Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A comparison of soil microbial communities in adjacent forest types that differ in nutrient cycling rates Leckie, Sara Elizabeth 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0261.pdf [ 6.91MB ]
Metadata
JSON: 831-1.0075021.json
JSON-LD: 831-1.0075021-ld.json
RDF/XML (Pretty): 831-1.0075021-rdf.xml
RDF/JSON: 831-1.0075021-rdf.json
Turtle: 831-1.0075021-turtle.txt
N-Triples: 831-1.0075021-rdf-ntriples.txt
Original Record: 831-1.0075021-source.json
Full Text
831-1.0075021-fulltext.txt
Citation
831-1.0075021.ris

Full Text

A C O M P A R I S O N O F SOIL M I C R O B I A L C O M M U N I T I E S I N A D J A C E N T F O R E S T T Y P E S T H A T D I F F E R I N N U T R I E N T C Y C L I N G R A T E S by S A R A E L I Z A B E T H L E C K I E B.Sc. M c G i l l University, 1998 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y OF G R A D U A T E S T U D I E S (Department of Forest Sciences) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A Apr i l 2003 © Sara Elizabeth Leckie, 2003 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of forc<=>\ Sciences The University of British Columbia Vancouver, Canada Date Apr," I 7 «300?3 DE-6 (2/88) A B S T R A C T Soil microorganisms have fundamental roles in terrestrial ecosystems yet little is known about the composition, activity, and dynamics of these communities. I investigated the soil microbial communities in two common coniferous forest types o f northern Vancouver Island, British Columbia. These two ecosystem types occur adjacently and are similar as coastal temperate coniferous forests with the same types of processes occurring. They differ, however, in nutrient availability and productivity. They thus provide an interesting contrast for exploring variability in forest soil communities and potential links between the organisms and soil processes. Microbial community composition was measured using several cultivation-independent approaches: denaturing gradient gel electrophoresis ( D G G E ) , ribosomal intergenic spacer analysis (RISA), internal transcribed spacer (ITS) and phospholipid fatty acid ( P L F A ) analyses. Although the communities in each forest type were found to be largely similar using D G G E , R I S A , and P L F A analyses, differences were detected using the ITS analysis of the fungal communities. P L F A analysis also detected subtle differences between the forest types in overall composition as well as within particular groups o f organisms. Fungal P L F A s were more abundant in the nutrient-poor C H forests. Bacteria were proportionally more abundant in H A forests than C H in the lower humus layer, and Gram-positive bacteria were proportionally more abundant in H A forests irrespective o f layer. Bacterial and fungal communities were distinct in the F, upper humus, and lower humus layers of the forest floor and total biomass decreased in deeper layers. These results suggest that the microbial communities in these two forest types are similar but that they do differ in detectable ways. These differences may relate to differences in ecosystem process rates, although, in this study, it was not possible to determine cause and effect. K e y W o r d s : bacteria, denaturing gradient gel electrophoresis ( D G G E ) , forest floor, fungi, internal transcribed spacer (ITS), microbial biomass, microbial community, phospholipid fatty acid ( P L F A ) , ribosomal intergenic spacer analysis (RISA) n TABLE OF CONTENTS Page A B S T R A C T .' i i T A B L E OF C O N T E N T S i i i L I S T OF T A B L E S v L I S T OF F I G U R E S v i L I S T OF A B B R E V I A T I O N S v i i A C K N O W L E D G E M E N T S v i i i I N T R O D U C T I O N 1 Rationale 1 Literature review 2 Methodological advances and limitations 3 Forest Type 9 Tree Species 11 Small Scale Spatial Variability 15 Distribution with Depth 16 Temporal Variability 18 Available Resources and p H 20 Biological Diversity of Forest Soils 23 Introduction to the study 25 M A T E R I A L S A N D M E T H O D S 29 Study Sites 29 Sampling 30 Microbial Biomass Carbon 31 Molecular Bacterial Community Fingerprints 32 Molecular Fungal Community Fingerprints 36 Phospholipid Fatty A c i d Fingerprints of Microbial Communities 36 P L F A Biomarkers 37 Statistical Analyses 39 R E S U L T S 42 Forest Floor Properties 42 Microbial Biomass 43 Microbial Community Composition 45 Within-plot Variability 58 D I S C U S S I O N 62 Do C H and H A forests differ in total microbial biomass? 62 i i i Page Do the biomass and composition of forest floor bacterial and fungal communities differ in C H and H A forests? 63 How different is the biomass and community composition in different layers of the forest floor? 72 Do composite samples adequately capture both the average community and the variability of a site? 74 Do C F E , D G G E , R I S A , ITS, and P L F A analyses give similar results for biomass and community composition patterns? 75 Future Research 81 C O N C L U S I O N S 85 L I T E R A T U R E C I T E D 87 iv L I S T O F T A B L E S Page Table 1. Primers used for P C R amplification of portions of the 16S r D N A genes for D G G E analysis of soil bacterial community D N A 34 Table 2. Phospholipid fatty acids used as biomarkers for different groups of microorganisms 39 Table 3. Analysis of variance table for split-plot randomized complete block design, showing the proper F-tests 41 Table 4. Analysis of variance table for split-plot randomized complete block design for multiple observations, showing the proper F-tests 41 Table 5. Moisture and p H of F , upper humus (Hu), and lower humus (H L ) layers in cedar-hemlock ( C H ) and hemlock amabilis-fir ( H A ) forest floors 42 Table 6. Analysis of variance results for nutrient characteristics of F, upper humus (Hu), and lower humus ( H L ) layers in cedar-hemlock (CH) and hemlock-amabilis fir (HA) forest floors 43 Table 7. Analysis of variance results for microbial biomass measurements for F, upper humus (Hu), and lower humus (H L ) forest floor layers in cedar-hemlock (CH) and hemlock-amabilis fir (HA) forests 44 Table 8. Analysis of variance results for total (nmol g"1 dry soil) and percent (% mol) fungal, arbuscular mycorrhizal fungal, and Gram-positive bacterial P L F A s and for the ratio of fungal-to-bacterial P L F A s in F, upper humus (Hu), and lower humus (H L ) layers in cedar-hemlock ( C H ) and hemlock amabilis-fir (HA) forest floors 56 Table 9. Analysis of variance results for total (nmol g"1 dry soil) and percent (% mol) bacterial, Gram-negative bacterial, and actinomycetal P L F A s in F, upper humus (Hu), and lower humus (H L ) layers in cedar-hemlock ( C H ) and hemlock amabilis-fir (HA) forest floors 57 Table 10. Summary of total P L F A s and ratio of fungal-to-bacterial P L F A s for forest and non-forest soils 71 v L I S T O F F I G U R E S Page Figure 1. Relationship between soil microbial biomass measured using chloroform fumigation-extraction (CFE) and total phospholipid fatty acids ( P L F A ) 44 Figure 2. U P G M A cluster analysis of bacterial community D G G E fingerprints from four sites in F, upper humus (Hu), and lower humus ( H L ) layers of cedar-hemlock (CH) and hemlock-amabilis fir (HA) forest floors, using primers 341F-926R 47 Figure 3. U P G M A cluster analysis of bacterial community D G G E fingerprints from four sites in F , upper humus (Hu), and lower humus (H L ) layers o f cedar-hemlock (CH) and hemlock-amabilis fir ( H A ) forest floors, using primers 43F-534R 48 Figure 4. U P G M A cluster analysis of bacterial community R I S A fingerprints from four sites in F, upper humus (Hu), and lower humus (H L ) layers of cedar-hemlock (CH) and hemlock-amabilis fir ( H A ) forest floors 49 Figure 5. U P G M A cluster analysis of fungal community ITS fingerprints from four sites in F, upper humus (Hu), and lower humus (H L ) layers of cedar-hemlock (CH) and hemlock-amabilis fir ( H A ) forest floors 50 Figure 6. P L F A profiles of F, upper humus, and lower layers o f C H and H A forest floors on a nmol g"1 dry soil basis 53 Figure 7. P L F A community composition profiles of F , upper humus, and lower layers of C H and H A forest floors on a % mol basis 54 Figure 8. Principal component plot of P L F A composition data for F, upper humus (Hu), and lower humus (H L ) layers of C H and H A forest floors 55 Figure 9. Principal component plot of P L F A composition data including the 10 H A H u samples that were analyzed separately 60 Figure 10. Principal component plot of P L F A composition data including the four H A H u samples from within 0.25 m 2 that were analyzed separately 61 vi L I S T O F A B B R E V I A T I O N S Forests C H cedar - hemlock forest H A hemlock - amabilis fir forest F layer partly decomposed litter with some recognizable plant structures Hu layer upper humus layer (well decomposed) H L layer lower humus layer (advanced stage of humification) M i c r o b i a l analysis C F E chloroform fumigation-extraction P L F A phospholipid fatty acid analysis P C R polymerase chain reaction D G G E denaturing gradient gel electrophoresis RIS ribosomal intergenic spacer (used for bacteria) ITS internal transcribed spacer (used for fungi) C L P P community-level physiological profiling T G G E temperature gradient gel electrophoresis A R D R A amplified ribosomal D N A restriction analysis T - R F L P terminal restriction fragment length polymorphism Data analysis U P G M A unweighted pair group method using arithmetic averages P C A principal component analysis v i i A C K N O W L E D G E M E N T S I would like to thank the many people who assisted and supported me through this process. Foremost I am grateful to my supervisor, Dr. Cindy Prescott, for allowing me the opportunity to delve into unknown territory and for being a dedicated teacher and mentor. I would like to thank Dr. B i l l Mohn for allowing a young forest ecologist into his lab and for his guidance. I would like to thank Dr. Susan Grayston for graciously allowing me the opportunity to visit the Macaulay Institute in Scotland and for giving me much of her time and energy. I would like to thank Dr. Colette Breuil for her time and insight as a member of my committee. Candis Staley and Solveig Adair provided untiring enthusiasm in the field. Western Forest Products Ltd. provided field accommodation and maps. Debbie Adams at the Nucleic A c i d and Protein Services Unit at U B C dedicated much effort and time to optimizing the GeneScan runs of my samples. Brian Ord exhibited great patience and good humour in teaching me the P L F A technique. Susan Harper and Paula Parkinson in the Environmental Engineering Laboratory, U B C , are thanked for completing the soluble organic carbon analyses. Dr. Tony Kozak provided kind assistance with the statistical analyses. I would like to gratefully acknowledge funding from the Natural Sciences and Engineering Research Council and the Edward W . Bassett Memorial Scholarship in Reforestation. This research was also supported by Forest Renewal British Columbia and Forestry Innovation Investment. I am indebted to the many people in both the Prescott and Mohn lab groups who have provided friendship, discussion, and tremendous support. Josh Neufeld patiently and expertly taught me the molecular biology techniques I used. I would also like to thank Nancy Smith, Gordon Stewart, and Daryl Smith for their ideas and encouragement in the lab. I am happy to have worked with and learned from Kirsten Hannam, David Blevins, Leandra Blevins, Candis Staley, Lucie Jerabkova, Yona Sipos Randor, L isa Zabek, Aimee Taylor, Rod Negrave, and M i k e Van Ham, who all made my experience as part of the Prescott group a positive and fun one. I am grateful to my parents for their understanding and their endless support, encouragement, and love. Finally, my greatest thanks to Yanik Berube for believing in me from the beginning and for his unwavering love, smile, and optimism. v i i i I N T R O D U C T I O N R A T I O N A L E Soil microorganisms are vital components of terrestrial ecosystems that have fundamental roles in decomposition and nutrient cycling as well as beneficial and antagonistic interactions with plants and animals. It is also now clear that soil is one of the most biologically diverse components of terrestrial ecosystems, with the majority of these organisms still undescribed (Torsvik et al. 2002). Due to their small size and the opaque and heterogeneous nature of soil, soil microorganisms are difficult to study. M u c h of the research in soil ecology has been focused on the physiology o f a very small number o f organisms or the measurement o f microbial processes, with limited consideration of the complex communities of organisms responsible for mediating those processes. Thus, little is known about the composition, activity, and dynamics of the microbial communities that inhabit forest soils. Recent methodological advances in molecular biology and biochemistry have helped to renew efforts to explore the composition and dynamics of soil microbial communities (Tiedje et al. 1999). This is being done within the broader context of understanding both how these communities are regulated and how they affect ecosystem processes. It is necessary to first describe how microbial communities are structured and function in space and time, and how they differ in ecosystems that are functionally different. Moreover, the molecular and biochemical methods that have been developed in the last 10 years have not yet been sufficiently assessed in well-designed ecological studies. This research investigated the soil microbial communities in two coniferous forest types that are common on northern Vancouver Island, British Columbia. The forests are both coastal 1 temperate coniferous forests and occur adjacent to one another on the same soil types, but they differ in nutrient availability and productivity. They thus provide an interesting context in which to explore forest soil communities and the links between the organisms and soil processes. LITERATURE REVIEW Traditional soil microbiological methods involve selectively cultivating organisms using specific media and conditions in the lab. Isolates can then be tested for physiological properties and identified and classified using morphological, physiological, biochemical, and genetic assessments. These techniques, however, have allowed only a glimpse of the extent of microbial diversity, as only a small fraction of bacteria are cultivable using current methods (Amann et al. 1995; Hugenholtz et al. 1998). Knowledge of the diversity and physiology of bacteria and fungi is thus biased towards the small percentage of organisms that can be successfully grown in the lab. The concept of microbial biomass was introduced in the 1970s and has been used as a holistic measure of soil organisms. It allows quantification of nutrients in the microbial pool as well as fluxes between the microbial and other pools. Despite suggestions that microbial biomass could be used as an indicator of decomposition and nutrient mineralisation processes, it has been difficult to establish these relationships (Bauhaus and Khanna 1999). The last fifteen years have brought many technological advances in developing cultivation-independent approaches to studying microorganisms and many researchers argue that it is now possible to explore the "black box" of soil microbial communities (Ritz et al. 1994; Tiedje et al. 1999). Molecular and biochemical tools are now being used to estimate the diversity of microbial communities, compare whole communities, and begin to detect differences in function, functional diversity, and activity of these communities. This information is contributing to our understanding of the controls on the distribution and abundance of soil microorganisms, the 2 effect of community structure and activity on ecosystem functioning, and the effect of disturbances on soil communities, and thus ecosystem processes. Methodological advances and limitations Analysis of microbial community composition using cultivation-independent methods generally involves the extraction and analysis of marker compounds from the community o f organisms. Compounds that are useful markers should a) be present in relatively stable amounts in a group of organisms, b) be variable in some measurable way across organisms of interest, and c) not exist in a free state in the soil (i.e. should degrade rapidly in soil upon death of the organism). One suitable candidate is deoxyribonucleic acid ( D N A ) , which is present in all organisms. Phospholipid fatty acids ( P L F A ) , components of cell membranes, have also proved to be useful and informative markers. Assessments of community structure, however, provide little information about the function of the microbial community. A different approach is to assess the potential functioning of the community by studying the community-level physiological profde (CLPP) of whole communities through measurement of the extent to which different carbon substrates can be utilized. The development of methods to extract and purify D N A from soil has enabled the investigation of the genetic diversity of microbial communities. In her landmark work, Torsvik (\990a;b) used reassociation kinetics of soil bacterial D N A to estimate that there was a minimum of 4,000 and up to 40,000 (Dykhuizen 1998) bacterial genome types in a single sample of forest soil. The reassociation kinetics method involves denaturing a sample of double-stranded D N A at high temperature and then allowing that D N A to reassociate in solution. The greater the complexity of D N A in solution, the longer it w i l l take for any two complementary strands to reassociate; the kinetics of D N A reassociation are thus proportional to D N A complexity. This method is not 3 routinely used for measuring microbial diversity because the estimate is quite approximate and the technique is laborious, time-consuming, and expensive (Ogram 2000). However, it became clear from this pioneering work that the diversity of forest soil bacteria far exceeds the total number of bacteria species previously cultivated and identified. Despite improvement in methods of extraction and purification of D N A from organisms in soil, there remain differences in the extractability of D N A from different organisms. Thus, studies of microbial genetic diversity in soil are likely biased to some degree towards organisms that are most easily lysed (Martin-Laurent et al. 2001). Efficiency of cell lysis differs among groups of organisms, based on the composition of the cell membrane as well as location of the cells within the soil matrix (Frostegard et al. 1999). Different methods of D N A extraction have been extensively compared (Zhou et al. 1996; Frostegard et al. 1999; Mi l l e r et al. 1999; Martin-Laurent et al. 2001), although it is difficult to determine the extraction bias of any particular technique. Most molecular methods now routinely used to describe soil diversity are based on work done by Norman Pace's group (Pace et al. 1986; Olsen et al. 1986). They applied Woese's (1990) concept of ribosomal R N A phylogeny to the analysis of natural microbial communities. Ribosomal R N A molecules, the sites of protein synthesis, are essential to all organisms. The r R N A molecules and their genes consist of highly conserved domains interspersed with variable regions; thus comparative analysis of the sequences is a powerful means to infer phylogenetic relatedness among organisms. Methods that have arisen from this work include cloning and sequence analysis of bacterial r R N A genes and a number of fingerprinting approaches for rapidly comparing communities. Borneman and Triplett (1997) found evidence of high diversity of soil bacteria in an Amazonian forest using cloning and sequence analysis of r R N A genes. O f 100 4 sequences analyzed, each was unique. Further attempts have been made to estimate species richness from this type of data, but rarely have enough sequences been analyzed to give a reliable estimate or to detect differences among treatments (Hughes et al. 2001). Sequence analysis has allowed new insight into the composition of