Soil and Biota of Serpentine: A World View 2009 Northeastern Naturalist 16(Special Issue 5):385–404 Nickel Hyperaccumulation by Brassicaceae in Serpentine Soils of Albania and Northwestern Greece Aida Bani1,2, Guillaume Echevarria2,*, Alfred Mullaj3, Roger Reeves4, Jean Louis Morel2, and Sulejman Sulçe1 Abstract - Ultramafic soils are widespread in the Balkans. Albania and Greece are the richest in the number of endemics, including several hyperaccumulator species, growing on serpentine. The objectives of this study were to understand the potential of Ni hyperaccumulation of these species in close relation with the characteristics of their native soil environments. Collection of both plant samples (analysis of element concentrations in aerial parts) and soil samples (analysis of total elements, DTPA-extractable Ni, Fe, and Ni distribution in mineral phases) allowed evaluation of phenotypic efficacy in hyperaccumulating Ni. Nickel availability in soils is controlled by soil weathering and mineral-bearing phases. Unsurprisingly, the highest levels of Ni availability were associated with amorphous Fe-oxides in moderately weathered Cambisols or with high-exchange capacity clays in wellevolved Vertisols. The highest Ni concentrations in leaves were found in Alyssum murale in Pojska (Albania; 2.0%), Alyssum markgrafii in Gjegjan (Albania; 1.9%), Bornmuellera baldacii subsp. markgrafii in Gramsh (Albania; 1.4%), and Leptoplax emarginata in Trigona (Greece; 1.4%). We identified a new member in the Albanian Ni-hyperaccumulator flora: Thlaspi ochroleucum in Pojska (Albania: 0.13% Ni) and in Pishkash (0.14% Ni). With regard to Ni availability in soils, A. markgrafii (Albania) is the most efficient Ni-hyperaccumulator among all species. Alyssum murale, which is widespread in the serpentines of the Balkans, accumulates Ni, with leaf concentration being negatively correlated to total Ca content of soils regardless of Ni availability (DTPA extractable Ni). If this relationship is confirmed, it would mean that genetic variability is not the main factor that explains the hyperaccumulation performance of this species. Introduction Hyperaccumulation of Ni, i.e., accumulation in aerial parts up to concentrations above 1000 mg kg-1 on a dry-matter (DM) basis (Brooks et al. 1977, Reeves 1992), was first discovered in Alyssum bertolonii Desv. in Italy (Minguzzi and Vergnano 1948). It has become recognized as an unusual response to the elevated Ni concentrations generally found in soils derived from ultramafic (i.e., Mg- and Fe-rich) rocks, such as peridotites and serpentinites, often referred to as serpentine soils. The extreme chemical nature of 1Agro-Environmental Department, Agricultural University of Tirana, Kamez, Albania. 2Laboratoire Sols et Environnement, Nancy-Université, INRA, 2 Avenue de la Forêt de Haye, B.P. 172 F-54505 Vandoeuvre-lès-Nancy, France. 3Science University, Tirana, Albania. 4School of Botany, University of Melbourne, Melbourne, Australia. *Corresponding - Guillaume.Echevarria@ensaia.inpl-nancy.fr. 386 Northeastern Naturalist Vol. 16, Special Issue 5 such soils, with abnormally high concentrations of Ni, Cr, and Co (in addition to Mg and Fe) and low concentrations of the important nutrients Ca, K, and P, often leads to such areas having a characteristic flora. The serpentine flora often consists of the limited subset of a region's species that can tolerate these unusual edaphic conditions, but may also contain a significant number of endemic species of very restricted range (Reeves et al. 1999), including Ni-tolerant and Ni-hyperaccumulating plants. Serpentine (ophiolithic) substrate covers large areas in the Balkans, more than in any other part of Europe. The number of Balkan endemics growing on serpentine is ca. 335 taxa (species and subspecies), of which 123 are obligate. Their distribution is presented in 50- x 50-km UTM squares as adopted in the Atlas Florae Europaeae project coordinated at Helsinki. The richest (in number of taxa) squares are situated in northwestern Greece (Epirus), the island of Evvia, northern Albania together with southwestern Serbia, and northern Greece (Vourinos). They indicate important centers of plant diversity in the Balkans, and thus areas to be noted for conservation attention. Serpentine areas in the Balkans exist in large blocks or as small outcrops separated from other geological formations in central Bosnia and western and central Serbia. They extend towards north-central and southeastern Albania to the serpentine formations of Epirus and Thessalia in Greece (Stevanovic et al. 2003). Although serpentine is inhabited by a smaller number of species than known from other types of geological substrates, nevertheless, the serpentine flora in the Balkans is characterized by a relatively high degree of endemism. Ultramafic outcrops in the Balkans host a rich vegetation (Stevanovic et al. 2003), including a number of endemics, subendemics, and Ni hyperaccumulators such as trans-regional endemics (Bornmuellera baldacii (Degen) Heywood [Greece and Albania], Leptoplax emarginata [Greece], Alyssum markgrafii [Albania, Serbia]) and regional endemics (Alyssum heldreichii [Greece], Bornmuellera tymphaea [Greece]). Previous studies clearly evidenced a great variability in Ni concentrations in serpentine soil and serpentine plants in Albania (Bani et al. 2008, Massoura et al. 2004, Shallari et al. 1998) and the Pindus Mountains of Greece (Chardot et al. 2005; R. Reeves, (unpubl. data). Using herbarium material collected in Greece and previous studies in Albania, the authors have detected the presence of Nihyperaccumulators: Alyssum markgrafii (Albania) with 13,700 mg Ni kg-1 DM (Brooks and Radford 1978); Bonmuellera baldaccii subsp. markgrafii (Albania) with 27,300 mg Ni kg-1 DM (Reeves et al. 1983); and Bornmuellera tymphaea (Greece) and Leptoplax emarginata (Greece) with more than 1% Ni DM. This extraordinary concentration of Ni hyperaccumulators in a limited geographical area and in a limited part of the plant kingdom has led us to investigate the hyperaccumulation of Ni by the ultramafic flora of Albania and Greece more widely. The objectives of this study were to identify collection sites in which Ni-hyperaccumulator species occur and to understand the relationships 2009 A. Bani, G. Echevarria, A. Mullaj, R. Reeves, J.L. Morel, and S. Sulçe 387 between Ni uptake by native species or populations