Soil and Biota of Serpentine: A World View 2009 Northeastern Naturalist 16(Special Issue 5):405–421 Ecological Studies on the Serpentine Endemic Plant Cerastium utriense Barberis Stefano Marsili1, Enrica Roccotiello1, Ivano Rellini2, Paolo Giordani1, Giuseppina Barberis1, and Mauro G. Mariotti1,* Abstract - Cerastium utriense Barberis (Caryophyllaceae) is an endemic plant growing on ultramafic outcrops in northwestern Italy. Despite its great phytogeographical importance, little is known about its ecological requirements and environmental range. Thus, the main objective of the present work was to examine and clarify these aspects. On the basis of a preliminary survey on its range, 28 plots were sampled, and Ellenberg ecological indices of the flora growing with C. utriense were defined. Furthermore, on the basis of the floristic diversity and physical, chemical, and biological properties of the soils, 10 of these plots were selected and more closely investigated. This preliminary study characterized C. utriense as a strictly Ni-excluding serpentinophyte with no apparent relationship with typical chemical characteristics of serpentine soils. On the contrary, the species showed a direct association with physical traits typical of serpentine substrates. Introduction The genus Cerastium (Caryophyllaceae) includes more than 100 species worldwide (Boscaiu et al. 1999, Gustafson et al. 2003) and presents a complex taxonomy that has long challenged botanists (Nyberg Berglund 2005). Natural hybridization between taxa within Cerastium followed by repeated backcrossing to the parental populations have created several complexes and groups with many intermediate forms (Nyberg Berglund 2005). There are many perennial European and American Cerastium endemics of ultramafic substrates, such as C. neoscardicum Niketic, C. vourinense Moschl & Rech. fil., and C. smolikanum Hartvig of the Balkans (Marin and Tatic 2001, Niketic 1994, Stevanovic et al. 2003); C. fontanum Baumg. subsp. scoticum Jalas e Sell of the United Kingdom (Nagy and Proctor 1997); and Cerastium velutinum Raf. var. villosissimum [Pennell] from North America (Gustafson et al. 2003; Morton 2004; Rajakaruna et al. in press, Tyndall and Hull 1999). None of these is a Ni-hyperaccumulator, but all are adapted to severe habitats, as those characterized by metalliferous soils. Cerastium utriense Barberis is an endemic perennial plant from a restricted area in northwestern Italy, between the regions of Liguria and Piedmont, described for the first time in 1988 and currently ascribed to the C. banaticum 1DIP. TE. RIS. Polo Botanico Hanbury, University of Genova, Corso Dogali 1M, I-16136, Genova, Italy. 2DIP. TE. RIS. University of Genova, Corso Europa 26, I-16132 Genova, Italy. *Corresponding author: m.mariotti@unige.it. 406 Northeastern Naturalist Vol. 16, Special Issue 5 group, pending further taxonomical investigations (Barberis 1988, Boscaiu et al. 1996). It grows on cliffs, debris, screes, and ultramafic rocky grasslands on ophiolitic rocks in the so-called “Voltri Massif” or “Voltri Group” (Chiesa et al. 1975, Vanossi et al. 1984), in an area of approximately 350 km2. The ecology of C. utriense is almost unknown. Because of its recent distinction as a species and the restricted range of this species, studies contributing to the conservation efforts of C. utriense are critical. Therefore, the aim of our study was to highlight for the first time ecological traits of the species, mostly focusing on its relationships with the soil, trying to provide useful information for the management of this edaphically restricted, rare plant. In this study, we investigated populations of C. utriense, taking into account the composition of plant communities associated with the species and the physical-chemical properties of the corresponding soils, in order to assess any specific relationships among these components. Field-site Description The study site was situated in the highlands of western Liguria (NW Italy), bordering the Tyrrhenian basin in the eastern Ligurian Alps (Fig. 1). The study site presents several summits over 1000 m a.s.l., with the highest at 1287 m a.s.l. (Mt. Beigua). The study area is located in the Voltri Massif, consisting of high-pressure meta-ophiolite and metasediments of the Ligurian-Piedmontese Domain of the Alps (Chiesa et al. 1975, Vanossi et al. 1984). The Massif includes slices of oceanic crust and subcontinental Figure 1. Lithological map of the study area. Black line encompasses the area in which plants and soils were sampled. 