Portal de Periódicos da CAPES

Trientalis europaea L.

2002; Wiley; Volume: 90; Issue: 2 Linguagem: Inglês

10.1046/j.1365-2745.2002.00644.x

1365-2745

K. Taylor, D. C. Havill, Jennifer Pearson, Janet Woodall,

Plant Pathogens and Resistance

Trientalis L. A perennial herb of the family Primulaceae, with a slender stoloniferous rootless rhizome, creeping horizontally close to the soil surface and an erect, glabrous stem 10–25 cm. Adventitious roots, developed on the tuber produced at the tip of each rhizome, occur in the soil surface organic layers. Leaves 10–90 × 5–15 mm, obovate to lanceolate, subsessile, stiff and shining, entire to minutely dentate at the apex, with cuneate base. Leaves (3–)5–8(–10) in a whorl at the top of the stem (occasionally a few alternate leaves below). Flowers usually 1–2 (rarely 3–4), erect (5–)7(–9)-merous, axillary, solitary on capillary pedicels (1–7 cm), ebracteate. Calyx 4–7 mm, deeply divided, teeth linear. Corolla rotate, 11–19 mm in diameter, white or tinged with pink; segments ovate, acuminate or acute; erecto-patent, longer than tube. Stamens (4–)6–7(–8) included; filaments adnate to top of tube. Superior ovary 1-celled, formed by the fusion of 5(–7) carpels; style 1 with capitate stigma. Capsule globose, c. 6 mm, average number two per plant; valves 5(–7) deciduous, leaving the seeds attached to the free-central placenta. The genus Trientalis is a complex of more or less boreal taxa with a circumpolar distribution (Hiirsalmi 1969; Anderson & Loucks 1973; Vergl. Chor., Vol. 2, 1978; Hultén & Fries 1986), including T. europaea, which is the native plant in the British Isles and in continental Europe. Phenotypically, T. europaea exhibits considerable plasticity which is expressed as variation in the morphology of all the vegetative and floral organs (Matthews & Roger 1941; Kovanda 1995). For a polyploid species with a vast geographical range, occurring in a variety of habitats, it is quite surprising that, from the taxonomic viewpoint, this variation appears to be negligible, not producing any appreciable infraspecific units (Kovanda 1995). However, both experimentally determined morphological and physiological differences appear to differentiate ecotypically a subarctic Swedish population from two British populations of T. europaea[see VI (E)]. Trientalis europaea is a native herb of open pine-woods, sessile oak-birchwoods and heather-moors, among moss in damp grassy places, usually rooting in humus, and mostly absent from both the driest and the wettest sites. Trientalis europaea is very local in England (Fig. 1): with a single record for Fritton in East Suffolk, now probably extinguished by flooding and felling (P. Lambley, personal communication); a single record in the Sheffield District of South Yorkshire (Clapham 1969), now effectively the southernmost locality; several localities in West and North Yorkshire northwards into Durham and Northumberland; with two records for north Lancashire (Livermore & Livermore 1987; P. Jepson, personal communication); and four sites in Cumbria (Halliday 1997). Mainly distributed throughout central and northern Scotland with an eastern bias (not recorded from the Inner or Outer Hebrides), locally common. Rare in Orkney (Bullard 1972); on Fair Isle; occurs locally in the Shetland Islands where it is often very dwarf (the so-called var. nana Druce), mainly found in the Dunrossness district of the southern Mainland (Palmer & Scott 1969), but also abundant on the hill behind Grutness, western Mainland (Scott & Palmer 1987). It occurs throughout northern Europe, extending southwards, mainly in the mountains, such as the Bohemian Massif especially in the Sudeten Mountains (Kovanda 1995), to the southern Alps, northern Romania and as far south as Corsica in the Mediterranean (Fig. 2). The distribution of Trientalis europaea in the British Isles. Each dot represents at least one record in a 10-km square of the National Grid. (○) pre-1950; (•) 1950 onwards. Mapped by Mrs J. M. Croft, Biological Records Centre, Centre for Ecology and Hydrology, mainly from records collected by members of the Botanical Society of the British Isles. The main distribution of Trientalis europaea in Europe and western Asia (modified from Vergl. Chor., Vol. 2, 1978; Danton & Baffray 1995; Kaesermann & Moser 1999). The base map is reproduced by permission of the Committee for the Mapping of the Flora of Europe (the bold dashed line shows the limits of Europe). Outside Europe, T. europaea extends eastwards throughout almost the whole of northern Asia to Okhotsk and Kamchatka; to Japan; in western North America, from Alaska southwards to British Columbia, Alberta and into California (Hiirsalmi 1969; Vergl. Chor., Vol. 2, 1978); and more recently recorded as a new species in Greenland (Bay 1993). Trientalis europaea is included in the northern montane element (Dist. Br. Fl.) and classified as circumpolar Boreal-montane by Preston & Hill (1997). The altitudinal range of T. europaea in the British Isles extends from sea level on the east coast in Aberdeenshire, Scotland to an upper limit of 1082 m on Lochnagar (Matthews & Roger 1941). It ascends to 600 m in northern Norway, to 1000 m in northern Sweden and to 1580 m on Jotunheimen, southern Norway (Atl. N.W. Eur.); up to 1380 m in the Hrubý Jeseník Mountains, Czech Republic, up to 1430 m in the Böhmerwald, Germany, to 1635 m in the Tatra, Slovenia, to 1400–1700 m in the Tirol, Austria (Vergl. Chor., Vol. 2, 1978), up to 1800 m in France (Danton & Baffray 1995), and in Switzerland in the cantons of Schwyz to 900 m, of Bern to 1400 m, of Ticino to 1700 m and of Grisons (= Graubünden) up to an altitude of 1900–2100 m (Kaesermann & Moser 1999). The 15.6 °C July mean isotherm approximates to the southern (lower) distribution limits of T. europaea in the British Isles (Atl. Br. Fl.). The geographical distribution and altitudinal restriction at the southern limits of its range in Europe correlate well with relatively high mean summer temperatures, i.e. limited by high temperatures. It has been demonstrated experimentally that the optimum temperature for carbon assimilation in T. europaea is 11–18 °C and that it declines at higher temperatures [see VI (E)]. Absent from the most open and most deeply shaded sites; characteristic habitats include moderately shaded sites in coniferous and deciduous woodland, and in treeless heathlands beneath the canopies of Calluna vulgaris, Juniperus communis, Myrica gale or Pteridium aquilinum. Less commonly, T. europaea occurs in open mountain mire communities and as part of a remnant ground flora after recent forest clear-felling. It has been shown that the growth performance of T. europaea is significantly reduced in open as compared to shaded sites [see V (B), VI (E)]. Trientalis europaea is a strict calcifuge, being found on soils which have highly acidic, humus-rich surface horizons (Table 1). The results of an assay of nitrate reductase activity in the leaves of T. europaea in native pinewood sites suggest that there is very little NO3− available during the growing season, but when fertilized with KNO3 the leaves do appear to possess the inducible isoform of the enzyme and have the capacity to assimilate NO3− (Table 2). However, in common with other non-nitrophilous species, T. europaea is more likely to utilize mixed sources of N (NO3−, NH4+ or organic-N) by root rather than shoot assimilation (Pearson et al. 1998). The species has also been reported to occur on a slightly podzolized brown loam (pH 5.2) in Vaccinium-rich birchwood (Pl. Comm. Scot.; Veg. Scot.), and on a humus iron podzol (pH 4.0–4.4) in a Betula pubescens community (Birse & Robertson 1976). In birch forest in northern Sweden, T. europaea also occurs on an iron podzol with a thick humus horizon (4–42 cm depth) of pH 3.7–4.3 in Vaccinium myrtillus dominated stands, and in Geranium–Vaccinium myrtillus stands with a thinner humus horizon (3–7 cm depth) of pH 4.2–5.2 (Sonesson & Lundberg 1974). The National Vegetation Classification records T. europaea in a range of communities (Rodwell 1991a,b). It is frequent in the Anemone nemorosa subcommunity of Quercus petraea–Betula pubescens–Oxalis acetosella woodland (W11), in north-east Scotland, where the oak when present shows some obvious robur characteristics. This vegetation type has affinities with communities described in earlier studies: in highland birchwood and oakwood (Tansley, Br. Isl.) and in the herb-rich birch and oakwood association of McVean (Veg. Scot.); in Aberdeenshire, north-eastern Scotland, T. europaea has been reported to