of selected Ni-hyperaccumulators and the corresponding Ni availability in soils across the two regions of Albania and Greece. This information is of particular interest now that several potential applications of hyperaccumulators, such as phytoremediation and phytomining, are being extensively discussed and tested (Brooks 1998). Materials and methods Sites investigated Several serpentine areas in Albania and the Pindus Mountains of Greece were surveyed because they host a significant proportion of regional hyperaccumulating endemics, with several common species in both areas. According to the geographical position of the areas, i.e., the grouping of serpentine formations as well as their spatial isolation, they fall in the same serpentine botanical areas in the Balkans: south-central Albania–northwestern Greece (Reeves et al. 1983 Stevanovic et al. 2003, Tan et al. 1999). Both regions are located in similar climate zones (Mediterranean climate with mountain influence). The Pindus Mountains (boundary between Epirus and Thessalia) are located in northwestern Greece and extend about 160 km from the border of Albania. Sampling of Ni hyperaccumulator species growing in those sites and their associated soils was conducted in May 2007 at eight serpentine sites in Albania, and in July 2006 at five serpentine sites in the Pindus Mountains (Greece). Leaf fragments of species growing on these sites were screened for Ni accumulation by a simple semi-quantitative test using filter paper impregnated with dimethylglyoxime. A positive result generally indicates a Ni concentration of >1000 mg kg-1 (DM basis; Reeves et al. 1996). Plants were identified using Flora Europea (Tutin et al. 1964, 1968, 1972, 1976, 1980). Nine taxa from eight species of hyperaccumulators were sampled on the two surveys, some of them sampled at various locations, including in both countries (Table 1). Shoots and leaf materials were washed in deionised water and dried at 75 °C for 48 h before elemental analyses. Serpentine sites in Albania. In the serpentine site of northeastern Albania, the climate data are: mean January temperature = 4.9 °C, mean July temperature = 23.8 °C, and total annual rainfall = 1800 mm. In the serpentine sites of southeastern Albania, they are respectively: for Prrenjas—mean January temperature = 5.6 °C, mean July temperature = 22.5 °C, and total annual rainfall = 744 mm; for Pojska—mean January temperature = 4.7 °C, mean July temperature = 22.3 °C, and total annual rainfall = 707 mm. Pojska (Pogradec district) is in the east of Albania, on the western shore of lake Ohrid (40°59'55.28"N, 20°38'0.92"E) at an elevation of 700 m above sea level, . It is a serpentine site with high contents of Fe, 388 Northeastern Naturalist Vol. 16, Special Issue 5 Ni, and Co silicates. Up to 2% Ni has been recorded in the iron-type laterites of the Librazhd-Pogradec region (Anonymous 1992). In this area, there are also Cr mines. The soils at the collection site were a Magnesic Hypereutric Cambisol and a Cambic Hypermagnesic Hypereutric Vertisol located close to the shores of Lake Ohrid at the piedmont of ultramafic hills (Serpentinized Harzburgite). The site at Prrenjas surrounds the city of Prrenjas at an elevation of 600 m. The parent material is rich in Fe, Ni, Cr, and Co (Anonymous, 1992). Sampling was conducted on serpentine rock of Prrenjas (Leptic Hypermagnesic Cambisol; 41°04'13"N, 20°33'53"E) and also on the Domosdova field, which contains a soil developed from a colluvium of ultramafic and magnesite origin (Cambic Hypermagnesic Hypereutric Vertisol; 41°04'08"N, 20°33'11"E). Gjegjan/Fushë-Arrëz is situated at the north of Albania at an elevation of 400–600 m, (41°55'47"N, 20°00'09"E). Silicate-type ores of this region consist of some laterites containing up to 2.6% Ni (Anonymous 1992). Sampling was conducted along the road from Gjegjan to Fushë-Arrëz on serpentine rock outcrops (villages of Gjegjan, Gjegjan-Gojan, and Fushë- Arrëz) (42°03'50"N, 20°01'26"E). Soils were mainly Hypermagnesic Hypereutric Cambisols. Gramsh (Grabova) is situated in the east of Albania at an elevation of 1594 m, (40°55'18"N, 20°21'13"E). The site of sampling is a pasture with low productivity that is currently being grazed. This is the only serpentine site where we have collected Bornmuelera baldacii subsp. markgrafii (Family Brassicaceae). Soil samples could not be collected on this survey. Serpentine sites in Pindus Mountains (Northwestern Greece). The four sites are located in the wide ultramafic outcrop of Pindus Mountains. The Pindus Mountain range has a sub-Mediterranean climate (mean January temperature = 4.6 °C, mean July temperature = 24.9 °C, and annual rainfall varies from about 1000 mm at the coast to 2500 mm on the mountains). The rainfall in Metsovo is 1512 mm (Metsovo—39°46N, 21°11E, elevation 1165m—is the nearest town center, a few kilometers away from Katara Table 1. List of Ni-hyperaccumulating taxa collected during the two surveys Country Species Site of collection GR Alyssum heldreichii Hausskn. Katara Pass, Malakasi AL A. markgrafii O.E. Schulz ex Markgraf Gjegjan AL, GR A. murale Waldst. & Kit. Katara Pass, Malakasi, Trigona, Gjegjan, Pishkash, Pojska, Prrenjas GR Bornmuellera baldacii baldacii (Degen) Heywood Vovoussa AL B. baldacii markgrafii O.E. Schulz Gramsh GR B. tymphaea (Hausskn.) Hausskn. Katara Pass, Malakasi, Trigona, GR Leptoplax emarginata (Boiss.) O.E. Schulz Katara Pass, Malakasi, Trigona AL Thlaspi ochroleucum Boiss. & Heldr. Pishkash, Pojska GR T. tymphaeum Hausskn. Katara Pass, Malakasi 2009 A. Bani, G. Echevarria, A. Mullaj, R. Reeves, J.L. Morel, and S. Sulçe 389 Pass). The bedrock is highly serpentinized and consists mainly of highly serpentinized harzburgite or lherzolite. All soils were all Hypermagnesic Hypereutric Cambisols. Malakasi is situated at an elevation of 1031–1065 m, (39°47.40'N, 21°17.57'E). The site of sampling is a site that is currently being used for pasture. In this site, Alyssum heldreichii, A. murale, Bornmuellera tymphaea, Thlaspi tymphaeum, and Leptoplax marginata were collected at fruiting-seeding stage. Katara Pass is situated at the elevation 1690 m, (39°47.77'N, 21°13.74'E). In this site, shoots of A. heldreichii, A. murale, T. tymphaeum, and L. emarginata were collected at flowering stage, and B. tymphaea was collected at fruiting stage. Vovoussa is situated at 1570 m elevation, (39°52.11'N, 21°02.99'E). In