2009 S. Marsili, E. Roccotiello, I. Rellini, P. Giordani, G. Barberis, and M.G. Mariotti 407 mantle involved in subduction and exhumation during the Alpine orogenic cycle, resulting in complex tectonic and metamorphic evolution (Capponi and Crispini 2002, Cortesogno et al. 1977, Spagnolo et al. 2007). Rocks from the oceanic crust are represented by serpentinite with metagabbro and metabasite, and metasediments dominated by calc-schists. Mantle rocks consist of lherzolite with minor pyroxenite and dunite bodies. The climatic features of Liguria are mainly determined by the topography and vertical relief, being either hilly or mountainous, and close to the sea. Liguria marks the transition between Mediterranean and sub-littoral climates on one side, and sub-continental ones on the other side, the latter being characteristic of the southwestern part of the Po plain. According to the Rivas-Martinez (2004) system, the bioclimate of the study area is classified as temperate continental with oceanic zones, and its thermotype is supratemperate and the ombrotype from humid to ultrahyperhumid. Data published from the Piedmont region (weather stations of Bric Berton, 773 m a.s.l., and Marcarolo, 780 m a.s.l.) show mean annual precipitation of 1000–1400 mm/year and mean annual temperatures of 9.1–10.2 °C. The present pedoclimate is characterized by a mesic, locally cryic, soil temperature regime associated with an udic soil moisture regime (sensu USDA 1998). Methods Sampling and vegetation analysis Twenty-eight plots hosting C. utriense were subjectively chosen in the geographical range of the species on the basis of habitat representativeness. The plot dimensions were defined according to the original phytosociological method, based on the biological minimum area (BMA) and the homogeneity of vegetation (Barkman 1989). Each plot measured about 20 m2. To characterize the plant communities, we used the phytosociological Braun-Blanquet method. The coverage of each species was assessed according to the Braun-Blanquet et al. (1952) scale (+ = less than 1%; 1 = 1–5%; 2 = 5–25%; 3 = 25–50%; 4 = 50–75%; 5 = 75–100%). Ecological indices of Ellenberg (1974) modified and adapted to the Italian flora by Pignatti et al (2005) have been assigned to each species recorded in the plots; the considered parameters were light (L), temperature (T), continentality (C), soil moisture (U), pH (R), and nutrients (N). This index shows the species requirements regarding these parameters, and it ranged from 1 to 9 or from 1 to 12. We calculated the average for each parameter for each plot and than the mean value for each parameter of the 28 plots (Fig. 2). Plant-soil relationship Pedology and soil analysis. Ten plots were randomly selected among the twenty-eight previously studied in order to better analyze plant-soil relationships. The soil profile investigated in each plot was situated close to the specimens of C. utriense in order to best reveal the relationships between 408 Northeastern Naturalist Vol. 16, Special Issue 5 soil and this species. The profiles consisted of small dug pits large enough to allow examination and description of the different horizons: They were dug and described until the bedrock or the depth of the first layer little affected by pedogenetic processes (C layer) and biological activity (rhizosphere) was reached. Field descriptions and soil classification were carried out according to the FAO methods and terminology (FAO 2006a, b). In order to characterize the physical and chemical properties of these soils, we sampled each solum horizon (approximately 1 kg) and only the superficial C layer of the unconsolidated deposits without a significant profile development. Laboratory routine analyses were performed in compliance with proposed Italian official methods (MiPAF 2000): soil samples were air-dried, particle size distribution analysis was carried out by wet-sieving for the fraction >50 μm and, the composition of the fine fraction (<50 μm) was determined by pipette procedure after dispersion of the sample with sodium hexametaphosphate, (NaPO3)6. The pH was measured with the potentiometric method in a 1:2.5 soil:water suspension, and electrical conductivity was measured in a 1:5 soil:water suspension. The total carbonate content was determined using the Dietrich Früling calcimeter, and the active carbonate content was determined with ammonium oxalate. The total organic C and N content were determined using an elemental analyzer based on Dumas’ (1831) methods; the soluble P was determined with NaHCO3 (Olsen et al. 1954). The cation exchange capacity (CEC) and exchangeable bases were determined with BaCl2-triethanolamine at pH 8.2; the concentrations of chemical elements extracted were determined by atomic absorption and flame emission spectrophotometry (FAAS). Figure 2. Ellenberg indices averages for the plots. T = temperature, L = light, C = continentality, U = soil moisture, R = pH (reaction), and N = nutrients. 