soil microbial communities (Niisslein and Tiedje 1998; McCaig et al. 1999; Krave et al. 2002) as there is a large and growing database of rRNA gene sequences to use for comparison. Given the high diversity of soil microbes, however, many sequences often show low similarity with sequences in the database and are added to the database as an unknown bacterium from an environmental sample. There are several molecular fingerprinting techniques that provide a rapid assessment of the community, particularly for comparison or monitoring purposes. Fingerprinting techniques based on rRNA genes include denaturing gradient gel electrophoresis (DGGE) and temperature gradient gel electrophoresis (TGGE), ribosomal intergenic spacer analysis (RISA), amplified ribosomal D N A restriction analysis (ARDRA), and terminal restriction fragment length polymorphism (T-RFLP). A l l of these techniques rely on the use of polymerase chain reaction (PCR) to make multiple copies of a particular region of D N A so that enough of the desired material is available for subsequent analysis. The PCR-amplified genes from the community of organisms are then separated based on length or sequence polymorphism using one of the above techniques, which produces a visual pattern, or fingerprint, of the community. Denaturing gradient gel electrophoresis (DGGE) was first applied to bacterial communities by Muyzer et al. (1993) and has since become a relatively common technique. The method involves extracting total community D N A from soil, amplifying a short region of the 16S ribosomal R N A gene that differs in sequence among organisms in the community, and resolving the mixture of gene fragments based on differential melting behaviour. Given that the fragments are equal in 5 length, melting behaviour is related to the sequence composition - melting w i l l occur more easily in fragments high in adenine and thymidine rather than guanine and cytosine, which have 3 rather than 2 hydrogen bonds. Fragments migrate through a gel with an electric current and along a gradient o f increased concentration of D N A denaturant (urea and formamide). Once a given fragment reaches a concentration of denaturant sufficient for denaturing to begin, the double strands begin to separate and this causes the fragment to cease migrating. The fingerprint is then comprised of different fragments migrating to different points in the gel. Ribosomal intergenic spacer analysis (RISA) and internal transcribed spacer (ITS) analysis are similar but involve fragments of D N A from spacer regions separating R N A genes for bacteria and fungi, respectively. Because the spacer regions do not code for a product, they are highly variable and can be resolved based on length polymorphism on a polyacrylamide gel. These techniques, though used less often than D G G E , have been previously shown to give reliable fingerprints of complex bacterial and fungal communities (Ranjard et al. 2001). Although P C R is a powerful technique, it has several limitations with respect to applications on communities of mixed D N A , such as that extracted from soil (Wintzingerode et al. 1997). The main concern relates to amplification bias. It is known that P C R efficiency varies depending on the sequence being amplified, so there is the possibility that some members of a community w i l l be preferentially amplified, thus biasing the results (Suzuki and Giovannoni 1996). Additionally, variability in the genome size and number of rrn operons (where the 16S r R N A gene is located) among different bacterial species has been shown to influence the P C R amplification of mixed D N A (Farrelly et al. 1995). It is therefore generally believed that PCR-based measures of community composition, whether used for fingerprints or clone library studies, are not 6 quantitative. Relative proportions of sequences from different species may be biased and not reflect the proportions of those species in the original sample. Phospholipid fatty acid ( P L F A ) profiling of soil microbial communities is a phenotypic fingerprinting method, based on variability in the types of fatty acids present in the cell membrane of different organisms. The profile of a sample consists of the abundance of each of the extracted P L F A s and can be compared with other samples using multivariate statistics. Phospholipids are thought to degrade rapidly in soils upon cell death (White et al. 1979; Federle 1986) and total P L F A content has been shown to correlate with other measures of microbial biomass (Zelles et al. 1992; Bailey et al. 2002). Variability in the types of P L F A s present in different organisms has been shown to be a quantitative and robust method for investigating soil microbial community composition and dynamics (Federle 1986). Certain P L F A s are used as markers for particular groups of organisms (e.g. Gram-positive and Gram-negative bacteria, actinomycetes, fungi), however the development of these markers is based on cultivated organisms. Most markers commonly used have not been thoroughly tested and it is known that some P L F A s used for as markers for one group of organisms occur in other groups as well (Zelles 1997, Zelles 1999). Therefore, this method is limited to some extent by cultivation-bias and results must be interpreted with care. The P L F A composition of organisms is also known to vary to a small degree with environmental conditions and activity o f the cells (Vestal and White 1989; Haack et al. 1994). Increases in the ratio of 16:lw7 trans-to-cis isomers have been associated with stress and starvation conditions for bacteria (Guckert et al. 1985; Heipieper et al. 1996). Cyclopropyl P L F A s have also been shown to increase during the stationary growth phase of many organisms as wel l as under stressful growth conditions of low carbon and oxygen concentrations, low p H , and high temperature (Guckert et al. 1985; Ratledge and Wilkinson 1988). Increases in cyl7 :0 and cyl9 :0 , relative to their respective metabolic precursors, 16:1OJ7C 7 and 18:lco7, therefore, may indicate physiological stress rather than a change in the composition of the community. It has been argued, however, that since physiological changes of organisms affects few P L F A s , the use of a large spectrum of P L F A s to profile a community is less sensitive to this type of change than to differences in community composition. Another common approach for characterizing and comparing soil microbial communities is based on the physiological capability of the heterotrophic community as a whole. Community level physiological profiles (CLPP) can be assessed directly on soil samples, by measuring substrate induced respiration (SIR), or C O 2 efflux following the addition of various individual carbon substrates (Degens and Harris 1997). A much less laborious method, first described by Garland and M i l l s (1991), uses Biolog® 96-well microtitre plates, which contain 95 single carbon substrates (and a control with no carbon). Biolog® plates contain the dye tetrazolium violet in each well , which is reduced during respiratory activity, causing a colour formation. Each well is inoculated with soil solution and the formation of colour is measured over time as absorbance on a spectrophotometer. The profile of colour intensity for each carbon substrate is then compared among samples. The Biolog® approach is currently very popular because many samples can be simultaneously analyzed for many different substrates. However, this method depends on the cultivability o f the organisms and profiles may primarily reflect the rapid growth of some organisms (Preston-Mafham 2002). There are also many inherent problems in analyzing and interpreting these data and it is unclear how the microplate technique relates to the actual functioning of microbial communities in the natural systems (Preston-Mafham 2002). While the limitations of the microplate method have been discussed at length in the literature, there has been no critical evaluation of the SIR method of measuring in situ catabolic diversity (Preston-Mafham 2002). Given that none of the above genotypic, phenotypic, or physiological methods 8 can completely describe soil microbial diversity or structure, the use of more than one approach is generally recommended. There are relatively few studies addressing the variability of microbial communities in forest soils using cultivation-independent approaches. Patterns of forest soil microbial communities have been investigated with respect to forest type and plant species, spatial and temporal variability, and chemical properties of litter and soil. A few studies have also addressed the effect of microbial community composition and diversity on ecosystem processes using an experimental approach. Following is a review of what is currently known about how forest soil bacterial and fungal communities are structured and how the dynamics of these communities affect the functioning of forest ecosystems. Forest Type There is evidence that distinct microbial communities exist in soils of different forest types. The degree to which bacteria or fungi are responsible for mediating decomposition and other soil processes has been shown to differ among forest types (Neely et al. 1991; Alphei et al. 1995). Fungal biomass is thought to exceed that o f bacteria in most soils, but particularly so in coniferous forest soils (Kil lham 1994). Fungi are generally more tolerant of acidic conditions than bacteria and can cause a decrease in soil p H through the release of organic acids. Forests with mull humus forms (Green et al. 1993) are thought to be dominated by bacterial and macrofaunal soil processes. M u l l humus forms are characterized by organic matter that is well incorporated into upper mineral soil horizons, presumably due to the movement and activity o f soil fauna. M o r humus forms, on the other hand, are characterized by large accumulations of partially decomposed organic matter on the mineral soil surface and are thought to be dominated by fungi and microarthropods. Actinomycetes (bacteria with a filamentous growth form) are 9 thought to become predominant only in soils of high p H or water stress (Kil lham 1994). Using P L F A markers for bacterial and fungal biomass, Frostegard & Baath (1996) estimated fungal-to-bacterial P L F A ratios for a range of soil types. They found ratios of 0.3 to 0.5 for acid coniferous forest soils, ratios of 0.05 to 0.1 for beech forest soils, and ratios of 0.03 to 0.13 for grassland and garden soils. These ratios of fungal-to-bacterial P L F A s agree with generalizations that fungal biomass is proportionally greater in soils of coniferous forests than broadleaf forests or other systems. Myers et al. (2001) showed that in three temperate forest ecosystems microbial biomass was equal, but the microbial communities, measured by P L F A and C L P P , were distinct. This suggests that there are landscape-level patterns o f microbial community composition and function in ecosystems with different plant communities. Pennanen et al. (2001) examined P L F A profdes along a transect from shore through dunes and young forest into a mature forest, interpreted as a chronosequence of succession. Microbial biomass (as measured by total P L F A s ) decreased towards the forest and was accompanied by a change in the structure of the Gram-negative bacterial community. A n increase in the ratio of fungal-to-bacterial P L F A s was evident in the forest ecosystems. The authors suggest that increases in the C : N ratio of the organic matter and a reduction in available carbon, perhaps exacerbated by lower p H , contribute to the decrease in microbial biomass and the change in community structure. The well-documented importance of mycorrhizal associations with the trees and ericaceous shrubs present in the forest ecosystems may explain part of the increase in fungal-to-bacterial ratios. Interestingly, the vegetation zones tended to reflect distinct microbial communities, rather than an increase or decrease along the transect. Therefore, the successional 10 stage seemed to be less important than the particular ecosystem characteristics of each zone for determining the soil microbial community composition. Along a transect of forest site fertility, Pennanen et al. (1999) found decreases in fungal P L F A s and increases in bacterial P L F A s as site fertility increased. Again, differences were detected in P L F A profiles along the transect, while microbial biomass and respiration did not differ. Whether site fertility was an effect of, or a contributing cause to, the microbial community structure (or both) is unknown. However, it has long been considered that bacteria rather than fungi mediate processes in soils of higher fertility (Wardle 2002). The above studies suggest that there is variability across the landscape in soil microbial communities at a scale at which patterns in vegetation and soil processes can be detected as well . However, with the exception of the study by Myers et al. (2001), none of these studies were replicated. These problems of experimental design make it difficult to determine whether microbial communities are distinct in each forest type or whether there are simply detectable differences among all samples. It is also unknown how the magnitude o f differences among forest types compares to the magnitude o f differences among different ecosystem types (i.e. forest, grassland, bog, tundra, alpine meadow). Tree Species The effect of plant species on soil processes and soil quality has long been of interest to ecologists and foresters. There are many mechanisms by which tree species may affect soils, including rates of nutrient inputs, outputs, and cycling, mycorrhizal associations, alterations of the microclimate, and water relations (Binkley and Giardina 1998). A l l o f these mechanisms are 11 potentially linked to the microbial community. Thus, tree species effects on soil microbes may involve numerous direct and indirect interactions and feedbacks. Relationships among tree species and the soil microbial community have been demonstrated in laboratory and field studies. In a pot experiment using Scots pine, Norway spruce, and silver birch planted in mineral soil, the presence of a tree changed the P L F A and C L P P profiles of the soil compared to a control soil, but there were no differences among tree species (Priha et al. 1999). However, the same experiment in humus showed an additional effect of tree species, with the birch community being the most distinct. Proportions of Gram-positive bacterial P L F A s and fungal P L F A s were greater in the birch microcosm soil than in the other treatments. The soil with birch also had a significantly higher microbial carbon and nitrogen biomass, but no difference in rates of carbon and nitrogen mineralisation. A related field study showed distinct P L F A profiles for humus and mineral soil for the three single-species stands at two sites (Priha et al. 2001). Microbial carbon and carbon mineralisation were higher under birch than spruce and pine. Proportions of Gram-negative bacteria tended to be higher under birch while Gram-positive bacteria tended to be higher under spruce and pine. Other studies have found that birch soils have faster decomposition and are associated with higher rates of soil respiration and enzymatic activities compared to other trees (Bradley and Fyles 1995). B i rch soil has also been shown to have a distinct microbial community compared to spruce (Saetre and Baath 2000). Priha et al. (1999) isolated the direct tree species effects (particularly root and rhizosphere effects) from long-term indirect effects, by growing the tree species in a common soil in pots. Their field study (Priha et al. 2001), however, included both the direct effects of tree roots and activity and the long-term indirect effects of trees on soil development. Birch rhizosphere soil had greater proportions of Gram-positive bacterial and fungal P L F A s than that of spruce and 12 pine in a common soil. However, the results were quite different in the field, as there was a greater proportion of Gram-negative bacteria in soil developing under birch. Saetre (1998) found that the microbial community of birch and spruce soil differed in laboratory incubations. The P L F A s that differed between spruce and birch soils were the same as those that differed between soil samples taken adjacently to birch and spruce trees in a mixed stand (Saetre and Baath 2000). They suggested that these differences in microbial community composition were thus likely related to the tree species' influence on soil organic matter rather than the effects of tree species on soil moisture, light, and ground vegetation. Grayston and Campbell (1996) showed differentiation of C L P P s of rhizosphere bacterial communities between hybrid larch and Sitka spruce, as well differentiation of three forest sites. However, these differences were obscured by a greater effect of soil type and crop history. Another study looking at agricultural species (Grayston et al. 1998) supported the hypothesis that plant species maintain distinct rhizosphere bacterial communities, based on C L P P . The authors suggested that differences in root exudates among plant species may, in part, control the composition o f the rhizosphere bacterial community. The effect of trees on soils and microbial communities may be a slow process, involving the accrual of indirect effects. Priha and Smolander (1997) found no significant tree species effect on soil chemical and microbiological characteristics in 24-year-old stands of pine, spruce, and birch. However, in two 60-year-old stands, there were differences in microbial processes among plots with the different tree species (Priha and Smolander 1999). Soi l p H , microbial biomass N , denitrification potential, and denitrification enzyme activity were highest under birch and lowest under spruce. This suggests that in the short term, one may be seeing the response to site factors 13 rather than the accumulated tree species effect seen in the long term. However, the detection of species effects on soils may be overwhelmed by other site characteristics (Prescott et al. 2000), so differences between young and old stands may simply reflect differences between the sites. Inhibitory compounds present in plant tissue or released by plants into the soil also affect soil microorganisms. Plant secondary metabolites present in litter and in rhizodeposits can affect microorganisms in many ways. Depending on the organisms, different compounds can be used as substrates, inhibit digestive enzymes, precipitate nutritional proteins, and have direct toxic effects on microorganisms (Benoit and Starkey 1968; Field and Lettinga 1992; Fierer et al. 2001). Thus, differences in polyphenol content among plant species seems to have an effect on soil organisms and nutrient cycling processes (Bradley et al. 2000a; Hattenschwiler and Vitousek 2000). Although it can be difficult to prove the specific mechanisms of interaction, the production and variation of secondary metabolites in different plant species seems to be important for soil-plant-microbe interactions and ecosystem nutrient cycling. Tree species l ikely influence the soil microbial biomass and community composition, and thus decomposition, through differences in litter quality. Indeed, microbial biomass has been found to vary up to 10-fold on litter of different herbaceous plants and tree species (Neely et al. 1991). Overall, the evidence suggests that the tree species effect on soil microbial communities is primarily mediated through their influence on organic matter quality, involving indirect and long-term mechanisms. Differences in the quantity and quality of root exudates may also play a role. Soil directly adjacent to roots has much greater microbial biomass and a different composition from bulk soil (Zak et al. 1996), but the relative importance o f rhizosphere and bulk soil microbial communities to ecosystem processes is unclear. 