MTT is publishing its research findings in two series of publications: MTT Science and MTT Growth. The MTT Science series includes scientific presentations and abstracts from conferences arranged by MTT Agrifood Research Finland. Doctoral dissertations by MTT research scientists will also be published in this series. The topics range from agricultural and food research to environmental research in the field of agriculture. MTT, FI-31600 Jokioinen, Finland. Tel. +358 29 5300 700, email julkaisut@mtt.fi 20 Selenium fertilization: plant uptake and residuals in soil Doctoral Dissertation Riikka Keskinen MTT CREATES VITALITY THROUGH SCIENCE www.mtt.fi/julkaisut Academic Dissertation: To be presented, with the permission of the Faculty of Agriculture and Forestry of the University of Helsinki, for public criticism in Arppeanum, Snellmaninkatu 3, Helsinki, on November 23rd, 2012, at 12 o’clock. 20 Selenium fertilization: plant uptake and residuals in soil Doctoral Dissertation Riikka Keskinen Supervisors: Professor Helinä Hartikainen University of Helsinki, Finland Professor Markku Yli-Halla University of Helsinki, Finland Doctor Päivi Ekholm University of Helsinki, Finland Pre-examiners: Professor Trine Sogn The Norwegian University of Life Sciences, Norway Doctor Graham Lyons University of Adelaide, South Australia Opponent: Professor Phil Haygarth Lancaster University, United Kingdom Custos: Professor Helinä Hartikainen University of Helsinki, Finland ISBN 978-952-487-408-3 (Print) ISBN 978-952-487-409-0 (Electronic) ISSN 1798-1824 (Printed version) ISSN 1798-1840 (Electronic version) www.mtt.fi/mtttiede/pdf/mtttiede20.pdf Copyright MTT Agrifood Research Finland Riikka Keskinen Distribution and sale MTT Agrifood Research Finland, Media and Information Services, FI-31600 Jokioinen, phone +358 29 5300 700, e-mail julkaisut@mtt.fi Printing year 2012 Cover photo Riikka Keskinen Printing house Tampereen Yliopistopaino Juvenes Print Oy MTT SCIENCE 20 3 Selenium fertilization: plant uptake and residuals in soil Riikka Keskinen The acidity and low redox potential in soils in Finland render selenium (Se) scarce- ly soluble, thereby severely restricting its availability to plants. To achieve an ade- quate intake level of this microelement es- sential for human beings and animals, a national Se fertilization programme has been in operation in Finland since the 1985 growing season. Nutritionally, sup- plementing compound fertilizers with sodium selenate has been a success. The practice is, however, challenged by the un- known cycle of annual Se additions in the environment. In this study, several soil ex- tractants were explored to find methods of sound basis for routine soil Se analyses and for a more detailed characterization of the soil Se status. To increase understanding of the behaviour of fertilizer Se in soil, both long- and short-term changes in soil Se re- serves in response to Se additions were ex- amined. The efficiency of plant uptake of fertilizer Se was also studied, emphasizing the influence of Se-adsorbing oxides and organic components on the retention of Se in soil. In addition, the uptake rhythm and translocation of Se within plants were determined. Sequential extractions of selenate-fertilized organic and mineral soils with varying Se content revealed a characteristic pattern in the distribution of Se. A mere 1% of the soil Se reserves were extracted with a KCl solution, nearly 20% were recovered in adsorbed, 40% in organically associat- ed, 15% in elemental and around 30% in recalcitrant fractions. The change in soil Se concentrations after a 13-year period of Se fertilization remained small. However, the data indicated that in mineral soils residual fertilizer Se had accumulated as adsorbed selenite and in the organically associated and recalcitrant pools. The minor fraction of salt-soluble Se reflected the instability of selenate in soils in Finland. Pot experi- ments, however, demonstrated that under moisture conditions drier than field capac- ity, easily soluble selenate may persist in acidic soils throughout the growing period. Since Se accumulates in poorly soluble pools, very weak extraction methods are not suitable for describing changes in soil Se reserves. Of the single extractions test- ed, acid ammonium oxalate or phosphate buffer with high pH appeared to be more suitable extractants for estimating the soil Se status than the hot-water extraction used in national monitoring at present. However, salt solutions, which were less efficient than hot water in extracting Se, recovered more Se than what was found within plants. These weak extractants can thus be used in assessing the immediately bioavailable soil Se pool, although the as- sociation between the amounts of Se ac- quired by soil analyses and that absorbed by plants needs to be further clarified. The Se uptake of wheat and ryegrass ranged between 5% and 50% of the amount of fertilizer selenate added in pot experiments. The Se uptake by wheat con- MTT Agrifood Research Finland, Plant Production Research, FI-31600 Jokioinen, riikka.keskinen@mtt.fi 4 MTT SCIENCE 20 tinued throughout the growth period and wheat transported over 50% of its Se con- tent into the grains. Ryegrass grown in soil with residual Se, however, accumu- lated over 80% of its total Se uptake in the roots. The translocation pattern with- in the plant can thus have a marked in- fluence on the apparent uptake efficien- cy of Se when only the harvested portion of the crop is considered. Since reduction of selenate to selenite was low during the growth period, the content of Se sorption components in the soil had minor influ- ence on the plant availability of Se. As sel- enite, Se was retained very efficiently in both peat and in peat enriched with iron hydroxides. On artificial mineral surfaces, Se was adsorbed by ligand exchange, but in pure peat an unknown mechanism of retention was operating. This study revealed that new approach- es are needed in Se research to unravel the nature of the unfamiliar forms of or- ganically bound Se that dominate Se re- serves in soil. The results of the pot exper- iments suggest that processes other than rapid reduction and subsequent fixation of the added selenate govern the uptake of the annual fertilizer Se addition by plants. Factors limiting plant Se uptake from the soluble pool in soil need to be addressed in future studies. Keywords: selenium, selenate, selenite, speciation, fertilization, soil extraction, adsorption, nutrient uptake MTT SCIENCE 20 5 Seleenilannoitus: otto kasviin ja jäännös maassa Riikka Keskinen MTT (Maa- ja elintarviketalouden tutkimuskeskus), Kasvintuotannon tutkimus, 31600 Jokioinen, riikka.keskinen@mtt.fi Maaperän happamuuden ja matalan ha- petus-pelkistyspotentiaalin johdosta ihmi- sille ja eläimille välttämätön hivenravinne seleeni (Se) esiintyy suomalaisissa mais- sa pääosin kasveille käyttökelvottomissa, niukkaliukoisissa muodoissa. Suomalaisten riittävä seleeninsaanti onkin varmistettu vuodesta 1985 lähtien kansallisella Se-lan- noitusohjelmalla. Ravitsemuksellisesti nat- riumselenaatin lisääminen lannoitteisiin on ollut menestyksekästä. Vuosittain toistuvi- en Se-lisäysten kiertoa ympäristössä ei kui- tenkaan tunneta tarkasti, mikä varjostaa lannoituskäytäntöä. Tässä tutkimuksessa tarkasteltiin erilaisia maan uuttomenetel- miä pyrkimyksenä löytää rutiinianalyysei- hin toiminnallisesti perusteltu, yksinker- tainen uuttotapa sekä maan Se-varannon yksityiskohtaisempaan tutkimukseen sopi- va menetelmä. Lannoiteseleenin maaperä- käyttäytymisen selvittämiseksi työssä tar- kasteltiin Se-lisäysten pitkä- ja lyhytaikaisia vaikutuksia maan Se-varantoon. Kasvien lannoiteseleenin ottotehokkuutta tutkittiin painottaen Se-pidättävien oksidien ja or- gaanisen aineksen merkitystä lisätyn Se:n käyttökelpoisuudelle. Lisäksi tarkasteltiin kasvien Se:n oton rytmiä ja Se:n kuljetus- ta kasvissa. Kokonaisseleenipitoisuudeltaan toisis- taan eroavilla Se-lannoitetuilla orgaani- silla ja kivennäismailla tehdyt perättäiset uutot osoittivat maan Se-varannolle tyy- pillisen jakauman. Vain 1 % maan Se:stä uuttui helppoliukoisena KCl-liuoksella, lä- hes 20 % oli adsorboituneena, 40 % or- gaaniseen ainekseen sitoutuneena, 15 % alkuainemuodossa ja noin 30 % niukka- liukoisena. Seleenilannoituksen aikaan- saamat muutokset maan Se-pitoisuuksissa 13 vuoden tarkastelujaksolla olivat pieniä, mutta kivennäismaalla Se:n havaittiin ker- tyneen adsorboituneeseen, orgaaniseen sekä niukkaliukoiseen fraktioon. Help- poliukoisen Se:n pieni osuus kuvastaa se- lenaatin pysymättömyyttä suomalaisissa maissa. Astiakokeissa havaittiin kuiten- kin, että kenttäkapasiteettia kuivemmis- sa olosuhteissa selenaatti voi säilyä maassa läpi kasvukauden. Koska Se kertyy maassa vaikealiukoisiin muotoihin, heikot uuttomenetelmät eivät sovellu maaperän Se-varannon muutos- ten seurantaan. Työssä testatuista yksit- täisuutoista hapan ammoniumoksalaatti ja fosfaattipuskuri korkeassa pH:ssa osoit- tautuivatkin Se-seurannassa käytettävää kuumavesiuuttoa lupaavammiksi menetel- miksi maan Se-varannon arviointiin. Te- hokkuudeltaan kuumaa vettä heikommal- la suolaliuoksella maasta uuttui kuitenkin huomattavasti kasvien ottamaa määrää enemmän Se:ä. Nämä heikot uuttomene- telmät ovat siten käyttökelpoisia välittö- mästi biosaatavissa olevan Se:n tarkaste- luun, joskin maauuttojen ja kasvien Se:n ottotehokkuuden välistä yhteyttä tulee sel- vittää tarkemmin. Vehnä- ja raiheinäkasvustot ottivat 5-50 % astiakokeissa lisätystä lannoiteselenaatista. Vehnän seleeninotto jatkui läpi kasvukau- den. Yli 50 % vehnäkasvuston ottamas- ta Se:stä päätyi jyviin, kun taas edellisen vuoden selenaattilannoituksen jäännök- sellä kasvaneen raiheinän ottamasta Se:stä 80 % kertyi juuriin. Seleenin kuljetuksel- la kasvissa voi siten olla merkittävä vaiku- 6 MTT SCIENCE 20 tus laskennalliseen Se:n ottotehokkuuteen, jos vain sadoksi korjattu osa kasvista tu- lee huomioiduksi. Astiakokeissa selenaatin pelkistyminen seleniitiksi oli vähäistä, jo- ten maan Se:ä pidättävillä komponenteilla ei ollut juuri vaikutusta Se:n käyttökelpoi- suuteen. Seleniittinä Se kuitenkin pidät- tyi tehokkaasti sekä puhtaaseen että rau- taoksideilla rikastettuun rahkaturpeeseen. Keinotekoisilla mineraalipinnoilla Se:n pi- dättyminen tapahtui ligandinvaihtomeka- nismilla, mutta puhtaassa turpeessa pidät- tymistapa jäi tuntemattomaksi. Tämä työ osoitti, että tulevaisuuden Se- tutkimuksessa tarvitaan uusia menetelmiä toistaiseksi tuntemattomien, mutta maassa hallitseviksi osoittautuneiden orgaanisten Se-muotojen tarkasteluun. Astiakokeiden tulosten perusteella muut tekijät kuin se- lenaatin nopea pelkistyminen ja sitä seu- raava pidättyminen vaikuttavat määräävän kasvien lannoite-Se:n ottotehokkuuden. Maan helppoliukoisen Se:n kasvien ottoa rajoittavia tekijöitä onkin jatkossa syytä selvittää. Avainsanat: seleeni, selenaatti, seleniitti, esiintymismuoto, lannoitus, maauutto, adsorptio, ravinteiden otto MTT SCIENCE 20 7 My work with selenium began in the pre- sent Department of Food and Environ- mental Sciences of the University of Hel- sinki in 2006, when I joined a project called Sorption and Bioavailability of Se- lenium in Soils of Finland, led by Professor Helinä Hartikainen. The three following years, mostly spent in the AAS laborato- ry and experimental greenhouses, were busy but rewarding times, and I feel priv- ileged to have shared them with such an enthusiastic team. I am grateful to Profes- sor Helinä Hartikainen, Professor Mark- ku Yli-Halla and Dr. Päivi Ekholm for the dedicated supervision of the work through to completion. I wish to thank Dr. Marja Turakainen for excellent cooperation and friendship over the years. For strong tech- nical support, I am grateful to Maija Yli- nen and Miia Collander. I sincerely thank Laura Pulli for the hard work with peat soils and all my coworkers in the Division of Environmental Soil Science for help and encouragement. In early 2009, I left Helsinki for a research position at MTT Agrifood Research Fin- land. The change might have been fatal for this thesis had Professor Martti Esala not valued the work and enabled its continu- ation. I owe