater, the Eldborg lava ring, considered as formed during the lOth century, i.e. shortly which is situated in the middle of the lava field. The after the settlement of Iceland. Eldborg lava ring is one of five craters which are situated on one km long volcanic fissure.

Fig. 1. Location of postglacial lava flows. Dark shading indicates histori­ cal lava fields, i.e. from later than 87 4

0 25 50 75 100 km AD. (from J6hannesson & Sremunds­ son 1998).

Acta Phytogeogr. Suec. 85 - Hookeria lucens, a new recordfo r the moss floraof Iceland - 81 •

Snj6fjOU 0

Fig. 2. Location of the three volcanic sys­ tems on the Snrefellsnes peninsula, W. Ice­ land (see Fig. 1). From west to east: the systems of Snrefellsjokull, Lysuskard and 50 km Lj 6sufjoll.

The peculiar habitat of Hookeria lucens Geothermal activity occurs in most parts of the coun­ try and it is therefore not surprising that there is also In the lava field Eldborgarhraun a small, isolated geothermal activity in this field, where plentiful ground­ geothermal area occurs in the centre. The lava fi eld is water is heated by the magma. The author, who has locally very rough and difficult to cross and very few studied many lava fields (Bjarnason 1991), learnt about people have known about this place, except some local this geothermal place (recently described by Torfason inhabitants, who have to control their sheep grazing in the 1998) and went there in the summer of 1999. The farmer area, and some geologists and hikers, who climb the crater at Snorrastaoir, Haukur Sveinbjomsson, showed him the along a path which has been formed in the lava field. place called I>j 6fhellisrj6our.

Legends: ,..,.,.,.,.. Perlplleryof lava -·-·- Perlphel"yof aa in Eklborgarhraun....__,.,...... "'""'"_.• ,...... , 0 . Crater

Fig. 3. Postglacial lava flows in Hnappadalur, W. Iceland (from J6hannesson 1978, revised).

Acta Phytogeogr. Suec. 85 82 A.H. Bjarnason

In the northeastern part the edge of the coria lava i found and below it a depre ion occur where team ri es at many place . The ize of the area was not urveyed exactly but it i e timated to be 6-8 m deep from the edge of the coria lava· the width i ea. 15-20 m and the length, more difficult to e timate because the boundary i un­ clear, may be 100 m. In the depre ion the vegetation i quite different from other place in the lava field. The exuberant fern Dryopterisfilix-mas, Athyrium filix-femina and Blechnum spicant and ome other herbaceou plant , including Dactylorhiza maculata, Geranium sylvaticumand Ranun­ culus acris, are important con tituents of the field layer of the vegetation wherea Betula pubescens, Salix lanata and Sorbus aucuparia dominate in the hrub layer. The e pecie are common in everal lava field but they are Dryopteri filix-ma eldom developed o luxuriantly (Fig. 4). It wa difficult to Fig. 4. Luxuriant fern vegetation with and tudy in detail the evaporation of heat without endangering Athyrium filix-femina in a depre ion at l>j 6fhelli tj6our in the the tall fern and therefore it wa not completed. In the lava field Eldborgarrhaun; 16 July 1999. Photo A.H. Bjarna on. ummer of 1977 the highe t mea ured temperature in a cavity proved to be 42 °C (J6hanne on per . comm.). A a triking difference with other lava fields in outhern and we tern Iceland and probably related to the heating effect of the team, the mo vegetation did not c n i t of a Racomitrium lanuginosum carpet, but in­ cluded other m os pecie dominating the bottom layer at the main urface of the lava field:

Dicranum majus Hypnum cupressiforme Hypnum jutlandicum Pohlia pp. Polytrichastrumfornzosum Polytrichum juniperinum Ptilidiun1 ciliare Rhizomnium pseudopunctatum Discussion Rhytidiade/phus loreus Sp hagnum ubnitens

Timmia bavarica The newly fo und locality of Hookeria lucens at Eldborga­ Near the opening of the hole in the lava field ome rhraun mean an exten ion of the northern limit of it pecie of the main urface occur abundantly, e.g. Sphag­ di tribution area in Europe. A hown in Fig. 6, the di tri­ num subnitens and Rhytidiadelphus loreus, but mo t of bution area i markedly di junct, with two centre of the pecie characteri tically grow on narrow ledge in di tribution, we tern Europe and the Pacific oceanic part the hole , a nutrient-poor ub trate in hady and wet of North America. H. lucens grow mainly under oceanic condition . Other pecie occur preferably on rock ledge condition , i.e. high precipitation and a mild winter. Out- of the uppermo t 10-50 cm of the hole , which are riche t ide it main European di tribution area the pecie ha in pecie , notably: been reported from Macarone ia and Tuni ia, from ingle