14 Small Scale Spatial Variability Spatial heterogeneity is characteristic of forest ecosystems and it is generally trees or understory vegetation that are thought to provide the spatial structure. Several studies have looked at identifying the spatial scales at which soil microbial communities and activity are structured, and how the structure relates to the distribution of other organisms or resources. Pennanen et al. (1999) found that the humus microbial community (PLFA profile) in a boreal coniferous forest was spatially autocorrelated at distances of 3-4 m (i.e. each humus sample was likely to have a similar community structure to another sample taken within 3-4 m but be different from samples taken further away). Patchiness in the humus microbial community correlated with the locations of trees and understory vegetation. Total microbial biomass was spatially autocorrelated at a smaller scale of up to 1 m. Similar results were found in a study of a young spruce-birch forest where individual PLFAs were spatially dependent at 1 to 11 m (Saetre and Baath 2000). The microbial community structure seemed to be related to the position of spruce trees more than birch trees, with a characteristic patch size of 4-5 m. This scale of spatial pattern differs from the <0.2 m found in an agricultural field (Cavigelli et al. 1995), which could be attributed either to differences between long-lived, widely-spaced trees and annual row crops or to differences in sampling and analytical methodologies. Wilkinson and Anderson (2001) investigated microbial communities in relation to tree locations in a young Norway spruce plantation. Samples taken 0 and 2 m from tree trunks had similar P L F A profiles, but those taken 1 m from tree trunks had a less fungal PLFAs. Fungal biomass (measured as ergosterol content) was found to vary at small spatial scales in the organic layer of a mature Scots pine forest, with 90% of sample variance being accounted for by spatial autocorrelation in a 4 m range (Mdttonen et al. 1999). Spatial variation in fungal biomass 15 correlated positively with soil p H , concentrations of carbon, M g and K and negatively with organic layer thickness. Variability of soil microbial communities within forests seems to correlate with the aboveground position of trees and other vegetation. The distribution of roots in the soil is also important, as rhizosphere communities are known to differ from those in bulk soil. However, few studies have investigated patterns at spatial scales smaller than 10 cm, despite indications that considerable variability exists below that scale (Nunan et al. 2002). For a bacterium, heterogeneity at the scale of millimeters or submillimeters may be equally important for structuring communities and affecting biotic interactions. For example, anaerobic processes are known to occur in aerobic soils due to the presence of anaerobic microsites at all moisture conditions (van der Lee et al. 1999). Similarly, the C : N ratios of substrates actually metabolized by microbes can differ significantly from a measurement made from bulk soil. Distribution with Depth Considerable attention has been paid to the vertical distribution of soil organisms. This has been done both as a surrogate for time in studying decomposition stages and in response to a bias in many studies to the top 10-20 cm of soil. With soil fauna, larger body size tends to correlate with presence in surface layers. For microorganisms, activity is generally higher near the surface, particularly in the organic layers (Fritze et al. 2000). Surface layers are refreshed with substrates through litter inputs, leachate, and rhizodeposition, while deeper layers generally have fewer fine roots and receive mainly decomposition products from above. Fritze et al. (2000) found reductions in fungal, bacterial, and total P L F A s , as well as changes in community composition, with depth. A higher ratio of fungal to bacterial P L F A s was found at the surface and again in the deepest layers (combined B and C horizons of the podzol). Actinomycetes 16 increased proportionally with depth, which was correlated with an increase in p H . Markers for Gram-positive and Gram-negative bacteria did not consistently correlate with depth. Several studies have reported different fungal communities associated with different forest floor horizons, which has been interpreted as a succession of fungi through stages of decomposition (Kjoller and Struwe 1982). Berg et al. (1998) found greater hyphal lengths in litter, rather than the F layer, and a dominance of bacteria in the humus layer. Moisture was positively correlated with fungal hyphal length in all layers, but only in the litter and F layer for bacterial counts. N o relationship was found with temperature. In an Indonesian pine plantation, D G G E profiles of bacterial communities were distinct among litter, fragmented litter, and mineral soil (Krave et al. 2002). Litter samples had the greatest fingerprint complexity (i.e. number of bands) while fragmented litter and mineral soil samples had more bands in common. The authors suggested that differences in organic matter content (and likely composition) are related to microbial community differences among forest floor layers. Trophic interactions also affect the vertical distribution of organisms. In three of four different forest ecosystems, Ekelund et al. (2001) found highest microbial activities in the top 20 cm of soil. In a spruce peat soil, however, highest bacterial abundance was found in lower layers, perhaps due to anaerobic conditions inhibiting protozoan grazers at that depth. In a beech forest mor, there was a slight increase in the abundance of microorganisms (bacteria, fungi, and protozoa) in the B soil layer compared with the A 2 , which may be related to an accumulation of leached substrates. Setala and Aarnio (2002) used 1 5N-isotope techniques to study the vertical stratification of microbes in a Douglas-fir forest floor. They showed evidence that fungi in the 17 litter layer obtained nutrients primarily from that layer and did not obtain significant resources from deeper layers. Microbes in the F and humus layers also obtained nutrients primarily from the respective layer, which is not surprising given that soil bacteria are generally not very motile. The distribution of soil microbes is thus strongly related to depth in forest floor and mineral soils. The composition and variability in the different layers seem to be related to resource availability and quality, moisture dynamics, and trophic interactions. It is also clear that there are substantial microbial populations in soil layers deeper than the top 10-20 cm that are typically studied. Microbial activity in these deeper layers, particularly i f they are highly organic, may play an important role in ecosystem processes. Temporal Var iab i l i ty Seasonal variations in microclimate and resource availability are expected to influence microbial activities in temperate forest soils. Seasonal changes in temperature and moisture availability may pose limits on microbial activity and flushes of resource inputs from litter or root exudates may allow dormant populations to become active. It is thought that the majority o f bacterial cells in soil are dormant; for example, Clarholm and Rosswall (1980) estimated that only 15-30% of bacterial cells were actively growing when conditions were temporarily improved through moisture or carbon increases. In a temperate forest and peatland, bacterial and fungal biomass peaked in autumn and spring (Clarholm and Rosswall 1980). Clarholm and Rosswall (1980) estimated that annual bacterial respiration was roughly equal to that of fungi, which is contrary to the belief that fungi dominate in coniferous forest soils. Bacterial activity, however, seemed to be restricted to brief but intense periods of growth when conditions were favorable. Short-term temporal variation in humus bacterial biomass correlated with rainfall events, even 18 when moisture was non-limiting. They suggested that this was due to leachable substrates present in rain. Krave et al. (2002) found that bacterial D G G E profiles of litter samples showed greater seasonal variation than fragmented litter and mineral soil samples, which showed almost no seasonal variation. Striking seasonal variation in litter samples seemed to relate to seasonal rain cycles, with the two samples taken at the end of the dry season bearing almost no similarity to the eight samples taken during the rainy season. Measurements o f moisture content, p H , ammonium, and nitrate concentrations were also most variable between wet and dry seasons in the litter and much more stable in the mineral soil. Rogers and Tate (2001) found evidence that the bacterial biomass and dehydrogenase activity of a pine forest soil varied seasonally, while the C L P P remained relatively stable. This agrees with other findings that the inactive microbial community can respond to better conditions by growing actively, but that seasonal fluctuations are not so great as to significantly alter the composition of the whole community. Grayston et al. (2001) showed temporal variability in microbial biomass and respiration as well as P L F A and C L P P of the community but the differences among treatments for P L F A profiles were clearly maintained over time. Peaks in microbial activity correlated with seasonal increases in temperature and plant growth. Temperature has been shown to influence the composition of a forest soil microbial community and to affect patterns of soil organic matter decomposition in laboratory incubations (Zogg et al. 1997). A laboratory incubation of spruce forest soil under different moisture regimes, however, had little effect on the original P L F A profile (Wilkinson and Anderson 2001). The authors suggest this reflects resource rather than moisture limitation of the communities and also 19 provides evidence that microbial communities may be stable in fluctuating environmental conditions. Fungal communities have been shown to change in composition in response to short-term (1-week) moisture fluctuations in forest humus microcosms (McLean and Huhta 2000). Variations in moisture, both spatial and temporal, influenced the fungal community more than moisture per se. There was also evidence o f increased fungal species richness with variation in the moisture regime. It therefore appears that microbial communities vary temporally, particularly in response to seasonal or other fluctuations in moisture, temperature, and substrate availability. It is possible that the temporal variability of biomass and activity may differ from that of the soil community structure. However, differences in community structure between ecosystems, sites, or treatments seem to be maintained over time. Therefore, although studies based on a single time-point need to be interpreted cautiously, clear patterns in community composition may be robust with respect to seasonal and temporal fluctuations. Avai lable Resources and p H There is tremendous physiological diversity among microorganisms, although individual species have distinct functions and a limited range of substrates they can use. The quality of litter inputs and character of soil organic matter are thus likely to be important factors for structuring the microbial community. The microbial community itself is also important for determining the character and fate of soil organic matter. Other factors such as p H and inorganic nutrient availability may also influence the presence and activity of different organisms in the soil. 20 Saprophytic soil microorganisms are generally believed to be adapted to nutrient-poor conditions and to have growth limited by carbon. Nitrogen or phosphorus generally limits plant growth in temperate forests. It has been suggested that the immobilization of nitrogen that is common in the early stages of litter decay indicates that microbes require more nitrogen than is available. However, there is little experimental evidence that nitrogen availability limits decomposition rates (Prescott 1995). Indeed, nitrogen fertilization of temperate coniferous forests has often resulted in decreased microbial activity and biomass (Sdderstrom et al. 1983; Smolander et al. 1994; Thirukkumaran and Parkinson 2000), while an increase has been invariably found with additions of simple carbon compounds to soil. Blagodatskaya and Anderson (1998) studied the ratio o f fungal-to-bacterial activity o f mineral soil from acidic and neutral p H spruce and beech forests. They aimed to investigate effects of substrate quality versus p H , although it was not stated what was responsible for the p H differences between the forests sampled. They found clear differences for both species between acidic and neutral forests. There was greater fungal activity and less bacterial activity at p H 3 rather than p H 6. There was also an effect of forest type, but the strength of this effect depended on the p H . Beech forest soil had higher fungal-to-bacterial activity than spruce at low p H , but at high pH, spruce had the higher fungal-to-bacterial activity. There was a larger difference between acidic and neutral forest for beech than spruce. The authors suggested that other studies that did not detect a relationship between p H and fungal-to-bacterial activities had a narrow p H gradient. However, an alternative interpretation is that soil type or site factors are responsible for this p H difference and that these factors have a greater effect than the tree species. Zak et al. (2000) measured the microbial community composition in Populus tremuloides microcosms with different soil treatments. One treatment consisted of solely A horizon soil and 21 a second consisted of A horizon soil mixed with C horizon soil, which reduced nitrogen availability, organic matter, and likely other chemical and microbiological components. Proportions of Gram-positive bacterial P L F A s were generally greater in the mixed soil and proportions of Gram-negative bacterial P L F A s were generally lower. N o differences were seen in actinomycetal or fungal P L F A s . The authors discussed the likelihood that the microbial community composition was not related to nitrogen availability (as intended) but to differences in the initial soils used. However, given that gross mineralisation and microbial immobilization per unit organic matter was significantly greater in the l o w - N treatments, they suggest a potential link between the community composition and soil nitrogen transformations. In addition to effects of substrate quality, the quantity of substrates is expected to influence the structure of microbial communities. There have been suggestions that increases in carbon substrates correlate with increases in certain P L F A s , particularly the monounsaturates, representing Gram-negative bacteria (Bossio and Scow 1998). Gram-negative bacteria also respond rapidly to substrate additions in enrichment studies. Branched fatty acids, indicative of Gram-positive bacteria, have been found to decrease in response to substrate increases in the form of straw incorporation (Bossio and Scow 1998). Griffiths et al. (19996) provided evidence that substrate quantity can be at least as important as substrate quality in structuring a soil microbial community. They tested different loading rates of the same synthetic root exudate mixture to soil in microcosms. Small increases in substrate loading rates had little effect on the microbial community, with small overall increases in biomass (total P L F A s ) . Bacteria tended to increase with increased substrate loading rates up to the highest rate at which there was a decrease. Fungi continually increased and dominated at the highest loading rate. Gram-negative and Gram-positive bacteria responded differently, with a greater increase of Gram-negative bacteria at higher loading rates. 22 Biological Diversity of Forest Soils Forests soils have very high diversity of organisms compared to other habitats (Torsvik et al. 2002). Biotic interactions resulting in competitive exclusion would be expected to prevent diversity reaching these very high levels. Several explanations have been offered to explain this "enigma o f soil diversity". Spatial and temporal heterogeneity are likely important factors as they allow spatial isolation and create an immense potential for resource partitioning. The complex structure of soil results in the isolation of organisms and partitioning of resources in different substructures of the soil environment, at different depths, and in relation to gradients in soil chemical properties and microenvironments. Phenological differences among organisms and the distinction between generalists and specialists could also increase isolation. Temporal heterogeneity in microclimatic conditions and resource availability should also increase diversity, given the ability o f many microorganisms to remain inactive for long periods o f time in poor conditions. Finally, high rates of speciation in bacteria, arising from short generation times, rapid accumulations of mutations, and low rates of extinction, may contribute greatly to the high levels of bacterial diversity described (Dykhuizen 1998; Torsvik et al. 2002). A different spatial perspective may also be useful in understanding how soil microbial diversity relates to the diversity of macroorganisms. Given the very small size of bacteria, the scale we generally consider for a soil sample (~100cm 3) may actually be comparable to the regional scale for macroorganisms (Godfray and Lawton 2001). It has also been suggested that, for microorganisms, local species richness is tightly linked to regional (and even global) species richness. This refers to the early hypothesis attributed to Beijerinck (~1901) that "everything is everywhere" and the environment selects which organisms are active (Godfray and Lawton 2001; Hillebrand et al. 2001; Finlay 2002). 23 While it is now recognized that the diversity of soil microorganisms, as wel l as soil fauna, is high, it is less clear how that diversity relates to the functioning of forest ecosystems. Most literature on the relationship between biodiversity and ecosystem functioning has focused on above-ground parts o f ecosystems, particularly grassland plants. Few studies have experimentally tested the hypothesis that ecosystem processes are inhibited when diversity in soil is reduced. Studies that have addressed this question have found little support for the hypothesis (Degens 1998; Griffiths et al. 2000; Griffiths et al. 2001). Indeed when two different mechanisms were used to manipulatively lower soil microbial diversity (chloroform fumigation and inoculation of sterile soil with serially diluted soil suspensions), there was no clear effect of diversity on soil processes. It has been suggested that in soil, there may not be a relationship between diversity and function both because diversity is so high and because there is thought to be considerable overlap or redundancy with respect to functions. Processes may only become affected in highly altered soils or when whole functional groups are eliminated. It has long been hypothesized, however, that more diverse communities might be resistant to fluctuating environmental conditions and show greater resilience (i.e. faster recovery following a disturbance). Griffiths et al. (2000) showed evidence that resilience may be decreased in soil microbial communities with diversity lowered by chloroform fumigation. However, when diversity was manipulated by dilution, there was no effect of diversity on the resistance or resilience of the community to a second disturbance (Griffiths et al. 2001). Thus, a relationship between diversity and function has not been demonstrated for soil communities, but it seems that the composition of the community is likely important for determining soil process rates. Patterns are beginning to emerge with respect to the structure of microbial communities in forest soils. The abundance and activity o f bacteria and fungi, as well as differences within each o f 24 these broad groups, are being correlated with different types of ecosystems and temporal and spatial dynamics. This information w i l l help in the development and testing of hypotheses about the controls on soil communities and the effects of these communities on ecosystem processes. M u c h of the work reviewed above has been useful as preliminary applications of recently developed methods and has provided a first look at the complexity and importance of soil communities. Most of the studies undertaken so far, however, have insufficient or no replication and there is a need for well-replicated studies in natural systems, using standard comparable methods. There is a need for further testing and comparison of the recently developed methods to gain assurance that the best available methods are being used for addressing particular questions and that we can correctly interpret the information. Whi le the black box of forest soil communities has been opened, the factors that control community composition and activity as well as the links to ecosystem processes remain largely unknown. INTRODUCTION TO THE STUDY In this study, I investigated the forest floor microbial communities in two forest types o f northern Vancouver Island, B C that have markedly different ecosystem process rates. Old-growth western redcedar {Thuja plicata Donn.) - western hemlock {Tsuga heterophylla (Raf.) Sarg.) ( C H type) forests have low nitrogen availability and seedling performance on cutovers is very poor (Prescott et al. 1993). Adjacent stands of western hemlock - amabilis fir {Abies amabilis (Dougl.) Forbes) ( H A type) have high nitrogen availability and good regeneration on cutovers (Prescott et al. 1993). Western redcedar, western hemlock, and Sitka spruce trees on C H cutovers all respond to fertilization, indicating nitrogen and phosphorus limitation (Weetman et al. 1989a; b), while trees on H A cutovers are less responsive to fertilization (Blevins and Prescott 2002). 