occur locally in planted Beechwood, essentially type W11 in which there has been a canopy replacement with beech, on deep fertile loam derived from Old Red Sandstone (Watt & Tansley 1932; Tansley, Br. Isl.); and T. europaea and Luzula pilosa differentiate the seminatural Trientali-Betuletum pendulae association, which is largely found in the eastern Grampian mountains, from the other communities in the order Quercion robori-petraeae (Birse 1982). On base-poor soils T. europaea is occasional in the Rhytidiadelphus triquetrus subcommunity of Quercus petraea–Betulapubescens–Dicranum majus woodland (W17), which occurs in the more continental parts of eastern Scotland. This is equivalent to the Trientali-Betuletum pendulae subassociation with Vaccinium vitis-idaea of Birse (1982). Trientalis europaea is frequent in Juniperus communis ssp. communis–Oxalis acetosella woodland (W19) at high altitudes within the east-central Highlands of Scotland, particularly the hills of the Cairngorm and Monadhliath ranges. This vegetation type unites a variety of previously described kinds of scrub or woodland, more and less calcifugous, including that dominated by Betula pubescens ssp. odorata in the Morrone Birkwoods National Nature Reserve, near Braemar in Upper Deeside, Grampian Region, Scotland, the finest example in Britain of a subalpine woodland comparable with and floristically similar to some Norwegian and northern Swedish subalpine birchwoods. In this habitat, T. europaea is a constant species which occurs under a more or less continuous but low-growing understorey of Juniperuscommunis ssp. communis, both where the tree canopy is open in the Juniperus communis-Vaccinium vitis-idaea nodum, and in the treeless Juniperus communis-Campanula rotundifolia nodum (Huntley & Birks 1979). Trientalis europaea is scarce in Pinus sylvestris–Hylocomium splendens woodland (W18), especially in the drier eastern Scottish Highlands in the Speyside and Deeside forests. This vegetation type includes native pinewoods in Scotland (Steven & Carlisle 1959); where T. europaea is occasional to locally frequent in a dry Deschampsia flexuosa–Hypnaceous moss community; occasional in a dry Vaccinium–Deschampsia–Hypnaceous moss community, in both a dry and a moist Calluna–Vaccinium–Deschampsia-moss community, and in a dry grass pinewood community. It also includes communities described by McVean & Ratcliffe (Pl. Comm. Scot.) in the Scottish Highlands where T. europaea is thinly distributed in the Pinetum Hylocomieto-Vaccinietum, and sparse in the Betuletum Oxaleto-Vaccinietum association. According to Birse & Robertson (1976), T. europaea, which is a character species of the order Vaccinio-Piceetalia, is scattered in the Betula pubescens community and scarce in Erica cinerea–Pinus sylvestris plantations at lower altitudes in Scotland. Trientalis europaea is occasional in Carex curta–Sphagnum russowii mire (M7), in the Carex aquatilis–Sphagnum recurvum subcommunity, which is concentrated around the Clova–Caenlochan area of the east-central Highlands of Scotland. It is scarce in Carexrostrata–Sphagnum warnstorfii mire (M8), mainly in the Central Highlands of Scotland essentially as described by McVean & Ratcliffe (Pl. Comm. Scot.); and also in Trichophorum caespitosum (Scirpus cespitosus)–Erica tetralix wet heath (M15), particularly in the western Highlands of Scotland. Trientalis europaea is also found sparsely in two heathland communities: Calluna vulgaris–Vaccinium myrtillus heath (H12), in the east-central Highlands, south-east Scotland and the North York Moors; and in Calluna vulgaris–Arctostaphylos uva-ursi heath (H16), which occurs widely but fairly locally through the east-central Highlands of Scotland, especially in Speyside. In the Shetland Isles, T. europaea occurs locally in a few sites, including a very exposed headland, Compass Head, near the southern tip of south Mainland c. 70–100 m above mean sea level, which in winter and summer storms is drenched in salt spray. At this site it is associated with at least 31 other species in a sheep and rabbit grazed habitat, including Armeria maritima ssp. maritima, Calluna vulgaris, Danthonia decumbens, Galium saxatile, Holcus lanatus, Juncus squarrosus, Molinia caerulea, Nardus stricta, Plantago maritima and Potentilla erecta (T. Russell, personal communication). Above the timberline in the mountain regions of Lappland, in the low-alpine belt, the vegetation is dominated by low growing plants including stunted forms of ‘lowland’ plants, such as Trientalis europaea, Vaccinium myrtillus, V. vitis-idaea and V. uliginosum (Rune 1965). In the subalpine belt in the Torneträsk area of northern Sweden, T. europaea occurs in open forest ecosystems in which Betula pubescens f. tortuosa is the most abundant tree. Where the ground vegetation is of the heath type, T. europaea is frequent in the Vaccinium myrtillus type and variants and in the Geranium–Vaccinium myrtillus type (Sonesson & Lundberg 1974). This continental subalpine birchwood subregion extends southwards in Fennoscandia on the eastern slopes of the Lapponian Scandes. Below these birchwoods, Piceeto-Vaccinietum myrtilli is the prevailing kind of forest on mesic soils with a typical mor humus, predominant in the north, but also widespread in southern Sweden in the Boreo–nemoral zone. In this type of conifer forest, the understorey vegetation is usually dominated by Vaccinium myrtillus with co-dominant V. vitis-idaea or Deschampsia flexuosa and Trientalis europaea a local constant (Sjörs 1965). In the Finnish forest zone, T. europaea occurs in spruce mires in which Picea abies is usually the dominant tree, but Betula pubescens is also important (Eurola et al. 1984). In Central Europe, the boreal species Picea abies, Rubus saxatilis and Trientalis europaea are differential species of the north-eastern spruce–lime–hornbeam woods (Ellenberg 1988). Trientalis europaea, an acid-tolerating plant of needle-leaved woodland and related communities, also occurs in the east of the northern plain in Central Europe on glacial sands, as in the spruce–oak–pine wood in the ancient forest of Bialowieža in Poland. It is described by Ellenberg (1988) as a character species of communities in the order Vaccinio-Piceetalia, class Vaccinio-Piceetea, together with Arctostaphylos uva-ursi, Corallorhiza trifida, Diphasium tristachyum, Epipogium aphyllum, Homogyne alpina, Huperzia selago, Juniperus communis ssp. alpina, Linnaea borealis, Listera cordata, Lonicera caerulea, Lycopodium annotinum, Melampyrum sylvaticum, Moneses uniflora, Monotropa hypopitys ssp. hypopitys, Orthilia secunda, Pyrola media, P. minor, P. rotundifolia, Vaccinium uliginosum and V. vitis-idaea. In several mountain systems of the Bohemian Massif, T. europaea occurs in spruce forest, Pinus mugo scrub and subalpine peat bogs (Kovanda 1995). In the Netherlands, Westhoff & den Held (1969) record T. europaea as differential in both Vaccinio-Piceetea and Quercetea robori-petraeae. Trientalis europaea is present in base-deficient fen Birch woods in the north-west of Central Europe, and as a distinguishing species of intermediate mire communities without trees, together with Molinia caerulea, as in the Harz at c. 800 m. It is also found in Calluna heaths in the class Nardo-Callunetea, in inland Jutland (Ellenberg 1988). When loamy Calluna heath in north-west Germany is afforested, T. europaea is a differential species in the Danthonia–pine planted-forest. If birch–oak wood (Betulo-Quercetum) in Central Europe is converted into a coniferous planted-forest, T. europaea is a differential species together with Chamerion (Epilobium) angustifolium, Frangula alnus, Galium hercynicum, Dryopteris carthusiana and Pseudoscleropodium purum in Dryopteris–pine forest (Ellenberg 1988). In the subarctic birch forests of northern Finland, in Utsjoki, an area of Betula pubescens var. tortuosa about 1350 km2 was damaged in 1965–66 by the geometrid larvae of the moth Oporinia (Epirrita) autumnata. In damaged mesic areas of forest, the ground vegetation became more vigorous during the first 4–5 years after the attack, when the loss of the canopy resulted in more light penetration to the communities containing Cornus suecica, Deschampsia flexuosa, Solidago virgaurea and Trientalis europaea. The mean cover values of the species then declined until the pre-damage situation was restored in the ground vegetation (Kallio & Lehtonen 1975). Vegetation data from permanent plots were collected in 1931, 1961 and 1991 in an area protected from logging in boreal forest 20 km north of Oslo, southern Norway. Major