this site, shoots of Bornmuellera baldacii subsp. baldacii were collected at the seeding stage. Trigona is situated at a much lower elevation (830 m) than the three other Greek sites, (39°47.29'N, 21°25.32'E). In this site, shoots of A. murale, B. tymphaea, and L. emarginata were collected at the seeding stage. Soil and plant sampling At each site and for each species, composite plant samples (whole plants) were taken with 3–5 specimens of each species on areas showing similar pedological conditions (same topography and soil type). Soil samples were also taken as composite samples collected randomly at the different precise locations of plant sampling (in a 0.5-m radius area around the plants). Soil was collected from the upper horizon at a depth of approximately 20 cm (when possible) or less, air-dried, and sieved to 2-mm. Soil and plant analyses Total major elements and total trace elements in soil were determined in the Soil Analyses Laboratory of Arras according to standardized methods in France (AFNOR 2004). Trace metal concentrations in shoots and leaves were analysed by plasma emission (ICP) spectrometry after digestion of plant samples in microwaves. A 0.25-g DM plant aliquot was digested by adding 8 ml of 69% HNO3 and 2 mL of H2O2. Solutions were filtered and adjusted to 25 mL with 0.1 M HNO3. Soil pH was determined in a soil/water (1:2.5) suspension with a pH-meter. Nickel availability in different soil samples was characterised by DTPATEA. DTPA-extractable Ni was determined using the method of Lindsay and Norvell (1978). Concentration of Ni in soil extracts was determined by plasma emission spectrometry (ICP-OES). Estimation of Fe and Ni partitioning was achieved through combined oxalate (McKeague and Day 1966) and dithionite-citrate-bicarbonate (Mehra and Jackson 1960) extractions. This procedure allowed us to estimate the different forms of iron (silicate, poorly crystallised, and well-crystallised Fe-oxides) in the soil and to evaluate the partition of Ni between those three 390 Northeastern Naturalist Vol. 16, Special Issue 5 mineral phases. Although some imprecision is associated with this method (as for all extraction methods), it has been successfully used previously to assess Ni availability (highly significant correlations were found between isotopically labile Ni and Ni fractions within Fe oxides extracted by oxalate “Feo”) in ultramafic soils worldwide, including Albanian soils (Massoura et al. 2006). We chose five distinct soils (Katara Pass, Trigona, Prrenjas-Rock, Pojska-Soil 2, Prrenjas Domosdova) from the different areas that represented different pedological conditions to assess Fe and Ni partition according to this method. Results Soil variability along the study sites Unsurprisingly, all the soils showed high total Fe contents, with values exceeding 7% except for one soil at Malakasi (Table 2). All Albanian soils had Fe concentrations higher than 9%. Total Mg and Ca contents in soils displayed more varied patterns among sites and soils. The richest soils had more than 10% Mg (with values higher than 12% at Gjegjan and Vovoussa), sometimes reaching the initial content of the bedrock which usually varies between 15 to 20%. The Ca total contents were even more variable between soils. Some had strong Ca deficiency (reaching less than 0.3% at Pojska), while others showed quite normal values (between 2 to 4%) for non-ultramafic soils. Accordingly, Mg:Ca ratios varied from 2.6 (Fushë-Arrëz) to 40 (Pojska), a range that is commonly reported in serpentine soil material (Proctor 1971). In Greek soils, they varied in a tighter range from 10.9 (Katara Pass) to 17.7 (Trigona) and did not reach the extreme values found in Albanian soils. Potassium total contents in soils were very low (between 0.1 to 0.8), except at the Malakasi-Soil 2 where it reached 1.0%. This level of K is usually a cause for nutrient deficiency in ultramafic soils (Proctor 1971). The Mn content in soils was quite homogenously distributed and varied between 1010 and 2860 mg kg-1. In general, trace element contents were variable, but the specific ultramafic metals (i.e., Co, Cr, and Ni) were found at high concentrations in all soils. Cobalt content in soils varied between 93 (Malakasi-Soil 2) and 280 mg kg-1 (Pojska-Soil 2). Chromium concentrations in soils varied from 667 mg kg-1 (Pojska-Soil 1) to 3250 (Pishkash) mg kg-1. The soils with highest Ni contents came from eastern Albania (Pojska [3240 mg kg-1], Prrenjas [3210 mg kg-1], and Pishkash [3240 mg kg-1]). The Ni concentration in soil samples that came from the Gjegjan area varied from 1070 to 2580 mg kg-1. In the Greek soils, it was also lower than that from eastern Albania and comprised between 1280 and 2660 mg Ni kg-1. Copper concentrations in soils were lower than 76 mg kg-1, except for the Gjegjan site where the Cu concentration was 1120 mg kg-1. Zinc concentrations in soils were lower than 130 mg kg-1, except for the Gjegjan soil. The latter is in use as 2009 A. Bani, G. Echevarria, A. Mullaj, R. Reeves, J.L. Morel, and S. Sulçe 391 Table 2. pH, total major elements, and total trace elements in Albanian and Greek soils. pH Total major elements (%) Total trace elements (mg kg-1) Site of collection (water) Ca Mg Al K Na Fe Co Cr Mn Cu Zn Ni Pojska-Soil 1 7.98 0.26 10.5 1.73 0.38 0.17 9.4 182 677 1760 15.7 102 3180 Pojska-Soil 2 6.81 0.45 7.8 2.62 0.53 0.26 11.0 280 2400 2610 21.6 131 3240 Prrenjas-Domosdova 7.36 0.60 5.4 3.33 0.36 0.26 9.6 267 2150 2860 21.8 104 3100 Prrenjas-Rock 7.44 0.72 9.4 2.31 0.42 0.22 9.3 222 2950 2160 24.2 112 3210 Pishkash-Soil 1 (A. murale) 8.18 2.33 9.5 1.96 0.31 0.14 10.3 176 3250 1330 20.8 101 3240 Pishkash-Soil 2 (T. ochroleucum) 6.73 0.57 10.5 2.43 0.43 0.15 10.2 175 2060 1330 22.5 114 2640 Gjegjan 7.08 3.35 9.9 3.89 0.15 0.32 10.9 104 1100 1010 1120.0 816 1070 Gjegjan-Rock 7.83 1.76 12.7 3.12 0.12 0.21 9.6 162 2020 1770 76.3 107 2580 Gjegjan-Gojan 6.61 1.75 5.8 5.65 0.39 0.39 9.9 169 1930 2550 66.9 107 1670 Fushë-Arrëz 7.74 3.46 9.1 6.17 0.19 0.26 7.0 100 1550 1220 56.3 87 1370 Vovoussa 6.72 1.33 16.9 1.77 0.19 0.11 9.2 201 2670 2160 - 78 2610 Trigona 7.02 0.44 7.8 2.63 0.42 0.16 10.0 175 2170 1620 - 107 2660 Katara Pass 6.30 0.95 10.4 4.00 0.77 0.45 7.5 102 1380 1820 - 104 1160 Malakasi-Soil 1 7.00 0.62 8.0 3.00 0.65 0.43 7.4 117 1100 1490 - 89 2340 Malakasi-Soil 2 6.76 0.54 7.0 3.89 0.97 0.74 5.8 93 1750 1310 - 82 1280 392 Northeastern Naturalist