my warmest thanks for that support. At the Soil Division of MTT, I found a lively work community filled with intriguing discussions on various aspects of science and everyday life, which have greatly lightened the long days of writing. I thank the whole Soil Team for that. I wish to express my special thanks to Jaak- ko Heikkinen for graphical assistance and to Johanna Nikama for being a heartfelt friend and an indispensable adviser from my first day at MTT onwards. I doubt this work would have ever been completed without the continuous sup- port and advice from my three muske- teers: Helena Soinne, Mari Räty and Kim- mo Rasa. I can never thank you enough for being there in the times of trouble and frustration, as well as in the times of laugh- ter and success. I am especially grateful to Mari Räty for participating in the soil Se fractionation work with a degree of com- mitment and endurance that I can only admire. I wish to thank the preexaminers of my thesis, Professor Trine Sogn and Dr. Gra- ham Lyons, for their valuable comments. For the funding of this research, I am grateful to the Finnish Ministry of Agri- culture and Forestry, August Johannes and Aino Tiura’s Agricultural Research Foun- dation, the Finnish Cultural Foundation, Suoviljelysyhdistys ry and the University of Helsinki Funds. Finally, I wish to thank my father Raimo Keskinen for the encouragement to study. I thank my sister Maria and her family Kimmo, Santeri and Niilo Palo for the true friendship which has carried me through many difficulties. Foremost, I am grateful to my mother Arja Virkki for the endless support regardless of the significance of the matter I have tried to accomplish. My pack of four-legged friends, who ultimately keep me going, would only tear any writ- ten acknowledgments to shreds, but will certainly appreciate the increased free time we can now spend roaming in the woods. Jokioinen, September 2012 Riikka Keskinen Acknowledgements 8 MTT SCIENCE 20 List of original publications This thesis is based on the following publications, which are referred to by their Roman numerals: I Efficiency of different methods in extracting selenium from agricultural soils of Finland Keskinen R, Ekholm P, Yli-Halla M and Hartikainen H Geoderma 153 (2009) 87-93 II Selenium fractions in selenate-fertilized field soils of Finland Keskinen R, Räty M and Yli-Halla M Nutrient Cycling in Agroecosystems 91 (2011) 17-29 III Plant availability of soil selenate additions and selenium distribution within wheat and ryegrass Keskinen R, Turakainen M and Hartikainen H Plant and Soil 333 (2010) 301-313 IV Retention and uptake by plants of added selenium in peat soils Keskinen R, Yli-Halla M and Hartikainen H Manuscript. Author’s contribution to the articles I The experimental design, including the selection of soils and extraction methods to be tested, was compiled jointly by Helinä Hartikainen, Markku Yli-Halla and Riikka Kes- kinen. The laboratory work was conducted by Riikka Keskinen under the supervision of Päivi Ekholm. Riikka Keskinen calculated and interpreted the data and prepared the manuscript. All authors commented on the manuscript. II The experimental design was built on the idea of Helinä Hartikainen using a set of paired soil samples originally accomplished by Markku Yli-Halla. All authors contributed to the planning of the study. Riikka Keskinen conducted the laboratory work together with Mari Räty. The calculations, interpretation of the data and preparation of the manuscript was done by Riikka Keskinen. All authors commented on the manuscript. III All authors contributed to the planning of the experimental design. The pot experiments and laboratory analyses were carried out by Riikka Keskinen and Marja Turakainen to- gether with technical assistance. Riikka Keskinen calculated and interpreted the data and prepared the manuscript. All authors commented on the manuscript. IV The experimental design was planned by Markku Yli-Halla and Riikka Keskinen. The pot experiments and laboratory analyses were carried out by Riikka Keskinen together with technical assistance. Riikka Keskinen calculated and interpreted the data and prepared the manuscript. All authors commented on the manuscript. Reprints of the original articles are published with the kind permission of Elsevier B.V. (I) and Springer Science+Business Media B.V. (II, III). MTT SCIENCE 20 9 Abbreviations AAS atomic absorption spectrometer APDC ammonium pyrrolidine dithiocarbamate AR aqua regia DW dry weight EDTA ethylenedinitrotetraacetate MIBK methyl isobutyl ketone NMD nutritional muscular degeneration SeCys selenocysteine SeMet selenomethionine SEP sequential extraction procedure 10 MTT SCIENCE 20 Contents 1 Introduction....................................................................................................11 1.1 Selenium in soil........................................................................................... 11 1.1.1 Selenium species and their characteristics........................................ 12 1.1.2 Extraction of soil selenium.............................................................. 13 1.2 Selenium uptake by plants........................................................................... 14 1.3 Selenium fertilization................................................................................... 15 1.4 Aims of the present study............................................................................. 17 2 Material and methods......................................................................................18 2.1 Experimental designs................................................................................... 18 2.1.1 Soils used in the studies................................................................... 18 2.1.2 Examination of selenium in field soil samples.................................. 19 2.1.3 Pot experiments............................................................................... 20 2.2 Soil selenium extractions.............................................................................. 23 2.3 Analyses of selenium.................................................................................... 25 2.3.1 Soil extracts..................................................................................... 25 2.3.2 Plant samples................................................................................... 26 2.3.3 Quality control................................................................................ 26 3 Results and discussion.....................................................................................27 3.1 Extraction methods for assessing the selenium status in soils of Finland....... 27 3.1.1 Comparison of methods.................................................................. 27 3.1.2 Selenium fractions of various field soils............................................ 29 3.1.3 Estimating plant-available selenium................................................. 30 3.2 Behaviour of added selenium in soil............................................................. 31 3.2.1 Monitoring short-term sorption...................................................... 31 3.2.2 Fertilization-induced long-term changes in soil selenium fractions.. 33 3.3 Efficiency of plant selenium uptake.............................................................. 35 3.3.1 Selenium uptake rhythm and distribution of selenium within the plant............................................................................... 35 3.3.2 Selenium uptake in various soils...................................................... 36 4 Conclusions.....................................................................................................37 5 References........................................................................................................38 MTT SCIENCE 20 11 1 Introduction Since its discovery in 1817, selenium (Se) was considered merely a toxic element causing a severe poisoning called alkali disease or blind staggers in livestock (Mox- on and Rhian 1943). The first evidence of its essentiality was presented as late as the 1950s when Schwarz and Foltz (1957) found that Se is an integral part of Factor 3, an agent protective against necrotic de- generations. Currently, dozens of seleno- cysteine (SeCys)- containing proteins, i.e. selenoproteins, have been identified (Lopez Heras et al. 2011). Their vital functions in antioxidant protection, energy metabo- lism, redox regulation and gene expression, providing protection against cancer, heart diseases, muscle disorders and weakening of the immune system, are being increas- ingly studied (Combs 2001, Lopez Heras et al. 2011). Dietary recommendations show that an intake of 40–80 µg Se d-1 is adequate for healthy adults, but it is argued that a higher level of 200–300 µg Se d-1 would be more appropriate for the main- tenance of health (Combs 2001, Schrauzer and Surai 2009). Combs (2001) estimated that hundreds of millions of people around the world are deficient in Se. In Finland, cases of nutritional muscular degeneration (NMD) in cattle are known to have existed since the early 1900s, but not until studies initiated by exceptionally high incidences of the disease in the 1950s was it shown that the disease is caused by deficiency of Se in the forage (Oksanen 1965, Oksanen and Sandholm 1970). The low concentrations of Se in crops were soon associated with low amounts and low solubility of Se in soil (Koljonen 1974, 1975, Sippola 1979). Concern about the ef- fects of low Se on public health increased after a survey of mineral elements in Finn- ish foods revealed that the average diet supplied merely 30 µg Se d-1 (Varo and Koivistoinen 1980). Yläranta (1985) con- ducted extensive studies on increasing the Se content of crops and reported that small additions of selenate (SeO4 2-)-Se were effi- cient at elevating the Se concentration of cereal and grass crops. In 1983, the Minis- try of Agriculture and Forestry set a work- ing group to draft a proposal for supple- menting compound fertilizers with Se and to develop a follow-up plan for monitoring fertilization-induced changes in soil, crops, feeds and foods and in the level of Se in- take (Ministry of Agriculture and Forestry 1984). In early July 1984, national Se fer- tilization practice was initiated by a deci- sion of the Council of State. After over 20 years of Se fertilization ac- companied by versatile monitoring, un- answered questions remain on the envi- ronmental aspects of the practice. The unknown fate of the residual Se in soil and considerable variation in the Se con- centrations of crops have puzzled research- ers (Yläranta 1990, Eurola et al. 2004, 2008). However, complexity of the Se spe- cies, trace amounts of additions in relation to the heterogeneity of soils and various interactions contributing to plant uptake continue to challenge the studies. In this thesis, the behaviour of the added selenate in soil is examined by characterizing the Se reserves and their changes in soils, us- ing several extraction methods. In addi- tion, the efficiency of Se fertilization is dis- cussed, considering the Se uptake rhythm, translocation within plants and the influ- ence of components retaining Se in soil. 1.1 Selenium in soil Se is very unevenly distributed in the soils of the earth’s surface (Oldfield 2002). Usu- ally, soil Se concentrations range between 0.1 and 2.0 mg kg-1 but in seleniferous ar- eas concentrations exceeding 1000 mg kg-1 are not uncommon, whereas low-Se soils 12 MTT SCIENCE 20 contain less than 0.04 mg Se kg-1 (Girling 1984, Oldfield 2002). In general, soils de- veloped from igneous rock are low in Se, while soils originating from sedimenta- ry rocks have higher Se contents (Girling 1984, Haygarth 1994). Studies by Koljonen (1973a-c, 1974, 1975) showed that the young soils of Finland re- flect the low content of Se in the underly- ing Precambrian bedrock. The lowest to- tal Se concentrations are found in coarse mineral soils containing predominantly quartz and feldspars and having a low ad- sorption capacity, whereas the highest con- centrations are found in clays, even though in Finland they are rich in illite with lim- ited adsorption capacity (Koljonen 1975). Surveys preceding Se fertilization reported average total Se concentrations for coarse mineral soils of < 0.010 (Koljonen 1974), 0.142 (Sippola 1979) and 0.172 mg kg-1 (Yläranta 1983a). For clay soils the corre- sponding values were 0.320 mg kg-1 (Kol- jonen 1974), 0.329 mg kg-1 (Sippola 1979) and 0.290 mg kg-1 (Yläranta 1983a). As for organic soils, Koljonen (1974) found an av- erage Se concentration of 0.180 mg kg-1 in mull soils and 0.120 mg kg-1 in peat soils, while noting that peats formed on lake shores were strikingly richer in Se than those not having received Se from the wa- ter flow. In Sippola’s data (1979), mull soils contained on average of 0.228 mg Se kg-1 and peat soils 0.093 mg Se kg-1. Yläranta (1983a) reported a somewhat higher av- erage for organogenic soils, 0.464 mg Se kg-1, but his data included an exception- al soil having Se concentrations as high as 1.281 mg kg-1. In a more recent sur- vey, Mäkelä-Kurtto et al. (2007) found a range of total Se concentrations between 0.03 and 5.4 mg kg-1 in cultivated soils of Finland, the average being 0.29 mg Se kg-1. Yläranta (1983a) showed that the con- tents of clay, organic carbon and alumin- ium oxides explained nearly 60% of the variation in total Se of mineral soils in Finland. Knowledge of the total soil Se content is, however, of limited value in bi- ological assessments, since the figure em- bodies Se fractions differing decisively in their bioaccessibility. 