changes were found in the vegetation composition during these 60 years. The main changes were a reduction in the frequency of species and the frequency of joint occurrences of species such as Andromeda polifolia, Calluna vulgaris, Cornus suecica, Eriophorum vaginatum, Maianthemum bifolium, Melampyrum pratense, Trientalis europaea, Vaccinium oxycoccos and V. uliginosum. The observed changes were interpreted as being induced by the increasing growth and dominance of Picea abies and Vaccinium myrtillus (Nygaard & Odegaard 1999). By marking individual plants in the field in a coastal locality and transplanting plants to a glasshouse, Wennström & Ericson (1990) have investigated the interaction between T. europaea and the host-specific smut fungus Urocystis trientalis. None of 100 healthy tagged plants died, but 50% of 100 tagged diseased plants died during the summer. Plants that succumbed produced neither stolons nor seeds. Surviving diseased plants produced significantly fewer winter buds than healthy plants (means ± SE of 1.12 ± 0.05 and 1.88 ± 0.07, respectively). Seed capsule production was low overall and did not differ between diseased and healthy plants. There were no signs of any disease symptoms in seedlings successfully raised in a potting compost in a glasshouse (n = 86). However, disease transmission in winter buds, collected in October from diseased mother plants of T. europaea, then transplanted and grown outdoors, was evident by the following spring in 33% (n = 98) of the buds. The degree of infection by U. trientalis on winter buds cultivated in a glasshouse under two light regimes, 150 µmol m−2 s−1 (moderate shade) and 50 µmol m−2 s−1 (deep shade), was 32% (n = 25) in the former treatment and 61% (n = 25) in the latter, a significantly higher disease level in the more shady situation. A further study (Carlsson et al. 1990) has determined the incidence of disease caused by U. trientalis in relation to the estimated age of T. europaea populations. Comparative studies of populations have been made in an archipelago area in the Gulf of Bothnia, Sweden. In this isostatic area subject to land uplift, primary succession is in strikingly different phases on different islands. The height above sea level of an island is correlated with age, and therefore with the time since it was first colonized by plants. Approximate ages of T. europaea could thus be estimated. Disease incidence peaked in 50–100-year-old populations (c. 40%), with only a low incidence of disease (< 10%) in > 400-year-old populations. The probability of epidemic infection seems to be confined to the early intermediate phases of population development in T. europaea. The results of a 4-year study of a population of T. europaea in an inland mesic spruce forest (Piqueras 1999b) showed that the incidence of disease caused by Urocystis trientalis was lower and the temporal fluctuations smaller than in the above experiment, and was not affected by ramet size. Disease reduced flowering, fruiting, stolon length and the number and size of daughter tubers, all of which were positively correlated with ramet size. Although fungal infection had a detrimental effect on ramet fitness, the low level of incidence and stability of the clone dynamics in simulation models suggest only a minor role of the disease in the regulation of this population. The interaction between Urocystis trientalis, T. europaea, and two herbivores (scale insects and voles) has been studied in a 2-year exclosure experiment on Sikskär Island on the Bothnian sea coast, Sweden, by Ericson & Wennström (1997). The results show that both the scale insects (Arctorthezia cataphracta) and the voles (Clethrionomys glareolus) preferred smut-infected shoots to healthy shoots. Fencing out the voles resulted in a significant increase in host density and a significantly higher disease level than were found in control plots. Thus the data suggest that the voles may not only reduce host densities, but may also have a cleansing effect by the consumption of diseased shoots. Two different types of herbivory, leaf damage caused by invertebrates (insects and slugs) and grazing of young shoots by vertebrates (mainly small rodents), have been examined in a population of T. europaea occurring in a mesic spruce forest in south-west