Welcome to Edinburgh and this second symposium on the plant life of SW Asia, supported by the University of Edinburgh, the Royal Botanic Garden and the Royal Society of Edinburgh. It is fifteen years since the first symposium was held as part of the Garden's tercentenary and now at this occasion you have an opportunity to consider progress, to renew old friendships and to make new ones. That should be easy, for the list of participants shows a wonderful representation from all the countries of SW Asia and also of the institutes in Europe and America involved in SW Asian studies. Unfortunately, not all of our friends are here for since we last met we have lost quite a few: we particularly miss Professor Michael Zohary and Professor Per Wendelbo, who died alas, a relatively young man, far too soon.

For earlier works see Annalen des Naturhistorischen Museums in Wien 75 (1971), 5–16; publications omitted from it are listed on p. 5. Flora Iranica fascicles are listed separately, see p. 4.

Geneva's contribution to this meeting in Edinburgh is the publication of a booklet dedicated to Boissier (Burdet 1985) on the centenary of his death—an anniversary which coincides with the completion, or approaching completion, of a series of major studies on the floras of SW Asia covering much of the area of Boissier's classic Flora Orientalis.

In SW Asia 260 species of Centaurea in 38 sections are recognised. Species numbers are highest in S and E Turkey and adjacent Iran and Iraq. East Iran and Afghanistan are poor in species. The number of sections is largest in NE Turkey (and adjacent Transcaucasia). Two hundred and one species are restricted to the area under study; the greatest number of narrow endemics is found in S and E Turkey. The subgenera and sections show quite different patterns of distribution and centres of diversity. The data presented point to E Anatolia and adjacent areas as the foremost centre of evolution for the genus and make probable a relatively recent speciation; an important secondary centre is S Turkey (and of course several other Mediterranean countries too). Some notes are added on the morphological variation in the genus and possible factors influencing its evolution.

The Labiatae of SW Asia arc analysed. There is a major development of the family in this area with c. 1100 species, much morphological diversity, many illuminating distribution patterns and very high endemism levels; species endemism averages over 70% in the medium-sized and large genera. Areas of high concentrations of endemic taxa are indicated. The very significant differences between the Labiate floras of Turkey, Iran and Afghanistan throw doubt on designating all the steppic part of the area as Irano-Turanian; there is a strong Mediterranean floristic connection throughout much of Turkey and little evidence of Irano-Turanian influence. There are numerous taxonomically very isolated genera and species; some of these genera have very restricted distributions, a few are almost cosmopolitan.

Astragalus, excluding Astracantha (formerly Astragalus subgenus Tragacantha), has a world total of c. 2500 species, of which c. 500 are in the New World; it is one of the largest genera of flowering plants. The estimated number of species in the Flora Iranica area is c. 900–1000. Because of the huge size of the genus, no one person is familiar with all its groups; although 150 sections have been described, many are poorly delimited, often not typified and many untenable. In the Old World the genus is almost entirely restricted to the northern hemisphere. Sections with primitive characters mostly contain mesophytic species widespread in mountains and relatively humid regions; in the more advanced sections the species are adapted to dry or arid conditions.

Only a few sections are strictly endemic to the Flora Iranica area (e.g. sects. Microphysa, Acanthophace, Leucocercis). In 2 of the better known large sections (Astragalus and Caprini) there are 2 main centres of diversity: the Armenian knot and the Zagros mountains; Afghanistan and the adjacent Pamir-Alai region. Their patterns of distribution and migration are probably repeated in several other sections in the genus.

Cousinia (Compositae) with over 400 species in SW Asia has its maximum concentration in the Flora Iranica area. Overall it has a remarkably limited distributional range compared with other genera of similar size. On the basis of the significant distribution patterns of Cousinia, it is argued that the highland part of the Flora Iranica area should phytogeographically be re-designated as the Irano-Turkestanian region, including the Central Asian mountain province of Soviet Central Asia. It is also proposed that the essentially lowland Turanian (Aralo-Caspian) area should be excluded from the Irano-Turanian region.