1.1.1 Selenium species and their characteristics Being a redox-sensitive element, Se occurs at four oxidation states in soils. Depend- ing on the pH and electron activity (pe) of the environment the predominant species can be fully oxidized selenate (VI), selenite (SeO3 2-) (IV), elemental Se (Se0) (0) or re- duced selenide (Se2-) (-II) (Geering et al., 1968, Elrashidi et al. 1987). In addition, Se exists in association with organic matter, either complexed or built into the molec- ular structures (Gissel-Nielsen et al. 1984, Abrams and Burau 1989). The different Se species have very different characteris- tics in terms of mobility and bioavailabil- ity, wherefore knowledge of the speciation is essential in environmental and nutri- tional contexts. Selenate is the anion of the strong selenic acid (H2SeO4), which occurs in solution either as biselenate (HSeO4 -) or SeO4 2-, the pKa2 for the dissociation of the proton being 1.9. Selenate exists only under high- redox conditions and becomes the pre- dominant species when pe + pH exceeds 15.0 (Elrashidi et al. 1987). In its behav- iour, selenate resembles sulphate (SO4 2-); it is weakly bound and thus relatively mo- bile in soil (e.g. Alemi et al. 1988). Its ad- sorption on oxide surfaces is influenced by the properties of the adsorbing surface, pH and ionic strength of the soil solution, but seems to occur mainly via an outer- sphere surface complexation mechanism, although inner-sphere complex formation can occur, especially at low pH (Peak and Sparks 2002, Peak 2006). In soil, the high solubility makes selenate the most available species to plants (Gissel-Nielsen and Bis- bjerg 1970, Eich-Greatorex et al. 2007), but on the other hand exposes it to leach- ing (Dhillon et al. 2008). In the labora- tory experiments of Yläranta (1982), se- MTT SCIENCE 20 13 lenate leached readily through columns of Carex peat, but in mineral soils its trans- port was slow. Selenite is the anion of the weak selenous acid (H2SeO3) and, depending on the soil pH, it occurs as H2SeO3, biselenite (HSeO3 -) or SeO3 2-. The pKa for dissociation of the first proton is 2.64 and that for the second 8.4. Selenite is the major species at medi- um redox range between pe + pH 7.5 and 15 (Elrashidi et al. 1987). It has a high af- finity for soil surfaces and its sorption onto minerals and oxides has been studied very extensively (e.g. Hingston et al. 1974, Rājan and Watkinson 1976, Parfitt 1978, Rajan 1979, Balistrieri and Chao 1987, Saha et al. 2004). The bonding occurs either via biden- tate, binuclear or mononuclear inner-sphere complexation mechanisms or through out- er-sphere complexation, depending on the environment and the structure of the sur- face (Peak et al. 2006 and references there- in). Mobility of selenite is enhanced by al- kaline pH, high total selenite concentration and competition with more strongly adsorb- ing anions such as phosphate (PO4 3-), arse- nate (AsO4 3-) or bicarbonate (HCO3 -) (Bal- istrieri and Chao 1987). Elemental Se occurs in low-redox soils as red crystalline or red or black amorphous granules that are scarcely soluble (Elrashidi et al. 1987, Haygarth 1994). Selenide like- wise precipitates in reduced soil, forming in- soluble metal Se2- compounds (Elrashidi et al. 1987). However, Se2--Se also exists in or- gano-Se compounds produced via biological assimilation (Sors et al. 2005). In addition to the Se biotically incorporated into the or- ganic structures, organically bound Se may occur as metal-humic complexes, Se0 col- loid-colloid associations and adsorbed onto oxides or clays fixed within the organic ma- trix (Bruggeman et al. 2007, Coppin et al. 2009). The behaviour of the organically as- sociated Se can be assumed to vary accord- ing to the characteristics of the carrier com- pound. Coppin et al. (2006) found that Se associated with humic substances is linked rather with the fulvic acid than the humic acid fraction. Overall, organically associat- ed Se has thus far been poorly characterized, even though it comprises a marked portion of the soil Se reserves (e.g. Gustafsson and Johnsson 1992, Coppin et al. 2006, Chris- tophersen et al. 2012). In soil, all the above-mentioned Se species can coexist, due to heterogenic micro-en- vironments and relatively slow transfor- mation rates from one species to another (Zawislanski and Zavarin 1996, Darche- ville et al. 2008). Changes in the oxidation state are predominantly biologically medi- ated via bacterial oxidation of the reduced species into selenate or, in contrast, assimi- latory reduction of the oxidized forms into organo-selenide compounds by plants and microorganisms or dissimilatory reduction of selenate via selenite into elemental Se by Se-respiring bacteria (Kulp and Pratt 2004, Stolz et al. 2006). However, abiological transformations through oxidative weath- ering and suboxic reduction occurring in the presence of reactive surfaces are also in- volved in the cycling of Se in the environ- ment (Myneni et al. 1997, Chen et al. 2009). 1.1.2 Extraction of soil selenium The distinct differences in solubility of the various Se species have enabled development of numerous single (e.g. Sharmasarkar and Vance 1995, Hagarová et al. 2005) and se- quential extraction procedures (SEPs) (e.g. Chao and Sanzolone 1989, Tokunaga et al. 1991, Séby et al. 1997, Martens and Suarez 1997a, Mao and Xing 1999, Wright et al. 2003) targeting specific operationally de- fined soil and sediment Se fractions. The most easily soluble forms, namely selenate and dissolved organically associated Se, are commonly extracted with water or a sim- ple salt solution, based on anion exchange and mass action (Wright et al. 2003). Sele- nite adsorbed onto soil surfaces is often ac- quired through ligand exchange (Hingston et al. 1967) by replacing selenite with more strongly binding phosphate anions (Rājan 14 MTT SCIENCE 20 and Watkinson 1976). The adsorbed forms of Se may also be attacked more aggressive- ly by dissolving the adsorbing oxide surfac- es, e.g. with acid ammonium oxalate (Jack- son et al. 1986) or concentrated HCl (Chao and Sanzolone 1989). Organically associat- ed Se is difficult to separate accurately from the inorganic Se species, but in SEPs dilute NaOH targets this fraction rather specifi- cally (Wright et al. 2003). In general, ex- tractions of organically bound or inorgani- cally associated reduced selenide-Se rely on oxidizing agents, such as H2O2 (Zhang and Frankenberger 2003), K2S2O8 (Zhang et al. 1999) or NaOCl (Zhang and Moore 1996). Elemental Se can be solubilized with CS2 (Yamada et al. 1999) or by forming soluble selenosulphate in a reaction between Se0 and Na2SO3 (Velinsky and Cutter 1990). When the total Se content of the soil is sought, a mixture of strong acids, such as HCl-HNO3 or HNO3-HClO4-HF is used in dissolving the entire soil matrix (Sharmasarkar and Vance 1995). The appropriate method for exploring soil Se is governed by the objective of the anal- ysis, as well as the type and nature of the soil under study (Dhillon et al. 2005, Yli- Halla 2005). In Finland, hot-water extrac- tion (Yläranta 1982) has been chosen for monitoring the fertilization-induced chang- es in soil Se status (Eurola et al. 2008). The method lacks a sound theoretical basis, but its boiling treatment likely solubilizes some weakly adsorbed selenite and organ- ic Se compounds, in addition to the solu- ble selenate pool (Sharmasarkar and Vance 1997). Roughly 4% of the total Se reserves in soils of Finland are extractable by hot wa- ter (Yläranta 1983a). Accordingly, hot-wa- ter extraction serves in examining chang- es in the immediately biologically available Se fraction. However, the method does not extract the weakly soluble but the poten- tially available pool, in which Se assumably accumulates under acidic and semireduc- ing conditions. Powerful acid digestions, in contrast, recover the majority of the soil Se, but substantial variation in the Se content of the soil parent material easily overpowers small changes in the soil Se status (Yli-Hal- la 2005). Therefore, intermediate extractions targeting the accumulating Se pool would be ideal in monitoring the fertilization-in- duced changes in soil Se. 1.2 Selenium uptake by plants In plants, unlike in humans and animals, essential functions have not been demon- strated for selenoproteins, wherefore Se is not considered a plant nutrient (Terry et al. 2000). However, Se-induced beneficial effects initiated by increased antioxidative capacity have been recorded, especially in stressed plants (Hartikainen et al. 2000, Xue et al. 2001, Seppänen et al. 2003, Lyons et al. 2009). Whether plants themselves prof- it from Se or not, they serve as an impor- tant link in transferring Se from soil into the food chain (Girling 1984, Combs 2001, Hawkesford and Zhao 2007). Plant Se uptake occurs via several mecha- nisms, depending on the species absorbed. Selenate is taken up through high-affinity sulphate transporters (Terry et al. 2000, Sors et al. 2005), whereas selenite likely enters the plant via a phosphate transport pathway (Li et al. 2008). Due to the entrances shared, there is competition in the uptake between sulphate and selenate and phosphate and sel- enite (Hopper and Parker 1999). The uptake of organic Se compounds is not well known, but the permeases specific for S-containing amino acids may mediate the uptake of se- lenoamino acids as well (Abrams et al. 1990 and references therein). The relative uptake rate of Se species has been addressed, both in hydroponic (e.g. Williams and Mayland 1992, Zayed et al. 1998) and soil studies (e.g. Gissel-Nielsen and Bisbjerg 1970, Eich-Greatorex et al. 2007). In the soil matrix, the dissimilar sol- ubility of the species governs their availabil- ity (see section 1.1.1), but in solution culture the differences between the species seem less MTT SCIENCE 20 15 pronounced. Li et al. (2008) found soluble selenite as available to plants as selenate, although they quote several studies show- ing inconsistent results, likely due to dif- ferent experimental conditions. The organ- ic forms selenomethionine (SeMet) and SeCys seem at least as available to plants as selenite when in solution (Williams and Mayland 1992, Zayed et al. 1998). All plant species likely access the same la- bile soil Se fraction, but some plants tend to accumulate Se, whereas others are able to discriminate against it (Mayland 1994, Terry et al. 2000, Goodson et al. 2003). Accumulator plants can contain several grams of Se kg-1 dry weight (DW), whereas the Se concentration of nonaccumulators rarely exceeds 