Aneura pinguis Bartramia ithyphylla place in northea tern Turkey and the we tern Cauca us alypogeia muel/eriana Diplophyl/um albicans and al o from many place in the Carpathian . In Scandi­ Fissidens osmundoide Frullania tamarisci navia the pecie i mo t common in outhwe tern Nor­ Hookeria /ucens Lejeunea avifolia way· it i found at ea. 30 localitie in we tern Sweden two Lophozia ventricosa Metzgeria fu rcata Plagiochi/a porelloides Plagiothecium cavifolium place in Denmark (Bohlin et al. 1977; Jannert 1996; Radula complanata Saelania glaucescen Lawton 1971; Potier de la Varde 1949). Weissia ontroversa The habitat of H. lucens i much the ame all over it Hookeria lucens i the only notable pecie in this other­ distribution area. In Norway it i mo t abundant in wet wi e common combination of mo e and liverwort habitat , in cavitie , on crag , hady rock , on rock ledge (Fig. 5). Thi pecie ha not previously been reported in along tream , primarily in fore t but al o in non-wooded Iceland. It grew often in patche intermingled with tuft areas. Mean temperature for January here i 0- +4 °C , mean of Fi sidens osmundoides. It had no cap ules but imma­ temperature for July +14 - +16°C and annual mean pre­ ture eta. Uni eriate, propaguliferou filament were pro­ cipitation i more than 1000 mm. In Sweden it grow duced from mall cell in the upper part of many leaves. al o in hadowy and moi t ituation near water often on

Acta Phytogeogr. Suec. 85 - Hookeria lucens, a new record fo r the moss flora of Iceland - 83

Fig. 5. Hookeria lucens forming a carpet in luxu­ riant bryophyte vegetation at Eldborgarhraun; 16 July 1999. Photo A.H. Bjarna on.

bare oil, trunk of tree and amid rock . Corre ponding ture in th Icelandic locality i con iderably lower but th climate figure for Swedi h localitie are: January and precipitation i a little higher. July temperature -2 --1 °C and + 15 - + 16°C , and annual It i well-known that condition in hole and depre - mean precipitation 650 -850 mm (Bohlin et al. 1977). ion of lava field are of a pecial nature, particularly the The meteor logical tation neare t to the ite of H. hot vapour emerging in the cavitie . Obviou ly tandard­ lucen in Iceland i the farm Haukatunga in Kolbein - ized meteorological ob ervation cannot adequately de- ta ahreppur, 5 km away. Continuou mea urement from crib thi environment. The atmo phere in ide the cavitie there are only available f r the year 1981 to 1983. Th i pre umably aturated with moi ture the whole year and mean temperature for January for tho e year wa -3.4 °C , temperature below zero are improbable. Mo t probably, the mean temperature for July +9.6°C and the annual thi geothermal ite i one of the mo t remarkable habitat mean precipitation wa 1275 mm . Clearly the tempera- in the country and de erve a much more exten ive tudy.

r' - .(ii� V� >.. <

Fig. 6. Di tribution area of Hookeria lucens. After Sttirmer (1969) and Bohlin et al. ( 1977), revi ed.

Acta Ph togeogr. uec. 5 84 A.H. Bja mason

Acknowledgements Volcanic Zone, Iceland. - Acta Nat. Isl. 26: 1-103 . Jannert, B. 1996. Hookeria lucens, skirmossa, vid Vattem. ­ The author thanks Haukur Sveinbjomsson, farmer at Sven. Bot. Tidskr. 90: 83-85. Johansson, T. 1980. Hookeria lucens, en tredj e lokal i Halland Snorrastaoir, for information on the locality, Halld6r samt en annorlunda biotop i Trollehallar. - Mossornas Kj artansson for an invaluable geological contribution and Vanner 10: 12. finally Sveinn Rognvaldsson for his help with the correc­ J6hannesson, H. 1978. Parvar ei brerinn, sem nu er borgin. ­ tion of thelanguage. NattUrufr. 47: 129-204. J6hannesson, H. 1998. islenskir mosar. - Fj olrit Natturu­ frre()istofnunar 36: 1-101. References J6hannesson, H. & Sremundsson, K. 1998. Geological Map of Iceland. 1: 500,000. Bedrock geology. - Icelandic Institute Askelsson, J. 1955. »Par var brerinn, sem nu er borgin.« - of Natural History. Natt6rufr. 25: 122-132. Lawton, E. 1971. Moss flora of the Pacific North-west. - Bjarnason, A.H. 1983. Islensk flora med litmyn()um. - Iaunn. Hattori Botanical Laboratory, Nichinan.