25 The two forest types occur adjacently in a matrix across the landscape of northern Vancouver Island and thus are exposed to the same climate. Most of the H A stands were established after a large wind storm in 1906 and the two types were initially classified as serai stages of the same ecosystem, based on observations that they did not differ in soil parent material or topography (Lewis 1985). H A stands are generally even-aged, with a sparse understory o f moss and ferns, while C H stands have an old-growth structure with abundant salal (Gaultheria shallon Pursh) in the understory. The forests are therefore quite similar wet coastal coniferous forests with overlap i n tree and understory plant community composition. However, they differ in age, composition, and structure and are dramatically different with respect to nutrient cycling and availability to plants. Thus, these two forest types provide an interesting context in which to study the variability in soil microbial communities within and between forest types and the potential link to nutrient cycling. I hypothesized that the microbial biomass and community composition differs between C H and H A forest floors, although I do not address whether the microbes are responding to different conditions in the forests or are a contributing cause. Differences between C H and H A forests' litter quality and forest floor chemistry may be correlated with differences in the microbial communities. C H forests floors tend to have greater C : N ratios, so fungi may be dominant, but they also have higher p H . Redcedar, present only in the C H forests, has also been previously associated with bacterially-mediated soil processes due to its ability to "pump" calcium, which can increase soil p H (Turner and Franz 1985; Collins et al. 2001). In this study, I could not distinguish effects of tree species from other factors, but differences between the forest types may be due to long-term accrual of direct and indirect effects of tree species. There is also some evidence suggesting that C H forests may be wetter than H A forests and that this greater moisture may inhibit decomposition processes (de Montigny et al. 1993; Battigelli et al. 1994; Prescott 26 and Weetman 1994). Inhibition of decomposition might be mediated through changes in the microbial community structure and/or their activity level in response to excessive moisture. Microbial community composition was measured using several cultivation-independent approaches: D G G E and R I S A analyses for bacteria, ITS analysis for fungi, and P L F A analysis for the whole community. Given the limitations of each method outlined above, the use of several approaches for this study should help to get a more complete and reliable description of the potential community differences between C H and H A forest types. Moreover, all these methods are relatively new and there is a need for further testing and comparison of the different methods in ecological studies. R E S E A R C H QUESTIONS: Do C H and H A forest floors differ in total microbial biomass? Chloroform fumigation-extraction (CFE) and total extractable P L F A s were used to estimate microbial biomass. In addition, extractable D N A was assessed as a possible measure of microbial biomass. Do the biomass and composition of forest floor bacterial and fungal communities differ in C H and H A forests? P L F A analysis was used to estimate relative bacterial and fungal biomass. Communities were profded using D G G E and R I S A for bacteria, ITS analysis for fungi, and P L F A analysis for the whole community. How different is the biomass and community composition in different layers of the forest floor? Three layers of the forest floor were sampled and compared in the two forest types: the F layer, upper-most humus layer and the deepest humus layer above mineral soil. 27 Do composite samples adequately capture both the average community and the variability of a site? Individual samples were usually composited for each site, but at one site, the 10 samples of the upper humus layer were analyzed separately and values were compared with that of the composite sample. Do C F E , D G G E , RISA, ITS and P L F A analyses give similar patterns for biomass and community composition in these forests? Patterns of biomass and community composition between forest types and among forest floor layers were determined using several techniques. Correlation analysis was used to find a relationship between relative biomass estimates derived from the different methods. Patterns of differentiation of forest types and forest floor layers derived from the different fingerprinting methods were also compared. 28 M A T E R I A L S A N D M E T H O D S Study Sites The sites were located within Western Forest Products Ltd. Tree Farm License (TFL) 6 on northern Vancouver Island, British Columbia, in the very wet maritime subzone of the Coastal Western Hemlock (CWHvm) biogeoclimatic zone (Green and Kl inka 1994). The climate is characterized by cool, moist summers and mild, wet winters, with an average annual precipitation of 1900 mm, 70% of which falls mainly as rain in the winter months (October to March). Mean daily temperatures range from 3.3 °C in January to 14.1 °C in August. The area is characterized by gentle topography with elevations no greater than 300 m (above sea-level). Mineral soils are well to poorly drained loamy Humo-Ferric Podzols which overlay unconsolidated morainal and fluvial outwash material (Prescott et al. 1993). Forest floors are up to 1 m thick and are predominantly humimors and lignomors, with well-developed humus horizons and large amounts of decomposing wood. Study sites were established in four adjacent uncut C H and H A forests. C H stands are old-growth, dominated by western redcedar with western hemlock as a co-dominant species and in the understory. These forests are uneven-aged and have a relatively open canopy. Salal dominates the understory with smaller amounts of Vaccinium spp., Rubus spectabilis (Pursh), Blechnum spicant (L.), Cornus canadensis (L.), and the mosses Hylocomium splendens (Hedw.) B . S . G . , Kindebergia oregana (Sull.) Ochyra, and Rhytidiadelphus loreus (Hedw.) Warnst. (de Montigny 1992). H A stands are predominantly second-growth, even-aged forests established after a large-scale wind event in the early 1900s. Western hemlock and amabilis fir dominate the dense canopy. 29 The sparse understory generally consists of Blechnum spicant, Vaccinium spp., and the mosses named above (de Montigny 1992). Sampling Forest floors were sampled in early October 2001. Each o f the 8 stands was sampled within an area of approximately 400 m 2 , always more than 50 m from the.transition between forest types or an edge. Ten sampling points were haphazardly located at least 3 m from one another, avoiding conspicuous mounds or depressions. A small pit was dug at each point and the F layer (Expert Committee on Soil Survey 1987), the uppermost humus (H) layer, and the deepest H layer directly above the mineral soil were sampled using a trowel. Total forest floor depth was measured at each sampling point. The 10 samples of each layer were composited within a plot, resulting in a total of 24 samples. Samples were placed in individual sterile bags and kept cool (~4 °C) during transport and processing. Each composite sample was sieved to < 2 mm to remove large materials (wood pieces, roots, or stones) and to homogenize. Duplicate subsamples were dried at 70 °C for 24 hr (or until constant mass) to estimate moisture content. Moisture was also measured on uncomposited, unsieved samples. Fresh material was used within one week for microbial biomass measurement (chloroform fumigation-extraction) and D N A extraction. Portions of each sample were frozen at -20 °C to be used later for P L F A , p H , carbon, and nitrogen analyses. The p H was measured on 5 g (fresh weight) thawed samples mixed with 20 ml of dFLO using a Hanna Instruments 9025 p H meter. A t one site, within the H A forest, the 10 samples of the upper humus layer were also analyzed individually and an additional 4 samples were collected within 0.25 m 2 , taken from the four sides 30 of an 11th sampling pit. These samples were used to address the degree of spatial variability within the site compared to between sites and to evaluate how effectively the composite sampling reflected the average conditions of the site. M i c r o b i a l Biomass Carbon Microbial biomass carbon was measured using the chloroform-fumigation extraction technique (Vance et al. 1987; Tate et al. 1988). Approximately 30 g (fresh weight) o f each sample was used. Ha l f of each sample was immediately extracted with 50 m l 0.5 M K2SO4 for one hour on a shaker table. Extracts were gravity-filtered with presoaked Whatman 42 filter paper and then vacuum-filtered with 0.45 um Mil l ipore filters. The other half o f each sample was fumigated with chloroform in a sealed dessicator for 5 days. At the end of the fumigation, the dessicators were flushed with air ten times and kept in the fume hood for an additional hour to completely remove all chloroform from the soil. Samples were then extracted and filtered as described above. A l l filtered extracts were immediately frozen for approximately 2 months until analyzed for organic carbon. Total organic carbon was analyzed using the high-temperature combustion method, with the Shimadzu TOC-500 Carbon Analyzer. Microbial biomass carbon was estimated as the difference between fumigated and unfumigated samples and was expressed as mg microbial carbon per gram of dry soil. No correction factor was used. Measurements from unfumigated samples are reported as extractable carbon. Total microbial, bacterial, and fungal biomass were also estimated using P L F A analysis (see below). 31 Molecular Bacterial Community Fingerprints 16S r D N A - Denaturing gradient gel electrophoresis analysis D N A Extraction. D N A was extracted from 400-500 mg fresh weight of each sample (equivalent dry weights). The B io 101 Fast D N A K i t for Soi l (La Jolla, California) was used for D N A extraction and purification, with some modifications from the manufacturer's directions based on empirical determination (as indicated below). This method involves direct extraction of D N A from soil after mechanically lysing cells with quartz beads in a buffer solution. Samples were vigorously shaken using a M i n i Bead-Beater (BioSpec Products, Bartlesville, Oklahoma) for 2.5 minutes at 5000 beats per minute. Although beating for a longer time resulted in a higher yield of extracted D N A , it also caused more shearing which can contribute to P C R artifact formation (Wintzingerode et al. 1997). Moreover, little difference in D G G E patterns was detected in a preliminary trial for determining optimum extraction procedures. After bead-beating, samples were centrifuged for 10 minutes (14000 rpm, Eppendorf Centrifuge 5415C), to maximize the accumulation of solid material into the pellet. A t that stage, the community D N A was in solution and underwent purification. Protein precipitation was followed by ethanol washing of D N A in spin filter columns with a DNA-bind ing matrix. The spin column filters frequently clogged and broke upon spinning, perhaps due to extractable compounds in these organic samples. To circumvent this problem, lOOpl of the supernatant (containing D N A ) was drawn off and further purified, rather than the whole volume. In a preliminary trial, no differences in D G G E patterns were detected between samples using the full volume for purification and those with the 100 pi . D N A was washed twice with the ethanol-salt solution, and eluted in lOOpl of water. 32 D N A quality and quantity was confirmed on a 1 % agarose gel, using ethidium bromide as a nucleic acid-binding stain and imaged using an Alphalmager 1200 (Alpha Innotech, C A ) . D N A was quantified by comparison with a range of known amounts of standard lkb D N A ladder. AlphaEase (version 3.3, Alpha Innotech, 1996) was used to estimate the intensities of ethidium bromide and a calibration curve was constructed. High molecular weight D N A indicated unsheared and, therefore, high quality D N A . A l l samples were diluted to give equal D N A concentrations (10 ng pi" 1), so that an equal amount could be used in all subsequent reactions. Polymerase Chain Reaction amplification. Two primer pairs were used to amplify different regions of 16S r D N A (Table 1). The lengths of P C R products were approximately 585 bp for primer pair 341F-926R and 491 bp for 43F-534R. A GC-r ich clamp, consisting o f 40 guanine and cytosine residues was used on the forward primer of each pair, to ensure a halt of migration of the fragments in the denaturing gradient gel (Muyzer et al. 1993). Polymerase chain reactions each consisted of l u l purified D N A template (lOng), 5pi 1 Ox P C R buffer (Qiagen) (final concentration of 1.5mM M g C b J , 200uM each deoxynucleoside triphosphate, 500 n M each primer, 670 pg ml" 1 bovine serum albumin, and 1.25 units Taq D N A polymerase in a final volume of 5 0 u l Reactions were carried out in a P T C 150 MiniCycle r ( M J Research, Waltham, Massachusetts). Following a simplified hotstart (samples loaded into thermal cycler at 95°C), an initial denaturation step was done at 94°C for 5 minutes. Subsequent cycles consisted of a 1-min denaturation step at 94°C, a 1-min primer-annealing step at 55°C, and a 1-min extension step at 72°C. A t the end of 25 cycles, a final 7-min extension step was used to ensure all P C R products were fully extended. Negative controls containing the same mixture but without any D N A template were included. P C R products were quantified on a 2% agarose gel, as described above 33 for genomic D N A . P C R products were purified using the QIAquick P C R Purification K i t (Qiagen Inc., Valencia,CA). Table 1. Primers used for P C R amplification of portions of the 16S r D N A genes for D G G E analysis o f soil bacterial community D N A . P r i m e r 3 Target Sequence (5' to 3') Specificity Reference site" 341F-GC C 341-357 C C T A C G G G A G G C A G C A G Bacteria Muyzer et al. 1993 926R 907-926 C C G T C A A T T C M T T T G A G T Universal Muyzer et al. 1995 T T 43F-GC 43-63 C A G G C C T A A C A C A T G C A A Bacteria Marchesi et al. G T C 1998 534R 518-534 A T T A C C G C G G C T G C T G G Universal Muyzer et al. 1993 a F (Forward) and R (Reverse) indicate the orientation of the primers with respect to the 16 r D N A sequence. ^Escherichia coli numbering of Brosius et al. (1981). c G C indicates a 40 bp GC-r ich sequence attached to the 5' end o f the primer. The sequence is 5 ' - C G C C C G C C G C G C C C C G C G C C C G T C C C G C C G C C C C C G C C C G - 3 ' Denaturing gradient gel electrophoresis. D G G E was performed using the Bio-Rad D-Code System (Bio-Rad, Hercules, C A ) , with modifications of the protocol of Muyzer et al. (1993). Approximately equal amounts of each P C R product were loaded with 10X loading buffer onto a 6% (37.5:1) polyacrylamide gel in I X T A E buffer (40mM Tris base, 2 0 m M acetate, I m M N a 2 -E D T A , p H 8). Gels were made with a denaturing gradient of 40-65%, where 100% denaturant contains 7.0 M urea and 40% deionized formamide. Standard markers were run on the outside and middle lanes of every gel to allow comparison of fingerprints within and among gels. Electrophoresis was carried out for 16 hrs at 60 °C and 75 V . After electrophoresis, gels were stained using S Y B R Green I nucleic acid gel stain (Molecular Probes, Eugene, OR) and immediately imaged using an Alphalmager 1200 (Alpha Innotech, C A ) . 34 Gel Analysis. D G G E fingerprint patterns were compared using Gel Compar II (Applied Maths, Belgium). Standard markers were used to normalize fingerprints within and among gels. Patterns were analyzed using Pearson's Product Moment Correlation, which gives pairwise percent similarity for the entire fingerprint image of all patterns. This method was used instead of band-matching approaches, to avoid the high subjectivity of identifying and matching individual bands in these complex fingerprints. Dendrograms were constructed using the unweighted pair group method using arithmetic averages ( U P G M A ) . Ribosomal intergenic spacer analysis The same D N A extracts (see above) were used for R I S A . Primers 1406-F (universal, 16S r R N A gene) and 23S-R (bacteria-specific 23S r R N A gene) were used to amplify the spacer region between the 16S and 23S ribosomal subunit genes (Borneman and Triplett 1997; Fisher and Triplett 1999). The forward primer 1406-F was 5' end labeled with the phosphoramidite dye 5-F A M . Reaction mixtures were as above except that they contained 2 u.1 purified D N A template (20ng). Cycl ing was as described above except for an annealing temperature of 58°C and the extension step was 1.5 min in each cycle to account for the longer product length. Controls and quantification were as described above, although no purification of P C R products was done. A RoboCycler gradient 96 (Stratagene, L a Jolla, C A ) thermal cycler was used for R I S A and ITS analysis PCRs . A n automated technique was used to generate R I S A profiles (Ranjard et al. 2001). R I S A fingerprints were resolved on 5% Long Ranger gels (6.0 M Urea, TTE) using an A B I Prism 377 at the Nucleic A c i d and Protein Services Unit, University of British Columbia. Mixtures of 0.5 ul sample, 0.3 ul Blue Dextran/EDTA loading buffer, 0.3 p i Size Standard (GS2500 Rox), and 1 pi deionized formamide were heated at 95 °C for 2 min and put on ice. 1 u.1 was loaded using a 35 porous membrane comb and the gel was run under denaturing conditions for 15 hrs at 4000 V . GeneScan software was used to convert fluorescence data to electropherograms and these data were imported and analyzed using Gel Compar II. Using this technique, very high resolution and extremely precise normalization among lanes could be achieved. Therefore, band-matching and Ochiai 's similarity among samples were used to analyze the fingerprints. Dendrograms were constructed using U P G M A (Sokal and Michener 1958). Molecular fungal community fingerprints: ITS analysis This analysis was almost identical to the R I S A except that fungal-specific primers were used in the P C R . Primers used were ITS1F-F (fungal-specific, 18S r R N A gene) and ITS4 (universal, 28S r R N A gene) (Gardes and Bruns 1993). Reactions included 2 m M M g C l and an annealing temperature of 56 ° C was used. Ge l analysis was as described above for R I S A . For D G G E , R I S A and ITS profiles, dendrograms were constructed including the 14 additional uncomposited samples to determine whether the uncomposited samples clustered with the composite samples of the same type. Phospholipid fatty acid fingerprints of microbial communities Phospholipid fatty acid analysis involves four stages: l ipid extraction, l ipid fractionation, alkaline methanolysis, and analysis by gas chromatography. A l l solvents and chemicals used were of analytical grade. Lipids were extracted using a B l igh and Dyer (1959) extraction, as modified by White et al. (1979) and Frostegard et al. (1991), with minor subsequent modifications. Briefly, approximately 650 mg (fresh weight) of thawed soil was extracted in a single-phase mixture consisting of chloroform : methanol: citrate buffer (including soil water content) (1:2: 0.8 v/v/v). 