2009 S. Marsili, E. Roccotiello, I. Rellini, P. Giordani, G. Barberis, and M.G. Mariotti 409 Ni detection with DMG screening test. Leaves of all studied species were submitted to the qualitative screening test of dimethylglyoxime (DMG 1% in ethanol 95%; sigma) for nickel detection (Charlot 1964, modified), allowing us to distinguish between Ni and Co accumulation. A positive reaction with DMG is evidenced by the development of a red Ni-DMG complex that is visible for concentrations of Ni in solution above approximately 100 mg L-1. Statistics We used a multivariate analysis to explore the relationship between the distribution and abundance of C. utriense and the physical-chemical characteristics of soils within the study area. We sampled plant species and soil main parameters in 10 plots randomly selected among the 28, finding a total of 63 species. Two matrices were considered in analyzing the data for detecting ecological trends between the gradient of selected environmental factors and the abundance of the species: (1) a matrix of sampling plots × species abundances, and (2) a matrix of sampling plots × environmental factors. Then, we analyzed these matrices using global non-metric multidimensional scaling (NMS) as the ordination technique (Kruskal 1964, Shepard 1962) with Sørensen distance. Non-metric multidimensional scaling analysis was run in autopilot mode, comparing 1- to 6-dimensional solutions. Pearson correlation of quantitative predictor variables with ordination axes were used to interpret relationships of these variables to community composition. Analyses were performed with PC-ORD version 4.25 (McCune and Mefford 1999). Results A total of 84 species were recorded. The plant communities where C. utriense lives are characterized by pioneer species of rocks and screes where Cerastium is associated with Euphorbia spinosa subsp. ligustica, Festuca ovina, Centaurea aplolepa subsp. aplolepa, Dianthus sylvestris subsp. sylvestris, Hieracium piloselloides, and Hypochaeris robertia. There are also present species frequently living on ultramafic soils such as Sesamoides interrupta, Linum campanulatum and, Asplenium cuneifolium. Where the vegetation coverage increases, the serpentinophytes species disappear, and there are more species typical of xerophilous grasslands such as Brachypodium genuense, Sesleria pichiana, and Bromus erectus. The phytosociological analysis of the 28 plots highlighted that in most cases the phytocoenosis are dominated by species of the Alyssion bertolonii Pignatti 1997 alliance (Table 1), which includes most of the italian communities of pioneer vegetation on ultramafic rocks (Furrer and Hofman 1969, Pignatti Wikus and Pignatti 1977, Vagge 1997). Many of the remaining species are typical of the Festuco-Brometea Br.-Bl. & Tx. 1943 ex Klika & Hadac 1944 class. The application of the Ellenberg indices showed (Fig. 2) that the plant communities were mainly characterized by species with low values of nutrients (N) and soil moisture (U), which is characteristic of pioneer and 410 Northeastern Naturalist Vol. 16, Special Issue 5 Table 1. Sampled species, their % presence within the plots, and minimum and maximum coverage (Min-max) in Braun-Blanquet scale. Species Presence Min-max Species of Alyssion bertolonii and other serpentinophytes Cerastium utriense Barberis 100.0 ±3 Euphorbia spinosa L. ligustica (Fiori) Pignatti 82.1 ±2 Centaurea aplolepa Moretti aplolepa 67.9 ±1 Minuartia laricifolia (L.) Schinz & Thell. ophiolitica Pignatti 32.1 ±1 Asplenium cuneifolium Viv. 25.0 + Potentilla hirta L. 25.0 ±1 Sesamoides interrupta (Boreau) G. López 21.4 + Alyssoides utriculata (L.) Medik. 14.3 ±3 Daphne cneorum L. 14.3 ±1 Scorzonera austriaca Willd. 14.3 ±1 Linum campanulatum L. 7.1 ±1 Linaria supina (L.) Chaz. 3.6 + Viola bertolonii Pio emend. Merxm. & W. Lippert 3.6 1 Species of Festuco-Brometea Festuca ovina L. 82.1 ±2 Brachypodium genuense (DC.) Roem. & Schult. 78.6 ±4 Dianthus sylvestris Wulfen sylvestris 75.0 ±1 Bromus erectus Huds. 50.0 ±2 Biscutella laevigata L. laevigata 39.3 ±1 Teucrium montanum L. 39.3 ±2 Asperula aristata L. f. oreophila (Briq.) Hayek 28.6 ±1 Carex humilis Leyss. 25.0 ±1 Trinia glauca (L.) Dumort. glauca 21.4 + Pimpinella saxifraga L. 17.9 ±2 Melica ciliata L. ciliata 14.3 + Avenula pratensis (L.) Dumort. 10.7 ±1 Leucanthemum vulgare Lam. 10.7 ±1 Thymus longicaulis C. Presl 10.7 ±1 Campanula glomerata L. 7.1 + Scabiosa triandra L. 7.1 + Artemisia alba Turra 3.6 + Carlina corymbosa L. 3.6 + Hippocrepis comosa L. 3.6 + Peucedanum officinale L. 3.6 + Polygala nicaeensis W.D.J. Koch mediterranea Chodat 3.6 + Other species Hieracium piloselloides Vill. 50.0 ±1 Hypochaeris robertia (Sch. Bip.) Fiori 46.4 ±2 Satureja montana L. montana 42.9 ±2 Sesleria pichiana Foggi, Gr.Rossi & Pignotti 42.9 ±4 Galium corrudifolium Vill. 39.3 ±1 Genista pilosa L. 28.6 ±2 Iberis umbellata L. 28.6 ±1 Sedum album L. 25.0 ±3 Thlaspi caerulescens J. & C. Presl 21.4 ±1 Thymus praecox Opiz polytrichus (Borbás) Jalas 21.4 ±1 Galium lucidum All. 17.9 ±1 Knautia arvensis (L.) Coult. 