Reykjavik, 352 pp. Nyholm, E. 1960. lllustrated moss flora of Fennoscandia. II. Bjarnason, A.H. 199l.Vegetation on lava fields in the Hekla Musci. Fasc. 4. - C.W.K. Gleerup, Lund. area, Iceland. - Acta Phytogeogr. Suec. 77: 5-114. Potier de la Varde, R. 1949. Nouveaux elements de la flore Bohlin, A., Gustafsson, L. &Hallingback, T. 1977. Skirmossan, tunisienne. - Rev. Bryol. Lichenol. 18: 82. Hookeria lucens, i Sverige. - Sven. Bot. Tidskr. 71: 273- Smith, A.J.E. 1978. The moss flora of Britain and Ireland. ­ 284. Cambridge University Press, Cambridge, 706 pp. Dixon, H.N. 1954. The student's handbook of British mosses. StOrmer, P. 1969. Mosses with a western and southerndistribu­ Wheldon & Wesley Ltd., London, 582 pp. tion in Norway. - Universitetsforlaget, Oslo. Einarsson, Th. 1970. Prettir urn jarfrre()i Hnappadalssyslu. - Thoroddsen, Th. 1911. Lysing Islands II. -Hio isl. b6k­ Arb6k Feroafelags Islands 1970: 105-123. menntafelag, Kaupmannahofn, 673 pp. Hallingback, T & Holmasen, I. 1982. Mossor. En fatthandbok. Torfason, H. 1998. J ar()hitasvreai-In: islensk votlendi - vemdun - Interpublishing, Stockholm, 220 pp. og nYtffig. - Ritstj6ri J6n S. 6lafsson, Reykjavik, pp. 89- Jakobsson, S.P. 1979. Petrology of recent basalts of the eastern 99.

Acta Phytogeogr. Suec. 85 Svenska Viixtgeografiska Siillskapet 85

SVENSKA AXTGEOGRAFIV SKA SALLSKAPET SOCIETAS PHYTOGEOGRAPHICA SUECANA Adress: Avdelningenfor viixtekologi, Upp sala Un iversitet, Villaviigen 14, SE-752 36 Uppsala, Sweden

Sallskapet har till andamru att vacka och underha.J.la intresse fOr The object of the Society is to promote investigations in flora vaxtgeografieni vidstrackastemenin g, att framj a utforskande av and vegetation, their history and their ecological background. flora och vegetation i Sverige och andra lander och att havda Through publication of monographs, and other activities, the geobotanikens praktiska och vetenskapliga betydelse. society tries to stimulate geobotanical research and its applica­ Sallskapet anordnar sammankomster och exkursioner samt tion to practical and scientific problems. utger en publikationsserie. Medlemskap kan erhallas efter Individual members and subscribers (societies, institutes, li­ anmalan hos sekreteraren. Foreningar, bibliotek, laroanstalter braries, etc.) receive Acta Phytogeographica Suecica on pay­ och andra institutioner kan inga som abonnenter. ment of the annual membership fee. There are additional fees in Sallskapet utger arligen Acta Phytogeographica Suecica. years when more than one volume is issued. For membership Medlemmar och abonnenter erhailer aretsActa mot postforskott please, apply to the Secretary. pa arsavgiften jamte porto och expeditionskostnader. The Society has also issued Studies in Plant Ecology (vols. 1- Vis sa ar utges extra band av Acta, som erhruls mot en tillaggs­ 20 Vaxtekologiska studier). All volumes are still available and avgift. can be ordered through OPULUS PRESS AB, Malmen, SE- Sallskapet har tidigareutgivit den ickeperiodiska serien Studies 740 11 Lanna, Sweden. in Plant Ecology (vols. 1-20 Vaxtekologiska studier). Den kan forvarvas efter bestallning hos OPULUS PRESS AB, Malmen, 74011 Lanna, Sweden.

ACTA PHYTOGEOGRAPHICA SUECICA

M. Wao m. 1939. Zur Kenntnis der Vegetation des Sees 1. E. Almquist. 1929. Upplands vegetation och flora. (Vege­ & tation and flora of Uppland.) Out of print. Tiikern. ISBN 91-7210-0 1 2-5. Price: SEK 160. 2. S. Thunmark. 193 1. Der See Fiolen und seine Vegetation. 13. Vaxtgeografiskastudier tillagnade CarlSkottsberg pasextio­ ISBN 91-7210-002-8. Price: SEK 240. arsdagen 1/12 1940. (Geobotanical studies dedicated to C. 3. G. E. Du Rietz. 1931. Life-forms of terrestrial flowering Skottsberg.) 1940. ISBN 91-7210-013-3. Price: SEK 290.

plants. I. ISBN 91-7210-003-6. Price: SEK 160. 14. N. Hy lander. 1941. De svenska formernaav Menthagentilis 4. B. Lindquist. 1932. Om den vildvaxande skogsalmens L. coli. (Zusammenf.: Die schwedischen Formen der Men­ raser och deras utbredning iNordvasteuropa. (Summ.: The tha gentilis L. sensu coli.) ISBN 91-7210-014-1. Price: races of spontaneous Ulmus glabra Huds. and their distri­ SEK 160.