Sweden (Piqueras 1999a). The small amounts of leaf area removed by invertebrate herbivores did not affect any measure of plant performance. In contrast, vertebrate grazing affected all phases of the pseudo-annual life cycle of T. europaea; grazing prevented flowering and fruiting, increased ramet mortality during summer and decreased tuber production; furthermore, grazed ramets produced shorter stolons and smaller, shallower placed tubers, which in turn had a lower winter survival and produced smaller ramets in the following growing season. Trientalis europaea spreads out in favourable sites to form a patch by vegetative means [see VI (C)]. A colony of this kind originates from the distal upward hooked overwintering bud of each new daughter tuber, or rarely may also originate from a seed. Dong et al. (1997) have assessed the clonal plasticity of plants of T. europaea developed from dormant hibernacles (tubers) in response to four levels of added NPK solution in sand culture in a garden experiment. The potted plants were shaded by a cloth which reduced PAR to 40% of full daylight without significantly altering the ratio of red to far-red light. At harvest, after 5 months of growth, there was a negative correlation between nutrient availability and primary spacer (rhizome) length, and a positive correlation with rhizome branching intensity. This morphological plasticity may contribute to spreading the risk of genet extinction as well as to an increased acquisition (foraging) of patchily distributed nutrients in their native environment. Tuber size also responded plastically to variation in nutrient supply; at higher levels of nutrient availability the plants produced more, but smaller, tubers than at lower levels. This result may also be of ecological significance for the survival and establishment chances of daughter ramets in the heterogeneous natural environment. The morphological response to increased nutrient availability has also been studied in a replicated field experiment in a mesic spruce forest in south-west Sweden by Piqueras et al. (1999). In marked plots individual ramets were watered every week, from the first week in June to the first week of August, with an NPK solution or tap water. In early September, all ramets with their corresponding rhizomes and roots were harvested. Fertilized plants produced more tubers and a larger main tuber. In contrast, rhizome length was not affected by the treatment. By comparing these results with those of Dong et al. (1997) the importance of environmental variation in the morphological response to nutrients can be seen. A spatially explicit simulation model [for details see Piqueras & Klimes (1998)] calibrated with data from the field experiment was used to examine the population dynamics of T. europaea ramets in a spatially heterogeneous, temporally constant, environment. The model showed that the species was effective at concentrating its ramets in favourable patches, but this process was strongly influenced by patch size. The analysis of this response at the clone level showed that ramet aggregation was mainly owing to the enhanced performance of clones located initially in the favourable patches, or clones that encountered a favourable patch by chance. In these clones, the simultaneous increase of ramet size and survival accelerated the production of ramets. The temporal and spatial scale at which the aggregation of ramets in favourable patches was manifested suggest that the effectiveness of the morphological response in T. europaea is favoured by a high spatio-temporal predictability in the environment. Long-term observations of the growth of T. europaea have been carried out on the Revack Lodge Estate, Nethy Bridge, in a semi-natural fragment of former Caledonian pine forest and also in an area clear-felled in the early 1980s and subsequently maintained by the removal of any regrowth (D. C. Havill, J. Pearson & J. Woodall unpublished). In both localities the plant was locally abundant over a wide area. Ordination (decorana version 1.0, NERC) of vegetation samples from the two areas showed considerable overlap. In early July, after the main period of vegetative expansion but before the normal onset of senescence in T. europaea, in most years between 1987 and 1997, shoot numbers were counted