Greece is estimated to have between 4,650 and 4,900 species of vascular plants, about 742 of which are endemic. The Mountain Flora of Greece covers about one-third of the total number of species and a considerably larger proportion of the endemics. Of 926 taxa (species and subspecies) treated in volume 1 (1986), no less than 304, or c. 33%, were either not recorded for Greece in Flora Europaea or were recorded under a different name or with a different rank or both. Phytogeographical analysis shows that 51% of the taxa are “narrows” (restricted to the Balkan Peninsula + Italy or the Balkan Peninsula + Anatolia), and 49% are more widely distributed. Crete is particularly rich in regional endemics, and Peloponnisos and Sterea Ellas have the highest numbers of Greek endemics. An Anatolian element is well-represented in Crete and in the north cast, reflecting the 2 principal routes of migration. Examples of Greek-Anatolian distributional patterns arc given, including some distributions that are widely disjunct.

An analysis of the 8 published volumes of the Flora of Turkey was made in order to determine if there was a factual basis for the concept of an Anatolian Diagonal affecting plant distribution in Turkey. The data accumulated indicated that a substantial number of species (c. 2600) have distributions associated with the Diagonal this total accounts for c. 32% of the number of species in the 8 volumes: 5% are ± restricted to the Diagonal; 14% are absent from the W of it and 13% from the E. Some suggestions are made for the reasons for this floral break.

The phytogeography of Rodhos is briefly sketched. The flora of Rodhos holds a dual position having affinities with the S Aegean islands as well as with the E Aegean islands and SW Anatolia. The present distribution of species on Rodhos corresponds with the palaeogeographic history of the area.

Differentiation has apparently taken place in several taxa since the separation of Rodhos from the S Aegean islands and SW Anatolia, which is shown by the occurrence of many allopatric taxa. The occurrence of many local variants without taxonomic rank shows that this process of speciation is still not complete. Areas of ophiolithic rock are shown to influence the species composition, since many species avoid or are more or less restricted to these areas. Many taxa of the ultrabasic areas show at least a slight morphological distinction from non-serpentine populations of the same species, and some are distinctive enough to recognise at species rank. Both these areas of ophiolithic rock and the limestone cliffs have apparently provided refugia for taxa with formerly wider distributions.

Four phytogeographical regions of Pakistan are recognised in an analysis of the phanerogams of Pakistan. Among the uniregionals, the most common element is Irano-Turanian (45.6%) followed by Sino-Japancse (10.6%), Saharo-Sindian (9.1%) and Indian (4.5%). Though in terms of the area, the Saharo-Sindian region occupies by far the biggest territory, the density of uniregional elements is lowest in this region. Overall there are only 6 endemic genera and an estimated 372 endemic species in Pakistan out of 4882 species. Highest numbers of uniregional endemics per unit area are met with in the Sino-Japanese region, followed by the Irano-Turanian and Saharo-Sindian region. Most of the endemics (78.22%) are confined to mountainous regions (c. 1200 m or above). Four areas, i.e. Sino-Japanese region of Kashmir (10.21% endemics), N Baluchistan (Irano-Turanian; 9.4% endemics) and Chitral (9.1% endemics), may be recognised as centres of radiation in Pakistan.

A broad outline is presented of forest distribution in Turkey during the last 15,000 years, based upon palynological evidence. The history of some selected taxa is given in more detail and compared with present day distributions. Considerable differences in forest composition were found between the mountain ranges N of the Anatolian plateau, the ranges S and SW of the plateau and the area SE of the eastern (Cilician) Taurus. Pinus dominated the northern part during most of the Late Glacial and the Holocene. In southwestern Anatolia a strong increase of Pinus halted SE of the Konya plain and this taxon did not cross the eastern Taurus. Cedrus and Abies fluctuated in relative importance during the period under discussion, but no marked shifts in the distribution are suggested by the pollen record. After c. 9000 BP, oak started to move into the treeless interior, especially in the SE where this tree settled on higher elevations in the steppe area. On the edges of the Anatolian plateau, oak acted as a pioneer, to be replaced by pine in the course of the Holocene. Birch, nowadays restricted mainly to higher elevations in eastern Turkey, was common in the Konya area during the last phase of the Late Glacial and the early Holocene. Beech increased in numbers during the second half of the Holocene. After the Late Glacial, mixed deciduous forest, rich in tree species, developed in the mountains of NW Turkey. It is assumed that in the Pleniglacial most of these species were present along the Black Sea coast, at lower elevations.