bution in NW Europe.) Out of print. 15. T. E. Hasselrot. 1941. Till kannedom om nagra nordiska 5. H. Osvald. 1933. Vegetation of the Pacific coast bogs of umbilicariaceers utbredning. (Zusammenf. : Zur Kenntnis North America. ISBN 91-7210-005-2. Price: SEK 160. der Verbreitung einiger U mbilicariaceen in Fennoscandia.) 6. G. Samuelsson. 1934. Die Verbreitung der hoheren Wasser­ ISBN 91-7210-015-X. Price: SEK 240. pflanzen in Nordeuropa. 1934. Out of print. 16. G. Samuelsson. 1943. Die Verbreitung der Alchemilla­ 7. G. Degelius. 1935. Das ozeanische Element der Strauch Arten aus der Vulgaris-Gruppe in Nordeuropa. ISBN 91- und Laubflechtenflora von Skandinavien. Out of print. 7210-016-8. Price: SEK 160. 8. R. Semander. 1936. Granskar och Fiby urskog. En studie 17. Th. Arwidsson. 1943. Studien i.iber die Gefasspflanzen in over stonnluckornas och marbuskarnas betydelse i den den Hochgebirgen der Pite Lappmark. ISBN 91-7210-017- svenska granskogens regeneration. (Summ.: The primitive 6. Price: SEK 240. forests of Granskar and Fiby. A study of the part played by 18. N. Dahlbeck. 1945. Strandwiesen am stidostlichen 6re­ storm-gaps and dwarftrees inthe regeneration of the Swed­ sund. ( Summ. : Salt marshes on the S. E. coast of 6resund.) ish spruce forest.) ISBN 91-7210-008-7. Price: SEK 240. ISBN 91-7210-01 8-4. Price: SEK 160. 9. R. Stemer. 1938. Flora der Insel CHand. Die Areale der 19. E. von Krusenstjema. 1945. Bladmossvegetation och blad­ Gefasspflanzen C>landsnebst Bemerkungen zu ihrer Oeko­ mossflora i Uppsalatrakten. (Summ.: Moss flora and moss logie und Soziologie. Out of print. vegetation in the neighbourhood of Uppsala.) ISBN 91- 10. B. Lindquist. 1938. Dalby Soderskog. En skansk lovskog i 7210-019-2. Price: SEK 290.

forntid och nutid. (Zusammenf. : Ein Laubwald in Schonen 20. N. Albertson. 1946. 6sterplana hed. Ett alvaromriide pa in der Vergangenheit und Gegenwart.) ISBN 91-7210- Kinnekulle. (Zusammenf. : 6sterplanahed. Ein Alvargebiet 010-9. Price: SEK 240. auf dem Kinnekulle.) ISBN 91-7210-020-6. Price: SEK

11. N. Stalberg. 1939. Lake Vattern. Outlines of its natural 240. history, especially its vegetation. ISBN 91-7210-011-7. 21. H. Sj ors. 1948. Myrvegetation iBergslagen. (Summ. : Mire Price: SEK 160. vegetation in Bergslagen, Sweden.) ISBN 91-7210-02 1-4. 12. G. E. DuRietz, A. G. Hannerz, G. Lohammar, R. Santesson Price: SEK 290.