17.9 ±1 Leontodon anomalus Ball. 17.9 + Silene saxifraga L. 17.9 + 2009 S. Marsili, E. Roccotiello, I. Rellini, P. Giordani, G. Barberis, and M.G. Mariotti 411 xerophilous communities. The medium values of temperature (T) indicated the presence of mesophilous species, and the medium values of pH (R) highlighted the presence of neutro-basophilous species, evidencing a good correspondence with the soils analysis. The high values of light (L) showed the dominance of heliophilous species, and the low values of continentality (C) characterized these plant communities as typical of temperate climate. Soils were mainly gravelly with continuous bedrock close to the surface presenting limited profile differentiation with weakly developed organic horizon, and lacking in aggregation (Table 2). The soils were classified as Lithic or Episkeletic Leptosols; in addition, soils with a base saturation greater than 80% were Hypereutric, and soils with a Ca/Mg ratio <1 were Magnesic. Soils were classified as Mollic Leptosols when they presented well-structured, dark-colored surface horizons with high base saturation and Table 1, continued. Species Presence Min±max Plantago holosteum Scop. 14.3 + Asperula purpurea (L.) Ehrend. purpurea 10.7 + Bupleurum ranunuculoides L. 10.7 + Echium vulgare L. 10.7 + Quercus petraea (Matt.) Liebl. 10.7 + Sorbus aria (L.) Crantz aria 10.7 + Armeria arenaria (Pers.) Schult. 7.1 + Aster alpinus L. 7.1 ±1 Carduus carlinifolius Lam. 7.1 + Ceterach officinarum Willd. 7.1 + Phyteuma scorzonerifolium Vill. 7.1 + Sedum dasyphyllum L. 7.1 + Senecio provincialis (L.) Druce 7.1 + Anthyllis vulneraria L. 3.6 + Arabis alpina L. caucasica (Willd.) Briq. 3.6 ±1 Asplenium trichomanes L. trichomanes 3.6 + Calluna vulgaris (L.) Hull. 3.6 + Cytisus hirsutus L. 3.6 + Cytisophyllum sessilifolium (L.) O. Lang 3.6 + Dittrichia viscosa (L.) Greuter 3.6 + Erica arborea L. 3.6 ±1 Festuca rubra L. 3.6 + Galium mollugo L. 3.6 + Helichrysum italicum (Roth) G. Don 3.6 + Herniaria glabra L. 3.6 + Lotus corniculatus L. 3.6 + Ornithogalum monticola Jord. & Fourr. 3.6 + Ostrya carpinifolia Scop. 3.6 + Peucedanum cervaria (L.) Lapeyr. 3.6 ±1 Pinus pinaster Aiton 3.6 + Potentilla erecta (L.) Raeusch. 3.6 + Rosa spinosissima L. 3.6 + Saponaria ocymoides L. 3.6 + Scrophularia canina L. 3.6 + Silene vulgaris (Moench) Garcke 3.6 + Teucrium chamaedrys L. 3.6 + 412 Northeastern Naturalist Vol. 16, Special Issue 5 Table 2. The main morphological and physical features of the pedological profiles (Pr). Aggregation (Aggreg.): SG = single grain, SB = subangular blocky, G = granular, m = medium, and f = fine. Abundance of stones (FAO 2006a): D = Dominant (>80%), A = abundant (40–80%), and M = many (15–40%). WRB = World reference base for soil resources (FAO 2006b). Fine earth fraction % Pr Horizon Depth (cm) Color Aggreg. Stones Sand Silt Clay Parent material WRB classification 1 C(A) 0–5 10YR 3/6 SG D 71.50 23.50 5.00 Lithic Episckeletic Leptosol R 5+ - - - - - - Serpentineschist 2 C(A) 0–20 10YR 3/2 fG D 68.00 25.80 6.20 Lithic Eutric Leptosol R 20+ - - - - - - Serpentineschist 3 C(B)1 0–20 7.5 YR 3/2 mG A 45.10 45.20 9.70 Pre-weathered soil material Colluvic Episkeletic Leptosol C(B)2 20+ 7.5 YR 3/4 mSB M - - - 4 C(A) 0–10 5 YR 3/2 mG A 58.40 31.60 10.00 Lithic Magnesic Leptosol R 10+ - - - - - - Serpentinite 5 A/C 0–10 5 YR 3/1 mG M 56.10 34.30 9.60 Lithic Magnesic Leptosol R 10+ - - - - - - Serpentinite 6 C1 0–8 5 YR 3/1 SG A 70.80 22.30 6.90 Technic Episkeletic Leptosol C2 8+ - SG D - - - Serpentine gravels 7 A/C 0–5 5 YR 3/1 fG A 75.40 21.50 2.90 Monogenic conglomerate Lithic Eutric Leptosol R 5+ - - - - - - 8 C(B) 0–4 10 YR 4/4 fG A 59.60 34.90 5.60 Pre-weathered soil material Lithic Colluvic Leptosol R 4+ - - - - - - 9 AO 0–10 5 YR 2.5/1 mG M 65.90 32.50 1.6 Mollic Magnesic Leptosol CA 10–17 5 YR 4/3 mG A 70.80 26.90 2.30 R 17+ - - - - - - Serpentinite 10 C1 0–5 2.5 Y 4/4 SG A 76.30 19.10 4.70 Episkeletic Magnesic Leptosol C2 5+ - SG A - - - Serpentinite 2009 S. Marsili, E. Roccotiello, I. Rellini, P. Giordani, G. Barberis, and M.G. Mariotti 413 high content of organic material, as in the case of profile 9 (Table 2). Some soils were formed by already highly weathered material, i.e., profiles 8 and 3 (Table 2). These materials consisted of sediments deposited by soil slip processes along the steep slopes subjected to erosion (Colluvic Leptosols). The complete absence presence of only a thin A horizons or their absence in some profiles occurred where plots were characterized by high slopes and therefore were subjected to high erosion. In some cases, the soils showed no significant profile development (profiles 6, 10) (Table 2) and consisted