The Turkish forests are a complex of Mediterranean, European, Euxine and Irano-Turanian influences. The boundaries between these 4 bioclimatic-biogeographical zones have, especially in the E of the country, been little studied and their limits are difficult to establish. Phytosociological and biogeographical aspects of these forest zones are discussed. In the bioclimatic-Mediterranean region, the forest and pre-forest structures are related to QUERCETALIA ILICIS in the thermo- and meso-Mediterranean stages, and to QUERCETALIA PUBESCENTIS in the supra-Mediterranean stage, except in southern Anatolia where they belong to QUERCO-CEDRETALIA LIBANI. In the prepontic bioclimatic region where the Mediterranean influence is still perceptible, they belong to QUERCO-CARPINETALIA ORIENTALIS and in the Euxine bioclimatic region to RHODODENDRO-FAGETALIA ORIENTALI and to PINOPICEETALIA ORIENTALIS. Some forest formations in NW Anatolia, with a European mountain influence, are related to FAGETALIA SYLVATICAE and to VACCINIO PICEETALIA.

There are 3 types of climate in Turkey: Mediterranean, Continental and Oceanic. The greatest part of the country, including much of the montane area, comes under the influence of various types of Mediterranean climate; the Continental climate occurs in two distinct areas in N and NE Anatolia; the Oceanic climate prevails in the region bordering the Black Sea. The forest vegetation of Turkey may also be divided into 3 main classes: QUERCETEA ILICIS; QUERCETEA PUBESCENTIS; QUERCO-FAGETEA. The first is confined to a coastal zone of the Mediterranean, the second is throughout the inner Mediterranean area, the last is in the Euxine zone of N Anatolia. Steppe vegetation belongs to the class ASTRAGALOBROMETEA widespread in C Anatolia and the Taurus.

An account is given of the distribution of the sand deserts in Iran, Afghanistan and Pakistan Baluchistan, where they cover large surfaces on the plains at low and medium altitudes. Climatologically they differ considerably from each other with respect to winter temperatures, but the latitudinal sequence is obscured by elevation effects. About one half to one third of the species are strict psammophytes; structurally they are the most important components of the different plant communities and in mobile sands usually no other plants occur. Their phytogeographical and taxonomic relationships are discussed in detail and attempts made to relate the distributional patterns to certain ecological factors. 66% of the species are truly Irano-Turanian, 23% are endemics, mostly derivatives of Irano-Turanian species, 10% are Saharo-Arabian and 7% are biregionals. Further aspects pointed out are the representations of life forms of families and genera. The results in the deserts studied give support to the concept of a coherent Irano-Turanian floristic region. The southern deserts form a sub-unit of their own, characterised by high numbers of endemics, some intruding Saharo-Arabian species and an absence of several otherwise common Irano-Turanian species.

The hyperarid climate of the Rub' al-Khali strongly conditions plant life in this subtropical hot desert. There is fossil evidence of a more humid climate here in late Tertiary and Pleistocene times and of a lakeforming phase as recent as 9000–6000 BP. Today rain may not occur for years in some parts, with annual means less than 25 mm, yet the extensive sands provide more scope for plant growth than do some rock and gravel tracts in less arid northern Arabia. The sand vegetation is characterised by very diffuse shrub communities led respectively by an endemic subspecies of Calligonum crinitum, the endemic saltbush Cornulaca arabica and, locally, by Haloxylon persicum. Cyperus conglomeratus, Dipterygium glaucum, Limeum arabicum, and Zygophyllum mandavillei are widespread associates. Sand-free gravel floors or salt flats, relatively small in extent, are nearly or entirely sterile. Tree forms are absent except for rare individuals on some borderlands. Floristic diversity is very restricted; only about 37 species (listed), of which hardly 10 are important in the vegetation, are recorded from an area of 500,000 km2. Annual plants are virtually absent, a situation contrasting sharply with that in Arabia to the N where more than half the desert species are therophytes.