Acta Phytogeogr. Suec. 85 86 Svenska Viixtgeografiska Siillskapet

22. S. Ahlner. 1948. Utbredningstyper bland nordiska barrtrads­ central Sweden. Microvegetation and pollen rain in inlets lavar. (Zusanunenf.: Verbreitungstypen unter fennoskan­ of the Baltic basin and in lakes from Preboreal time to the dischen N adelbaumflechten.) ISBN 91-7210-022-2. Price: present day.) ISBN 91-7210-038-9. Price: SEK 160. SEK 240. 39. M. Fries. 1958. Vegetationsutveckling och odlingshistoria 23. E. Julin. 1948. Vessers udde, mark och vegetation i en i V arnhemstrakten.En pollenanalytisk undersokning i Vas­ igenvaxande lovang vid Bjarka-Saby. (Zusammenf. : tergotland. (Zusammenf. : Vegetationsentwicklung und Vessers udde. Boden und Vegetation in einer verwachsen­ Siedlungsgeschichte im Gebiet von Varnhem. Eine den Laubwiese bei Bjarka-Saby in bstergotland, Stid­ pollenanalytische Untersuchung aus Vastergotland (Si.id­ schweden.) ISBN 91-7210-023-0. Price: SEK 240. schweden).) ISBN 91-7210-039-7. Price: SEK 160. 24. M. Fries. 1949. Den nordiska utbredningen av Lactuca 40. B. Pettersson. 1958. Dynamik och konstans i Gotlands alpina, Aconitum septentrionale, Ranunculus platanifolius floraoch vegetation. (Zusanunenf.: Dynamikund Konstanz och Polygonatum verticillatum. (Zusammenf. : Die in der Flora und Vegetation von Gotland, Schweden.) nordische Verbreitung von Lactuca alpina, Aconitum ISBN 91-7210-040-0. Price: SEK 400. septentrionale,Ranunculus platanifolius und Polygonatum 41. E. Uggla. 1958. Skogsbrandfrut i Muddus nationalpark. verticillatum.) ISBN 91-7210-24-9. Price: SEK 160. (Summ. : Forest fire areas inMuddus National Park, north­ 25. 0. Gjre revoll. 1949. Sn11ileievegetasjonen i Oviksfjellene. ern Sweden.) ISBN 91-7210-041-9. Price: SEK 160. (Summ.: The snow-bed vegetation of Mts Oviksfjallen, 42. K. Thomasson. 1959. Nahuel Huapi. Plankton of some Jamtland, Sweden.) ISBN 91-7210-025-7. Price: SEK lakes in an Argentina National Park, with notes onterre st­ 160. rial vegetation. ISBN 91-7210-042-7. Price: SEK 160. 26. Osvald. 1949. Notes on the vegetation of British and H. 43 . V. Gillner. 1960. Vegetations- und Standortsuntersuch­ Irish mosses. ISBN 91-7210-026-5. Price: SEK 160. ungen in den Strandwiesen der schwedischen Westki.iste. 27. S. Selander. 1950. Floristic phytogeography of southwest­ ISBN 91-7210-043-5. Price: SEK 240. em Lule Lappmark (Swedish Lapland). I. 1950. ISBN 91- 44. E. Sj ogren. 1961. Epiphytische Moosvegetation in Laub­ 7210-027-3. Price: SEK 240. waldem der Insel bland, Schweden. (Summ.: Epiphytic 28. S. Selander. 1950. Floristic phytogeography of southwest­ moss communities in deciduous woods on the island of em Lule Lappmark (Swedish Lapland). Il. Karlvaxtfloran (Hand, Sweden.) ISBN 917210-044-3 (ISBN 91-7210- i sydvastra Lule Lappmark. ( Summ.: Vascular flora.)ISBN 444-9).Pr ice: SEK 160. 91-7210-028- 1. Price: SEK 160. 45. G. Wistrand. 1962. Studier i Pite Lappmarkskarlvaxtflora, 29. M. Fries. 1951. Pollenanalytiska vittnesbord om senkvartar med sarsk.ild hansyn till skogslandet och de isolerade vegetationsutveckling, sarskilt skogshistoria, i nordvastra fjallen. (Zusammenf.: Studien iiberdie Gefasspflanzenflora Gotaland. (Zusammenf.: Pollenanalytische Zeugnisse der der Pite Lappmark mit besonderer Beri.icksichtigung des spatquartaren Vegetationsentwicklung, hauptsachlich der Waldlandes und der isolierten niederen Fj elde.) ISBN 91- W aldgeschichte, im nordwe stlichen Go tal and, 7210-045-1 (ISBN 91-7210-445-7). Price: SEK 240. Stidschweden.) ISBN 91-7210-029-X. Price: SEK 240. 46. R. Ivarsson. 1962. Lovvegetation i Mollosunds socken. 30. M. Wrem. 1952. Rocky-shore algae in the bregrund Archi­ (Zusammenf.: Die Laubvegetation im Kirchspiel Mol­ pelago. ISBN 91-7210-030-3. Price: SEK 290. losund, Bohuslan,Schweden .) ISBN 91-7210-046-X (ISBN 31. 0. Rune. 1953. Plant life on serpentines and related rocks 91-7210-446-5). Price: SEK 160. in the Northof Sweden. 1953. ISBN 91-7210-03 1-1. Price: 47. K. Thomasson. 1963. Araucanian Lakes. Plankton studies SEK 240. in North Patagonia, with notes on terrestrial vegetation. 32. P. Kaaret. 1953. Wasservegetation der Seen Orlangen und ISBN 91-7210-047-8. Price: SEK 240. Trehorningen. ISBN 91-7210-032-X. Price: SEK 160. 48. E. Sj ogren. 1964. Epilitische und epigaische Moosvege­ 33. T. E. Hasselrot. 1953. Nordliga lavar i Syd- och Mellan­ tation in Laubwaldem der Insel (Hand, Schweden. (Summ.: sverige. (Nordliche Flechten in Stid- und Mittelschweden.) Epilithic and epigeic moss vegetation in deciduous woods ISBN 91-7210-033-8. Price: SEK 240. on the island of bland, Sweden.) ISBN 91-7210-048-6 34. H. Sjiirs. 1954. SlatterangariGrangarde Finnmark. (Summ.: (ISBN 91-7210-448-1). Price: SEK 240. Meadows in Grangarde Finnmark, SW Dalarna, Sweden.) 49. 0. Hedberg. 1964. Features of afroalpine plant ecology. ISBN 91-7210-034-6. Price: SEK 160. (Resume fran�ais.) ISBN 91-7210-049-4 (ISBN 91-7210- 35. S. Kilander. 1955. Karlvaxternas ovre granser pa fj rul i 449-X). Price: SEK 240. sydvastra Jiimtland samt angransande delar av Harjedalen 50. The plant cover of Sweden. A study dedicated to G. Einar och Norge. (Summ.: Upper limits of vascular plants on Du Rietz on his 70thbirthday by his pupils. 1965. ISBN mountains in southwesternJamtland and adj acent parts of 91-7210-050-8. Price: SEK 460. Harjedalen (Sweden) and Norway.) ISBN 91-7210-03 5-4. 51. T. Flensburg. 1967. Desmids and other benthic algae of Price: SEK 240. Lake Kavsjon and Store Mosse. SW Sweden. ISBN 91- 36. N. Quennerstedt. 1955. Diatomeerna i Langans sjo­ 7210-05 1-6 (ISBN 91-7210-45 1-1). Price: SEK 240. vegetation. (Summ.: Diatoms in the lake vegetation of the 52. E. Skye. 1968. Lichens and airpollution. A study of cryptoga­ Langan drainage area, Jamtland, Sweden.) ISBN 91-7210- mic epiphytes and environment in the Stockholm region. 036-2. Price: SEK 240. ISBN91-7210-052-4(ISBN91-7210 -452-X). Price: SEK240.

