Cyclamen parviflorum

Findings of the 1987 Cyclamen Society Expedition to Turkey

Introduction

Cyclamen parviflorum Pobed. is probably the least known species of its genus. Probably endemic to the mountains of north eastern Turkey the species was first collected, near Artvin, in 1914 and subsequently described by Pobedimova (1946). Since its first discovery other localities have been noted and Meikle (1978) records its occurrence at the Zigana Pass (the best known location), Karakapan and Ikizdere - all sites in the Pontic Alps south and east of Trabzon - at altitudes of no less than 1,200 metres. Other localities have, however, been known for some years.

The Society's expedition to Turkey in 1987 investigated one of these sites, Meryemana below the Sumela Monastery, and also the Ikizdere area from which the species was reported by Meikle. Other sites where it might have been found, or from which it had previously been recorded, could not been studied owing to the depth of snow still lying. Plants studied at Meryemana appeared to be more variable than is apparent from the literature and to be flourishing under conditions differing from those usually described. Above Ikizdere both plants and habitat conformed more closely to previous accounts. The following descriptions of the plants and habitats examined are intended to amplify existing knowledge and to provide information for those attempting to cultivate the species to best effect.

The Plants

1. Flowers

C. parviflorum is usually described as having small, pale purple or lavender pink blooms with a purple basal blotch lacking any pale edge or eye. The range of colours recorded for the main part of the petal is shown in Table 1 where colours are noted according to the Royal Horticultural Society Horticultural Colour Chart, 1966. All these colours were assessed in the field. The basal blotches were more uniform in colour, usually 77A (purple) or 80A (paler purple) although one plant had a blotch of 79A - so dark a purple as to be almost black.

Flower sizes often exceeded those previously reported. Saunders (1973) and Meikle (1978) suggest a maximum petal length of 10mm. The largest flower recorded had, in the field during March 1987, a width of 25mm and height of 17mm. In as much as these dimensions could be remeasured on the same flower as a herbarium specimen in October 1987 they had shrunk to 22mm and 12mm respectively. These dimensions are not easily related to petal length, however, the petal length of this flower in October was 14mm.

The most noticeable feature of the majority of plants examined at Meryemana was not size or colour but that they bore no flowers with completely reflexed petals. The flower shape characteristic of these plants is shown in Figure 1.

Figure 1. Flower shape of Cyclamen parviflorum at Meryemana. (A - seen from above)

Nearly 89% of plants examined (47 out of 53) bore only blooms of such a shape. Flowers kept overnight did not reflex their petals which, together with the fact that plants carrying pollinated blooms and coiled seed pods bore none with reflexed petals, suggests that this flower shape is characteristic. That there were in the same area a minority of plants with normal reflexed petals might indicate two genetically fixed strains. The flower shape illustrated by Pobedimova (1946) is, in fact, closer to that of Figure 1 than that shown by the authors cited above. It was noticeable that of 29 flowers examined, 25 were scented and that most plants were extremely floriferous, tubers with up to 37 fully opened blooms were noted. Contrary to many reports the flowers were borne on long stalks (often 6-9cm) well above the foliage. At Ikizdere the plants studied conformed closely to existing descriptions.

Figure 2. Leaf shapes of Cyclamen parviflorum recorded at Meryemana. (Figures inside each shape show the number of that shape noted)

2. Leaves Small (1-3cm diameter), suborbiculate, uniformly plain green leaves are reported for this species. Meikle (1978) notes that the margins can be entire or subentire and Buglass (1981) illustrates an interesting, but atypical leaf - marbled with a scalloped edge. Figure 2 shows the range of leaf shapes noted. The length and width of leaves measured in the field are shown in the left hand columns of Table 2 while the right hand columns show the same dimensions of certain leaves (those retained as herbarium specimens) remeasured in October. As with flower size some shrinkage occurred in the interim period of drying. Four plants had leaves with scalloped edges. All leaves examined were plain green above with beetroot red undersides.

3. Tubers

The tubers are described as green or brown, pubescent, globose, small (1-2cm diameter) and rooting from the base. Illustrations are provided by Pobedimova (1946) and Buglass (1986). Fourteen tubers of mature, flowering, plants were examined and measured. Details of size, shape, colour and (for some only) the depth at which they grew are shown in Table 3. From this it would appear that mature tubers can be of substantial size, are usually brown and flattened when 12mm or over in diameter and when below this diameter can be, at least partially, green or colour.

Insofar as could be ascertained without uprooting this rare species the roots emerge from the base of the tuber. Floral trunks did not appear frequently and only two of the tubers examined in detail bore them - the fifth and tenth of those listed in Table 3 which had trunks of 15 and 3mm respectively. Pobedimova (1946) illustrates a tuber with a long floral trunk. The surface texture of the tubers was uniformly smooth.

Localities and Habitats

Most of the detailed information on C. parviflorum was gathered around Meryemana, some 35km south of Trabzon and northwards along the road to Macka and Trabzon. This area is treated as one locality. The second locality visited lay above Ikizdere, beyond Derekoy.

1. Meryemana

The area studied lay along an 8km stretch of road extending north and south of the car park at Meryemana. The highest point examined was at an altitude of 1,370m (4,500ft) and the lowest 700m (2,300ft). Higher sites could not be studied as snow still lay deep. Although no sites were examined below 700m it is considered unlikely that C. parviflorum would occur at substantially lower altitudes in the same variety.