36 Lipids were fractionated using SPE silica columns (International Sorbent Technology Ltd, U K ) into neutral, glyco-, and polar (phospho-) lipids (Frostegard et al. 1991). The polar l ipid fraction was then subject to a mild alkaline methanolysis to yield fatty acid methyl esters ( F A M E s ) . F A M E s were separated and quantified by gas chromatography (Hewlett Packard E5895 Series II) using splitless injection, helium as a carrier gas, and a polar column, and were identified using standard bacterial acid methyl ester mix (Supelco; Supelco U K , Poole, Dorset, U K ) . Fatty acid nomenclature follows Frostegard et al. (1993). Designations are made in terms of the total number of carbon atoms : number of double bonds, with w indicating the position of double bonds from the methyl end of the molecule. Double bond configurations cis and trans are designated by ' c ' and ' t ' respectively. Prefices ' i ' and ' a ' refer to iso and anteiso branching respectively, 'br' refers to a methyl branch at an unknown position, ' l O M e ' indicates a methyl group on the 10 t h carbon atom from the carboxyl end o f the molecule, and 'cy ' refers to cyclopropane fatty acids. P L F A biomarkers The composition of phospholipid fatty acids is diverse among microorganisms and certain P L F A s are commonly used to mark different groups of organisms. However, the biomarkers used are not always consistent among studies. Bacteria commonly contain a variety o f normal, straight-chain, monounsaturated, branched-chain and cyclopropane fatty acids (Lechevalier and Lechevalier 1988). Polyunsaturated fatty acids are rare in bacteria. Monounsaturated fatty acids are generally lw7, but double bonds also occur at other positions (5,9, 11) (Lechevalier and Lechevalier 1988). Gram-positive bacteria are characterized by having a high proportion of branched-chain fatty acids, which are less common in Gram-negative bacteria (O'Leary and Wilkinson 1988). Gram-negative bacteria generally have straight-chain fatty acids with even 37 numbers of carbon (Wilkinson 1988). The fatty acid 18:2w(9,12) is commonly used as a marker for fungi. The former has been shown to correlate well with ergosterol, a well-established marker used to estimate fungal biomass (Frostegard and Baath 1996). The P L F A 10Mel8:0 is used as a biomarker for actinomycetes (Lechevalier 1977; (Frostegard et al. 1993) and 16:lo)5 is used for arbuscular mycorrhizal ( A M ) fungi (Olsson 1999). It is generally considered more robust to use a group of P L F A s as markers for large groups, as there is considerable variability within groups. Table 2 shows the particular P L F A s chosen to represent the various groups of organisms, based on those found in other investigations (Federle 1986; Frostegard et al. 1993; Frostegard and Baath 1996). Only P L F A s with less than 20 carbons are included, as those with greater chain length are thought to arise predominantly from plants. For the bacterial markers, only those found in considerable quantities in my samples (>~10 nmol g"1 dry soil) were used. One very abundant P L F A , 18:lco7, is commonly used as a marker for Gram-negative bacteria although it is also found in arbuscular mycorrhizal fungi. 38 Table 2. Phospholipid fatty acids used as biomarkers for different groups of microorganisms. Group of organisms P L F A s Reference Gram-negative bacteria 16:lw7c 1 6 : l « 7 t (Wilkinson 1988) Gram-positive bacteria cy l7 :0 18:lw7 i l 5 :0 al5:0 i l 6 : 0 al7:0 (O'Leary and Wilkinson 1988) br(2)17:0 10Mel7:0 Total bacteria Above plus, (Lechevalier and Lechevalier 1988) Actinomycetes Fungi 15:0 cyl9 :0 minus 10Mel7:0 10Mel8:0 18:2u(9,12) (Lechevalier 1977) (Federle 1986; Frostegard and Baath 1996) (Olsson 1999) Arbuscular mycorrhizal fungi 16:1 a 5 Note: The ratio of fungal-to-bacterial P L F A s was also calculated without 16:lco7c in the measure of bacterial biomass. This was to make the ratios comparable to those originally presented by Frostegard and Baath (Frostegard and Baath 1996). Statistical Analyses To test for differences in microbial biomass carbon, p H , moisture, extractable carbon and P L F A biomarkers, I used analysis of variance of a split-plot experiment in a randomized complete block design, with forest type as the whole-plot factor, and forest floor layer as the subplot factor (Tables 3 and 4). Site (blocks) and forest floor layer were considered random effects and forest type was considered a fixed effect. Depth of forest floor effect was analyzed as a completely randomized block design. Data for fungal P L F A s and Gram-negative bacterial P L F A s were log-transformed, extractable carbon data were square-root-transformed, and microbial carbon data were inverse square-root transformed to meet the assumptions of homogeneity of variance and normality. A n alpha value o f 0.05 was considered significant for all analyses. Analysis of variance was followed by pairwise t-test comparisons of the least-square means, with the alpha 39 level (0.05) adjusted for the number of comparisons made using Bonferonni's adjustment. Statistical analyses were performed using S A S (version 8.2, S A S Institute Inc., 1999, Cary, N C ) To address variability among the uncomposited samples within the plot, I posed the following questions with respect to the estimates of microbial biomass, the abundance of marker groups of P L F A s , and other forest floor measurements: 1. Does the value from the composite sample differ from the population as sampled by 10 separate samples? (t-test) 2. Does the variance for the four sites differ from the variance of 10 samples within the one plot? (F-test) 3. Does the mean and variance differ between four samples from within 0.25 m and 10 samples from within the whole plot? (t-test) Principal component analysis was used to explore the large P L F A data set for similarities among samples in their overall P L F A profde. Proportional P L F A data (% mol) were log-transformed prior to analysis to remove the effect of skewed distributions. The proportion data represents a closed data set, so it is subject to problems associated with compositional data. Because the data are summed to a constant, in this case 100%, an increase in one P L F A necessitates decreases in others, which could give rise to spurious negative correlations (Aitchison 1986). Aitchison (1986) proposed the log-ratio transformation to deal with this problem. Few studies use this transformation (Frostegard et al. 1997) despite ecologists being aware of the problem of closure with compositional data for several decades. Thus, principal component analysis was conducted on the compositional (and log transformed) data as well as on log-ratio-transformed data. Correlation matrices were used and the plots were scaled to show maximum variation of the samples. P L F A s with missing values were used as supplementary data that did not influence 40 sample scores. Data from the uncomposited samples were also used in the principal component analysis for the P L F A data. The uncomposited samples were used as supplementary samples and thus did not influence the analysis. Principal component analyses were done using C A N O C O (Version 4, Microcomputer Power, Ithaca, N Y ) . Table 3. Analysis of variance table for split-plot randomized complete block design, showing the proper F-tests. Source of variation Degrees of freedom Mean Square Ftest Site (block) 4-1=3 M S B M S B / M S E 2 Forest 2-1=1 M S F . M S F / M S E I * Site x Forest (Error 1) (4-l)(2-l)=3 M S E i M S E I / M S E 2 Layer 3-1=2 M S L M S L / M S E 2 Forest x Layer (2-l)(3-l)=2 M S F X L M S F X L / M S E 2 Site x Forest x Layer (Error 2) 2(4-l)(3-l)=12 M S E 2 (no test) Total (4)(3)(2)-l=23 M S T * provided that there is no significant effect of E l Table 4. Analysis of variance table for split-plot randomized complete block design with multiple observations, showing the proper F-tests. Source of variation Degrees of freedom Mean Square Ftest Site (block) 4-1=3 M S B M S B / M S E 2 Forest 2-1=1 M S F M S F / M S E I * Site x Forest (Error 1) (4-l)(2-l)=3 M S E I M S E I / M S E 2 Layer 3-1=2 M S L M S L / M S E 2 Forest x Layer (2-l)(3-l)=2 M S F X L M S F X L / M S E 2 Site x Forest x Layer (Error 2(4-l)(3-l)=12 M S E 2 M S E 2 / M S S E Sampling Error (4)(3)(2)(10-1)=216 M S S E (no test) Total [(4)(3)(2)(10)]-1=239 M S T * provided that there is no significant effect of E l 41 R E S U L T S Forest floor properties There was no difference in depth of forest floor in C H (lsmean =32.46 cm, 95% confidence interval 28.79-36.60) and H A (lsmean=29.07 cm, 95% confidence interval 26.05 - 32.78) forests. Forest floor moisture at the time of sampling increased in deeper forest floor layers in C H forests but did not differ among layers in H A forests (Table 5). In the lower humus layer, C H forests had higher moisture content than H A . The p H was significantly higher in the F layer than the humus layers in C H forest but did not differ among layers in the H A forests (Table 5). The two forests differed in p H , with C H having higher p H in the F layer. Extractable carbon was greater in F than humus layers (Table 6). The difference between C H and H A forests appeared large but was not significant. Carbon concentrations did not differ among forest floor layers or forest types (Table 6), although there was a site effect. Nitrogen concentrations differed among forest floor layers, with F-layer samples having significantly greater nitrogen concentrations than the upper humus samples (Table 6). Carbomnitrogen ratios showed both a site effect and a layer effect, with F samples having greater C : N ratios than humus samples (Table 6). Table 5. Moisture and p H of F, upper humus (Hu), and lower humus (H L ) layers in cedar-hemlock ( C H ) and hemlock amabilis-fir (HA) forest floors. Interactions between forest type and forest floor layer were significant for both measurements. The means o f four sites are presented, with standard errors of the mean in parentheses. Values with the same letter are not significantly different, within each forest type. Underlined values indicate significant differences between forest types. Forest floor layer moisture (%) p H C H H A C H H A F 76.1 a 76.9 x 4.37 a 3.92 x (1.3) (1.0) (0.09) (0.03) H U 78.4 ab 77.8 x 3.92 b 3.64 x (1.3) (1.0) (0.08) (0.04) H L 80.3 b 77.4 x 3.79 b 3.71 x (1.1) (1.0) (0.09) (0.09) 42 Table 6. Analysis of variance results for nutrient characteristics of F, upper humus (Hu), and lower humus (H L ) layers in cedar-hemlock (CH) and hemlock-amabilis fir (HA) forest floors. The means of four sites are presented. Values with the same letter are not significantly different, within a column for each effect. Forest Type effect Extractable carbon (mg C g"1 dry soil) % C % N C : N C H 0.99 a 50.75 a 1.22 a 41.82 a H A 0.66 a 50.77 a 1.27 a 40.08 a F(p) df=l 6.33 (0.086) 0.08 (0.798) 1.14(0.365) 1.13 (0.365) Layer effect F H U H L F(p) df=2 1.36 a 0.71 b 0.49 b 20.08 (0.0001) 51.28 a 49.48 a 51.52 a 2.45 (0.128) 1.19a 1.31 b 1.23 ab 4.98 (0.027) 43.20 a 39.37 b 40.29 b 4.29 (0.039) Microbial biomass Microbial biomass was estimated using the chloroform fumigation-extraction technique and as total P L F A s . It has also been suggested that total extracted D N A can be used as a measure of biomass (Marstorp et al. 2000). Relative estimates were compared since correction factors have not been developed for any o f these techniques for forests o f northern Vancouver Island. The estimates using C F E and P L F A show a similar pattern, with decreasing biomass in deeper layers (Table 7). A significant positive correlation was found between microbial biomass carbon, as estimated by C F E , and total P L F A (Figure 1). D N A yield, however, did not differ among layers or forest types and did not correlate with the other two measures of microbial biomass (Table 7). 43 Table 7. Analysis of variance results for microbial biomass measurements of F, upper humus (Hu), and lower humus (H L ) layers in cedar-hemlock (CH) and hemlock-amabilis fir (HA) forest floors. Means with the same superscript letters are not significantly different within a column for each effect. C F E Total P L F A D N A (mg C g"1 dry soil) (nmol g"1 dry soil) (lag g"1 dry soil) Forest Type effect C H 2.75 a 1325.2 417 H A 2.53 a 1250.2 484 F(p) df=l 0.77 (0.4449) Site x Forest interaction Site x Forest interaction F=4.14 p=0.0314, at 13.88 0.0005, at site 3, site 3, C H > H A H A > C H Layer effect F 4.00 a 1638.0 a 427 a H U 2.83 b 1297.9 b 452 a H L 1.77 c 927.2 c 473 a F(p) df=2 83.96 (O.0001) 80.12 (O.0001) 2.78 (0.1055) • C H F © HA F T C H Hu V H A H u • C H HI _ HA HI 600 800 1000 1200 1400 1600 1800 2000 2200 Total PLFA (nmol / g dry soil) Figure 1. Relationship between soil microbial biomass measured using chloroform fumigation-extraction (CFE) and total phospholipid fatty acids ( P L F A ) (R 2=0.72, pO.0001) . 44 Microbial community composition D G G E profiles The two different primer pairs used to amplify regions of the 16S r D N A gave broadly similar results. However, primer pair 43F-534R seemed to be less reliable and produced 2-3 intense bands in the D G G E profile that appeared to be artifacts. The position of these bands changed in relation to the standard marker for the same sample run on different gels and thus the top section of the fingerprints containing these bands was not analyzed. There was no detectable clustering based on forest type although distinct clusters separated the F layer samples from the humus layer samples using both primer pairs (Figures 2 and 3). Within the humus cluster, there was separation of lower and upper humus layers using the primer pair 341F-926R (Figure 2). Fingerprints of each forest floor layer showed very little variability and no more than variability between different runs of the same sample (data not shown). R I S A profiles o f bacterial community Bacterial community fingerprints using R I S A showed a similar pattern as that seen with D G G E , but there was much greater variability among samples. Reproducibility of the P C R and the electrophoresis was extremely high, with two runs of the same sample being near identical (data not shown). The R I S A patterns clustered loosely based on F and H forest floor layers, although the discrimination of layers was less clear and there was much greater variability in all the profiles for the R I S A compared to D G G E (Figure 4). There was no clear distinction between the two forest types. The similarity among clustered samples for R I S A was much lower (35-55% similarity) than for D G G E (80-95% similarity). 45 ITS profiles of the fungal community The fungal community fingerprints were differentiated by both forest type and forest floor layer although there was relatively low similarity overall among the clustered samples (Figure 5). Fungal community fingerprints also showed very high reproducibility for runs o f the same or different P C R products from the same sample. The profiles discriminated the F and humus layer samples, within each forest type (Figure 5). There was less distinction, however, between upper and lower humus samples. For the humus samples, there was some overlap between both the forest types and the upper and lower layers. Similarity among clustered samples was 45 - 60%, similar to that seen with R I S A and again much lower than with D G G E . 46 Percent similarity '-ft j -—U1 • : i : MM! -* 1 « 11] * * ? z i kirn, in I -1 .X i 1 H J i 1 1 2 3 4 3 2 4 1 1 3 2 2 3 4 4 3 4 2 4 3 2 1 1 CH HA CH CH HA HA HA CH CH HA HA CH CH CH HA HA HA CH HA CH HI H! HI HI HI HI HI HI Hu Hu CH Hu CH Hu HA Hu HA Hu Hu Hu F F F F F F F F Figure 2. U P G M A cluster analysis of bacterial community DGGE fingerprints from four sites in F, upper humus (Hu), and lower humus (HL) layers of cedar-hemlock (CH) and hemlock-amabilis fir (HA) forest floors, using primers 341F-926R. Only the top portion of the fingerprint was analyzed, as indicated by bracket above images. Each image is labeled with the site, forest type, and forest floor layer to the right. 47 Percent similarity r f . . . . y . J r r . i 3 HA HI 4 CH HI 2 HA HI 3 C H HI 3 HA Hu 4 HA HI 4 HA Hu 2 HA Hu 4 C H Hu 3 CH Hu A -l CH HI 1 HA HI 1 HA Hu 2 CH HI 1 CH Hu 2 CH F 1 HA F 4 HA F 1 CH F 3 CH F 4 CH F 2 HA F 3 HA F 2 CH Hu Figure 3. U P G M A cluster analysis of bacterial community DGGE fingerprints from four sites in F, upper humus (Hu), and lower humus (HL) layers of cedar-hemlock (CH) and hemlock-amabilis fir (HA) forest floors, using primers 43F-534R. Only the middle portion of the patterns was analyzed, as indicated by bracket above images. Each image is labeled with the site, forest type, and forest floor layer to the right. 48 Percent similarity III! I Ml i! Ill 2 HA HI 1 HA Hu 3 CH Hu 4 CH Hu 4 HA Hu 3 HA Hu 2 HA Hu 3 CH HI 2 CH HI 4 CH HI 3 HA HI 4 HA HI 1 CH F 1 CH Hu 3 HA F 4 HA F 1 CH HI 1 HA F 2 HA F 3 CH F 4 CH F 2 CH F 2 CH Hu 1 HA HI Figure 4. U P G M A cluster analysis of bacterial community R I S A fingerprints from four sites in F, upper humus (Hu), and lower humus (H L ) layers of cedar-hemlock (CH) and hemlock-amabilis fir (HA) forest floors. Each sample is labeled with the site, forest type, and forest floor layer. 49 Percent similarity I I) I I I 11 mu II in ii in i i in i II ii II i i 3 2 4 1 1 1 3 4 2 2 4 4 3 3 1 2 4 4 2 2 3 1 1 3 CH CH CH CH HA HA HA HA HA HA HA CH HA HA HA HA CH HA CH CH CH CH CH CH F F F F F Hu F F F Hu Hu Hu Hu HI HI HI HI HI HI Hu Hu Hu HI HI Figure 5. U P G M A cluster analysis of fungal community ITS fingerprints from four sites in F, upper humus (Hu) , and lower humus (H L ) layers of cedar-hemlock (CH) and hemlock-amabilis fir (HA) forest floors. Each sample is labeled with the site, forest type, and forest floor layer. 