0.1 g kg-1, even when grown on seleniferous soils (Terry et al. 2000). Within plants, selenate and selenite are readily metabolized into various organic Se compounds via the S-assimilation pathway (Terry et al. 2000, Sors et al. 2005). The oxidized species are first reduced into sele- nide and thereafter assimilated into SeCys and further to SeMet. These selenoamino acids may then be nonspecifically incor- porated into proteins in place of Cys and Met, which can lead to alterations in pro- tein structure and thus weakening of its functions. Plants accumulating Se can ex- clude this substitution, e.g. by metaboliz- ing Se into nonprotein selenoamino acids, such as Se-methylselenocysteine, seleno- cystathionine, Se-methylselenomethionine and γ-glutamyl-Se-methylselenocysteine (Brown and Shrift 1981, Whanger 2002, Sors et al. 2005). In Se accumulators, these nonprotein selenocompounds can com- prise the major portion of the total plant Se, whereas SeMet is the predominant form of Se in cereal grains, soybeans and grassland legumes (Stadlober et al. 2001, Whanger 2002). By methylation, plants can convert SeMet into volatile dimethyl- selenide, the main contributor to atmos- pheric distribution of Se (Chasteen 1998, Sors et al. 2005). Of the inorganic species, selenate predominates over selenite within plants since reduction of selenate into sel- enite limits the biosynthesis rate of organic Se compounds (Ellis and Salt 2003). The transport, distribution and speciation of Se within plants is ultimately governed by the plant species, its developmental stage and the species of Se absorbed (Bis- bjerg and Gissel-Nielsen 1969, Williams and Mayland 1992, Terry et al. 2000, Whanger 2002). Selenate, which is highly mobile in plants, is transported into leaves to be reduced in the chloroplasts, whereas selenite can be reduced into selenide non- enzymatically in the roots, where it tends to accumulate as does Se absorbed in or- ganic form (Terry et al. 2000). 1.3 Selenium fertilization In vast areas of the world, food systems are too low in Se to support the maximal expression of the SeCys enzymes (Gissel- Nielsen et al. 1984, Combs 2001, Ray- man 2002). Accordingly, increasing the intake of Se would have beneficial ef- fects on health in most populations, even though actual Se deficiency diseases are rare. The use of Se medication in the form of high-Se-containing supplements (Ray- man 2004), fortifying foods, such as ta- ble salt (Yu et al. 1997) or bread (Rayman 2002) with Se or supplementing animal feeds, thus increasing the level of Se in foods of animal origin (Aro et al. 1998), are possible remedies for Se undernourish- ment. However, probably the best way to extend the measure throughout the popu- lation in a safe way is to introduce Se gen- erally into the food chain through fertili- zation (Broadley et al. 2006). Feasible application techniques and sourc- es of Se for field treatment of crops have been studied since the 1960s (summarized by Gissel-Nielsen et al. 1984, Gissel-Niels- en 1998). Most studies have focused on various selenate and selenite salts applied to soil as such or incorporated into com- pound fertilizers. In general, selenates have 16 MTT SCIENCE 20 proved to be more efficient than selenites in increasing plant Se content (e.g. Bisb- jerg and Gissel-Nielsen 1969, Gissel-Niels- en and Bisbjerg 1970, Yläranta 1983a,b). Additions of elemental Se, thought to function as a slow-releasing fertilizer, and organic forms of Se have shown minor in- fluence on the Se concentration of har- vested crops (Gissel-Nielsen and Bisbjerg 1970, Eich-Greatorex et al. 2007). Fer- tilization via foliar spraying and coating of seeds with Se has also been explored. In seed treatment, Se amounts equal to those in direct soil application are need- ed to attain the desired Se concentration of crops, whereas with appropriately timed foliar spraying including the use of a sur- factant, the plant Se concentration can be increased with a small Se supply (Yläran- ta 1984a–c, Gissel-Nielsen 1998). Howev- er, Yläranta (1985) concluded his extensive studies on Se fertilization by specifying soil application of Se-supplemented nitro- gen-phosporus-potassium (NPK) fertiliz- er to be the most reliable and cost-effec- tive method of increasing the Se content of all crops. Encouraging results of Se fertilization trials (Gissel-Nielsen and Bisbjerg 1970, Yläranta 1985) persuaded the Finnish au- thorities to launch a national Se fertiliza- tion programme to improve the Se nutri- tion of the entire population (Ministry of Agriculture and Forestry 1984). The aim was to increase the Se concentrations of crops from the extremely low level of 0.01 mg Se kg-1 (DW) (Oksanen and Sand- holm 1970, Sippola 1979, Yläranta 1990) up to 0.1 mg Se kg-1, which is defined as adequate for animals and humans (Gis- sel-Nielsen et al. 1984). The first fertiliz- ers supplemented with sodium selenate (Na2SeO4) were available for farmers for the third cut of grass in 1984. Se fertili- zation has been in general use since the 1985 growing season. At first, only com- pound fertilizers (NPK and NK) were sup- plemented with Se, but since 1996 N fer- tilizers have also been amended. To control the safety of the Se fertiliza- tion practice, a group of experts has regu- larly monitored the average daily Se intake level and concentrations of Se in foods, feeds and human serum according to spe- cific sampling policies (Ministry of Agri- culture and Forestry 1985, 1990a). Based on follow-up data, the level of Se supple- mentation in fertilizers has been adjusted as follows: 1984 – 1990 6 mg Se kg-1 in fertilizers intended for grasslands 16 mg Se kg-1 in fertilizers intended for cereal crops 1990 – 1998 6 mg Se kg-1 in all fertilizers 1998 – 2007 10 mg Se kg-1 in all fertilizers 2007 onwards 15 mg Se kg-1 in all fertilizers, with the exception that those used in comple- menting manure may contain 25 mg Se kg-1 The two initial supplementation levels were based on the studies of Yläranta (1985), showing the Se concentration of grasses in- creasing more efficiently than that of cere- als. Doubts that Se fertilization causes in- creased algal growth in waterways together with detection of some high Se concentra- tions of crops led to the adoption of the lower supplementation level for all crops (Eurola et al. 2003). However, decreases in the Se intake level of people and livestock resulting mainly from a trend toward de- crease in the amount of fertilizers used ne- cessitated subsequent increases in fertilizer Se concentrations (Eurola et al. 2011). On a hectare basis, the annual Se addition is now typically around 7.5 g. By means of Se fertilization, the average daily Se intake in Finland has increased from as low as 25–30 µg Se d-1 (Koivis- MTT SCIENCE 20 17 toinen 1980, Varo and Koivistoinen 1981, Mutanen and Koivistoinen 1983, Eurola et al. 2008) to around 80 µg Se d-1 (Euro- la et al. 2011). Nutritionally, the fertiliza- tion practice has thus proved successful. Yet, concern over health and environmen- tal risks related to relatively small differ- ences between Se essentiality and toxicity has thus far restricted implementation of the measure to Finland and low-Se areas of China, New Zealand and the UK, where Se fertilization is used in preventing Se de- ficiency of grazing livestock (Combs 2001, Broadley et al. 2006). Aquatic ecosystems are especially sensitive to elevated Se lev- els, wherefore leaching of fertilizer Se pos- es a particular threat to waterways (Maier et al. 1998). The efficiency of fertilizer Se uptake by plants is usually low, between 5% and 20% of the annual selenate appli- cation (Yläranta 1985), ranging, however, from below 1% to nearly 50% (Yläranta 1985, Tveitnes et al. 1996, Eich-Greato- rex et al. 2007). In Finland, residual Se is assumed to be fixed in the soil (Yläran- ta 1985), in which case the computation- al fertilization-induced increase in the Se concentration of the plough layer of Finn- ish agricultural soils would be around 20% (Eurola et al. 2003). However, hot-water- extractable Se concentrations have not in- creased from the prefertilization level of 0.01 mg Se l-1 soil (Eurola et al. 2008). In acid digestion, deviation in the inherent Se concentration of the soil parent mate- rial overwhelms the small fertilization-in- duced changes (Yli-Halla 2005). 1.4 Aims of the present study Due to the narrow margin of safe Se con- centrations, impacts of Se fertilization on the Se content of foods, feeds and daily Se intake level have been regularly monitored in Finland since commencement of the supplementation practice in 1985 (Min- istry of Agriculture and Forestry 1986a,b, 1987, 1988, 1989, 1990b, 1994, Eurola and Hietaniemi 2000, Eurola et al. 2003, 2008, 2011). Follow-up of the effects of added Se on soils and the environment has been less thorough and comprehensive un- derstanding of the cycling of the added se- lenate in the soil-plant system is lacking. About 90% of the annually applied Se is assumed to be immobilized in the soil (Eu- rola et al. 2003), but the fate of this residu- al Se has not been verified by soil analyses. The purpose of this thesis was to increase the level of understanding of the behav- iour of fertilizer Se in the acidic soils of Finland. The specific aims were to: Find suitable extraction methods of sound basis for both detailed examination of the Se reserves in soil and for routine analy- ses of the Se pool likely available to plants (I,III) Characterize the distribution of soil Se fractions in Se-fertilized mineral and or- ganic field soils in different parts of Fin- land and assess long-term changes in these fractions (II) Examine plant availability of added Se in mineral and organic soils with emphasis on Se retention in soil (III, IV) Gather information on the total uptake, rhythm of uptake and distribution of Se within crops (III, IV) 18 MTT SCIENCE 20 2 Material and methods In this thesis, the results of two laborato- ry studies (I and II) and four pot experi- ments (III and IV) are presented. The lab- oratory studies consist of extractions of soil Se. First, various extractants were tested in two soils (I) and, thereafter, Se reserves of nine soil pairs from different parts of Fin- land were fractionated (II). In the first pot experiment, hereafter named EXP1, the uptake and translocation of Se in spring wheat (Triticum aestivum L. cv. Manu) was investigated concomitantly with chang- es in soil Se fractions (III). The second pot experiment, EXP2, aimed to clarify the Se uptake efficiency of Italian ryegrass (Lolium multiflorum Lam. cv. Meroa) and the congruence between plant Se uptake and soil Se extractions (III). The third and fourth pot experiments, EXP3 and EXP4, were conducted to determine the uptake of added Se by plants and its retention in peat soils, which are often low in adsorbing ox- ide components (IV). 2.1 Experimental designs 2.1.1 Soils used in the studies The soils used in the experiments were mostly collected from fields in different parts of Finland (Figure 1, Table 1). Var- ious Se extraction methods were com- pared in two mineral soils, sand and silty clay, collected from Viikki Research Farm, Helsinki. The same two soils were used in EXP2 and the sand soil also in EXP1. The sand was included as a control in EXP3, in which the uptake and fate of added Se was studied in two peat soils from Jaak- kola Farm (Hausjärvi) and on a commer- cial sphagnum (Sphagnum L.) peat. EXP4 was conducted solely with a commercial sphagnum peat substrate manipulated chemically with iron (Fe). For character- izing the soil Se reserves, nine sample pairs from research stations of MTT Agrifood Research Finland (Jokioinen, Maaninka, Mietoinen, Pälkäne, Rovaniemi, Ruukki and Ylistaro) were fractionated. The field soils, except the peat soils of Hausjärvi, were collected from the plough layer, approximately 0–20 cm from the soil surface, using an auger or spade. The Hausjärvi topsoil peat was composed of the 30-cm surface layer, whereas the sub- soil peat was collected from depths of 50– 80 cm. The sand and silty clay soil were taken in May 2006 for soil extraction tests and for use in EXP2. For EXP1 and EXP3, peat soils and sand from the same location as in 2006 for EXP1 were col- lected in spring 2007. The first set of the paired MTT research station samples were taken in autumn 1992 (Urvas 1995) and Figure 1. Sampling locations for the soils used in the study. MTT SCIENCE 20 19 the same locations were resampled in au- tumn 2004 (Yli-Halla 2005). These soils were stored air-dried in cardboard boxes until taken to the fractionation analyses in summer 2009. 