of unconsolidated, coarsely grained material resulting from anthropogenic activity (Techinc Leptosols). The physical and chemical analyses (Tables 2, 3) documented an overall pedogenetically similar environment of the soils examined, belonging to land with high or medium altitude, well drained and steep slopes (strongly dissected topography), and strongly eroding area, influenced by a temperate climate on parent rock rich in magnesium (serpentinite). Particle size analysis showed that these soils were rich in sand, often exceeding 60%. At the same time, most profiles were characterized by a low clay content (Table 2). The texture ranged from loamy sand in profiles 7 and 10 to sandy loam in the other profiles (Fig. 3). An exception was represented by profile 3, where we registered an increase of fine particles (silt and clay) due to the pre-weathered nature of the parent materials. The weak aggregation in these soils could also be linked to the loamy sandy texture and also to the high amount of Mg in the CEC. Also, the soils always showed a great stoniness (Table 2), with stones nearly unweathered. The organic material content greatly increased in A horizons of the moredeveloped soil (i.e., Mollic Leptsol), while on average reached 3–4%. The C/N ratio around 10 and the neutral pH of the organic horizons were typical of the Mull humus form, and these horizons consisted of well-humified organic matter with stable mineral-organic complexes. In fact, soil reaction was generally neutral to subalkaline, with the highest pH(H2O) value which was 8, recorded in profile 10 (Table 3). The pH of the soil developed from ultrabasic rocks was often neutral because of the high MgO content; magnesium ions were predominant instead of aluminum and hydrogen (Kataeva et al. 2004). The total CaCO3 content was variable; some samples of the shallow and well-drained soils showed percentages <5, denoting complete leaching of carbonates (this may explain their low pH and low CEC), while higher values, ranging from 9 to 15, occurred in the other soil profiles. Cation Exchange Capacity (CEC) was low to moderate, ranging between 4.0 and 28.6 cmol(+)/kg-1 (Table 3). CEC could be related to the amount of organic matter content in some A horizons (profiles 7, 9) or to the amount of clay (profiles 4, 5) in others. Ca and Mg were the dominant exchangeable cations in all samples, whereas K was much lower and Na negligible (Table 3). These element concentrations were low because of their low presence in the parent rock 414 Northeastern Naturalist Vol. 16, Special Issue 5 Table 3. Main chemical characteristics of the selected soil horizons. CND = conductivity, TC = total carbonates, AC = active carbonate, CEC = cation exchange capacity, BS = base saturation, OM = organic matter, TN = total nitrogen, and C/N = carbon/nitrogen ratio. Profile and horizon 9 Characteristic 1 C(A) 2 C(A) 3 C(B)1 4 C(A) 5 A/C 6 C(A) 7 A/C 8 C(B) AO AC 10 C1 pH (H2O) 6.9 6.7 7.5 7.4 7.3 7.8 7.1 7.5 7.2 7.4 8.0 CND (μS/cm) 34.4 94.0 41.7 37.5 51.7 27.0 63.6 34.0 34.4 51.1 24.4 TC (g/100g) 0.9 1.1 4.5 14.4 12.5 15.6 9.1 9.0 13.8 14.5 10.4 AC (g/100g) - - - 0.1 0.2 0.2 0.4 0.1 0.2 0.0 0.0 CEC (cmol(+)/kg) 4.5 4.0 7.4 15.6 24.0 15.6 18.1 14.7 28.6 7.0 9.0 BS (%) 54 100 100 86 77 82 80 70 67 100 97 OM (g/100g) 2.4 3.3 1.7 3.2 3.1 1.1 4.6 2.4 12.7 4.1 0.6 TN (g/kg) 1.3 1.7 1.1 1.6 2.4 0.7 3.0 2.0 9.7 3.9 0.6 C/N 10.8 11.3 9.0 11.6 7.4 9.1 8.8 6.9 7.6 6.1 6.0 K (mg/kg) 44.2 59.8 29.8 47.2 70.6 49.4 63.5 58.9 73.7 46.5 30.0 Ca (mg/kg) 261.1 414.6 587.4 879.4 1117.6 1039.3 1127.8 872.6 1418.3 434.3 291.2 Mg (mg/kg) 125.9 230.0 691.2 1038.2 1542.0 900.3 1055.5 700.0 1472.5 819.4 880.4 Ca/Mg 2.07 1.8 0.84 0.84 0.72 1.15 1.06 1.24 0.96 0.5 0.3 Na (mg/kg) - - - - - - - - - - - P (mg/kg) 3.3 3.8 0.5 0.5 0.5 0.5 1.6 0.5 2.8 1.2 0.8 Zn (mg/kg) 54.7 2.3 0.9 2.4 1.6 1.2 1.5 1.2 2.3 0.1 0.1 Cd (mg/kg) 0.013 0.021 0.0018 0.076 0.060 0.024 0.057 0.017 0.41 0.013 0.041 Cr (mg/kg) <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 Ni (mg/kg) 22.4 57.1 31.1 57.6 70.4 24.5 33.7 47.7 122.7 49.7 26.0 2009 S. Marsili, E. Roccotiello, I. Rellini, P. Giordani, G. Barberis, and M.G. Mariotti 415 (Alexander 2004). The majority of the profiles (60%) present low Ca/Mg ratios unfavorable for plant growth , which is typical of soil formed on ultramafic rock where abundant Mg is released from serpentine weathering (Brooks 1987). But it should be noted that some soil profiles in the research area are characterized by higher Ca/Mg ratios. High Ca/Mg ratios matched with soils which were very young and weakly developed (C[A]-R profile; Table 2). In these soils, at the early stage of soil formation, the process of weathering of serpentine mineral, and subsequent release of large amounts of Mg, was limited; moreover, they presented a profile rich