Most of Sinai belongs to the Saharo-Arabian phytogeographical region. It is covered with a sparse vegetation of semishrubs, restricted to wadis or growing on slopes of rocky hills and in sand fields. The high mountains of southern Sinai support a mainly Irano-Turanian steppe vegetation. Smooth-faced rock outcrops supply sufficient run-off water to permit the survival of trees, shrubs, many relict species, and most of the endemics of Sinai. There are 28 endemic species in Sinai, of which 25 occur in the mountainous districts of Sinai. The foothills of eastern Sinai and the coastal area of the Gulf of Elat have a savanna-like vegetation and support thermophilous Sudanian trees and shrubs in wadis and rock crevices.

The distribution and ecological characteristics of the thorn-cushion formations (Tragacanthic vegetation) in the western part of SW Asia are discussed. The typical structure of this formation is illustrated by 2 examples from the mountains of the Antilebanon. These show a clear zonation in relation to the growth form. This, and the available micro-climatological data, show that thorn-cushions reach their ecological optimum on windswept slopes and hill tops. This growth form is obviously an adaptation for sites exposed to strong winds. Taxa with this growth form have a competitive advantage in such habitats.

Senecio vernalis and Calendula arvensis (Asteraceae) are two annuals which grow together in mixed populations and flower in Israel soon after the first winter rains. They were found to be potential selfers, without a self-incompatibility system. The average pollen production per floret and per flowering head was compared with the percentage of pollen with a known function or allocation: pollen grains deposited on stigmas, grains remaining in wilted florets, and non-viable pollen—overall totals of 24% in S. vernalis, 36% in C. arvensis. The remaining pollen (76% & 64%) is distributed according to fluctuating ecological factors, biotic (insects, etc.) and abiotic (rain, wind etc.). The outbreeding potential was assessed by study of the flowering phenology of florets and heads and the behaviour of pollinating insects, mainly flies. Considerable differences were found between the two species in most parameters. In both, especially C. arvensis, the probability of foreign pollen (i.e. pollen not from the same head) reaching the stigmas is very small. Pollenovule ratios for heads were found to be comparatively higher than those recorded for flowers with a similar rate of outbreeding. The efficiency of pollen transfer in the capitulum of the Asteraceae and the large investment in pollen are discussed.

Hypotheses concerning diaspore behaviour are presented. It is suggested that in many plants, dispersal strategy involves alternative agents (2 or more vectors) and combination of distances. The hypotheses are substantiated by data compiled from observations of dispersal modes and mechanisms in herbaceous plant species of the Middle East. Two special modes, heterocarpy and amphicarpy, represent complex strategies with different propagules on the same individual. Amphicarpy resembles cryptophytic life-form in many respects. In several other cases the same propagule can be dispersed for various distances by a single vector or through alternative modes. Combination of long-range, short-range, and in situ dispersal is described for many species; Compositae of Turkey serve as a study case of the possible abundance of this strategy.

Several combinations of different vectors that can distribute the same diaspores, alternately or successively, are recorded. It seems that the combination of wind and temporary water currents is quite common in our region. Cruciferae of Israel serve as a study case of the possible abundance of this strategy and its various combinations. It is concluded that dispersal strategies with alternate modes are typical of many herbaceous species and probably adopted by most angiosperms. The one-to-one conservative relationships between adaptation and mode, mechanism and agent, are rather misleading. An adaptation or a syndrome should not be rigidly defined as they may function in more than one way. Often, the alternate mode can be more effective than the mode attributed to that particular syndrome. By such strategy a plant can better guarantee its distribution in space.