2.1.2 Examination of selenium in field soil samples After a literature search, the performance of seven single extractants and two SEPs were chosen for testing on two dissimilar mineral soils (sand and silty clay). Of sin- gle extractants, aqua regia (AR), acid am- monium oxalate, hot water and phosphate buffer in four different concentrations and pH combinations were compared. The SEPs included were a four-step method developed by Chang and Jackson (1957) for soil P and a five-step procedure devel- oped for Se by Zhang and Moore (1996) Table 1. Description of the experimental soils. For more detailed properties, see the appropriate articles. Location Identificationa Typeb pHc Clay < 0.002 mm (%)d Corg (%)e Article Helsinki Sand Sand soil from an uphill field 5.9 3 2.2 I,III,IV Helsinki Silty clay Field soil composed of brackish water sediment of the former Baltic Sea 5.4 45 5.1 I, III Jokioinen JKA3 Clay soil 5.2 71 3.5 II Jokioinen JKA20 Organic field soil 4.7 72 17 II Maaninka PSA10 Organic field soil 4.9 24 11 II Mietoinen LOU5 Clay loam soil 5.3 39 2.2 II Pälkäne HÄM7 Sandy loam soil 5.0 10 2.7 II Pälkäne HÄM10 Organic field soil 4.6 40 22 II Rovaniemi LAP3 Silt loam soil 5.5 10 3.4 II Ruukki PPO2 Organic field soil 4.1 5 28 II Ylistaro EPO2 Silty clay loam soil 4.0 29 7.4 II Hausjärvi Topsoil peat Peat from the surface layer of a moderately well humi- fied field 5.2 na 30 IV Hausjärvi Subsoil peat Raw peat from the subsoil layer of a peat field 4.4 na 53 IV Sphagnum peat Commercial light-coloured sphagnum peat, Kekkilä A0 3.8 na na IV na = not analysed aRefers to the nomenclature used in the individual articles. bThe mineral soils are classified according to the USDA textural triangle. cMeasured in 0.01-M CaCl2. dThe clay percentages are calculated from the inorganic matter of the soil. eAnalysed by dry combustion (LECO® CHN 900 analyser) except for Hausjärvi soils in which the percentage refers to loss of weight on ignition/1.724. 20 MTT SCIENCE 20 and modified by Wright et al. (2003). For details of the methods, see section 2.2. The SEP of Wright et al. (2003) was se- lected for further use in characterizing the Se reserves of field soils. Paired sam- ples from research stations of MTT Agri- food Research Finland by Yli-Halla (2005) also enabled investigation of changes in the soil Se status between 1992 and 2004. Nine pairs of samples were carefully se- lected from the entire set of 48 locations to include both organic and mineral soils. In addition, the crop rotation and com- putational Se balances were considered. Grassland systems prone to uneven dis- tribution of Se in soil due to topdressing of fertilizers and sparse tillage, and fields with Se balances below 30 g Se ha-1 were avoided. A further criterion was the ac- curacy of the sampling location, which was assessed in advance, based on volume weights. A considerable change in this var- iable reflecting the organic matter content of the soil would have been unlikely dur- ing the study period, wherefore an ine- quality in volume weight within the sam- ple pair was assumed to signify differing sampling locations. The uniformity of the chosen pairs was confirmed with analyses of the organic carbon content and parti- cle-size distribution of the inorganic mat- ter. In one of the soils chosen (JKA20), these characteristics deviated considera- bly between the 1992 and 2004 samples. Therefore, this soil was omitted from pair- wise comparisons. 2.1.3 Pot experiments The batches of field soils collected for the pot experiments were first passed through a 10-mm sieve and mixed thoroughly. In EXP2, the residual effect of Se fertiliza- tion was exploited by reusing soils from a three-week pot experiment conducted with oilseed rape the previous year (re- sults not shown). The soils were packed in plastic bags, each pot separately, and stored overwinter in a cold greenhouse. These preprepared EXP2 soils and com- mercial sphagnum peat of EXP4 were not sieved, but mixed. After homogenization of the soils, plastic pots of 1.5 l in EXP1, 2 l in EXP2 and 3.5 l in EXP3 and EXP4 were filled with soil by weighing (Table 2). In EXP4, the Se sorption capacity of the peat was manipulated by iron hydrox- ide enrichment. This was done by mix- ing FeCl3 solution and solid Ca(OH)2 into the soil in the pots. Control pots (Fe0) re- ceived only Ca(OH)2, whereas the moder- ately enriched pots (Fe1) received 2 g Fe kg-1 peat DW and the amply enriched pots (Fe2) 20 g Fe kg-1 peat DW. In all experiments, the pots were individ- ually fertilized by the same method (Ta- ble 3). In EXP2, however, no S was added, due to high S addition (200 mg l-1 soil) in the previous year’s experiment. The macro- nutrients were applied separately as solid compounds and micronutrients dissolved in a single combined solution, after which all nutrients were hand-mixed evenly into the soil volume. The peat soils in EXP3 were limed with Ca(OH)2 to elevate their pH to 5.5. In EXP4, enough Ca(OH)2 was added with the FeCl3 to increase the peat pH to 4.7 in the Fe0 and Fe1 treatments and to 4.5 in the Fe2 treatment. Se fertili- zation was conducted simultaneously with application of the nutrients according to Table 2. Se was applied as an Na2SeO4 so- lution. In EXP4, Se addition of 1 mg l-1 soil was applied also as a sodium selenite (Na2SeO3) solution. In EXP1 and EXP3, spring wheat and in EXP2 and EXP4 Italian ryegrass were sown in the pots (Table 2). The EXP4 in- cluded plant-free pots for observation of the behaviour of Se in soil without the in- fluence of plants. After sowing, the pots were completely randomized onto green- house tables in EXP1, EXP2 and EXP4. In EXP1, the border effect was avoided by surrounding the experimental pots with nurse pots, whereas in EXP2 and EXP4 the pots were rearranged periodically. MTT SCIENCE 20 21 EXP3 was arranged in a complete rand- omized block design and bordered with nurse pots. The greenhouse settings were identical in all experiments. The temperature was set to 20 °C by day and 16 °C by night and Table 2. Summary of the four pot experiments EXP1–EXP4 conducted in the study. EXP1 EXP2 EXP3 EXP4 Objectives Defining the uptake rhythm and translo- cation of Se within wheat, and the short- term changes in soil Se frac- tions. Comparing the efficiency of ryegrass and soil extractants in the with- drawal of Se and defining the distribu- tion of Se in ryegrass. Examining the efficiency of selenate ferti- lization in soils increasing in the content of organic matter. Clarifying the effects of organic and inorganic com- ponents on the retention of Se in soil. Pot volume (l) 1.5 2.0 3.5 3.5 Soil; g DW per pot Sand; 1510 Sand; 2484 Silty clay; 1545 Sand; 3147 Topsoil peat; 920 Subsoil peat; 393 Sphagnum peat; 262 Sphagnum peat; 356 Fe0; Fe1; Fe2 enrichments Se treatments (mg Se l-1 soil) 0.1 Residuals of 0; 0.0025; 0.005 0; 0.0025; 0.005; 0.1; 1 0; 0.005; 1 Replicates 47 pots in total 6 5 4 planted; 3 unplanted Design Completely randomized Completely randomized Randomized blocks Completely randomized Plant Spring wheat (Triticum aestivum L. cv. Manu); 10 plants per pot Italian ryegrass (Lolium multi- florum Lam. cv. Meroa); 1.3 g seeds per pot Spring wheat (Triticum aestivum L. cv. Manu); 20 plants per pot Italian ryegrass (Lolium mul- tiflorum Lam. cv. Meroa); 2 g seeds per pot, plantless pots included Sampling Soil and plant samples after 1, 2, 3, 4, 6, 8 and 10 weeks of growth. Plant samples separated into leaves, stems, spikes, (grains) and roots. Soil samples at the begin- ning of the experiment. Ryegrass leaf cuttings 4 and 6 weeks after sowing and roots after the second cut. Soil and plant samples after ripening of the crop (10 weeks of growth). Plant samples separated into leaves, stems, grains and chaff. Ryegrass leaf cuttings 3 and 6 weeks after sowing. Soil samples after the second cut. 22 MTT SCIENCE 20 the relative humidity to 40%. Natural day- light was supplemented with high-pressure sodium lamps at 300 W m-2 from 6 a.m. to 10 p.m. to maintain the photosynthetic photon flux density in the range of 250– 300 μmol m-2 s-1. Excessive light intensity ( > 500 W m-2) was prevented by closing shading curtains. Deionized water (electri- cal conductivity < 1 μS cm-1) was added as needed to keep the pots moist, yet avoid- ing leaching. In EXP1, samples were collected after 1, 2, 3, 4, 6, 8 and 10 weeks of growth. Sam- pling was conducted by random selection of seven pots (five only in the first sam- pling), which were harvested after spec- ifying the growth stage of wheat on Za- doks’ scale (Zadoks et al. 1974). The aboveground plant mass was separated into leaves, stems and spikes, and in the last sampling of the ripend crop also into grains. From four pots (three in the first sampling), roots were elutriated, using a root washer (Model 13000 Gillison’s Vari- ety Fabrication Inc.). In EXP3, wheat was harvested at maturity, separating leaves, stems and spikes that were further as- sorted into grains and chaff. The Italian ryegrass in EXP2 and EXP4 was cut twice. In EXP2, ryegrass was harvested 4 and 6 weeks after sowing, and in EXP4 3 and 6 weeks after sowing. In EXP2, the roots were elutriated after the second cutting from three replicates of two treatments (0 and 0.005 mg Se l-1) of both soils with a root washer (Gillison). All plant material, except the harvest-ripe wheat, was frozen in liquid nitrogen im- mediately after harvesting. The ripe wheat was dried directly at 60 °C. The frozen samples were stored at -20 °C until lyophi- lized. The dry mass of each plant sample was weighed and the samples milled (Cy- clotec Sample Mill 1093, Kika Labortech- Table 3. Amounts of nutrients added (mg l-1 soil) in pot experiments EXP1–EXP4 Nutrient Compound Amount of nutrient added (mg l-1 soil) N NH4NO3, KNO3 150 P Ca(H2PO4)2 × H2O 50 K KNO3 125 Mg MgCl2 × 6H2O 25 Sa CaSO4 × 2H2O 40 Mo Na2MoO4 × H2O 1 Cu CuCl2 × 2H2O 2 Zn ZnCl2 5 Feb FeCl3 × 6H2O 5 Mn MnCl2 × 4H2O 5 B H3BO3 1 Cac Ca(H2PO4)2 × H2O, CaSO4 × 2H2O, Ca(OH)2 30–1000 aNo S was added in EXP2. bIn EXP4, 0.2 g and 2 g Fe l-1 soil as FeCl3 were introduced in Fe enrichment treatments. cWithin Ca(H2PO4)2 × H2O 32 mg Ca l-1soil, and within CaSO4 × 2H2O 50 mg Ca l-1 soil were added. The peat soils of EXP3 and 4 received an additional 250 – 1000 mg Ca l-1 soil as Ca(OH)2 to elevate the soil pH. MTT SCIENCE 20 23 nik A10). The dried samples were stored at room temperature until analysed for total Se concentration (see section 2.3.2). In EXP1, the soil samples were collected at every sampling from the replicates not used for root elutriation and subjected di- rectly to analyses. In EXP2, the soils were sampled at the beginning of the experi- ment from the batches remaining after fill- ing of the pots. In EXP3 and EXP4, the soil samples were taken after the final har- vest. For the soil extractions and Se analy- ses, see sections 2.2 and 2.3.1. 