in stones and gravel, resulting in better drainage and increased Mg leaching (Lee et al. 2001). Soils formed from ultramafic rock generally displayed high contents of heavy metals (Ni and Cr) compared with more common soils on alumiminosilicate rocks (Kierczack et al. 2007). Within the studied plots, the soils were very shallow, young pedons, and their concentrations of Ni and Cr were very low due to the short time of weathering and/or pedogenesis. In fact, chromium, which is predominantly contained in chromites, which were highly resistant to weathering (Oze et al. 2004), appeared in insignificant proportion (<0.05), as did Cd. Nickel concentrations were also low relating Figure 3. Chart of the basic textural classes. Dots represent soil profiles. 416 Northeastern Naturalist Vol. 16, Special Issue 5 to Ca and Mg, and the mean concentrations were lower than phototoxic level (<0.02 cmol/kg, Proctor and Woodell 1975). Finally, Ni increased clearly in soil profiles (9, 5) with high values of organic C and CEC, as suggested by Gasser and Dahlgren (1994). The best solution for NMS was a bi-dimensional configuration (maximized difference between the best of 40 runs of real data and 50 randomized runs, P < 0.05 from Monte Carlo test; average stress = 10.6) (Fig. 4). Cumulative Pearson r2 between distances in the original space and distances on the first two ordination axes was 0.825. Axes were rotated on the response variable “Cerastium utriense” Axes 1 and 2 each accounted for about 40% of the total variation of the dataset (Axis 1 r 2 = 0.391; Axis 2 r 2 = 0.434) (Figs. 4 and 5 and Table 3). Axis 1 was associated with a gradient of increasing available P (r = 0.853) and decreasing pH (r = - 0.670). Plots with negative scores on Axis 2 were characterized by high levels of available Ni (r = -0.735), Cd (r = -0.724), exchangeable Mg (r = -0.761) and Ca (r = -0.662), Figure 4. NMS ordination of plots based on species composition. Lengths of arrows for predictive factors represent strength of correlations; directions represent signs. 2009 S. Marsili, E. Roccotiello, I. Rellini, P. Giordani, G. Barberis, and M.G. Mariotti 417 and high values of CEC (r = -0.772). The total lime was associated with negative scores of both Axis 1 (r = -0.761) and Axis 2 (r = -0.668). It is noteworthy that the abundance of C. utriense in the plots was generally rather low (Fig. 5A). Nevertheless, the species showed increasing cover with high levels of available P (Fig. 5B), low pH (Fig. 5C) and low CEC (Fig. 5D). Among the species screened, only Thlaspi caerulescens and Alyssoides utriculata gave positive reaction to the DMG screening test for nickel detection. Despite the high concentration of trace metals in the substrate, C. utriense did not show any reaction to the DMG test. Figure 5. NMS ordination of plots based on species composition. Triangle sizes are proportional to the values of the considered variable: A) Cerastium utriense coverage, B) concentration of available P, C) soil pH, and D) cation exchange capacity (CEC). 418 Northeastern Naturalist Vol. 16, Special Issue 5 Discussion The present study provided the first phytosociological descriptions of plant communities where C. utriense grows These plant communities seem to belong to Italian vegetation typical of serpentine soils represented by species of the Alyssion bertolonii alliance characterizing poorly developed soils with the strongest influence of parent rock; other important and frequently occurring species represent the link with the steps characterized by xerophilous grasslands of the Festuco-Brometea class, living in more developed soils. More phytosociological studies are needed to evaluate the existence of an independent C. utriense-community (association), characterized by the same C. utriense and other species. The first quantitative analysis of the ecology of such plant communities permitted us to indirectly extrapolate informations about the ecological requirements of C. utriense. According to the presence of xerophilous species, heliophilous species, and species adapted to live in nutrient-poor soils, we can indirectly infer that C. utriense has the same requirements. According to the NMS ordination and the granulometric analysis, the distribution of C. utriense seems to be related to physical parameters of serpentine soils. In fact, C. utriense shows a strong preference for soils with low percentages of silt, characterized by low water retention and higher aeration, whereas there were no significant relationships with typical chemical factors, such as low Ca/Mg ratio and high concentration of heavy metals. Furthermore, the species does not show a tendency to hyperaccumulate Ni, to the contrary of data reported for other serpentinophytes in Italy, such as Alyssum bertolonii Desv. (Minguzzi and Vergnano 