Vol. 16, Special Issue 5 a garden soil and has probably suffered from anthropogenic inputs because of the high concentrations in Cu and Zn, although these two elements can sometimes be found at high concentrations in ultramafic soils. The lowest trace metal concentrations were found for the soil that had the lowest concentration of total Fe (i.e., Malakasi-Soil 2). All the soils were neutral to alkaline. The range of pH values among the soils was quite narrow. The most acidic soil had a value of 6.3 (Katara Pass), which was also the soil at the highest elevation (more than 1600 m above sea level). The most alkaline soil had a value of 8.2 (Pishkash-Soil 1). Distribution of Fe and Ni within the different mineral compartments in soils In all soils, both total Fe and Ni contents were much higher than in typical non-ultramafic soils. Since the distribution of Fe in the mineral phases seems to influence Ni chemical availability in soils (Chardot et al. 2007, Massoura et al. 2006), we characterized the mineral pools under which Fe and Ni were present in these soils. The results from oxalate and DCB extractions (Fig. 1) showed that the fraction of free Fe of the total Fe content in surface horizons of the Albanian soils represented 48% in Prrenjas- Rock, 56% in Pojska-Soil 2, and 20% in Prrenjas-Domosdova (a soil with less total Fe content and high secondary clay content than other soils, as it was a Vertisol). In these soils, free Fe was mainly present as well-crystallized Fe oxides (i.e., Fe-included in the crystal lattice of oxides) extracted by DCB. The fraction of amorphous Fe-oxides (oxalate-extracted) Figure 1. Distribution of Fe within the different mineral compartments in the surface horizons of five representative serpentine soils from Albania and Pindus Mountains (Greece). 2009 A. Bani, G. Echevarria, A. Mullaj, R. Reeves, J.L. Morel, and S. Sulçe 393 represented less than 15% of total Fe in the three soils. The fraction of free Fe in the surface horizon of Pindus Mountains soils represented 47% of total Fe in Katara Pass and 25% in Trigona. In these soils, free Fe was mainly present as well-crystallized Fe-oxides, but the fraction of amorphous Fe-oxides was more important than in the Albanian soils. In contrast with Fe distribution, most of the Ni associated with Fe oxides (Fig. 2) in Albanian soils was found in the amorphous phase (two-thirds or more). The latter fraction represented 30% of total soil Ni in Prrenjas-Rock, 34% in Pojska-Soil 2, and 26% in Prrenjas-Domosdova). In Greek soils, this trend was even more pronounced than in Albanian soils. In the Greek soils, more than two-thirds of total soil Ni was found with the clay fraction. Nickel associated with well-crystallized oxy-hydroxides was also quite low compared to Albanian soils (5% of total Ni in Trigona and 0% in Katara Pass). Variability of Ni availability along the Albanian and Pindus Mountains soils DTPA-extractable Ni (Ni DTPA) was chosen as an estimate of soil Ni chemical availability. Although it only reflects the potential pool of available Ni (R. Chaney, USDA-Agricultural Research Service, Animal Manure and By-products Lab, Beltsville, MD, USA, 2008 pers. comm.), it allows relative comparison of soils. It has also been correlated to isotopically labile Ni (Echevarria et al. 1998). We used this measurement as the indicator of the relative variability of Ni availability of the different serpentine soils (Fig. 3). Figure 2. Distribution of Ni within the different mineral compartments in the surface horizons of five representative serpentine soils from Albania and Pindus Mountains (Greece). 394 Northeastern Naturalist Vol. 16, Special Issue 5 The amounts of DTPA-extracted Ni were higher for the soils of southeastern Albania: Pojska-Soil 2 (285 mg kg-1), Pojska-Soil 1 (96 mg kg-1), Prrenjas-rock (81 mg kg-1), and Pishkash (95 mg kg-1). The amounts of DTPA-extracted Ni were lower in soils of the Gjegjan area in northern Albania, although the soil of Gjegjan-Gojan was an exception (66 mg kg-1). The amounts of DTPA-extracted Ni were also high in Pindus Mountains soils. DTPA-extracted Ni in soils varied from 126 mg kg-1 (Malakasi-Soil 1) to 48 mg kg-1 (Malakasi-Soil 2). Uptake of Ni and other elements by hyperaccumulator plants Sampling at the Albanian and Greek sites allowed for collection of nine taxa: eight plant species and two subspecies of one of them. Species were each identified according to the Flora Europaea. Analysis of metals performed on plant bulk shoots (i.e., stems, leaves and flowers/ fruits) and leaf material showed different plant responses to the presence of trace metals in soils collected at the different sites (Table 3). Leaves were preferred because they reveal the highest Ni (and possibly Co) concentrations in hyperaccumulating Brassicaceae and especially the genera studied here (Broadhurst et al. 2004, Psaras et al. 2000). For the first time, we have analysed Ni in freshly collected plant parts of the taxon Bornmuellera baldacii subsp. markgrafii from Albania; Ni concentration in bulk shoots was 12,215 mg kg-1. With the help of the dimethylglyoxime filter paper test (Reeves et al. 1996), we also identified a new member of the Figure 3. Soil Ni availability assessed by DTPA-TEA extraction in the surface horizons of serpentine soils from Albania and Pindus Mountains (Greece). Letters indicate significant difference between soils at the 5% confidence interval (ANOVA Newman-Keuls test). 