2.2 Soil selenium extractions In comparing the extraction methods for soil Se (I), single extractants of varying ag- gressiveness (Table 4) and two SEPs frac- tionating soil Se into separate operation- ally defined chemical pools (Table 5) were included. The strong acids in AR disin- tegrated the soil matrix nearly complete- ly, thus producing semitotal Se concentra- tions. Acid ammonium oxalate extraction performed in the dark dissolves noncrys- talline metal oxides (Jackson et al. 1986) and thereby releases Se, which is retained by poorly crystalline Al and Fe oxyhydrox- ides. Phosphate buffers recover ligand-ex- changeable Se, since phosphate replaces selenite, due to its higher sorption strength (Rājan and Watkinson 1976). Hot-water extraction targets the easily soluble Se pool. The successive extraction steps of SEP1 (Chang and Jackson 1957) separat- ed soil Se into four fractions: 1) the most easily soluble Se was extracted with a sim- ple salt solution (NH4Cl), removing Se in soil water and Se bound unspecifically by Table 4. Summary of the single extraction methods used in the study. Extractant Description Reference Article Aqua regia A suspension of 3 g soil and 43 ml of a mixture of HCl and HNO3 was allowed to stand at room temperature for 16 h, after which it was boiled under reflux for 2 h. The cooled extractant was filtered (Whatman Grade 589/2, White ribbon) into a 100-ml volumetric flask, which was filled to the mark with 0.5 M HNO3. ISO 11466 I, II Acid ammonium oxalate A suspension of 2.5 g soil and 50 ml of oxa- late solution (0.18 M (NH4)2(COO)2 + 0.1 M (COOH)2, pH 3.3) was shaken in a reciprocat- ing shaker for 2 h in the dark. Thereafter, the suspension was centrifuged (15 min, 3000 x g) and the supernatant filtered (Whatman Grade 589/3, Blue ribbon). I Phosphate buffer A suspension of 2–10 g soil and 50 ml of either 0.1 or 1 M phosphate buffer solution (K2HPO4 -KH2PO4) at pH 4.4 or 7.0 was shaken in a reciprocating shaker for 4 h, then centrifuged (15 min, 3000 x g) and the supernatant filtered (Whatman Grade 589, Black ribbon). e.g. Rājan and Watkin- son (1976) I, III Hot water A suspension of 25 g soil and 100 ml of milli-Q water was boiled for 30 min under reflux, then centrifuged (15 min, 3000 x g) and the super- natant filtered (Whatman Grade 589, Black ribbon). Yläranta (1982) I 24 MTT SCIENCE 20 anion exchange and mass action, 2) Se as- sociated with Al oxides was dissolved by NH4F, assuming that fluoride complexes the Al selectively (Turner and Rice 1952), 3) Se bound by Fe oxides was obtained Table 5. Summary of the sequential extraction procedures (SEPs) used in the study. Extractant Description Reference Article SEP1 Step 1) A suspension of 10 g soil and 50 ml of 1-M NH4Cl was shaken in a reciprocating shaker for 30 min, then centrifuged (15 min, 3000 x g) and the supernatant filtered (Whatman Grade 589, Black ribbon) and collected for analyses. Step 2) 50 ml of 0.5-M NH4F, pH 8.5 was added to the residual soil and the suspension shaken in a reciprocating shaker for 1 h and thereafter centrifuged as before. The residual soil was rinsed with 25 ml of saturated NaCl solu- tion, which was discarded. Step 3) 50 ml of 0.1-M NaOH was added to the rinsed soil and the suspension shaken in a reciprocating shaker overnight and then centrifuged and rinsed as before. Step 4) 50ml of 0.25-M H2SO4 was added to the residual soil, shaken in a reciprocating shaker for 1 h and thereaf- ter centrifuged as before. Chang and Jackson (1957) I, III SEP2 Step 1) A suspension of 10 g soil and 50 ml of 0.25 M-KCl was shaken in a reciprocating shaker for 2 h, and then centrifuged (15 min, 3000 x g). The supernatant was filtered (Whatman Grade 589, Black ribbon) and collected for analyses. Step 2) 50 ml of 0.1-M K2HPO4, pH 8.0 was added to the residual soil and the suspension shaken in a reciprocating shaker for 2 h, then centrifuged as before. The superna- tant was collected and the extraction repeated. Thereaf- ter, the residual soil was rinsed with 10 ml of 0.1-M KCl, which was combined with the extracts. Step 3) 50 ml of 0.1-M NaOH was added to the residual soil and the suspension shaken in a reciprocating shaker for 4 h, then centrifuged and rinsed as before. Step 4) 50 ml of 0.25 M-Na2SO3, pH 7.0, was added to the residual soil and the suspension disrupted with a 2-min sonication, then held in an ultrasonic bath for 4 h. The samples were centrifuged as before and thereafter the residual soil was rinsed twice with 20 ml of 0.25-M Na2SO3 and once with KCl as before. Step 5) 40 ml of 5% NaOCl, pH 9.5, was added to the residual soil and the suspension placed in a 90 ºC water bath for 30 min. The samples were centrifuged as before and thereafter the extraction was repeated. Wright et al. (2003) I, II by NaOH via ligand exchange (Hingston et al. 1967) and 4) Se bound in acid-sol- uble form was dissolved with H2SO4. In SEP2 (Wright et al. 2003), five pools of soil Se were segregated: 1) salt-soluble Se MTT SCIENCE 20 25 was recovered by KCl, 2) adsorbed Se was displaced with a phosphate buffer, 3) or- ganically associated Se was extracted with NaOH, 4) elemental Se was released by oxidation with Na2SO3 and 5) recalcitrant forms of organic Se and metal selenide compounds were oxidized with NaOCl. The SEP2 procedure was also used in frac- tionating the Se reserves of field soil sam- ples (II). In EXP1, the hypothesized re- duction of the fertilizer selenate to selenite and subsequent adsorption onto soil sur- faces was followed closely throughout the growing period. The first three steps of the SEP1 shown to remove selenite add- ed to soils very efficiently (Nakamaru et al. 2005) and a single 4-h phosphate buf- fer extraction (1 M KH2PO4/K2HPO4, pH 7.0) were carried out at 1-week inter- vals during the first month after sowing and thereafter at 2-week intervals until the wheat harvest. For the soils in EXP2, EXP3 and EXP4, salt solution extraction (step 1 of SEP1) and a phosphate buffer ex- traction (step 2 of SEP2 as one continuous 4-h extraction) were combined in a two- step sequential procedure. In EXP4, the concentration of the phosphate buffer was increased to 1 M. To standardize contrast- ing samples to correspond to the same soil volume, aliquots of 2 g sphagnum peat, 7 g humified peat and raw peat, and 10 g sand were weighed for the analyses. Prior to the extractions, the soil samples were air-dried at room temperature and passed through a 2-mm sieve while remov- ing and discarding visible plant roots. The only exception was EXP1, in which the soils were not dried but sieved and taken for analyses directly after sampling. The AR and hot-water extractions and step 5 of SEP2 were conducted in boiling flasks. All other extractions were carried out in plastic 100-ml tubes. The extractions were done in three replicates in all experiments, except EXP4, in which four replicates were analysed. 2.3 Analyses of selenium The total Se concentration of the soil ex- tracts and plant samples was analysed by a graphite furnace atomic absorption spectrometer (AAS) (PerkinElmer Zee- man 5100), using an electrothermal AAS method for food samples (Kumpulainen et al. 1983, Ekholm 1996). In the meth- od, Se was concentrated in an organic sol- vent, methyl isobutyl ketone (MIBK) as an ammonium pyrrolidine dithiocarba- mate (APDC) chelate. Selenite is the only chelateable species, wherefore an oxida- tion-reduction procedure was included for transforming all Se in the sample into an analysable form. In EXP1, an attempt to separate selenite from the other Se species in the extracts was made by also conduct- ing the analysis without the oxidation-re- duction step. 2.3.1 Soil extracts Aliquots of 2–15 ml of the soil extracts were taken into the Se analyses. First, all Se in the samples was oxidized to selenate by adding 10 ml of concentrated HNO3 and digesting the solutions for 45 min at 120 ºC. Thereafter, 10 ml of 4-M HCl were added and the digestion continued for 20 min at 130 °C to reduce the se- lenate to selenite. Subsequently, the sam- ples were allowed to cool at room temper- ature, after which the pH of the solutions was adjusted, according to a bromphenol blue indicator, to a range between 4 and 5 with NH3 or HNO3. The pH was buffered to this range for optimal selenite chelation with 5 ml of 0.1-M ammonium citrate solution. To prevent competition for the chelating agent between Fe and selenite, Fe was first chelated by a 5-ml addition of 5% ethylenedinitrotetraacetate (EDTA), after which the selenite was bound with a 2-ml addition of 2% APDC. Next, the APDC-chelated Se was extracted to 2.5−10 ml of MIBK. After the MIBK addition, the samples were first shaken in a recipro- cating shaker for 5 min and thereafter al- lowed to stand for 20 min to let the phas- 26 MTT SCIENCE 20 es separate. A sample from the upper-layer organic phase was then collected for anal- yses. For instrumental parameters, see I. 2.3.2 Plant samples The dried and milled plant samples were further dried at 70 °C overnight prior to the Se analyses. After removal of the ad- sorbed moisture, aliquots of 0.1–1.0 g of each sample were weighed into wet-diges- tion tubes. Next, 10 ml of an acid mix- ture (HNO3-HClO4-H2SO4 in a ratio of 2.5:1.5:1) were added, after which the sam- ples were allowed to stand at room tem- perature overnight. The acid digestion was completed according to a five-step temper- ature programme: 1) 30 min at 70 °C, 2) 3 h at 120 °C, 3) 30 min at 160 °C, 4) 30 min at 190 °C and 5) 5 h at 220 °C. There- after, the selenate was reduced to selenite with HCl, chelated with APDC and ex- tracted with MIBK in the same way as for the soil samples. In EXP1 and EXP2, the plant samples analysed for Se represented the replicates from which the roots were elutriated. Thus, the wheat samples were analysed in quad- ruplicate and the ryegrass samples in trip- licate. In EXP3 and EXP4, the Se analy- ses of the plant samples were carried out in quadruplicate. In EXP4, the total Se con- centration of the unplanted peat soil sam- ples (0.5–1 g soil aliquots) were analysed in three replicates in a way similar to that of the plant samples. 2.3.3 Quality control In the analyses of the soil extracts, blank samples were included in all series as a contamination control. Throughout the study, the Se concentration of the blanks remained below the detection limit of 1 µg Se l-1, defined in the testing protocol (I) as the mean + 3 × standard deviation of 37 blank samples. Known Se additions were also used and the recovery of spiked Se was on average 97 ± 15% (n = 57). The extracts high in Fe needed dilution and an extra EDTA addition to avoid interference. For the soil Se analyses, an expanded un- certainty of 8% was defined by propagat- ing uncertainties according to the Inter- national Organization for Standardization – Guide to the Expression of Uncertainty in Measurement (ISO-GUM). In the plant analyses, a standard reference material (wheat flour 1567a, National In- stitute of Standards & Technology) and an in-house wheat flour control material were included in all sample series to assess the accuracy of the analyses. For the standard wheat flour with a certified Se concentra- tion of 1.1 ± 0.2 mg kg-1 DW, an Se con- centration of 1.2 ± 0.2 mg kg-1 DW (n = 40) was obtained. The in-house reference with a known Se content of 0.040 ± 0.008 mg kg-1 DW produced an average concen- tration of 0.045 ± 0.008 (n = 98). MTT SCIENCE 20 27 3 Results and discussion 3.1 Extraction methods for assessing the selenium status in soils of Finland 3.1.1 Comparison of methods In comparison to the standard method of AR extraction, which produces Se concen- trations similar to the total amounts re- covered by HNO3-HClO4-HF digestion (Yli-Halla 2005), the efficiency of the ex- tractants tested in dissolving soil Se var- ied greatly (Figure 2, I). In Se acquisition, SEP2 was equally as efficient as AR. Due to this high efficiency and good reported specificity (Wright et al. 2003), the SEP2 method was considered suitable for charac- terization of the soil Se reserves (see 3.1.2). However, the five subsequent steps make it too laborious for use in routine analyses. The next most efficient methods, SEP1 and acid ammonium oxalate, recovered rough- ly half of the amount of Se obtained with AR. Both of these methods target the ox- ide-associated Se pool, the oxalate by dis- -1.16 -1.17 -1.38 -1.47 -1.67 -1.82 -2.11 -2.16 -2.15 nd -3.05 -0.46 -0.42 -0.86 -0.76 -1.01 -1.16 -1.44 -1.79 -1.88 -2.65 -2.64 Aqua regia SEP2 total SEP1 total Acid ammonium oxalate Phosphate buffer 1 M pH 7.0 Phosphate buffer 1 M pH 4.4 Phosphate buffer 0.1 M pH 7.0 Phosphate buffer 0.1 M pH 4.4 Hot water Ammonium chloride (step 1 of SEP1) Potassium chloride (step 1 of SEP2) Relative efficiency of soil Se extractions Sand Silty clay Figure 2. Relative efficiency of different soil Se extraction methods in comparison to aqua regia extraction in sand and silty clay soils. The bars are averages of three replicate extractions. The log-transformed means for the corresponding Se concentrations are reported at the end of each bar. Standard error for the log-transformed means is 0.03 and the least significant difference 0.08. 