1948). A relatively complex relationship was found between C. utriense and the main soil nutrients. C. utriense was preferentially found at sites with a very low concentration of available P. This observation, linked with the phytosociological data and the relationship with poorly developed soils, suggests a pioneer character for the species. Nevertheless, the NMS analysis partially contradicted this vision, pointing out that the coverage of C. utriense strongly increases with increasing levels of available P; this finding suggests that factors other than chemicals may control the species distribution. It is clear that C. utriense is a competition-avoider because it is constantly present on under-developed soils, but it grows well without competitors on soils relatively rich in nutrients. Nevertheless, this species has been found only on ultramafic substrates. From its discovery (Barberis 1988) until the last years (Barberis et al. 2004), all the field research (including this study) showed that the species lives on peridotites, serpentinites, and serpentinoschists, but it disappears where calceschists, metagabbrous, and eclogites are present. Hence, it is difficult to examine the exact nature of C. utriense’s relationship to ultramafic soil. In forthcoming works, we will analyze soils with comparable vegetational and lithological characteristics but lacking C. utriense in order to evaluate alternative hypotheses. 2009 S. Marsili, E. Roccotiello, I. Rellini, P. Giordani, G. Barberis, and M.G. Mariotti 419 Acknowledgments Thanks are expressed to Dr. Nishanta Rajakaruna and to two anonymous reviewers for the helpful comments on this manuscript. Literature Cited Alexander, E.B. 2004. Serpentine soil redness, differences among peridotite and serpentinite materials, Klamath Mountains, California. International. Geology Review 46:754–764. Barberis, G. 1988. Cerastium utriense sp. nov. (Caryophyllaceae) nell’Appennino ligure – piemontese. Webbia 42:153–160. Barberis, G., S. Marsili, and F. Orsino. 2004. Stato delle conoscenze della flora del Parco Naturale di Capanne di Marcarolo (AL). Revue Valdotaine d’Histoire Naturelle 58:77–102 Barkman, J.J. 1989. A critical evaluation of minimum area concepts. Plant Ecology 85:89–104. Boscaiu, M., K. Marhold, and F. Ehrendorfer. 1996. Karyological and nomenclatural notes on the Cerastium banaticum group (Caryophyllaceae). Flora Mediterranea 6:71–79. Boscaiu, M., O. Vicente, and F. Ehrendorfer. 1999. Chromosome numbers, karyotypes, and nuclear DNA contents from perennial polyploid groups of Cerastium (Caryophyllaceae). Plant Systematics and Evolution 218:13–21. Braun Blanquet, J., M. Roussine, and R. Negre. 1952. Le Groupements Végétaux de la France Méditerranéenne. Centre Nationale Récherche Scientifique, Montpellier, France. Brooks, R.R. 1987. Serpentine and its Vegetation. Discorides Press, Portland, OR, USA. Capponi, G., and L. Crispini. 2002. Structural and metamorphic signature of alpine tectonics in the Voltri Massif (Ligurian Alps, northwestern Italy). Eclogae Geologicae Helvetiae 95:31–32. Charlot, G. 1964. Colorimetric Determination of Elements. Elsevier Scientific Publishing Co., Amsterdam, The Netherlands. Chiesa, S., L. Cortesogno, F. Forcella, M. Galli, B. Messiga, G. Pasquare, G.M. Pedemonte G.B. Piccardo, and P.M. Rossi. 1975. Assetto strutturale ed interpretazione geodinamica del Gruppo di Voltri. Bollettino Società Geologica Italiana 94:555–581. Cortesogno, L., W.G. Ernst, M. Galli, B. Messiga, G.M. Pedemonte, and G.B. Piccardo. 1977. Chemical petrology of eclogitic lenses in serpentinite, Gruppo di Voltri, Ligurian Alps. Journal of Geology 85:255–277. Dumas, J.B.A. 1831. Procédés de l'analyse organique. Annales des Chimie et des Physique 247:198–213. Ellenberg, H. 1974 Zeigerwerte der Gefässpflanzen Mitteleuropas. Scripta Geobot. 9. Göttingen, 1974. 2. Aufl. (1979). 3. Aufl. (1992), In H. Ellenberg et al. (Ed.). Scripta Geobotanica 18:9–166 (rist. 2004). FAO, 2006a. Guidelines for Soil Description. Fourth Edition. Rome, Italy, 97 pp. FAO, 2006b. World Reference Base for Soil Resources. 2006 Edition. Rome, Italy, 128 pp. Furrer, E., and A. Hofman. 1969. Das Euphorbietum spinosae-ligusticae, eine serpentingesellschaft in Ligurien. Acta Botanica croatica XXVIII:81–91. 420 Northeastern Naturalist Vol. 16, Special Issue 5 Gasser, U.G., and R.A. Dahlgren. 1994. Soil-phase speciation and surface association of metals in serpentinitic soils. Soil Science 158:409–420. Gustafson, D.J., G. Romano, R.E. Lathman, and J.K.Morton. 2003. Amplified fragment-length polymorphism analysis of genetic relationships among the serpentine barrens endemic Cerastium velutinum Rafinesque var. villosissimum Pennell (Caryophyllaceae) and closely related Cerastium Species. Journal of the Torrey Botanical Society 130:218–223. Kataeva, M.N., N.V. Alexeeva-Popova, I.V., Drozdova, and A.I Beljaeva. 