2009 A. Bani, G. Echevarria, A. Mullaj, R. Reeves, J.L. Morel, and S. Sulçe 395 Table 3. Concentration of Ni, Co, Mn, Zn, Fe, Ca, and Mg in leaves of the different taxa of Ni hyperaccumulators from Albanian and Greek serpentine soils. Results are given as mean values of three replicates ± standard deviations. Species Soil Ni (103 mg kg-1) Ca (103 mg kg-1) Mg (103 mg kg-1) Co (mg kg-1) Fe (mg kg-1) Mn (mg kg-1) Zn (mg kg-1) A. heldreichi Katara Pass 11.8 ± 0.4 14.4 ± 1.2 7.0 ± 0.5 16.0 ± 2.9 687 ± 323 39.0 ± 5.7 173 ± 25 Malakasi-Soil 1 5.4 ± 0.1 26.9 ± 1.7 6.5 ± 0.6 16.0 ± 0.5 832 ± 114 73.0 ± 3.7 102 ± 5 A. markgrafii Gjegjan-Gojan 16.0 ± 1.5 32.0 ± 3.6 4.3 ± 0.1 12.0 ± 1.1 792 ± 463 70.0 ± 5.4 332 ± 20 Fushë-Arrëz 19.1 ± 0.7 25.0 ± 1.6 5.6 ± 0.1 25.0 ± 1.0 556 ± 133 69.0 ± 4.8 250 ± 22 Gjegjan 15.4 ± 4.7 29.3 ± 1.7 4.6 ± 1.1 19.0 ± 6.0 1351 ± 407 103.0 ± 20.0 376 ± 21 A. murale Pojska-Soil 1 20.1 ± 1.4 19.1 ± 4.2 5.7 ± 0.5 90.0 ± 28.0 562 ± 248 30.0 ± 2.0 310 ± 26 Prrenjas-Domosdova 15.6 ± 1.1 34.8 ± 0.7 4.6 ± 0.8 70.0 ± 5.6 923 ± 300 121.0 ± 6.4 208 ± 14 Prrenjas-rock 8.1 ± 4.1 24.5 ± 2.2 5.7 ± 1.6 28.0 ± 14 1138 ± 715 55.0± 32.0 135 ± 8.3 Pishkash-Soil 1 9.3 ± 1.1 34.2 ± 3.3 4.7 ± 0.2 38.0 ± 8.9 1261 ± 141 68.0 ± 4.0 183 ± 21 Gjegjan 4.7 ± 2.3 22.2 ± 1.4 3.5 ± 0.3 15.0 ± 6.3 792 ±107 31.0 ± 2.6 292 ± 83 Katara Pass 11.3 ± 1.4 42.9 ± 5.4 2.1 ± 0.1 12.0 ± 1.4 535 ± 149 442.0 ± 11.0 192 ± 63 Malakasi-Soil 2 10.7 ± 0.5 27.9 ± 1.0 5.6 ± 0.06 14.0 ± 0.4 879 ± 1 38.0 ± 27.0 211 ± 2 Trigona 13.5 ± 0.1 9.2 ± 0.1 7.1 ± 0.06 8.8 ± 0.9 450 ± 32 17.0 ± 1.5 163 ± 4 B. baldacii Gramsh 14.0 ± 4.3 11.1 ± 3.0 7.3 ± 2.4 9.9 ± 2.6 380 ± 418 42.8 ± 9.6 161 ± 72 Vovoussa 10.9 ± 0.2 15.8 ± 0.2 12.0 ± 0.7 43.0 ± 3.5 940 ± 586 85.6 ± 5.9 188 ± 6 B. tymphaea Malakasi-Soil 1 9.6 ± 2.0 16.7 ± 0.8 3.6 ± 0.3 8.3 ± 4.6 667 ± 39 63.0 ± 36 211 ± 3 Trigona 13.8 ± 0.2 16.4 ± 0.6 6.6 ± 0.2 7.9 ± 0.7 444 ± 447 34.8 ± 2.3 168 ± 5 L. emarginata Katara Pass 5.6 ± 0.4 23.3 ± 0.8 2.5 ± 0.2 11.0 ± 0.7 417 ± 20 53.4 ± 2.3 134 ± 7 Malakasi-Soil 1 11.7 ± 0.2 10.1 ± 0.2 6.7 ± 0.2 6.9 ± 0.2 249 ± 45 31.9 ± 1.6 117 ± 2 Trigona 13.6 ± 0.1 8.8 ± 0.1 3.7 ± 0.1 5.9 ± 0.2 359 ± 99 30.8 ± 1.0 140 ± 3 T. ochroleucum Pishkash-Soil 2 1.1 ± 0.1 9.1 ± 1.5 9.4 ± 0.6 8.2 ± 1.7 2904 ± 145 55.9 ± 1.5 113 ± 3.5 Posjka-Soil 2 1.3 ± 0.1 8.1 ± 0.7 7.9 ± 0.9 16.0 ± 3.9 1587 ± 114 107.0 ± 9.4 112 ± 11 T. tymphaeum Katara Pass 6.4 ± 0.2 13.8 ± 0.7 3.2 ± 0.6 6.4 ± 2.9 570 ± 274 30.0 ± 2.3 300 ± 39 Malakasi-Soil 2 7.0 ± 0.3 55.1 ± 0.5 3.9 ± 1.7 5.7 ± 1.5 1688 ± 804 37.0 ± 4.5 219 ± 54 396 Northeastern Naturalist Vol. 16, Special Issue 5 Albanian Ni-hyperaccumulating flora which has already been recorded in neighboring countries (e.g., Greece): Thlaspi ochroleucum. Analyses of acid-digested shoots by inductively coupled plasma (ICP) emission spectroscopy confirmed Ni concentrations of 1355 mg kg-1 in Pojska-Soil 2 and 1134 mg kg-1 in Pishkash-Soil 2 in T. Ochroleucum. The plant specimens collected from the field were already in the end of fruiting stage, and most of the leaves were dry or had fallen from stems. This fact may explain such low concentrations of Ni in shoots of T. ochroleucum. Highest Ni values in plants were recorded for Alyssum murale, which was the dominant species on all serpentine sites in Albania and Greece, but Ni concentration in this species was highly dependent on the site of collection. The highest Ni concentration in bulk shoots of A. murale was observed in Pojska-Soil 1 (13,676 mg kg-1), followed by Prrenjas-Domosdova (11,234 mg kg-1). For A. markgrafii, which was the species second-richest in Ni, the highest Ni concentration was found in Gjegjan-Gojan (12,387 mg kg-1). High Ni values in plants were also recorded for B. tymphaea in Malakasi-Soil 1 (11,774 mg kg-1) and L. emarginata also in Malakasi-Soil 1 (9509 mg kg-1DM). Thlaspi species—T. ochroleucum and T. tymphaeum— had much lower concentrations of Ni than the other taxa, but we were not able to perform any statistical comparison between species due to the limited number of sites of collection for each species. Leaves exhibited the highest concentrations among all taxa. For example, A. murale leaves had 20,135 mg kg-1 in Pojska-Soil 1 and 15,698 mg kg-1 in Prrenjas-Domosdova. The concentration of Ni in leaves of A. murale was the lowest in Gjegjan (4729 mg kg-1). It was intermediate in the three Greek soils, ranging from 10,740 mg kg-1 in Malakasi-Soil 2 to 13,595 mg kg-1 in Trigona. The highest Ni concentration in leaves of A. markgrafii was close to that of A. murale and was observed at Fushë-Arrëz: 19,145 mg kg-1, followed by Gjegjan-Gojan with a Ni concentration of 16,094 mg kg-1. Alyssum heldreichii was sampled at Katara Pass (11,819 mg kg-1) and Malakasi-Soil 1 (5463 mg kg-1), and its Ni concentrations were lower than those of A. murale from the same locations. Leaves of B. tymphaea showed Ni concentrations from 9641 mg kg-1 in Malakasi-Soil 1 to 13,806 mg kg-1 in Trigona. For B. baldacii, Ni concentration in leaves of the Albanian subspecies from Gramsh was 14,055 mg kg-1; Ni concentration in leaves of the Greek subspecies baldacii from Vovoussa was 10,956 mg kg-1. Leptoplax emarginata was also a substantial hyperaccumulator and exhibited its highest concentration of Ni in leaves at Trigona (13,674 mg kg-1) and its lowest at Katara Pass (5682 mg kg-1). It is also the tallest and greatest in biomass of all eight species. The ratio of Ni concentration in plant shoots to its concentration in soil varied from 1.0 for T. ochroleucum (Pojska-Soil 2) to 13.9 for A. markgrafii (Fushë Arrëz). This ratio represented the transfer factor since soil was sampled around the roots of plant species and was in a range reported for serpentines hyperaccumulator plants (Nanda Kumar et al. 1995, Reeves 2009 A. Bani, G. Echevarria, A. Mullaj, R. Reeves, J.L. Morel, and S. Sulçe 397 2002, Shallari et al. 1998). For the three species of Alyssum (highest ratios), the ratios reached 10.2 for A. heldreichii (Katara Pass), 6.3 for A. murale (Pojska-Soil 1), and 13.9 for A. markgrafii (Fushë Arrëz). This ratio can be interpreted as an indication of optimal hyperaccumulation conditions due to high availability of Ni in the soil of origin and high accumulation potential of the plant. Moreover, Ni accumulation in leaves of hyperaccumulators did not respond to Ni chemical availability in soils or to any other expected parameter (e.g., soil pH). There was, however, a significant correlation between Ni concentration in leaves of A. murale and total Ca contents in soils (Fig. 4). Calcium concentrations in leaves were extremely high despite low concentrations of this element in all soils: about 0.9–4.3% for A. murale. For A. markgrafii, it was also extremely high (2.5–3.2%), and for B. baldacii it was high too with 1.2% (Table 3). In general, Ca concentrations were higher for the Alyssum species and lower for the Bornmuellera, Leptoplax, and Thlaspi species, but with large variability and not enough samples for statistical comparisons. However, resulting Ca concentrations in leaves were not correlated with Ni concentrations. For Mg, the mean concentration for A. heldreichii was 0.65–0.71%, for A. murale 0.46–0.57%, for A. markgrafii 0.43–0.52%, and the highest values were found in B. baldacii subsp. baldacii (1.2%) and T. ochroleucum (0.9%). Hyperaccumulator plants showed a different uptake pattern in response to the high concentration of Mg and the low concentration of Ca in soils than the other non-accumulator species (Bani et al. 2007), especially the species from the genus Alyssum. All nine hyperaccumulators had a much lower Mg:Ca ratio in leaves than in the soil, with specimen sometimes showing a ratio lower than 0.1 (except for T. ochroleucum at Pojska-Soil 2 for which Mg:Ca ratio was close to 1, a value that is still much higher than that of the corresponding soil). All these plants Figure 4. Relationship between Ni concentration in leaves of A. murale from all locations in Albania and Greece and total concentration of Ca in soils. 