28 MTT SCIENCE 20 solution of the oxides and the NH4F (step 2) and NaOH (step 3) of SEP1 by ligand exchange on the Al and Fe oxide surfaces, respectively. Most of the Se acquired with the SEP1 was recovered in these oxide-as- sociated fractions. In clay soil, 60% of the total amount of Se extracted with this pro- cedure was associated with Al oxides and 35% with Fe oxides. In the sand soil, the corresponding percentages were 50% and 50%. Accounting for the oxalate-extracta- ble Fe/Al molar ratio of the soils, Fe oxides in the sand seemed to be preferred in Se sorption, but no such trend was seen in the clay. Yläranta (1983a) showed that Se was associated rather with acid ammonium ox- alate-extractable Al than with Fe, but John et al. (1976) showed contrasting results. All three extractants targeting the ox- ide-associated Se were particularly dark in colour, with dissolved organic mat- ter suggesting the presence of organical- ly associated Se. In preliminary tests, hu- mic Se was precipitated from the NH4F and NaOH extracts by acidifying them with H2SO4. This procedure reduced the Se concentration of the NH4F extract by 40% in the sand and 30% in the clay soil. In the NaOH extract, the corresponding reductions were 50% and 40%. Séby et al. (1997) showed that the nonacid-soluble fraction may contain considerable amounts of selenite in addition to the humic acid- bound Se. However, organic Se seems to be preferentially associated with the ful- vic acid fraction (Séby et al. 1997, Coppin et al. 2006). The aggressiveness of the phosphate buff- ers in releasing soil Se was dependent on the phosphate concentration and pH of the buffer solution. The neutral concentrated buffer dissolved around 30% of the AR-ex- tractable Se, whereas the corresponding re- covery for the acidic and dilute buffer was merely 5–10%. The adsorption-desorption behaviour of selenite is dependent on the pH (Balistrieri and Chao 1987). Selenite adsorption is enhanced at low pH and de- sorption is known to increase with increas- ing pH (Neal et al. 1987). An increase in the phosphate concentration of the des- orbing solution increases competitive ad- vantage for phosphate over selenite for the binding sites, resulting in increased ac- quisition of Se (Balistrieri and Chao 1987, Saha et al. 2005). These phenomena ena- ble weakly adsorbed selenite to be distin- guished from that bound to higher en- ergy sites by adjusting the concentration and pH of the buffer extractant (Saha et al. 2005). However, in addition to the in- organic selenite, the phosphate buffer-ex- tractable adsorbed Se fraction may contain organically associated Se, especially sele- noproteins solubilized from plant material (Martens and Suarez 1997b, Wright et al. 2003). In the phosphate extracts of EXP1 soils, which were analysed both with and without the oxidation-reduction pretreat- ment to separate selenite from the other species, merely 20% of the total Se con- tent of the extracts was recovered directly as selenite (II). Hot water and salt solutions were placed at the weakest end in the efficiency of soil Se extractants (Fig. 2). The hot-water ex- traction dissolved over five times more Se than did the salt solutions, likely due either to the disruption caused by the boiling treatment or the low ionic strength in the water, which enhances desorption (Ryden and Syers 1975, Hartikainen and Yli-Halla 1982). Wright et al. (2003) found KCl ex- traction efficient at removing easily soluble selenate. The higher efficiency of the hot water in comparison to the KCl suggests that the method removes weakly bound selenite or organic Se in addition to the selenate. Sharmasarkar and Vance (1997) showed that hot-water extracts contain considerable proportions of organic Se in addition to the inorganic selenate. Excluding the SEP2 and acid ammonium oxalate, the extractants tested were rela- tively more efficient in the sand than in the silty clay soil. This phenomenon result- MTT SCIENCE 20 29 ed likely from the fact that in the clay soil the proportion of poorly soluble Se, possi- bly occluded within the mineral structures out of the reach of the extractants operat- ing on the soil surfaces, was greater than in the sand soil. The absolute amount of Se extracted with AR from the silty clay (0.35 mg Se kg-1) was fivefold higher than that acquired from the sand (0.07 mg Se kg-1), an order of total Se contents typical between these soil types (Koljonen 1975). Hot-water extraction has been used as the follow-up method in the Se monitor- ing programme in Finland (Eurola et al. 2008). It may be useful in assessing the readily bioavailable Se pool, but it is clear- ly too weak for observing the soil Se re- serves. With the more aggressive methods, the accumulation of fertilizer Se could be observed before it is reflected in the easi- ly soluble Se concentrations. Of the single extractions, AR reaches the majority of the Se reserves, but embodies fractions out- side the biological cycles. The acid ammo- nium oxalate or strong phosphate buffers thus appear most promising for monitor- ing the changes in the potentially bioac- cessible soil Se pool. 3.1.2 Selenium fractions of various field soils In the laboratory studies, the Se reserves of 11 field soils were fractionated with the SEP2 method (I,II). The soils, including the sand and silty clay from Viikki Re- search Farm, Helsinki and the nine MTT Research Station samples, differed great- ly in total Se, pH, organic matter content and texture, but produced noticeably simi- lar Se distributions (Figure 3). The pattern of 1% salt-soluble, 20% adsorbed, 40% or- ganically associated, 15% elemental and 30% recalcitrant organic Se or metal sele- nides can thus be considered characteris- tic for the agricultural soils of Finland. A rather similar Se distribution was report- ed for a silty clay loam soil from Rotham- sted, UK (Coppin et al. 2006). The small size of the salt-soluble Se fraction consisting of soluble selenate is in accord- ance with the thermodynamic equilibria of Se speciation showing selenite, elemental Se or metal selenides predominating under acidic and reducing conditions (Koljonen 1975, Elrashidi et al. 1987). The pool of 0 10 20 30 40 50 60 Soluble Adsorbed Organically associated Elemental Recalcitrant organic / metal selenides D is tri bu tio n of s oi l S e (% ) Figure 3. Distribution of the soil Se reserves in the field soils of Finland into soluble (KCl), adsorbed (K2HPO4), organically associated (NaOH), elemental (Na2SO3), and recalcitrant organic / metal selenides (NaOCl) according to the Wright et al. (2003) sequential extraction procedure. The bars are averages of 11 soils ± standard deviation. 30 MTT SCIENCE 20 adsorbed selenite recovered by phosphate buffer extraction remained low in compar- ison to the prominent role it has achieved in explaining the poor bioavailability of Se in soil. The NaOH-extractable organi- cally associated Se, which in contrast has attracted very little attention, stood out as the major fraction. Overall, the predomi- nance of the organically bound Se reserves over the inorganic pool was marked, even though there is some inaccuracy in the se- lectivity of the extraction steps. Mainly, the recalcitrant reduced inorganic and or- ganic Se forms cannot be separated; the organically associated pool may contain some carry-over of the inorganic selenite, and on the other hand all of the inorgan- ic fractions may contain some organically associated Se (Wright et al. 2003). The forms of organically bound Se have thus far not been identified. The reten- tion may occur through microbiological immobilization (Vuori et al. 1994, Stolz et al. 2006) or through the incorporation of Se into the molecular structures of var- ious organic compounds by plants (Sors et al. 2005). Suggestively, Se may also be abiotically occluded or adsorbed onto or- ganic particles (Hamdy and Gissel-Niels- en 1976, Bruggeman et al. 2007). Fulvic acids in particular have been linked to or- ganically associated forms of Se (Coppin et al. 2006). However, the organic associ- ation of Se may alternatively occur indi- rectly via surface Fe oxides or clays (Cop- pin et al. 2009). 3.1.3 Estimating plant-available selenium In EXP2, the applicability of soil extrac- tions in estimating the Se uptake by plants was assessed by comparing the amounts of Se withdrawn from soils by a salt solution (1 M NH4Cl), phosphate buffer (0.1 M K2HPO4, pH 8.0) and ryegrass (III). Rye- grass proved to be the weakest of the three in acquisition of Se (Fig. 4). The two sub- sequent leaf cuttings and roots had taken ND 0 10 20 30 40 50 60 70 80 90 100 Se0 Se+ Se0 Se+ Sand Silty clay S e w ith dr aw al (µ g pe r p ot ) Ryegrass Salt solution Phosphate buffer Figure 4. Withdrawal of Se (µg per pot) by two leaf cuttings and roots of ryegrass and the Se quantity extracted by a salt solution (NH4Cl) and subsequent phosphate buffer (0.1 M K2HPO4, pH 8.0) extraction in sand and silty clay soils without added Se (Se0) or supplied with 0.0175 mg selenate-Se per pot (Se+) the previous year. The bars are averages of three replicates ± standard deviation. ND = no data, due to contamination of the leaf samples. MTT SCIENCE 20 31 around 2 µg Se from those pots not fertil- ized with selenate the previous year and roughly 3 µg Se from pots supplied with 17.5 µg selenate-Se prior to the preced- ing 3-week oilseed rape experiment. The amount of Se recovered with the salt solu- tion was on average 2.4 times that found in the plant. In the soil columns of Good- son et al. (2003), the cumulative Se with- drawal by plants was likewise less than 50% of the salt-soluble Se fraction. With- in the Se treatments, no significant differ- ences between the sand and silty clay soil were found in the total Se uptake by rye- grass or the recovery of Se by the salt so- lution. However, the residuals of supple- mental selenate-Se increased significantly Se both in the plant and in the salt-solu- ble fraction in soil compared with the con- trol not fertilized with Se. The simple salt extraction was thus relatively sensitive to small fertilization-induced changes and ef- ficient enough to give a good estimation of the immediately bioavailable soil Se pool. As for the phosphate buffer-extractable Se, the relationship with plant uptake was not as straightforward as with the salt solution, since the phosphate buffer dissolves non- labile adsorbed Se fractions. The clay soil was considerably richer in the adsorbed Se reserves than the sand. Thus, the plant uptake in the clay corresponded to mere- ly 4 ± 1% of the amount of Se acquired with the phosphate buffer, whereas in the sand the corresponding proportion was 18 ± 4%. The phosphate buffer extraction thus appears less suitable for estimating the plant-available Se than the salt solution, but clearly more suitable in monitoring the potentially phytoavailable pool of Se. In estimating the availability of an element to plants by soil extractions, a good corre- lation between the plant and soil analy- ses is more important than the quantita- tive difference between them (Sillanpää 1982). Since different soil features may have contrasting effects on the amounts of elements acquired by the plant and by the extractant, the most applicable meth- od would likely vary according to the soil properties (Sillanpää 1982, Dhillon et al. 2005). The ability of various extractants to imitate plant Se uptake has been studied, with inconsistent results (e.g. Williams and Thornton 1973; Sippola 1979; Wang and Sippola 1990; Goodson et al. 2003; Dhill- on et al. 2005; Zhao et al. 2005). Factors affecting the availability of Se in the sol- uble pool, such as the amounts of anions competing for plant Se uptake (Hopper and Parker 1999) also need to be consid- ered. Weng et al. (2011) found a major pro- portion of salt-soluble Se being incorporat- ed