2004. Chemical composition of soils and plant species in the Polar Urals as influenced by rock type. Geoderma 122:257–268. Kierczak, J., C. Néel, H. Bril, and J. Puziewicz, 2007. Effect of mineralogy and pedoclimatic variations on Ni and Cr distribution in serpentine soils under temperate climate. Geoderma 142:165–177. Kruskal, J.B. 1964. Nonmetric multidimensional scaling: A numerical method. Psychometrika 29:115–129. Lee, B.D., R.C. Graham, T.E. Laurent, C. Amrhein, and R.M. Creasy, 2001. Spatial Distributions of Soil Chemical Conditions in a Serpentinitic Wetland and Surrounding Landscape. Soil Science Society of America Journal 65(4):1183–1196. Marin, P.D., and B. Tatic. 2001. Serpentine soil and plant diversity, with emphasis on the Balkan Peninsula. Bocconea 13:145–150. McCune, B., and M.J. Mefford. 1999. Multivariate Analysis of Ecological Data. Version 4.25. MjM Software, Gleneden Beach, OR, USA. Minguzzi, C., and O. Vergnano. 1948. II contenuto di nichel nelle ceneri di Alyssum bertolonii. Atti Società Toscana Scienze Naturali 55:49–74. Ministero delle Politiche Agricole e Forestali (MiPAF). 2000. Osservatorio nazionale pedologico e per la qualità del suolo, International Society of Soil Science, Società Italiana della Scienza del Suolo. Metodi di analisi chimica del suolo. F. Angeli (Ed.). Milano, Italy. Morton, J.K. 2004. New combinations in North American Caryophyllaceae. SIDA Contributions to Botany 21:887–888. Nagy, L., and J. Proctor. 1997. Soil Mg and Ni as causal factors of plant occurrence and distribution at the Meikle Kilrannoch ultramafic site in Scotland. New Phytologist 135:561–566. Niketic, M. 1994. Cerastium neoscardicum, a new species of ser. Alpina from Mt. Shar-Planina, Serbia. Bulletin of the Institute of Botany and Botanical Garden, University of Belgrade, Belgrade, Yugoslavia 26–27:63–70. Nyberg Berglund, A.B. 2005. Postglacial colonization and parallel evolution of metal tolerance in the polyploid Cerastium alpinum. Ph.D Dissertation. Department of Plant Biology and Forest Genetics, SLU Service/Repro, Upsalla, Sweden. Acta Universitatis Agriculturae Sueciae, 2005(65):1–51. Olsen, S.R., C.V. Cole, F.S. Watanabe, and L.A. Dean. 1954. Estimation of available phosphorus in soils by extraction with sodium bicarbonate. USDA circular 939. US Governement Printer Office, Washington, DC. Oze, C., S. Fendorf, D.K. Bird, and R.G. Coleman, 2004. Chromium geochemistry in serpentinized ultramafic rocks and serpentine soils from the Franciscan complex of California. American Journal of Science 304:67–101. Pignatti Wirkus, E., and S. Pignatti. 1977. Die vegetation auf serpentin-standorten in den nordlichen apenninen. Studia Phytologica in honorem jubilantis A.D.Horvat 14:113–124. Pignatti, S., P. Menegoni, and S. Pietrosanti. 2005. Valori di bioindicazione delle piante vascolari della flora italiana. Braun-Blanquetia 39:3–97. 2009 S. Marsili, E. Roccotiello, I. Rellini, P. Giordani, G. Barberis, and M.G. Mariotti 421 Proctor, J., and S.R.J. Woodell. 1975. The ecology of serpentine soils. Advances in Ecolological Restoration 9:255–366. Rajakaruna, N., T.B. Harris, and E.B. Alexander. In press. Serpentine geoecology of eastern North America: A review. Rhodora. Rivas-Martinez, S. 2004. Synoptical Worldwide Bioclimatic Classification System. Available on line at http://www.globalbioclimatics.org/book/claves.htm. Accessed 2008. Shepard, R.N. 1962. The analysis of proximities: Multidimensional scaling with an unknown distance function. I. Psychometrika 27:125–139. Spagnolo, C., L. Crispini, and G. Capponi. 2007. From Alpine to Apennine orogeny: Structural records in the Voltri Massif (Ligurian Alps, Italy). Geodinamica Acta 20:21–35. Stevanovic, V., K. Tan, and G. Iatrou. 2003. Distribution of the endemic Balkan flora on serpentine I. Obligate serpentine endemics. Plant Systematics and Evolution 242:149–170. Tyndall, R.V., and J.C. Hull. 1999. Vegetation, flora, and plant physiological ecology of serpentine barrens of eastern North America. Pp. 67–82, In R.C. Anderson, J.S. Fralish, and J.M. Baskin (Eds.). Savannas, Barrens, and Rock Outcrop Plant Communities of North America. Cambridge University Press, New York, NY, USA. United States Department of Agriculture (USDA). 1998. Keys to Soil Taxonomy, 8th Edition. US Department of Agriculture, Soil Survey Staff, Natural Resources Conservation Service, Washington DC, 326 pp. Vagge, I. 1997. Le garighe a Genista desoleana Valsecchi ed Euphorbia spinosa L. subsp. ligustica (Fiori) Pign. della Liguria orientale (Italia NW). Fitosociologia 32:239–243 Vanossi, M., L. Cortesogno, B. Galbiati, B. Messiga, G. Piccardo, and R. Vannucci. 1984. Geologia delle Alpi Liguri: Dati, problemi, ipotesi. Memorie Società Geologica Italiana 28:5–75.