in 100 marked quadrats (0.5 × 0.5 m) arranged in a 5 × 5 m contiguous block in each of the two sampling plots. On each occasion sufficient, previously unsampled 0.5 × 0.5 m quadrats, chosen at random from outside the plots, were harvested to provide at least 50 shoots for the determination of mean shoot biomass. Both density (43%) and mean shoot biomass (65%) at the clearing site declined significantly (P < 0.05), and during the same 10-year period there were also smaller, but still significant (P < 0.05), decreases in the equivalent values for the native pine site (20% and 21%, respectively). The seasonal growth patterns in leaves collected from the same two sites, together with samples from an open Sphagnum bog and a pine plantation situated nearby, are given in Table 3. Leaf biomass reached a peak between July and August and was not significantly different in all four sites. However, leaf area was generally smaller in the open sites than in the afforested areas. Hence, the lower values of specific leaf area in the open habitats implies an adaptation to the higher levels of irradiance, expressed in the production of thicker leaves. In Table 3, the leaf nitrogen concentrations in all four sites show a trend similar to that described for another Boreal–montane species, Rubus chamaemorus (Marks & Taylor 1972), and for two subalpine birchwood species, Deschampsia flexuosa and Vaccinium myrtillus, at Kevo, northern Finland (Wielgolaski et al. 1975). At full leaf expansion by mid July, percentage N in T. europaea ranges from 1.14 to 1.68, comparable with the acidic, nutrient-poor end of the scale reported for Rubus chamaemorus over a wide range of habitats (Taylor 1989). Other field observations were carried out by D. C. Havill, J. Pearson and J. Woodall (unpublished) at Tullochgrue, Rothiemurchus, Speyside, in an area of wet heath dominated by Calluna vulgaris, rising to drier ground dominated by Juniperus communis. The mean shoot density of T. europaea in 25 × 25 cm sample areas in the shade (n = 47) was 37.6 ± 6.3 m−2 and in open areas (n = 24) was 23.2 ± 3.1 m−2 (difference significant at P < 0.05). The mean oven-dried weight (mg) of shoots harvested from shaded sampling areas (n = 20) was 48 ± 3.9 compared with 29 ± 3.8 from open areas (n = 20), the difference also significant at P < 0.05. Reddened or necrotic (> 20% leaf area) shoots accounted for 84% of the shoot biomass samples from the open areas, but only 36% from shaded areas. In both these and the pinewood habitats above, the growth performance of T. europaea was significantly reduced in open as compared to shaded areas, and was associated with the higher levels of irradiance and temperature in the open, leading to a decline in carbon assimilation on warm, sunny days [see VI (E)]. The effects of air pollution (mainly SO2 and heavy metals) on the structure of forest floor vegetation beyond the treeless zone, which surrounds metal, chemical and fertilizer factories at Kokkola, west Finland has been assessed by Vaisanen (1986); the most tolerant forest floor species, in terms of distance from the factories, were Arctostaphylos uva-ursi, Cornus suecica, Deschampsia flexuosa, Empetrum nigrum, Trientalis europaea and Vaccinium uliginosum. Observations made during the past 7 years at Compass Head, Shetland Isles, on an exposed sea cliff top (see III) show that T. europaea had an average population of 1500–2000 plants, of which only about 5% flowered in any given year and the plants in flower were no more than 6 cm high (T. Russell, personal communication). Night frosts in the early growing season may damage the flowering shoots of T. europaea. According to Hiirsalmi (1969) frost mainly affects the carpel; an overnight exposure to ground level temperatures of –3.5 to –4.0 °C in southern Finland in May, damaged well-developed carpels, such that when the flowers opened they were small, brown and withered. A succession of slightly frosty nights apparently has a hardening effect upon the plants and is less harmful than an isolated occurrence in the middle of a warm period. At the end of the growing season, daughter tubers produced in the previous late summer become separated from the mother ramet on the death of the connecting rhizomes and aerial stem. In summer, each daughter tuber gives r