398 Northeastern Naturalist Vol. 16, Special Issue 5 are therefore Ca accumulators. However, the Mg:Ca ratios were markedly lowest for the Alyssum species. No particular trend was observed concerning Fe and Mn accumulation by the nine taxa despite reported Mn accumulation in Alyssum leaves (Broadhurst et al. 2004). Zinc absorption was also very similar to what can be observed in crops. Cobalt concentration in shoot tissues varied from 8.2 to 90 mg kg-1, typical values of accumulating plants. Maximum concentrations were observed with A. murale collected at Pojska- Soil 1 (90 mg kg-1), at Prrenjas-Domosdova (70 mg kg-1), and at Pishkash (38 mg kg-1). Lower values for Co were recorded from A. murale in Prrenjas. Alyssum markgrafii at the Gjegjan site varied from 12 to 27 mg kg-1. Co concentration in plant tissues of H. ochroleucum varied from 8 to 15 mg kg-1 from Pishkash to Pojska. Bornmuellera baldacii subsp. baldacii (Vovoussa) was the highest Co accumulator in Greek sites with 42.7 mg kg-1. Discussion Soil characteristics and pedological evolution The serpentine soils of Albania and the Pindus Mountains (Greece) are ultramafic soils that contain elevated levels of metals, such as Mn, Ni, Cr, Co, and Fe. Surface horizons of the soils we collected were characterized by extremely high Mg:Ca ratios, which are toxic to unadapted plant species. In general, the Mg:Ca ratio observed in ultramafic soils is reported within the range from 2.5 to 47 (Proctor 1971). Mg:Ca ratios in the ultramafic soils in this study reached 30 in Albania and 17.7 in Greece, and were typical of serpentine soils (Reeves et al. 1997). Levels of Co in these soils were high and also typical of serpentine soils in the Mediterranean area (Reeves et al. 1997, Wenzel and Jockwer 1999). Besides the fact that all soils have a marked ultramafic origin, the presence of such typical elements (i.e., Mg, Fe, Ni, Cr) can also inform about the pedogenesis of the soil since their behavior during weathering is quite peculiar for most of them. Therefore, they are indicators of pedological processes (Massoura et al. 2006, Kierczak et al. 2007). The soils of the three areas have a similar geological background. Well-serpentinized harzburgite seems to dominate on most of the three areas. However, climate conditions are quite different, mostly because of the variation in average elevation of the sites. The sites in Albania (apart from Gramsh) are from relatively low elevation, whereas Greek sites are from moderate to high elevation. This aspect has a direct consequence in weathering intensity of the soils and therefore on their mineralogy, pH, nutrient fertility, and Ni availability. The analysis of Fe partitioning clearly shows some distinct stages of weathering among the five soils. Slightly weathered soils have almost all Fe included in primary silicate minerals. Moderately weathered soils have a much higher proportion of free Fe, but also have a large proportion of amorphous Fe oxides (soils from Katara Pass and Trigona). More intensely 2009 A. Bani, G. Echevarria, A. Mullaj, R. Reeves, J.L. Morel, and S. Sulçe 399 weathered soils have a higher proportion of free Fe which is also dominated by well-crystallized Fe oxides (Massoura et al. 2006); this is the case of two Albanian soils (Prrenjas-Rock and Pojska-Soil 2). Another possibility for well-evolved soils is the accumulation of secondary clay minerals (e.g., smectites) in Vertic soils (Hseu et al. 2007, Lee et al. 2003). These minerals then hold most of the Fe. The soil at Prrenjas-Domosdova is a clear example of such soils. According to the Fe distribution pattern, we can say that the Greek soils are less evolved Cambisols than the soils collected in Albania. The soils here do not show lateritic features, which would mean a complete weathering of primary minerals and accumulation of well-crystallized Fe oxides. Control of Ni availability in soils Although soil pH was quite high (neutral to slightly alkaline) and varied moderately among soils, Ni availability varied considerably and therefore could have been mainly influenced by pedogenesis and Ni-bearing minerals (Massoura et al. 2006). Nickel availability as assessed by DTPA extraction was variable among soils. The pH values of the Greek soils (6.3–7.0) in general were lower than Albanian serpentine soils (6.7–8.2) probably because of the higher elevation and rainfall intensity affecting soil processes. Interestingly, Ni chemical availability in these Greek soils was often higher than in Albanian soils, but with at least several exceptions. In this case, the direct effect of pH on Ni availability could be interpreted as the main factor. However, Ni partition between mineral phases plays the most important role in ultramafic soils (Massoura et al. 2006). In particular, the association of Ni with amorphous Fe oxides (slightly weathered soils) and with secondary high-exchange clays (Vertisols) are the main factors of Ni availability in ultramafic soils (Massoura et al. 2006). In our study, the Greek soils are also the least weathered soils of all in the point of view of Fe speciation. This factor is acting in the same way as pH, which is probably not the dominant factor. Nickel is clearly associated with unweathered clay minerals and amorphous Fe oxides in the Cambisols. It seems that the crystallization of Fe oxides during pedogenesis is a process that segregates Ni (divalent ion as opposed to FeIII), as Ni seldom enters significantly into goethite particles (Bani et al. 2007). The soil with the highest amount of Ni associated with the amorphous Fe oxides is unsurprisingly the soil with the highest chemical availability of Ni. For the rest of the soils, there is no clear link between the two parameters. Ni accumulation