37. M. -B. Florin. 1957. Planktonof fresh andbrackish waters in 53. J. Lundqvist. 1968. Plant cover and environment of steep the Sodertalje area. ISBN 91-7210-037-0. Price: SEK 160. hillsides in Pite Lappmark. (Resume: La couverture vege­ 38. M.-B. Florin. 1957. lnsjostudier i Mellansverige. Mikro­ tale et l'habitat des flancs escarpes des collines de Pite vegetation och pollenregn i vikar av bstersjobackenet och Lappmark.) ISBN 91-7210-053-2 (1SBN 91-7210-453-8). insjoar fran preboreal tid till nutid. (Summ.:Lake studies in Price: SEK 240.

Acta Phytogeogr. Suec. 85 Svenska Viixtgeografiska Siillskapet 87

grazing season. Growth and nutritional content in relation 54. Conservation of vegetation in Africa south of the Sahara. Proceedings of a symposium held at the 6th Plenarymeeting to climatic conditions. ISBN 91-7210-070-2 (ISBN 91- of the AETFAT, Uppsala Sept. 12-16, 1966. I. & 0. Hedberg 7210-470-8). Price: SEK 240.

(eds.) 1968. ISBN 91-7210-054-0 (ISBN 91-7210-454-6). 71. C. Johansson. 1982. Attached algal vegetation in running Price: SEK 290. waters of Jamtland, Sweden. ISBN 917210-071-0 (ISBN 55. L.-K. Konigsson. 1968. The Holocene history of the Great 91-7210-47 1-6). Price: SEK 240. Alvar of Oland. ISBN 91-7210-055-9 (ISBN 91-7210- 72. E. Rosen. 1982. Vegetation development and sheep graz­ 455-4). Price: SEK 290. ing in limestone grasslands of South Oland, Sweden. ISBN

56. H. P. Hallberg. 1971. Vegetation auf den Schalenablager­ 91-7210-072-9 (ISBN 91-7210-472-4). Price: SEK 290. ungen in BohusHin, Schweden. (Summ.: Vegetation on 73. Zhang Liquan. 1983. Vegetation ecology and population shell deposits in Bohuslan, Sweden.) ISBN 91-7210-056- biology of Fritillaria meleagris L. at the Kungsangen N a­ 7 (ISBN 91-7210-456-2). Price: SEK 240. ture Reserve, eastern Sweden. ISBN 91-7210-073-7 (ISBN 57. S. Fransson. 1972. Myrvegetation i sydvastra Varmland. 91-721 0-473-2). Price: SEK 240. (Summ.: Mire vegetation in southwestern Varmland, Swe­ 74. I. Backeus. 1985. Aboveground production and growth den.) ISBN 91-7210-057-5 (ISBN 91-7210-457-0). Price : dynamics of vascular bog plants in central Sweden. ISBN SEK 240. 91-7210-074-5 (ISBN 91-7210-474-0). Price: SEK 240. 58. G. Wa llin. 1973. Lovskogsvegetation i Sj uharadsbygden. 75. E. Gunnlaugsd6ttir. 1985. Composition and dynamical (Summ.: Deciduous woodlands in Sjuharadsbygden, Vas­ status of heathland communities in Iceland in relation to tergotland, southwestern Sweden.) ISBN 91-7210-058-3 recovery measures. ISBN 91-7210-075-3 (ISBN 91-7210- (ISBN 91-7210-458-9). Price: SEK 240. 475-9). Price: SEK 240. 59. D. Johansson. 1974. Ecology of vascular epiphytes in 76. Plant cover on the limestone Alvar on Oland. Ecology­ West African rain forest. (Resume: Ecologie des epiphytes sociology-taxonomy. E. Sj ogren (ed.) 1988. ISBN 91- vasculaires dans la foret dense humide d' Afrique 7210-076-1 (ISBN 91-7210-476-7). Price: SEK 320. occidentale.) ISBN 91-7210059-1 (ISBN 91-7210-459- 77. A. H. Bj arnason. 1991. Vegetation on lava fields in the 7). Price: SEK 290. Hekla area, Iceland. ISBN 91-7210-077-X (ISBN 91-