50 Phospholipid fatty acid profiles P L F A data showed general similarity among all sample types. However, there were clear differences between the forest floor layers and subtle differences between the two forest types. These data can be considered as total amounts (nmol g"1 dry soil) or, alternatively, as proportions of the total P L F A s extracted (% mol), to address the composition of the community irrespective of the total biomass. For P L F A abundance, there was a general reduction in all P L F A s in the deeper forest floor layers but few apparent differences between C H and H A forests within each layer (Figure 6). Considering just the community composition, the overall pattern of relative abundances of each P L F A appear highly similar among either forest floor layers and forest types (Figure 7). In the principal components analysis, the two forest types were discriminated along the second principal component (14% of variability explained) while forest floor layers were discriminated by the first principal component (42.7% o f variability explained) (Figure 8a). There was separation of both the layers and forest types, with minimal overlap o f the six groups. The P C A based on log-ratio-transformed data showed a similar pattern, with clear discrimination of forest floor layers along the first principal component (57% of variability explained) and forest types along the second principal component (11.7% of variability explained) (Figure 8b). In this analysis, however, there was much greater differentiation between the two forest types in the humus samples than the F-layer samples. Analysis of variance results for fungal and Gram-positive bacterial as well as the ratio of fungal-to-bacterial P L F A s are presented in Table 8. Results for general bacterial, Gram-negative bacterial, and actinomycetal P L F A s involved significant interactions between forest type and forest floor layer and are presented separately in Table 9. Fungal P L F A s , both as total amount 51 and as a proportion of the total, decreased with depth and there were significantly more total fungal P L F A s in C H forests than H A (Table 8). P L F A s marking A M fungi were proportionally more abundant in C H than H A forests and in the upper humus than lower humus layer (Table 8). The total amount of A M fungal P L F A s declined in the lower humus layer compared to the surface layers but not differ between the forest types (Table 8). Bacterial P L F A s were proportionally more abundant in the lower humus layer than the other layers in the H A forest type and, in this lowest layer, bacterial P L F A s were proportionally more abundant in the H A forest type than C H (Table 9). The ratio of fungal-to-bacterial P L F A s decreased with depth, and although there was a tendency for the ratio to be greater in C H forests, there was a significant effect at only one site (Table 8). Gram-positive bacterial P L F A s were proportionally more abundant in H A forests than C H and in the F layer than in the lower humus layer. The total abundance of Gram-positive bacterial P L F A s declined with depth (Table 8). For Gram-negative bacteria, the P L F A s decreased in abundance with depth, particularly in the C H forest. A s a proportion of the total, there were no significant differences among layers in H A forests, but in C H forests, there was a higher proportion in the F layer than the humus layers (Table 9). The difference among forest types in the F layer seemed large but was not significant, both proportionally (p=0.0057 vs adjusted p value of 0.0027, equivalent to p=0.10) and as total abundance (p=0.0048 vs adjusted p value of 0.0027,equivalent to p=0.09). The marker for actinomycetes did not meet the assumption of homogeneity of variances and could not be corrected by transforming the data. Data were particularly variable in the C H forests. There was a trend towards increased abundance and proportion of actinomycetal P L F A s with depth, although it was not tested statistically (Table 9). 52 CD > * CD CO CD < I H E o o o o m o C N T - • « -Cr c c G < x H I 0-o o C O o o o o m o C N T -o m cu < I HZ < d K c-K C £ C H Z [ 1 C KZ HZ: c o o co IIP* ft o o o o o m o in CM T " T -o o in o CNI C N o o o o o o o m o in o m C O C N C N t - 1-I O w o o co o o o o o m o in o m C N C N i -IX, 53 54 a) 1.0 0.5 • CH F o HAF T CH Hu V HAHu • CH HI HA HI 0.0 o o -0.5 H -1.0 -i . , , . , 1 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 PC1 (42.7% of variation) b) • • • • T • • T • o T o „ T o • O V V • • V V • • CH F o HAF T CH Hu V HA Hu • CH HI • HA HI -1.5 -1.0 -0.5 0.0 0.5 1.0 PC1 (57% of variation) 1.5 2.0 Figure 8. Principal component plot of PLFA composition data for F, upper humus (Hu), and lower humus (HL) layers of C H and H A forest floors. In a) % mol data were log-transformed and in b) abundance (nmol g"1 dry soil) data were log-ratio transformed. 55 ta •^ J f-i > I. s o s s « S '3D — — "a "« « N •< § Sf ^ s a C3 '£ 5 13 s « = 5 gj-g •S an « O o ft H a x ! £ QJ C3 u .— v o O R W) e © ox c 3 ta <U a H in i-o ta os © 03 O C N co eS .© oo co © o\ CO C N i 00 C N CN uri oo o - - i C N C3 ^3 O N VO CO N O o3 O O O N H O N O NO C N K 3 CO co" S °. © CO o O N 2 § CN —^» O £ ^ < ! X g iO A CO £ 2-> A - S K .3 S o CO ^ 2 § «» _ ta y ^ rj x 2 A 2 £ <=> < co .3 S f f i NO 0 O o CO w ^ o OO o CN w 00 00 _ H 00 CN S & ta C3 r-- O N OO C O Os O N C N SD od C O © C N , 0 03 C N IT) NO i—H 03 O N O N NO j^-' u >> ta ta P ta o o o o V a cS o 00 CO o CO so CO CO CN SO 03 o O 00 o CO rt O o o o o o o V O O N O o © '—J CO o r—1 CN CN <—i CO © © © © V Tf © oo © os o od ^ Oa o oo SO SO T-H CO CN < 00 CO »—1 © © © ^ © C N T3 56 cd 2 « ~ S CD O to S i ) B <T cd w •° * > • .23 CO - H g CO O _o o c C D u. ca I-I CD -4—» > CD CD I cn CD s B w CD 8 6 CD J OH i o CO CO T3 cd -2 •fj o O g co r l CO ' — 1 S co cd <U „ •-i ^ U J ? O W .2 CO fa 3 * § •si <u >-. co CC3 CD "~ P-CD * ^ « s s i • CD CD - f j CD C CD * 3 CD cd I r ° I B • I-H C T3 § § s « c 'co CD -4—» cd CD -3 c CO i - i O £ CD cd fO SX >> a 3 cd ON [o CD < J IS _1 H Pi CO CD •r CD ^ .£ ^ 15 g £ co —— CD 1-.2 § co cu CD W +^  C4_( co O ™ « c C + H cd ° « S 'co C U +^ 0> o £ _ © « c H eg CO C U CU £ o c 3^ CU « 6X1 CU e • a "a o H cu > « CU C S « v O cs .2 "C cu cu Xi "e« o H .2 CU CU ca PQ « i , h © 5 cs ta tt, :-l 2 oo oo cd c n o n m NO x i n ON CN O o CN © cd ON ^ 7^ 2S CN r<-N Q © NO — H P , ON o cd r o LO r o CN NO CN r o o o 00 r o o d C O 00 © © CN o> . - H 0 u o <~i - H 0 ^ r o N ON ^2 ^ cd cd cd o c o 1/0 o o ON o — H CN CN X X >» p CN u o o O o 00 CN CN —-< cd cd 00 © r o •^ r © ' 00 X X X O CO 00 r o CN CN 1^-S H fcs. M CN © CN NO NO NO C O r o NO NO © ^ — ' c/0 ^ — ' N ^ c^ C O w 57 Within-plot variabi l i ty To address whether the composite samples were representative of uncomposited samples, uncomposited samples were analyzed separately for one sample type ( H A Hu). Although there is no replication of the test, it provides some insight into the variability within a site. For estimates of microbial biomass, P L F A markers, and other forest floor measurements, there were no differences between the value of the composite sample and the mean value for the 10 uncomposited samples, except for p H . The composite p H value was greater than the mean of 10 within-plot samples. For all data except actinomycetal P L F A s , the variance was greater within the one site than among four sites (composite sampling). There were no significant differences in mean values for four samples taken within 0.25 m 2 and 10 samples from the whole plot. The variance, however, was unequal between the two groups of samples for microbial biomass carbon and extractable carbon. For microbial biomass carbon, the variance of the four close samples was greater than that of the 10 individual samples (p = 0.01) and the opposite was true for extractable carbon (p=0.04). For D G G E fingerprints, there was high similarity among all samples from the same layer, regardless of forest type or site. The 14 uncomposited samples were similar to the other humus samples, regardless of sampling location. The four samples from within 0.25 m 2 tended to cluster, but not consistently with both primer sets. The variability among the 10 uncomposited samples was similar to that among the composited samples clustered by forest floor layer. The same pattern was found for bacterial community R I S A fingerprints, although the variability was greater. 58 For ITS fingerprints, the uncomposited samples were similar to each other and clustered with the composited humus samples from the H A forest type. Variability among these 10 samples was similar to that among the other clusters of forest floor layer and type, indicating that there was considerable variability among the 10 samples. The four samples from within 0.25 m 2 did not cluster tightly together. The extra 14 within-plot samples were used as supplementary sites in the P C A of the P L F A data, so they did not influence the analysis but could be plotted with the results of the main samples. In general, the ten samples fell within the group of H A upper humus samples (Figure 9). They showed more variability than the four composite samples for the four sites, but did not generally overlap with the other samples types. The four samples from within 0.25 m 2 showed greater variability on principal component 1, with two and one samples overlapping with the H A - F samples in the composition analysis and log-ratio analysis respectively (Figure 10). This suggests that there was high variability within the single site, but without replication, it could also be due to peculiarity of the particular sampling locations. 59 a) 1.0 -1.0 -0.5 0.0 0.5 1.0 PC1 (42.7% of variation) 1.5 b) 1.0 -0.5 0.0 0.5 1.0 PC1 (57% of variation) 1.5 • C H F o H A F • C H Hu V H A H u • C H HI • HA HI V HA Hu 10 uncomposited • CH F o H A F • CH Hu V H A H u • CH HI • HA HI V HA Hu 10 uncomposited Figure 9. Principal component plot of P L F A composition data including the 10 H A Hu samples that were analyzed separately. In a) % mol data were log-transformed and in b) abundance (nmol g" dry soil) data were log-ratio transformed. Circle indicates the value for the composite H A H u sample. 60 a) 1.0 -0.5 b) 1.0 0.5 0.0 CM o D- -0.5 -1.0 • CH F o H A F T CH Hu V H A H u • CH HI • HA HI V HA H y 4 uncomposited -1.0 -0.5 0.0 0.5 1.0 PC1 (42.7% of variation) 1.5 • • • • T • • T • o T ® T o ^ • • o V V • V 1 (v) ^7 • -1.5 -1.0 -0.5 0.0 0.5 1.0 PC1 (57% of variation) 1.5 • C H F o H A F T C H Hu V HA Hu • C H HI • HA HI V HA Hu4 uncomposited Figure 10. Principal component plot of PLFA composition data including the four H A Hu samples from within 0.25 m 2 that were analyzed separately. In a) % mol data were log-transformed and in b) abundance (nmol g"1 dry soil) data were log-ratio transformed. Circle indicates the value for the composite H A Hu sample. 61 DISCUSSION Do C H and H A forests differ in total microbial biomass? Total microbial biomass was not different between C H and H A forests at the time of sampling. Microbial biomass estimated using C F E was very similar to that measured in C H forest floors by Chang et al. (1995). Their measurements were made on four dates (May, July, August, and October 1992), with overall means of 3.8 and 2.1 mg C g"1 dry soil in the F and H layers respectively. They found no significant differences among dates. However, it is still not known i f microbial biomass in H A forests differs seasonally or whether differences between the two forests might be larger during winter or spring months. Microbial biomass carbon in C H and H A forest floors was within the range reported by others for forest humus (Ross and Tate 1993; Bauhaus and Khanna 1999; Priha et al. 2001). The P L F A concentration was similar to that found for spruce forest humus and at the low end of reported values for other forest humus (Baath et al. 1995; Frostegard and Baath 1996; Pennanen et al. 1999; Priha et al. 2001). P L F A concentration of C H and H A F layer and humus was generally 1-2 orders of magnitude greater than values reported for non-forest and forest mineral soils (Frostegard and Baath 1996; Bossio and Scow 1998; Myers et al. 2001; Bailey et al. 2002). D N A concentration was generally an order of magnitude greater than values reported for other soils on a dry weight basis (Harris 1994; Zhou et al. 1996; Frostegard et al. 1999; Bundt et al. 2001), but none of these studies included forest humus which is lighter than mineral soil. Given that rates of nitrogen mineralisation differ in forest floors of C H and H A forests (Prescott et al. 1993) the similarity in microbial biomass suggests that the microbial communities may differ in composition and/or activity levels. Other studies have found that microbial biomass 62 was equal in two ecosystems while microbial community composition differed (Pennanen et al. 1999; Myers et al. 2001). Those studies are similar to the present one in that they investigated forest stands that differed in nutrient cycling process rates and nutrient availability. Most other studies that reported differences in both biomass and composition among sites compared dramatically different or altered systems. Our result illustrates that measurements of microbial biomass do not adequately explain differences in process rates among forests; rather the answer may lie in the composition of the microbial communities. Do the biomass and composition of forest floor bacterial and fungal communities differ in C H and H A forests? Although the microbial communities in C H and H A forest floors were largely similar, there were detectable differences. Profiling by D G G E and R I S A showed similar patterns in the two forest types. ITS fingerprints of fungal communities showed clear differences between C H and H A forests, but overall variability was very high. This suggests that the variability in fungal communities due to forest type were small compared to overall variability among all samples. The forest types were distinguished in the P C A of the P L F A data (Figure 8), but while roughly half of the variability was explained by P C I (separating samples by layer), only 11-14 % of the variability was explained by PC2 (separating samples by forest type). Forest type, therefore, did not account for a high proportion of the variability in the P L F A dataset. There were, however, significant differences in the abundance and proportion of groups of microorganisms, as measured by marker P L F A s . In C H forests, there was a significantly greater abundance of the P L F A 18:2w(9,12), which is representative of fungal biomass. This result is striking because there was no interaction with forest floor layer. C H forest floors have greater C : N ratios than H A forests (Prescott et al. 1993; 63 Keenan et al. 1993), although the data in this study do not clearly show that pattern. A s well , both C H and H A forests are relatively acidic but the F layer in C H forests has a higher p H than H A forests. It has been assumed that fungal biomass is greater in nutrient poor forests because fungi are better able to use poor quality substrates (with a low C : N ratio) and recalcitrant materials than bacteria. However, dominance by fungi resulting from other site factors could also have a role in causing low site fertility. It seems unlikely that greater fungal biomass in C H forests is directly caused by the different tree species in C H and H A forests. Western redcedar, found in C H forests but not H A , has been associated with lower forest floor fungal biomass, and lower ratios of fungal-to bacterial biomass than hemlock, spruce and Douglas-fir (Grayston et al. unpublished). Forest floor and mineral soil under western redcedar has also been found to have lower fungal spore counts, and higher bacterial counts and populations of ammonia-oxidizers compared to hemlock (Turner and Franz 1985). These differences were attributed to the greater acidity of hemlock samples. It has long been believed that soil processes are dominated by bacteria in cedar stands and that this is associated with greater nutrient availability (Turner and Franz 1985; Collins et al. 2001). Amabilis fir, present only in H A stands, has been little studied. Thus microbial communities found in C H forests are not consistent with those seen in other western redcedar forests and may be more related to site factors than effects of tree species. Salal is the dominant understory species in C H forests and is much less abundant in H A forests. Salal has been implicated in the problem of low productivity on C H cutovers, as it is thought to directly compete with trees for nutrients and to potentially cause binding of enzymes and organic nutrients into tannin-complexes, due to the high tannin content of its tissues (Bradley et al. 64 20006). It is possible that the presence of salal and/or it roots may influence the composition of the microbial community as well , but this has not been sufficiently investigated. It must also be remembered that the estimate of fungal biomass does not distinguish among saprotrophic and mycorrhizal fungi, such that the greater fungal biomass of poor sites could simply be a result of only highly mycorrhizal plants being able to grow there. There is strong evidence that mycorrhizal fungi have saprotrophic capabilities (Cairney and Burke 1998a; Cairney and Burke 19986; Chen et al. 2001) and may be short-circuiting the nutrient cycle. I f mycorrhizal fungi mobilize nutrients directly from organic matter, the net result could be a buildup of poor quality humus without hindrance of growth of established plants. Fungi, especially white-rot fungi, are thought to be the main decomposers of lignin, the most recalcitrant component of plant litter and woody debris. Although there is some evidence of bacterially-mediated lignin decomposition, it is thought to be a different and much slower process (Blanchette 1995). The greater fungal biomass in C H forests could be associated with the greater amount of decaying wood in these stands (Keenan et al. 1993). There was also a slightly greater proportion of A M fungal P L F A s in C H forests than H A . Because the difference was very small, it does not explain the greater difference in total fungal biomass, but may relate to the different mycorrhizal associations of the dominant trees in C H and H A forests. Cedar trees have A M mycorrhizal associations while hemlock and fir have mainly ectomycorrhizal associations. Salal, which dominates the understory of C H forests but is much less abundant in H A forests, has ericoid mycorrhizal associations. Salal fine roots (Bennett et al. 2002), and presumably the fungal mycelium associated with them, however, are most abundant in the F layer of C H forests, which does not explain the greater fungal biomass in C H compared to H A forests in deeper layers. 