At the higher sites C. parviflorum was growing in deep shade beneath Picea orientalis on both the west and east facing slopes of a steep sides narrow gorge running roughly north south. Plants were to be found in rock crevices, sometimes on vertical rock faces or amongst stones of a dry stone wall and in deep but stony soil rich in leaf mould. In may places leaves and flowers were pushing through un-melted snow. In one place, C. parviflorum was growing on a large boulder with Galanthus rizehensis, in living moss. The tubers and bulbs were 1cm deep in the moss, below which there was no true soil.

On the western slopes of the gorge it was noticeable that C. parviflorum grew extensively up to an altitude of 1,160m (3,800 ft) while from 1,220m (4,000 ft) it was replaced by Cyclamen coum. On the eastern slopes C. parviflorum appeared to be the only Cyclamen to a height of some 1,220m (4,000 ft) above which the depth of snow obscured all low vegetation. On the western slopes the change from C. parviflorum to C. coum appeared to correspond to a change in the tree cover when deciduous trees, notably Acer cappadocicum, Corylus avellana, Juglans regia, and Tilia sp., begin to appear in numbers that would affect the winter light intensity and, perhaps, humidity. Close to Meryemana the two species did not seem to mix, although they could in places be found within 30m (100 ft) one of the other.

Further north, at an altitude of 1,020m (3,350 ft) C. coum reappeared, usually to the east of the road in open situations. Lower still some sites supported both C. coum and C. parviflorum growing intermingled. Such sites were always close to the river bed and most often on flat areas of alluvial gravel lying on bends of the river.

It was noticeable that all plants could easily be identified as C. coum or C. parviflorum. No evidence of hybridisation could be found here, or elsewhere. Populations with the two species mixed were always in the valley bottom beside the river and were noted from an altitude of 920m (3,000 ft) to 730m (2,400 ft) - although at the highest point it is possible that C. coum had been introduced with ballast for road building. However, by 880m (2,895 ft) both species were certainly growing naturally together (or had been introduced sufficiently long ago to have become naturalised).

2. Ikizdere

C. parviflorum was reported from this area by Meikle (1978). The site visited was some 50km from the coast and lay at an altitude of 1,310m (4,300 ft). C. coum was common, constituting some 15% of the cover in some places, but C. parviflorum was scattered in groups which seldom contained more than two or three tubers. These tubers lay some 6cm deep at the junction of the surface layer of humic, peaty, soil and an underlying layer of coarse alluvial sand. Tubers of C. coum at this depth usually had floral trunks but none were noted on C. parviflorum. The very limited area in which this species was found was part of a level grassy area below Picea orientalis in an open river valley.

The area approximated to the habitat from which C. parviflorum has previously been described. Patches of snow lay around and unbroken snow was less than 1km distant. Plants associated with C. parviflorum here included C. coum, Ajuga reptans, Bellis perennis, Cerastium fontanum, Hieracium sp., Lapsana communis, Sedum reflexum, Sedum spurium, Trifolium Sp., the lichens Cladonia chlorophaea, C. subcervicornis and Peltigera canina together with many bryophytes. As at Meryemana there was no sign of hybridisation between the two cyclamen species.

Discussion

The plants studied, especially those at Meryemana, show far greater variation in form and habitat than published records suggest. Some differences in flower and leaf size could be due to the extensive reliance placed by taxonomists on herbarium material (and previous publications) rather than on field observations or living material. However, even the herbarium measurements of leaves (Table 2) and the measurements of tubers (Table 3) suggest that more variation is to be found than has been previously reported from wild plants. The range of variation in the population as a whole is not, however, such as would justify the subdivision of the species (although the flower shape merits investigation). Nonetheless variation within the species would suggest that selective breeding might produce plants superior to many of those now grown - especially as forms from lower altitudes, although little known, are generally considered easier in cultivation than those from other locations.

The observations made would also support the accepted view that this species would benefit from cool moist conditions throughout the year. In addition they would suggest that conditions of low light intensity (C. parviflorum tended to occur below Picea) might be beneficial both in winter and summer. Perhaps a site below the staging of an alpine house would be suitable.

At all locations investigated the surface soil would be classed as a sandy peat according to the criteria of the Soil Survey of England and Wales (U.K. Ministry of Agriculture, Fisheries and Food, 1984). As important, however, is the fact that in all cases this surface soil lay above soils with extremely sharp drainage - coarse alluvial sands and gravels or rock debris. The classic state of a well drained moisture retentive soil would seem to have been achieved.

Most of the plants noted were blooming through snow or in areas from which snow could only recently have cleared. Undoubtedly flowers, leaves and tubers will withstand sub-zero temperatures. However, it is unlikely that the peaty soil would freeze, below snow, to the depth at which most tubers were found - drainage of melt water would therefore be rapid.

C. parviflorum and C. coum are often considered to be closely related species. The fact that at several sites both were flowering in close proximity to each other but showed no signs of interbreeding is powerful evidence that they are truly distinct. Insects were not seen pollinating either plant where they grew together (although ants, Lasius niger, were active in C. coum elsewhere) and the flowers are not adapted for wind pollination. It seems therefore likely that both species are self-pollinated in nature as is usual in cultivation and that this maintains their genetic isolation. While the production of artificial hybrids under garden conditions would be of extreme interest it would not necessarily mean that the two species should be merged.

Herbarium specimens under the collection numbers of Bailey, White, Wiltshire and Wood have been deposited at the University of Istanbul (ISTE) and at the Royal Botanic Gardens, Kew (K).