in colloidal-sized organic matter, the mineralization of which seemed to control the availability of Se. 3.2 Behaviour of added selenium in soil 3.2.1 Monitoring short-term sorption In EXP1, the distribution of the added Se between the salt-soluble and adsorbed pools in soil was closely followed through- out the 10-week growing period of wheat, using the first three steps of SEP1 (III). In EXP3 and EXP4, in contrast, sequential soil extractions with a salt solution and phosphate buffer were carried out only at the end of the experiments (IV). The out- comes of the soil analyses were, however, uniform in all studies, showing that most of the added selenate persisted in salt-sol- uble form in soil kept moist but drier than at field capacity (Fig. 5). In EXP1, 84 ± 8% of the ample addition of 0.1 mg SeO4 2--Se kg-1 soil was recovered in the salt-soluble pool in the first sam- pling and 72 ± 1% in the final sampling. The decrease was mostly accounted for by the Se uptake of wheat. The Al oxide-asso- ciated Se fraction showed an increase cor- responding to 13 ± 3% of the amount of Se added at the first sampling, after which no further increase was detected. The Se concentration of the Fe oxide-associated 32 MTT SCIENCE 20 pool fluctuated during the monitoring, but even at its highest level the increase was less than 10% of the Se addition. At the end of EXP3, 60−70% of the se- lenate added was recovered in salt-solu- ble form in the sand and on average 40% in the peat soils, where more Se was with- drawn by the wheat. Furthermore, in all soils the proportion traced as adsorbed was merely 14%. In the peat soils of EXP4, the same pattern was repeated; the av- erage recoveries of the selenate added as salt-soluble were around 70% in the low- er and 40% in the higher Se addition lev- el, whereas the proportion recovered as ad- sorbed was roughly 10%. The persistence of the added selenate in the acidic soils throughout the 1.5–2.5-month experiments contrasted with the hypoth- esis of rapid reduction of selenate to sele- nite and subsequent adsorption onto soil surfaces. However, Yläranta (1983a) found that selenate added to mineral and peat soils from Finland also remained hot-wa- ter-soluble throughout the 3-month incu- bation. In their studies, Vuori et al. (1994) identified four field soil groups differing in the sorption behaviour of selenate, but in most of their soils collected from different parts of Finland, over 50% of the amount of selenate added remained water-soluble after 75 days of incubation. Selenate re- duction can, however, occur very rapid- ly under favourable conditions. Sposito et al. (1991) reported that without an oxygen supply, soluble selenate disappeared from a soil suspension amended with starch with- in 1 week. Since the reduction of selenate is predominantly a microbially mediated process, easily degradable organic matter enhances it markedly (Garbisu et al. 1996, Camps Arbestein 1998, Hunter and Man- ter 2008, Camps Arbestein and Rodríguez Arós 2001). The aeration status of the soil, i.e. the supply of oxygen, controls the re- dox conditions. In wet or waterlogged soil, the redox potential decreases below the stability limit of selenate (Koch-Steindl and Pröhl 2001). In the pot experiments, the soils were kept drier than field capaci- 0 20 40 60 80 100 120 Se3 Se1 Se2 Se3 Se4 Se1 Se2 Se3 Se4 Se2 Se4 Sand Sand Peat Peat EXP1 EXP3 EXP4 R ec ov er y (% ) Plant uptake Soluble in soil Adsorbed in soil Figure 5. Recovery of the added selenate within the plant and as soluble and adsorbed in soil in three individual pot experiments (EXP1, EXP3 and EXP4). The Se addition levels correspond to 0.0025 mg Se l-1 soil (Se1), 0.005 mg Se l-1 soil (Se2), 0.1 mg Se l-1 soil (Se3) and 1 mg Se l-1 soil (Se4). MTT SCIENCE 20 33 ty, which facilitated the persistence of se- lenate. Under field conditions, heavy show- ers may well cause waterlogging during the growing period. As long as the added selenate remains unal- tered, its availability to plants is controlled by factors other than the Se sorption ca- pacity of the soil. After reduction to sele- nite, however, sorption reactions emerge, as evidenced in EXP4, which included a treatment with selenite (IV). In peat soils fertilized with selenite, some of which con- tained artificially constructed Fe hydrox- ide surfaces, merely 2% of the selenite-Se addition was recovered in salt-soluble form after the 6-week study. In pure and mod- erately Fe- enriched peat, the recovery in the adsorbed fraction was not much high- er, however, since it contained roughly 5% of the selenite addition. In peat soils am- ply enriched with Fe, in contrast, around 70% of the selenite addition was acquired as adsorbed. High recovery of the added selenite (92 ± 15%) in the total Se analy- ses of the unplanted soils revealed that the selenite added was retained by the soil in the pure and moderately Fe-enriched peat as well. It thus seems that in soils rich in oxide surfaces, adsorption occurs by ligand exchange, but evidently in organic soils an alternative, yet unknown, mechanism of retention is operating. 3.2.2 Fertilization-induced long- term changes in soil selenium fractions The set of paired MTT Agrifood Research Station samples was used in examining the residual fertilizer Se accumulated in soil over a 13-year period (II). The paired sam- pling scheme and an additional verification of the accuracy of the resampling location by checking the physical similarity of the soils within a sample pair provided control over the heterogeneity of soil, thus render- ing tracing of the small amount (27–67 g ha-1) of residual Se meaningful. The draw- back of the careful preselection was that it biased the data. The previous study by Yli-Halla (2005) evidenced, however, that spatial variability would have overpowered the small temporal changes induced by fertilization. Selection of representative sample pairs from the organic soils proved to be diffi- cult. Two of the four soils chosen for the study, JKA20 and PPO2, showed mis- matched properties within the pair, JKA20 to the extent that it had to be omitted. Thus, the data on organic soils remained limited and showed no statistically signif- icant changes in any of the soil Se frac- tions between 1992 and 2004. The set of five mineral soils, in contrast, was found satisfactory in terms of compatibility of the paired samples. At two locations, howev- er, the Se accumulation obtained by the fractionation procedure was over two-fold higher than the expected value calculat- ed by Yli-Halla (2005), based on detailed records of the crops grown and fertiliz- ers used over the study period. This dif- ference may, to some extent, be accounted for by inaccuracy in the transformation of soil Se concentrations (µg kg-1 DW) into a hectare-based balance and the roughness of the Se balances calculated, due to use of estimates of crop yields and Se concentra- tions. Most likely, the disparity between the measured and calculated balances was due to an uneven distribution of Se in the soil, as influenced by the use of placement fertilization, i.e. localized application of the fertilizer. The data on mineral soils showed small but statistically significant increases from 1992 to 2004 in the Se concentrations of the adsorbed, organically associated and recalcitrant fractions, as well as in the to- tal amount of Se acquired by the SEP2 method (Fig. 6). The changes in the salt- soluble and elemental fractions were not significant, but the soluble Se evidenced a clear trend to decrease. Accumulation of the fertilizer Se in the soluble pool would have been unexpected, since selenate is not 34 MTT SCIENCE 20 a stable species in the acidic and reduc- ing soils of Finland (Koljonen 1975). It may, however, persist for lengthy periods, as Yläranta (1984c) found in field experi- ments. His ample selenate addition showed a residual effect in timothy grass in the sec- ond year after application. Sposito et al. (1991) demonstrated that the reduction of selenate follows the theoretical reduc- tion sequence of nitrate (NO3 -) > selenate > manganese oxide (MnO2) at pH > 5, the reduction of selenate somewhat coinciding with denitrification. Several bacteria are known to be capable of reducing Se (Stein- berg and Oremland 1990, Siddique et al. 2006, Hunter and Manter 2008). Weather conditions likely govern the reduction rate of the added selenate and thus the size of the soluble soil Se fraction. The partitioning of the added Se into sev- eral species was found already by Cary et al. (1967). Darcheville et al. (2008) used a modification of the SEP2 method in ex- amining the retention of added selenite in soil in a 6-d incubation and found that the Se retained was distributed mainly among the adsorbed and organically associated pools, but to a lesser extent also among the elemental and residual fractions. In the mineral soils of Finland, the greatest ferti- lization-evoked increase in Se was found in the organically associated pool, which was closely followed by the recalcitrant fraction embodying organic Se (Fig. 6). This find- ing indicates that much of the Se added is incorporated into the biological cycle. Or- ganic Se compounds are returned to the soil in plant residues, but soil microorgan- isms may likewise produce organically as- sociated Se (Stolz et al. 2006). Darcheville et al. (2008) found that microbial activity increased both the amount of Se retained by the soil and the strength of retention. In the studies of Gustafsson and Johnsson (1992), the selenite added was predomi- 0.06 0.03 0.01 0.84 0.01 0.01 -5 0 5 10 15 20 25 30 35 40 Soluble Adsorbed Organically associated Elemental Recalcitrant organic / metal selenides Total A ve ra ge c ha ng e in s oi l S e be tw ee n 19 92 a nd 2 00 4 (µ g S e kg -1 ) Figure 6. Change in Se concentration (µg Se kg-1 soil) between 1992 and 2004 in the soluble (KCl), adsorbed (K2HPO4), organically associated (NaOH), elemental (Na2SO3) and recalcitrant (NaOCl) fraction of selenate-fertilized mineral field soils. The bars are averages of five soil pairs ± standard error. At the end of each bar is given the P value of pairwise t test at the 0.05 significance level. MTT SCIENCE 20 35 nantly fixed in organic matter by an un- known mechanism. Abiotic retention of Se by organic compounds via chelation or by mineral particles carried within the organ- ic matrix cannot be excluded (Hamdy and Gissel-Nielsen 1976, Coppin et al. 2009). 3.3 Efficiency of plant selenium uptake 3.3.1 Selenium uptake rhythm and distribution of selenium within the plant The 10-week follow-up in EXP1 showed that the Se uptake of wheat continued throughout the period of growth (III). The Se content of wheat was highest at 8 weeks after sowing and decreased by 15% during the last 2 weeks of ripening, when the plant dry mass decreased by 10% as well. In the Se concentrations of the veg- etative organs (roots, leaves and stems), a trend to decrease over time was observed, whereas in the spikes a contrasting trend was shown. The highest Se concentrations, around 6 mg kg-1 DW, were attained in the young leaves and harvest-ready grains, whereas the lowest, around 2 mg kg-1 DW, were found in the roots and stems of the mature plants. The proportion of Se locat- ed in the roots remained low throughout the experiment. During the first half of the growth period, Se accumulated in the leaves and emerging stems and during the last half in the spikes. Clearly, Se was re- mobilized and translocated from the leaves and stems into the filling grains known to act as sinks for carbohydrates and proteins (Evans et al. 1975, Simpson et al. 1983). At harvest, 55% of the total Se content in the wheat was found in grains, 10% in the head chaff, 15% in the stems, 15% in the leaves and 5% in the roots (Fig. 7). In EXP3, the wheat roots were not collect- ed, wherefore the Se distribution within the plant could be determined only for the aboveground plant mass (IV). However, the results demonstrated a similar high ef- ficiency in the translocation of Se into the grains, since roughly 50% of the plant Se was recovered in the grains, 35% in the leaves