65 Bacterial communities also differed between C H and H A forests. Bacteria, as a group, were proportionally more abundant in the lower humus layer of H A than C H forests, and the proportion of P L F A s marking Gram-positive bacteria was 20% greater in H A forests than C H . These results seem surprising at first because Gram-positive bacteria are generally associated with stressful environments and it may be assumed that nutrient-poor C H forests would be more stressful. Many Gram-positive bacteria can form endospores and are very good at surviving poor conditions. Gram-negative bacteria, on the other hand, generally respond faster to nutrient enrichments (Griffiths et al. 19996) and are thought to be very competitive in nutrient rich environments. This reasoning, however, assumes that growth of bacteria and plants is limited by the same nutrient in these forests. There have been many studies addressing whether carbon or nitrogen limits microbial growth in forest soils, but the answer remains unclear. It has generally been thought that heterotrophic soil bacteria and fungi are carbon-limited (Alden et al. 2001; Ekblad and Nordgren 2002) although there is evidence that soil microbes can be limited by nitrogen (Kaye and Hart 1997; Hart and Stark 1997; Ekblad and Nordgren 2002). I f bacteria are limited by carbon in a usable form then the inorganic nitrogen status of C H and H A forest floors may not be important for structuring the microbial communities. Indeed, Gram-negative bacteria were most abundant in the F layer of C H forests, where soluble organic carbon was also most abundant. Zak et al. (2000) found proportions of Gram-positive bacteria under poplar trees to be higher in non-rhizosphere (where carbon substrates are presumably less abundant) than rhizosphere soil, and in soils with higher nitrogen concentrations. Higher proportions of fungal biomass, however, were associated with rhizosphere soil (Zak et al. 2000). This suggests that greater root biomass, and thus greater proportions of rhizosphere soil in C H forests, regardless of a species effect, could be related to the patterns in bacterial and fungal communities. In this 66 study, samples were sieved to remove roots but the sieved material likely included both rhizosphere and non-rhizosphere soil. Bennett et al. (2002) measured the fine root distribution of cedar, hemlock, and salal in C H forests. They found that the density of cedar roots was not significantly different among the forest floor and upper mineral soil horizons, but that the greatest mean density was in the deepest humus horizon. Hemlock and salal roots were less dense in deeper humus layers and mineral soil than the surface humus layer. If hemlock roots were less abundant in deeper layers in H A forests, then there could be a difference in total root density in different forest floor layers between C H and H A forests. Root distributions of these species may not, however, be consistent on other sites and are likely affected by the presence of other roots, or by environmental and nutrient conditions. Hemlock roots may be distributed at greater depth in H A than C H forests i f C H forests are indeed wetter. It is possible, however, that root density is greater in C H forest floors, at increased depth due to redcedar and at the surface due to the abundance of the understory shrub salal. A greater proportion of rhizosphere soil in C H forests could relate to the greater abundance of fungi and lower abundance of Gram-positive bacteria. Moreover, the concentration of salal fine roots in the F layer of C H forests (Bennett et al. 2002) could be related to the high proportion of Gram-negative bacterial P L F A s in the F layer of C H forests. Food web interactions are important for both structuring biotic communities and influencing soil processes. There is evidence from two soil microcosm experiments that some Gram-positive bacteria become proportionally more abundant in the presence of protozoan grazers while some Gram-negative bacteria become less abundant (Griffiths et al. 1999a; Rann et al. 2002). Moreover, Ronn et al. (2002) demonstrated that the composition of the protozoan community is important for determining the structure of the bacterial community. Although there is evidence 67 that soil fauna communities differ between C H and H A forests (Battigelli et al. 1994), it is unknown whether protozoan abundance or species composition differs. Protozoan grazing of bacteria has also been shown to be important for bacterial turnover, nutrient mineralisation, and plant growth (Brussaard 1998). Protozoan distribution may be limited in the very wet soils of C H and H A forests. If waterlogged soil conditions occur more in C H forests, there may be reduced protozoan numbers, which could affect the bacterial community structure and nutrient mineralisation processes. Nematodes are also important grazers of fungi and bacteria, and differences in feeding habits are exhibited among nematode families and genera. Forge and Simard (2000; 2001) found differences in the abundance of fungivorous and bacterivorous nematodes, but not in the abundance of protozoa, in clearcuts and forests of British Columbia and in response to fertilization of young stands. Clearcuts had greater abundance of bacterivorous nematodes and lower abundance of fungivorous nematodes than forests (Forge and Simard 2000). Fertilization increased the ratio of bacterial to fungal biomass, the abundance of bacterivorous nematodes and the ratio of bacterivorous-to-fungivorous nematodes (Forge and Simard 2001). In these studies, N mineralisation was positively correlated with increased abundance of total nematodes, increased ratios of bacterial-to-fungal biomass, and increased abundance and proportion of bacterivorous nematodes. They suggested that increases in N mineralisation in fertilized stands was related to increased bacterial decomposition and energy flow through bacterial rather than fungal energy channels of the soil food web. However, they also indicate that nematodes may have indirect effects on N mineralisation through altering the structure of microbial communities or that site factors that foster large populations of nematodes may also foster microbial populations that mineralize more nitrogen (Forge and Simard 2000). Battigelli et al. (1994) did not find significant differences in nematode abundance in C H and H A forests, despite a tendency 68 for numbers to be higher in HA. They did not distinguish, however, between fungivorous and bacterivorous nematodes. It is possible that increased rates of nutrient mineralisation and increased nutrient availability are related to differences in cycling through bacterial versus fungal energy channels in C H and H A forests. Although the abundance and community structure of soil food web interactions are not well characterized in C H and H A forests, they are evidently important in soil processes. Given the broad similarity between C H and H A forests, it is important to put detectable differences into context by comparing the community composition with other forests and ecosystems. It was expected that soils of these cool, wet forests with large accumulations of organic matter would be dominated by fungi. However, ratios of fungal to bacterial PLFAs seem to be at the lower end of the range of values reported for other coniferous forest soils (Table 10). Humus layers of C H and H A forests of northern Vancouver Island are much deeper than other coniferous (generally boreal) forests studied in the literature, so the lower humus layer (H L) samples are probably not comparable to values in most other studies. The F layer and upper humus layer, however, are at depths comparable to humus layers in other studies. Several values for coniferous forest humus are lower than those from this study, although they are all from a single study. The lower humus layers in C H and H A forests seem to be more similar to values reported for mineral soil in forests and other ecosystems. This suggests that while fungi may be abundant in surface organic layers, bacteria dominate in deeper layers whether they are mineral soil or humus. Differences detected between C H and H A forests support the idea that different forest types have distinct soil microbial communities, as suggested by Myers et al. (2001). It must be remembered, however, that community composition differences detected in C H and H A forests 69 do not reflect the actual activity levels of the whole microbial biomass or different components. If activity levels of different groups were known, similarities between the two communities may be much greater or much less than the composition alone. 70 Table 10. Summary of the ratio of fungal-to-bacterial P L F A s for forest and non-forest soils. V a l u e s i n b o l d are f rom this study. Ecosystem Soil type Ratio of fungal: bacterial PLFAs Reference Forest Floors poor Scots pine forest F / H layer 0.59 Pennanen et al. 1999 Scots pine forest humus layer (3 cm) 0.56 Baath etal . 1995 Spruce and pine forest humus layer (6 cm) 0.52 Baath et al. 1995 Scots pine forest F / H layer 0.49 Pennanen et al. 1999 Norway spruce forest Organic layer (4cm) 0.43 Siira-Pietikainen et al. 2001 Norway spruce forest F / H layer 0.42 Pennanen et al. 1999 Spruce forest Forest floor 0.42 Frostegard & Baath 1996 Norway spruce forest Organic layer (4cm) 0.39 Siira-Pietikainen et al. 2001 rich Norway spruce forest F / H layer 0.38 Pennanen et al. 1999 Norway spruce (fertile site) Humus 0.33 Fritze et al. 2000 Scots pine (poor site) Humus 0.33 Fritze et al. 2000 Hemlock-fir forest F Layer 0.31 This study Spruce forest Forest floor 0.30 Frostegard & Baath 1996 Pine spruce forest Humus 0.30 Fritze et al. 2000 Cedar-hemlock forest F Layer 0.29 This study Spruce forest (poor site) Humus layer 0.25 Priha etal . 2001 Birch forest (poor site) Humus layer 0.23 Priha et al. 2001 Cedar-hemlock forest Upper humus layer 0.21 This study Pine forest (poor site) Humus layer 0.20 Priha etal . 2001 Hemlock-fir forest Upper humus layer 0.16 This study Birch forest (fertile site) Humus layer 0.15 Priha et al. 2001 Cedar-hemlock forest Lower humus layer 0.14 This study Spruce forest (fertile site) Humus layer 0.13 Priha etal . 2001 Pine forest (fertile site) Humus layer 0.12 Priha et al. 2001 Hemlock-fir forest Lower humus layer 0.07 This study Mineral Soils Norway spruce (fertile site) Mineral soil (A) 0.16 Fritze et al. 2000 Birch forest (poor site) Mineral soil 0-3 cm 0.14 Priha et al. 2001 Scots pine (poor site) Mineral soil (A) 0.13 Fritze et al. 2000 Pine forest (poor site) Mineral soil 0-3 cm 0.12 Priha et al. 2001 Spruce forest (poor site) Mineral soil 0-3 cm 0.12 Priha et al. 2001 Grassland Top 5 cm 0.12 Frostegard & Baath 1996 Birch forest (fertile site) Mineral soil 0-3 cm 0.11 Priha etal . 2001 Pine spruce forest Mineral soil (A) 0.11 Fritze et al. 2000 Spruce forest (fertile site) Mineral soil 0-3 cm 0.08 Priha etal . 2001 Spruce forest Mineral soil (B) 0.08 Frostegard & Baath 1996 Arable field Top 5 cm 0.08 Frostegard & Baath 1996 Grassland Top 5 cm 0.06 Frostegard & Baath 1996 Pine forest (fertile site) Mineral soil 0-3 cm 0.05 Priha et al. 2001 Beech forest Mineral soil (40 cm) 0.05 Frostegard & Baath 1996 Beech forest Top 5cm 0.04 • Frostegard & Baath 1996 Beech forest Top 5 cm 0.04 Frostegard & Baath 1996 Grassland Top 5cm 0.03 Frostegard & Baath 1996 71 How different is the biomass and community composition in different layers of the forest floor? There were large differences in microbial biomass and community composition among forest floor layers that were consistent in C H and H A forests. Bacterial, fungal and total microbial biomass decreased in deeper layers. Only actinomycete biomass showed a tendency to increase in deeper layers. Decreases in microbial biomass with depth have been found in many forest soils (Kj0ller and Struwe 1982; Federle 1986; Berg et al. 1998; Fritze et al. 2000; Ekelund et al. 2001). This pattern may be associated with a decrease in organic matter and root density with depth in mineral soils. In this study, however, the deepest layers (at an average of 30 cm and a maximum of 60 cm deep) were humus and still had significant amounts of plant fine roots (Bennett et al. 2002). Decreases in abundance of microorganisms with depth in these forests are more likely related to changes in the quality of the organic matter, microenvironment, and trophic interactions. In addition, surface layers are supplied with fresh carbon from litterfall and leachate in rainfall while deeper layers receive the products of decomposition from above. In C H forests, moisture increased, p H decreased, and soluble organic carbon decreased with depth. In H A forests, moisture and p H did not differ significantly among layers but soluble organic carbon was again lower in the humus than the F layer. The composition of the bacterial community differed among forest floor layers, with the F layer generally being the most distinct. Based on P L F A markers, Gram-positive bacteria (including actinomycetes) comprised a larger component of the bacterial community in the deep humus layers, while Gram-negative bacteria were proportionally less abundant in the lower humus layer than the F Layer in C H forests. Fritze et al. (2000) found somewhat contradictory results, with greater proportions of Gram-negative bacterial P L F A s in the elluvial layer than the humus above. Their study involved sampling layers down to about 45 cm, although the humus layer was at 72 most 5 cm deep. Patterns of microbial community structure related to depth are likely strongly affected by substrate type (humus type, mineral soil layer) as well as differences in available substrates and moisture as affected by depth per se. Fritze et al. (2000) also found greater proportions of actinomycetes with depth and a reduction in fungal P L F A s with depth, in agreement with the present study. The fungal community composition also differed among forest floor layers as seen in the ITS analysis. In C H forests, the F layer was distinct from the humus layers, while in H A forests the F and upper humus layers were more closely clustered and the deepest humus layer was clustered separately. This difference could be related to differences in mycorrhizal associations and rooting distributions among the plant species in the two forest types. The dominant plants in C H forests have arbuscular, ecto-, and ericoid mycorrhizal plant species, while H A forests are dominated by ectomycorrhizal plant species. P L F A analysis indicated at least small differences in the fungal community by layer, with P L F A s representing arbuscular mycorrhizal fungi being more abundant in upper than lower layers of the forest floor. However, fungal community differences detected with ITS analysis may include both mycorrhizal and saprotrophic fungi and it cannot be distinguished which groups are responsible for the patterns detected. Landeweert et al. (2003) used ITS clone library analysis to identify ectomycorrhizal fungal mycelia in different soil layers of a Swedish boreal forest. They found differences in the community composition in different layers, with 16 of 25 operational taxonomic units (OTUs) being found exclusively in mineral soil layers below the 3 cm deep humus layer. There were also 3 O T U s found exclusively in humus, 3 O T U s found exclusively in the lower elluvial horizon, and 4 OTUs found exclusively in the il luvial horizon. Therefore, differences in fungal functional types, forms of mycorrhizae, and fungal species identities may contribute to the different fungal communities found in the different forest floor layers. 73 Do composite samples adequately capture both the average community and the variability of a site? For the HA forest at one site, the composite sampling seemed to adequately characterize the site compared to the uncomposited samples. The molecular fingerprinting methods generally showed good clustering of the uncomposited samples with the composite samples of the same type. However, the four samples taken within 0.25 m2 did not consistently cluster closely together. This is in agreement with the considerable variability among the four samples found with the PLFA data and suggests high variability at small spatial scales. In fact, two of the four samples clustered with the HA F-layer samples in the PCA of the PLFA data, while none of the 10 main uncomposited samples showed overlap with a different sample type. This potentially high variability at small spatial scales, however, did not mask clear patterns in the communities based on forest floor layer and forest type when composite sampling techniques were used. Because there was no replication of this test of within-plot sampling, it is possible that those four samples were simply from an "unusual" area and do not represent the average variability at that scale. Moreover, the four extra samples were taken from an "11th" sampling point so they are not represented in the composite sample for the plot. The arithmetic mean of the uncomposited samples was not significantly different from the value of the composite for the biomass or proportion of any of the PLFA marker groups. Of the soil properties, only the pH values were significantly different. This suggests that the composite sampling did adequately represent the mean of the plots as measured by analyzing all the samples separately. However, analyzing all samples individually would have captured more variability. The variability within the one HA forest site was greater than the variability among the composite samples from four sites (except for actinomycete PLFAs, which had high and 74 heterogeneous variability in the composite samples). This agrees with other studies that have shown the scale of variability of forest soil microbial communities to be around 3-4 m (Pennanen et al. 1999; Saetre and Baath 2000) - approximately similar to the distances between sampling points in any one plot in this study. The composite sampling was therefore likely adequate to characterize the average conditions in each site. It must be remembered, however, that there likely exists much greater variability at the micro-scale. Most studies looking at spatial variability in soil communities have not looked at scales of less than 10 cm. Given the size of individual bacterial cells, and the heterogeneity of soil, organic matter particles, and distribution of fine roots, the variability in species composition and activity may be high at the scale of pm and mm. For a study trying to inventory species in a site, for example, it would be appropriate to sample at those very small spatial scales. Do C F E , D G G E , RISA, ITS, and P L F A analyses give similar results for biomass and community composition patterns? Microbial Biomass Results from the different analyses generally agree with each other, although only the P L F A and fungal ITS analyses discriminated samples from the different forest types. There was a very good relationship between total extracted P L F A s and microbial biomass carbon measured using chloroform fumigation-extraction. Microbial biomass has been notoriously difficult to measure and it is mainly through agreements among different methods that we can gain confidence in any of the measurements (Martens 1995). Bailey et al. (2002) found a strong relationship between CFE-flush of carbon and total P L F A s and proposed the following general equation to convert P L F A s to the more common measure of microbial carbon: C F E f l u s h = 2.4 (totPLFA) + 46.2 75 Although this study included a range of different mineral soil types, the soils were all low in biomass compared to forest humus. Values of P L F A s were all below 200 nmol g"1 and values of microbial carbon were all below 500 u.g carbon g~'soil, an order of magnitude smaller than the values reported here for humus. The linear relationship presented in Figure 1 had a slope of 3.3 and may be useful as a general equation for forest humus, complementing the relationship demonstrated by Bailey et al. (2002) for mineral soils. While there was good agreement between total P L F A s and chloroform-labile carbon, extracted D N A did not show the same pattern and thus did not seem to be a good measure of microbial biomass in these soils. Marstorp et al. (2000) found a strong correlation between microbial carbon measured by C F E and extracted double-stranded D N A in soils differing in p H and organic matter content and suggested that D N A could be used as a measure of microbial biomass. However, the soils they analyzed were all agricultural soils with low microbial biomass and less than 3% organic matter. Fungal biomass was also very low as measured by ergosterol. The D N A content of fungi per unit biomass is both lower and much more variable than that of bacteria (Harris 1994). Hyphal compartments of fungi can contain a variable number of nuclei ranging from 1 up to about 50 (Carlile and Watkinson 1994), and thus the D N A content is not a good indicator of active fungal biomass. D N A therefore may be a poor measure of total microbial biomass in soils with a higher proportion of fungal biomass. It is also thought that fungal D N A is more difficult to extract from soil than bacterial D N A because of the tough cell wall enclosing fungal cells (Harris 1994). Unfortunately there have been no studies explicitly addressing the efficiency or recovery of fungal versus bacterial D N A from soils. V a n Elsas et al. (2000) found no difference in fungal D N A recovery from mineral soils between a bead-beating procedure and a protocol designed specifically to enhance recovery of fungal D N A from soil. They estimated that both methods induced lysis of 99.9% of both spores and mycelia. 