60. H. Olsson. 1974. Studies on South Swedish sand vegeta­ 7210-477-6). Price: SEK 290. tion. ISBN 91-7210-060-5 (ISBN 91-7210-460-0). Price: 78. Algological studies of nordic coastal waters -Afestschrift SEK 240. to Prof. Mats Wrem on his 80th birthday-. I. Wallentinus & 61. H. Hy tteborn. 1975. Deciduous woodland at Andersby, P. Snoeijs (eds .). 1992. ISBN 91-7210-078-8 (ISBN 91- eastern Sweden. Above-ground tree and shrub production. 7210-478-3 ). Price: SEK 290. ISBN 91-7210-061-3 (ISBN 91-7210-46 1-9). Price: SEK 79. TamratBekele. 1994. Vegetation ecology ofremnantAfro­ 240. montane forests on the CentralPlateau of Shewa, Ethiopia. 62. H. Persson. 1975. Deciduous woodland at Andersby, east­ ISBN 91-7210-079-6 (ISBN 91-7210-479-1). Price: SEK em Sweden. Field-layer and below-ground production. 290. ISBN 91-7210-062-1 (ISBN 91-7210-462-7). Price: SEK 80. M. Diekmann. 1994. Deciduous forest vegetation in Boreo­ 160. nemoral Scandinavia. ISBN 91-7210-080-X (ISBN 91- 63. S. Brakenhielm. 1977. Vegetation dynamics of afforested 7210-480-5). Price: SEK 290. farmland in a district of south-eastern Sweden. ISBN 91- 81. Plant root systems and natural vegetation. H. Persson & 7210-063-X (ISBN 91-7210-463-5). Price: SEK 240. I.O. Baitulin (eds.) 1996. ISBN 91-7210-081-8 (ISBN 91-

64. M. Y. Ammar. 1978. Vegetation andlocal environment on 7210-081-3). Price: SEK 290. O shore ridges at Vickleby, land, Sweden. An analysis. 82. R. Virtanen & S. Eurola 1997. Middle oroarcticvegetation ISBN 91-7210-064-8 (ISBN 91-7210-464-3). Price: SEK in Finland and middle-northern arctic vegetation on 240. Svalbard. ISBN 91-7210-082-6. (9 1-7210-482-5). Price: 65. L. Kullman. 1979. Change and stability in the altitude of SEK 290. the birch tree-limit in the southern Swedish Scandes 1915- 83. E. Kaimierczak. 1997. The vegetation of kettle-holes in 1975. ISBN 91-7210-065-6 (ISBN 91-7210-465-1). Price: central Poland. ISBN 91-7210-083-4. (9 1-7210-483-X). SEK 240. Price: SEK 290. 66. Waldemarson Jens en. 1979. Successions in relationship 84. Swedish plant geography - dedicated to Eddy van der E. to lagoon development in the Laitaure delta, North Swe­ Maarel on his 65th birthday-. H. Rydin, P. Snoeijs & M. den. ISBN 91-7210-066-4 (ISBN 91-7210-466-X). Price: Diekmann (eds.) 1999. ISBN 91-7210-084-2 (ISBN 91- SEK 240. 7210-484-8). Price : SEK 400. 67. S. Tuhkanen. 1980. Climatic parameters and indices in 85. Succession and zonation on mountains, particularly on plant geography. ISBN 91-7210-067-2 (ISBN 91-7210- volcanoes - Dedicated to Erik Sjogren on his 65th birth­ 467 -8). Price: SEK 240. day-. E. van der Maarel (ed.)ISBN 91-7210-085-0 (ISBN 68. Studies in plant ecology dedicated to Hugo Sj ors. E. Sjogren 91-7210-485-6). Price: SEK 290. (ed.) 1980. ISBN 91-7210-068-0 (ISBN 91-7210-468-6). Price: SEK 290. A limited number of cloth-bound copies of Acta 44, 69. C. Nilsson. 1981. Dynamics of the shore vegetation of a 45, 46, 48, North Swedish hydro-electric reservoir during a 5-year 49, 51, 52, 53, 56, 57, 61, 63, 66,67, 6� 69, 70, 71, 72, 73,74, period. ISBN 91-7210-069-9 (ISBN 91-7210-469-4 ). Price: 75, 76, 77, 78, 79, 80, 81, 82, 83, 84 and 85 is available at an SEK 240. additional cost of 75 SEK per volume. (Use ISBN nos. within 70. K. Wa renberg. 1982. Reindeer forage plants in the early brackets to order.)