by plants Distribution in different areas of Ni hyperacumulators depends mainly on the biology of the species. Therefore, it is influenced by edaphic factors such as climate, soil pH, etc. Bornmuellera baldacii is a high-elevation species in both countries and only found at elevations higher than 1500 400 Northeastern Naturalist Vol. 16, Special Issue 5 m. Alyssum markgrafii is endemic to Albania and is only found in the North and the center of Albania. It appears from the work of Brooks et al. (1979) and some other studies that many of the Alyssum species in section Odontarrhena are both serpentine-endemic and invariably Ni-hyperaccumulators. This is the case for A. markgrafii in Albania, which strongly accumulates Ni despite being in the presence of limited available pools (DTPA-extractable Ni) in soils. Alyssum murale is also reported in Albania elsewhere than on serpentines, such as on calcareous soils (National Herbarium of Albania, Tirana). The taxa Bornmuellera baldacii subsp. markgrafii and Thlaspi ochroleucum in Albania contain high concentrations of Ni and, according to the definition of Brooks et al. (1977), are classified as Ni hyperaccumulators. We report Ni hyperaccumulation by Albanian specimens of T. ochroleucum in two different sites for the first time. Nickel concentrations in leaf tissues of the Ni hyperacumulator flora of Albania and the Pindus Mountains was high, but did not reflect the availability of Ni in soils as assessed by DTPA (no statistical relationship was found). It could also have depended on the abilities of different populations of plants. All plants collected in this study accumulate Ni at extremely high concentrations, except the two species of Thlaspi that did not reach 10,000 mg kg-1 on any of the sites. All species had an opposite Mg:Ca ratio in their leaves compared to that of their respective soils, because of the exceptional ability of Alyssum species to take up Ca, even from soils that have low total Ca concentrations. This point was made by Reeves et al. (1997), and reiterated by others (Lombini et al. 1998, Shallari et al. 1998). This remarkable Ca uptake ability is probably an important feature in Ni-hyperaccumulator physiology (Broadhurst et al. 2004, Chaney et al. 2008). Alyssum murale showed differences in uptake between the two countries and this could be due to genetic variation within this species, which is very widespread across the Balkans. The highly significant correlation we found suggests that the genetic difference among sites is not as important as the nutrient balance of the soils. Unlike what was found with Thlaspi caerulescens (Chardot et al. 2007), A. murale does not fully respond to chemical availability of Ni, but instead mainly to the Ca contents of soils. In particular, the role of Ca in Ni uptake probably needs more thorough attention in future ecophysiological research on Ni hyperaccumulators. Previous work (Kukier et al. 2004) showed that in Ni-rich non-ultramafic soils, increasing soil pH through liming increased Ni uptake by two species of Alyssum (including A. murale). However, the same authors showed a decrease of Ni uptake by these plants when liming an ultramafic soil. The Ca competition in such Ca-deficient soils, and probably the Mg:Ca ratio of the soil solution in these soils, must have a strong influence in Ni absorption, translocation, and hyperaccumulation. Our plants reached extremely high Ca concentrations in leaves (>4%), especially on such low-Ca soils. 2009 A. Bani, G. Echevarria, A. Mullaj, R. Reeves, J.L. Morel, and S. Sulçe 401 Consequences for phytoextraction Some soils (Trigona, Pojska-Soil 2, and Prrenjas-Domosdova) are highly suitable for profitable phytoextraction (e.g., phytomining), and it appears that these soils display two distinct situations that favor Ni availability: the moderately weathered soils, in which the available Ni pool is controlled by poorly crystallized Fe-oxides (e.g., Trigona), and the more evolved soils, with significant accumulation of Ni in secondary high-charge phyllosilicates (e.g., Prrenjas-Domosdova). These last soils are also Ca-deficient, which seems to enhance Ni accumulation in leaves. The maximum ratio of Ni concentration in plants/Ni concentration in soil varied from 13.9 for A. markgrafii to 6.3 for A. murale, 6.9 for T. tymphaeum, 5.0 for B. tymphaea, and 4.0 for L. emarginata. Alyssum markgrafii (Albania), T. tymphaeum (Greece), and B. tymphaea (Greece) are the most efficient Ni-hyperaccumulators among the species. It is clear that the endemic A. markgrafii and also some efficient populations of A. murale, L. emarginata, and B. tymphaea (Greece), displayed the best Ni-efficiency for use in phytomining. Acknowledgments The authors wish to thank the French Embassy in Tirana (Albania) for the doctoral scholarship of Aida Bani. Literature Cited AFNOR. 2004. Evaluation de la qualité des Sols. Volumes 1 and 2. AFNOR. Saint- Denis, France. 461 pp. and 486 pp. Anonymous. 1992. Advertisement supplement. Mining Journal (London) 318(8172):8. Bani, A., G. Echevarria, S. Sulçe, A. Mullaj, and J.L. Morel. 2007. In-situ phytoextraction of nickel by native population of Alyssum murale on ultramafic site in Albania. Plant and Soil 293:79–89. Bani, A., G. Echevarria, R. Reeves, S. Sulçe, A. Mullaj, and J.L. Morel. 2008. Nickel hyperaccumulation in the serpentine flora of Albania and Greece. Poster in Sixth International Conference on Serpentine Ecology. College of Atlantic. Bar Harbor, ME, USA. 15–20 June. Broadhurst, C.L., R.L. Chaney, J.S. Angle, T.K. Maugel, E.F. Erbe, and C.A. Murphy. 2004. Simultaneous hyperaccumulation of nickel, manganese, and calcium in Alyssum leaf trichomes. Environmental Science and Technology 38:5797–5802. Brooks, R.R., 1998. Plants that Hyperaccumulate Heavy Metals. CAB International, Wallingford, UK. Brooks, R.R., and C.C. Radrord. 1978. Nickel accumulation by European species of the genus Alyssum. Proceedings of the Royal Society of London B 200:217–224. Brooks, R.R., J. Lee, R.D. Reeves, and T. Jaffré. 1977. Detection of nickeliferous rocks by analysis of herbarium specimens of indicator plants. Journal of Geochemical Exploration 7:49–57. 402 Northeastern Naturalist Vol. 16, Special Issue 5 Brooks, R.R., R.S. Morrison, R.D. Reeves, T.R. Dudley, and Y. Akman. 1979. Hyperaccumulation of nickel by Alyssum Linnaeus (Cruciferae). Proceedings of the Royal Society of London B 203:387–403. Chaney, R.L., K.Y. Chen, Y.M. Li, J.S. Angle, and Baker, A.J.M. 2008. Effects of calcium on nickel tolerance and accumulation in Alyssum species and cabbage grown in nutrient solution. Plant and Soil 311:131–140. Chardot, V., S.T. Massoura, G. Echevarria, R.D. Reeves, and J.L. Morel. 2005. Phytoextraction potential of the nickel hyperaccumulators Leptoplax emarginata and Bornmuellera tymphaea. International Journal of Phytoremediation 7:323–335 Chardot, V., G. Echevarria, M. Gury, S.T. Massoura, and J.L. Morel. 2007. Nickel bioavailability in an ultramafic toposequence in the Vosges Mountains (France). Plant Soil 293:7–21. Echevarria, G., J.L. Morel, J.C. Fardeau, and E. Leclerc-Cessac. 1998. Assessment of phytoavailability of nickel in soils. Journal of Environmental Quality 27:1064–1070. Hseu, Z.Y., H. Tsai, H.C. His, and Y.C. Chen. 2007. Weathering sequences of clay minerals in soil along a serpentinic toposequence. Clay and Clay Minerals 55(4):389–401. Available online at DOI: 10.1346/CCMN.2007.0550407. 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. Kukier, U., C.A. Peters, R.L. Chaney, J.S. Angle, and R.J. Roseberg. 2004. The effect of pH on metal accumulation in two Alyssum species. Journal of Environmental Quality 33:2090–2102. Lee B.D., K.S. Sears, R.C. Graham, C. Amrhein, and H. Vali. 2003. Secondary mineral genesis from chlorite and serpentine in an ultramafic soil toposequence. Soil Science Society of America Journal 67:1309–1317 Lindsay, W.L., and W.A. Norvell. 1978. Development of DTPA soil test for zinc, iron, manganese, and copper. Soil Science Society of America Journal 42:421–428 Lombini, A., E. Dinelli C. Ferrari, and A., Simoni. 1998. Plant-soil relationships in the serpentinite screes of Mt Prinzera (Northern Apennines, Italy). Journal of Geochemical Exploration 64:19–33. Massoura, S.T., G. Echevarria, E. Leclerc-Cessac, and J.L. Morel. 2004. Response of excluder, indicator, and hyperaccumulator plants to nickel availability in soils. Australian Journal of Soil Research 42:933–938. Massoura, S.T., G. Echevarria, T. Becquer, J. Ghanbaja, E. Leclerc- Cessac, and J.L. Morel. 2006. Nickel-bearing phases and availability in natural and anthropogenic soils. Geoderma 136:28–37. McKeague, J.J., and J.A. Day. 1966. Dithionite and oxalate extractable Fe and Al as aids in differentiating different classes of soils. Canadian Journal of Soil Science 46:13–22. Mehra, O.P., and M.L. Jackson. 1960. Iron oxides removal from soils and clays by dithionite-citrate systems buffered with sodium bicarbonate. Pp. 317–342, In A. Swineford (Ed.). Proccedings of the 7th National Conference on Clay Minerals. Pergamon Press, Elmsdorf, NY, USA. Minguzzi, C., and O. Vergnano. 1948. Il contenuto di nichel nelle ceneri d’Alyssum bertolonii Desv. Atti della Societa Toscana di Scienze Naturali, Memorie Serie A 55:49–77. 2009 A. Bani, G. Echevarria, A. Mullaj, R. Reeves, J.L. Morel, and S. Sulçe 403 Nanda Kumar, P.B.A., H. Dushenkov, H. Motto, and I. Raskin. 1995. Phytoextraction: The use of plants to remove heavy metals from soils. Environmental Science and Technology 29(5):1232–1238. Proctor, J. 1971. The plant ecology of serpentine. III. The influence of a high magnesium/ calcium ratio and high nickel and chromium levels in some British and Swedish serpentine soils. Journal of Ecology 59:827–842. Psaras, G.K., Th. Constantinidis, B. Cotsopoulos, and Y. Manetas. 2000. Relative abundance of nickel in the leaf epidermis of eight hyperaccumulators: Evidence that the metal is excluded from both guard cells and trichomes. Annals of Botany 86:73–78. Revees, R.D. 2002 .Hyperaccumulation of trace elements by plants. Pp. 29, In J.L. Morel, G. Echevarria, and N. Goncharova (Eds.). Phytoremediation of Metal- Contaminated Soils, NATO Science Series (IV): Earth and Environmental sciences Volume 68. Reeves, R.D. 1992. The hyperaccumulation of nickel by serpentine plants. PP. 253–277, In A.J.M. Baker, J. Proctor, and R.D. Reeves (Eds.). The Vegetation of Ultramafic (Serpentine) Soils. Intercept Ltd., Andover, UK. Reeves R.D., R.R. Brooks, and T.R. Dudley. 1983. Uptake of nickel by species of Alyssum, Bommuellera, and other genera of Old World tribus Alysseae. Taxon 32:184–192 Reeves, R.D., A.J.M. Baker, A. Borhidi, and R. Berazain. 1996. Nickel-accumulating plants from the ancient serpentine soils of Cuba. New Phytologist 133:217–224. Reeves, R.D., A.J.M. Baker, and A. Kelepertsis. 1997. The distribution and biogeochemistry of some serpentine plants of Greece. Pp. 205–207, In T. Jaffre, R.D. Reeves, and T. Becquer (Eds.). Ecologie des Milieux sur Roches Ultramafiques et sur Sols Metalliferes, ORSTOM, Noumea, Documents Scientifiques et Techniques No. III/2 Reeves, R.D., A.J.M. Baker, A. Borhidi, and R. Berazain. 1999. Nickel hyperaccumulation in the serpentine flora of Cuba. Annals of Botany 83:29–38. Shallari, S., C. Schwartz, A. Hasko, and J.L. Morel. 1998. Heavy metals in soils and plants of serpentine and industrial sites of Albania. The Science of the Total Environment 209:133–142. 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. Tan, K., A. Mullaj, B. Ruci, and G. Iatrou 1999. Serpentine endemics in Albania and NW Greece. Abstacts of the Vth Conference on Plant Taxonomy. Lisboa 1999 s.68. A. Correia et al. (Eds.). Museu, Laboratorio, e Jardim Botanico da Universidade de Lisboa, Lisboa, Portugal. Tutin, T.G., V.H. Heywood, N.A. Burges, D.H. Valentine, S.M. Walters, and D.A. Webb. 1964. Flora Europaea, Vol. 1. Cambridge University Press, Cambridge, UK. Tutin, T.G., V.H. Heywood, N.A. Burges, D.M. Moore, D.H. Valentine, S.M. Walters, and D.A. Webb. 1968. Flora Europaea, Vol. 2. Cambridge University Press, Cambridge, UK. Tutin, T.G., V.H. Heywood, N.A. Burges, D.M. Moore, D.H. Valentine, S.M. Walters, and D.A. Webb. 1972. Flora Europaea, Vol. 3. Cambridge University Press, Cambridge, UK. 404 Northeastern Naturalist Vol. 16, Special Issue 5 Tutin, T.G., V.H. Heywood, N.A. Burges, D.M. Moore, D.H. Valentine, S.M. Walters, and D.A. Webb. 1976. Flora Europaea, Vol. 4. Cambridge University Press, Cambridge, UK. Tutin, T.G., V.H. Heywood, N.A. Burges, D.M. Moore, D.H. Valentine, S.M. Walters, and D.A. Webb. 1980. Flora Europaea, Vol. 5. Cambridge University Press, Cambridge, UK. Wenzel, W.W., and F. Jockwer. 1999. Accumulation of heavy metals in plants grown on mineralized soils of the Austrian Alps. Environmental Pollution 104:145-155.