Acta Phytogeogr. Suec. 85 88 Svenska Viixtgeografiska Siillskapet

STUDIES IN PLANT ECOLOGY (VOL. 1-20)

Riversidevascular flora in the upper and middle catchment 1. S. Bn1kenhielm& T. lngelOg. 1972. Vegetationen i Kungs- hamn-Morga naturreservat med forslag till skotselplan. area of the River Skelleftea.Iven, northern Sweden.) ISBN (Summ. : Vegetation and proposed management in the 91-7210-809-6. Price: SEK 112. Kungshamn-Morga Nature Reserve south of Uppsala.) 10. A. Miiller-Haeckel. 1976. Migrationsperiodik einzelliger ISBN 91-7210-801-0. Price: SEK 112. Algen in Fliessgewassern. ISBN 91-7210-810-X. Price: SEK 112. 2. T. IngelOg & M. Risling. 1973. Kronparken vid Uppsala, historik och bestindsanalys av en 300-fu:igtallskog. (Summ. : 11. A. Sjodin. 1980. Index to distribution maps of bryophytes Kronparken,history and analysis of a 300-year oldpinewood 1887-1975. I. Musci. (hard-bound). ISBN 91-7210-811-8. nearUppsala, Sweden.) ISBN 91-7210-802-9. Price: SEK Price: SEK 160. 112. 12. A. Sj odin. 1980. Index to distribution maps of bryophytes 1887-1975. Hepaticae. (hard-bound). ISBN 91-7210- 3. H. Sjors et al. 1973. Skyddsvarda myrari Kopparbergs Hin. II. [Summ.: Mires considered for protection in Kopparberg 812-6. Price: SEK 112.

County (Prov. Dalama, central Sweden.)] ISBN 91-7210- 13. 0. Eriksson, T. Palo & L. Soderstrom. 1981. Renbetning 803-7. Price: SEK 112. vintertid. Undersokningarrorande svensk tamrensnar ings- 4. L. Karlsson. 1973. Autecology of cliff and scree plants in ekologi under snoperioden. ISBN 91-7210-813-4. Price: Sarek National Park, northern Sweden 1973. ISBN 91- SEK 112. 7210-804-5. Price: SEK 160. 14. G. Wistrand. 1981. Bidrag till Pite lappmarks vaxtgeo- 5. B. Klasvik. 1974 .. Computerized analysis of stream algae. grafi. ISBN 91-7210-814-2. Price: SEK 112. ISBN 91-7210-805-3. Price: SEK 112. 15. T. Ka rlsson. 1982. Euphrasia rostkoviana i Sverige. (Eng- lish summ.). ISBN 91-7210-815-0. Price: SEK 160. 6. Y. Dahlstrom-Ekbohm. 1975. Svensk miljovards- och omgivningshygienlitteratur 1952-1972. Bibliografi och 16. Theory and models in vegetation science: Abstracts. R. analys. ISBN 91-7210-806- 1. Price: SEK 112. Leemans, I.C. Prentice & E. van der Maarel (eds.) 1985. ISBN 91-7210-816-9. Price: SEK 160. 7. L. Rodenborg. 1976. Bodennutzung, Pflanzenwelt und ihre Veranderungen in einem alten Weidegebiet auf Mit- 17. I. Backeus. 1988. Mires in theThaba-Put soa Range of the tel-Oland, Schweden. ISBN 91-7210-807-X. Price: SEK Maloti, Lesotho. ISBN 91-7210-817-7. Price: SEK 160. 112. 18. Forests of the world - diversity and dynamics (Abstracts) E. Sjogren (ed.) 1989. ISBN 91-7210-818-5. Price: SEK 8. H. Sj ors & Ch. Nilsson. 1976. Vattenutbyggnadens effek- ter pa levande natur. En faktaredovisning overvagande 290. 19. fran Umea.Iven. (Summ.: Bioeffects of hydroelectric de- E. Sj ogren. 1994. Changes in the epilithic and epiphytic velopment. A case study based mainly on observations moss cover in two deciduous forest areas on the island of along the Ume River, northern Sweden.) ISBN 91-7210- Gland (Sweden). - A comparison between 1958-1962 and 808-8. Price: SEK 160. 1988-1990. ISBN 91-7210-819-3. Price: SEK 200. 20. Vegetation science in retrospect and perspective (Abstracts) 9. J. Lundqvist & G. Wistrand. 1976. Strandflora inom ovre och mellersta Skelleftea.Ivens vattensystem. Med en sam- 1998. E. Sj ogren, E. van der Maarel & G. Pokarzhevskaya manfattningbetr affande botaniska skyddsvarden. (Summ.: (eds.) ISBN 91-7210-820-7. Price: 160 SEK.

Distributor: OPULUS PRESS AB Malmen, 740 11 Lanna, Sweden Tel. + 46 174 65025 Fax + 46 174 22400 www .opuluspress.se

Acta Phytogeogr. Suec. 85

Distributor: OPULUS PRESS AB , Uppsala, Sweden

ISBN 91-721 0-085-0 (ISBN 91-7210-485-6) ISSN 0084-5914