76 Regardless, it appears that for soils with high microbial biomass and/or high fungal biomass, there is no reliable relationship between extracted D N A and chloroform-labile carbon. The fungal component of microbial biomass has long posed a problem for measurement (Newell 1992). Fungi can transfer cytoplasmic material from older portions of the mycelial mass to younger active hyphae and thus a considerable portion of a fungal mass can be inactive empty hyphae, consisting of only a cell wall (Cooke and Raynor 1984). Thus, cytoplasmic components of fungal tissue are likely a better measure of active fungal biomass (Newell 1992). P L F A s , which are components of the cell membrane, are likely dismantled from senescing hyphae and relocated to l iving tissue, making them a reasonable indicator of active fungal biomass. Community fingerprints It has been suggested that R I S A should detect more variability in a sample than D G G E and therefore offer greater resolution when comparing samples (Garcia-Martinez et al. 2002). This is because the ribosomal spacer region has greater variability among organisms than the R N A genes because the spacers do not code for transcribed products and thus accumulate mutations more rapidly. However, R I S A detects length heterogeneity of the spacer but not the additional sequence differences. Moreover, because both D G G E and R I S A depend on P C R , potential biases and limitations at that first level apply to both techniques. Each analysis also depends on detection limitations of the gel and visualizing process. R I S A fingerprints did show more variability and complexity (i.e. number of bands) than D G G E fingerprints, but I do not know i f this was because of differences in the targeted D N A regions or i f it was related to the very different electrophoresis systems used to resolve the banding patterns. Sequencer systems, as used for the R I S A , offer high resolution and, for small fragments, can detect single base pair differences in length (McEvoy et al. 1998). This is considerably greater resolution than 77 conventional agarose or polyacrylamide gel electrophoresis. Given that ITS and R I S A fingerprints showed similar variability and were both run using the sequencer system, it seems likely that differences between D G G E and R I S A were due to the electrophoresis systems used. It is also possible that microbial community patterns are consistent at the different taxonomic levels thought to be distinguished with D G G E and R I S A . For example, i f D G G E profiles resolve at the generic level and R I S A profdes resolve at the species level, the patterns of community similarity among the forest types and layers may be similar at the generic and species level. Two primer pairs were used for D G G E analysis of the forest floor bacterial communities. It has previously been shown that different variable regions of the 16S r R N A gene can give different results in community fingerprinting (Schmalenberger et al. 2003). It was thought that this is due to intraspecies operon heterogeneity detected in some variable regions as well as different specificity of the "universal" primers for different variable regions. D G G E fingerprints using primer pairs 341F-926R and 43F-534R both produced results that distinguished forest floor layers but not forest types. Thus, it appears, in this study at least, that either of these primer pairs was adequate for the purpose of comparing samples. However, one of them, 43F-534R, produced spurious bands and more variable results and thus seemed to be less reliable. Broadly similar results were obtained with the genetic ( D G G E and RISA) and P L F A fingerprinting techniques. Both methods indicated that there were different communities in each forest floor layer and that there was similarity between the two forest types. The P L F A technique, however, distinguished differences between the two forest types with the overall fingerprint as well as for particular groups of marker fatty acids. The P L F A analysis does not, however, detect different components of the fungal biomass, so the molecular ITS technique was 7 8 valuable. Neither technique effectively discriminates between mycorrhizal and saprotrophic fungi, although there is a P L F A marker used for A M fungi. Genetic fingerprinting and P L F A fingerprinting are based on a similar approach but differ greatly in the types of compounds studied and the inherent variability of those compounds. A s described above, regions of D N A are conserved at different phylogenetic levels so it is possible to find regions that are universally common, or include variability at the species or even individual level. P L F A s , as phenotypic products of gene expression, however, are much less variable. Fatty acid analysis was originally developed for aiding species identification (Lechevalier 1977; Lechevalier & Lechevalier 1988) and broad patterns among groups of microorganisms were described. The whole community P L F A profile allows one to infer abundance of those groups. However, it must be noted that those markers have been developed on the basis of relatively few cultured organisms and they may, in themselves, be biased (Zelles 1997). Measures of total microbial, bacterial, and fungal biomass have been tested against independent measures of those populations in a range of soils (Frostegard & Baath 1996; Bailey et al. 2002), but there has not been appropriate testing for the other groups. Comparison of community profiles using multivariate statistics, allows the detection of overall differences, without the use of markers. However, it is possible that physiological differences (rather than community composition differences) would explain some discrimination of samples. Given that few P L F A s are thought to change in abundance to a relatively small degree, it is expected that community composition differences are largely responsible for patterns seen. Community P L F A profiles were analyzed by principal component analysis using both log-transformed data and log-ratio-transformed data, to assess whether the "closure" effect associated with compositional data sets was a concern. The use of multivariate analyses for 79 compositional data sets can often produce spurious results (Aitchison 1986). The broad patterns were similar with the two analyses (Figure 8), however, there was little discrimination of F-layer samples from C H and H A forests and much greater discrimination of lower H samples when the log-ratio transformation was used compare with the log-transformation. Given that the overall pattern of clear discrimination of forest floor layers and less dramatic, but clear, differences between forest types was found with both analyses, it seems that the "closure" effect was not a serious problem with this data set. It has been suggested that D G G E fingerprints represent the "dominant" bacterial community in a sample, but this has not been tested nor has a solid definition of "dominant" been offered (Muyzer and Smalla 1998). However, given that D G G E fingerprints generally contain far fewer resolved bands than the estimates of species richness for forest soil, then this pattern clearly represents only part of the community. The number of bands may be low due to 1) bias in the P C R , 2) few species being very abundant in the soil sample, 3) the extent of denaturant gradient and size of gels (resolution), 4) detection limits for different nucleic acid staining systems, or 5) overlap in melting behaviour for many organisms. Currently, it is unknown which factors contribute most to the patterns seen. It is tempting to conclude that D G G E patterns show the most abundant organisms while P L F A s show the total community; differences in P L F A fingerprints in this study would then suggest that there are differences in the composition of the less abundant organisms. However, there is no strong evidence that this is true. Given that the P L F A profiling gave more information than the molecular fingerprinting of the bacterial communities, it may be redundant to do both. Most studies using D G G E have focused on agricultural plants and soils, and there have been relatively few properly replicated studies. O f these, many have found no difference in D G G E profiles among different soils or treatments 80 (Felske and Akkermans 1998; Smit et al. 2001; Duineveld et al. 2001) while those that have detected differences have compared very different soil types (Gelsomino et al. 1999; Nakatsu et al. 2000), environmental conditions (0vreas et al . 1998), or soil layers (Krave et al. 2002). D G G E may be more useful for comparing very different systems, monitoring after a large perturbation, or comparing a small portion of the community using taxon-specific primers, than for studying the extent of variability within and between ecosystems. Future Research Patterns in soil microbial communities described in this and other studies need to be further examined to determine the factors controlling those patterns as well as to begin to understand the effect of microbial community structure on ecosystem processes. The detection of differences between C H and H A forests needs to be put into context by further study of similar and very different ecosystem types. B y understanding the extent to which communities differ in ecosystems with similar and very different process rates, we w i l l be able to interpret the potential functional implications of the community differences documented in this and other studies. A weakness of this study was that forests were only sampled a single time. Sampling throughout the year would confirm whether the differences detected between C H and H A forests are maintained throughout the year. This is also crucial in elucidating the causes to differences in soil process rates and productivity in C H and H A forests. Although the two forest types are exposed to the same climate, it is still unclear whether C H forests may be wetter than H A forests due to slight differences in slope position. Addressing this question through comprehensive soil moisture measurements would aid in understanding underlying differences between these two ecosystems. The effect of excessive moisture on soil communities could be investigated experimentally. Lab incubations of C H and H A soil under different moisture conditions may 81 allow some insight into the potential effect of moisture on soil communities, however, they may not be very comparable to what would happen in undisturbed field conditions. Large-scale irrigation experiments would be difficult but may provide more relevant results. Reciprocal transplants of soil cores between C H and H A forests may be useful for addressing the question of whether the communities are controlled by the substrate (chemical characteristics of organic matter) or the environment (moisture levels, interactions with other organisms). Although the presence of western redcedar in C H forests did not seem to explain the differences in microbial communities between C H and H A forests, it is unknown whether other aspects of the vegetation, including the presence of salal in C H forests, may be important. Tree species effects can be investigated in common-garden experiments, but the understory is more difficult to manipulate. Separating direct effects (e.g. due to root presence and activities) and indirect effects (e.g. litter quality, effects on soil pH) of plant species also seems to be important. Differences in root density and distribution associated with the different tree and understory species in C H and H A forests may be important. Quantifying the abundance of fine roots in the soil and learning about which materials organisms are actually using w i l l be important for determining how important rhizosphere versus bulk soil microbial communities are for mediating different soil processes. While rhizosphere communities are undoubtedly important for the immediate nutrition of plants, decomposition of soil organic matter may be primarily mediated through rhizosphere or non-rhizosphere organisms. The fungal component of the soil communities in C H and H A forests requires particular attention, given that clear differences were seen both in biomass and in community composition between the two forest types. Symbiotic and saprotrophic functions of fungi are now known to overlap amongst species and it is within this context that perspectives in nutrient cycling and 82 ecosystem functioning of forests have begun shifting to "a less phytocentric perspective" (Lindahl et al. 2002). Research investigating fungal community structure and distribution and biomass of fungi should be complemented with studies on the functional significance of those organisms. Moreover, while there are now several methods for measuring bacterial and fungal biomass, there is a need to test assumptions of the roles of these organisms and their importance in different ecosystems. In this study, it seemed that the ratios of fungal-to-bacterial biomass were much lower than would be expected for a cool, moist northern forest with large accumulations of humus. The generalization that fungi dominate soil processes in these types of forests may need to be revisited. The alternative explanation is that the P L F A method of measuring bacterial-to-fungal ratios is not comparable among studies. Although the method has now been applied in many studies, it is unknown how meaningful comparisons among studies w i l l be. This study offered a comparison of several methods, but more rigorous testing of methods and their interpretation is necessary. Future research on soil microbial communities should focus on two important, but challenging questions: 1) which factors are most important for controlling the abundance, composition, and activity of microbial communities, and 2) how does the community composition and functioning influence ecosystem processes. Studies characterizing the composition of soil microbial communities can be complemented with work on the function and activity of the communities using C L P P and assays of important microbial enzymes. Further refinement and use of stable isotope techniques to learn which organisms are using which substrates, as well as tracing the movement of material through the soil system wi l l increase our understanding of soil biological processes. Finally, continued rapid development of techniques in molecular microbial ecology promise new and powerful approaches to studying community composition and gene expression 83 in the future. In the age of genomics, the development of microarray technology holds promise for vastly increasing our knowledge of how these invisible systems are functioning. Only once we better understand the answers to the questions posed above wi l l we be able to predict the effect of human activities and natural disturbances on forest soil processes and the entire ecosystems. 84 CONCLUSIONS Microbial biomass, as measured by C F E and total P L F A s did not differ between C H and H A forests at the time sampled. Given that there was just one sampling time, it is possible that the biomass differs between the forest types during different times of the year. The microbial communities in C H and H A forest floors were shown to be quite similar. P L F A profiles, however, distinguished subtle differences in the overall community. Fungal P L F A s were significantly more abundant in C H forests than H A and the fungal community differed between the two forest types, as shown by ITS fingerprinting. Arbuscular mycorrhizal P L F A s were also proportionally more abundant in C H forests. Bacterial P L F A s were proportionally more abundant in H A forests than C H in the lower humus layer and there were differences in the bacterial community, including proportionally more Gram-positive bacteria in H A forests. A l l analyses confirmed that microbial biomass decreased in deeper forest floor layers and that the community structure differed among layers. Most P L F A s decreased in abundance in deeper layers, except actinomycetal P L F A s , which showed an increasing trend. Fungal P L F A s were proportionally less abundant with depth, while the opposite was seen with bacterial P L F A s in H A forests. P L F A s marking Gram-positive bacteria increased proportionally with depth, while those marking Gram-negative bacteria were proportionally more abundant in the F Layer of C H forests than the lower humus layer. The composite samples seemed to adequately characterize the average site conditions, as compared to analyzing 10 subsamples separately. There was evidence that within-plot variability was quite high, although that variability did not seem to overshadow the differences seen between forest types and forest floor layers. There was very good agreement between C F E and P L F A estimates of microbial biomass, however, extracted D N A did not seem to be a useful 85 technique for biomass measurements in these forests. P L F A , D G G E , R I S A , and ITS analyses all clearly showed differences in the microbial communities in different forest floor layers. Only P L F A and ITS analyses detected differences between the forest types. R I S A fingerprints showed much greater variability among all samples for the bacterial community than D G G E , however, it seemed to be due to the very different electrophoresis systems used rather than to differences in the variability of the targeted D N A . This study provided information about the variability and distribution of soil microbial communities in an interesting ecological situation. It also provided a good comparison of several of the newer methods available in the field of soil ecology and microbiology. Results from this and other studies w i l l be put into better context with continued investigation of the distribution and variability of soil microbial communities in a wide range of ecosystem types. Future research should be directed to both understanding the factors controlling the composition and activities of these communities as well as experimentally testing the link between the microbial community and ecosystem functioning. 8 6 LITERATURE CITED Aitchison, J. 1986. The statistical analysis of compositional data. Chapman & Hal l , London. Alden, L . , Demoling, F. , and Baath, E . 2001. Rapid method of determining factors limiting bacterial growth in soil. App l . Environ. Microbiol . 67: 1830-1838. Alphei , J., Bonkowski, M . , and Scheu, S. 1995. Application of the selective inhibition methods to determine bacterial: fungal ratios in three beech wood soils rich in carbon - optimization of inhibitor concentrations. B i o l . Fertil. Soils 19: 173-176. Amann, R.I., Ludwug, W. , and Schleiffer, K . - H . 1995. Phylogenetic identification and in situ detection of individual microbial cells without cultivation. Microbiol . Rev. 59: 143-169. Baath, E . , Pennanen, T., and Fritze, H . 1995. Microbial community structure and p H response in relation to soil organic matter quality in wood-ash fertilized, clearcut or burned coniferous forest soils. Soil B i o l . Biochem. 27: 229-240. Bailey, V . L . , Peacock, A . D . , Smith, J .L. , and Bolten Jr., H . 2002. Relationships between soil microbial biomass determined by chloroform fumigation-extraction, substrate-induced respiration, and phospholipid fatty acid analysis. Soil B i o l . Biochem. 34: 1385-1389. Battigelli, J.P., Berch, S .M. , and Marshall, V . G . 1994. Soil fauna communities in two distinct but adjacent forest types on northern Vancouver Island, British Columbia. Can. J. For. Res. 24: 1557-1566. Bauhaus, J. and Khanna, P . K . 1999. The significance of microbial biomass in forests soils. In Going underground - ecological studies in forest soils. Edited by N . Rastin and J. Bauhaus, Research Signpost, Trivandum, India, pp. 77-110. Bennett, J .N. , Andrew, B . , and Prescott, C E . 2002. Vertical fine root distributions of western redcedar, western hemlock, and salal in old-growth cedar-hemlock forests on northern Vancouver Island. Can. J. For. Res. 32: 1208-1216. Benoit, R . E . and Starkey, R . L . 1968. Enzyme inactivation as a factor in the inhibition of decomposition of organic matter by tannins. Soil Sci. 105: 203-208. Berg, M . P . , Kniese, J.P., and Verhoef, H . A . 1998. Dynamics and stratification of bacteria and fungi in the organic layers of a Scots pine forest soil. B i o l . Fertil. Soils 26: 313-322. Binkley, D . and Giardina, C. 1998. Why do tree species affect soils? the warp and woof of tree-soil interactions. Biogeochemistry 42: 89-106. Blagodatskaya, E . V . and Anderson, T . - H . 1998. Interactive effects of p H and substrate quality on the fungal-to-bacterial ratio and q C 0 2 of microbial communities in forest soils. Soil B i o l . Biochem. 30: 1269-1274. 87 Blanchette, R . A . 1995. Degradation of the lignocellulose complex in wood. Can. J. Bot. 73 (Suppl. 1): S999-S1010. Blevins, L . L . and Prescott, C E . 2002. Salal Cedar Hemlock Integrated Research Program-Research Update #2: Silvicultural practices for regeneration of cedar-hemlock sites in coastal British Columbia. University of British Columbia , Vancouver, B . C . Bl igh, E . G . and Dyer, W.J . 1959. A rapid method of total l ipid extraction and purification. Can. J. Biochem. Physiol. 37: 911-917. Borneman, J. and Triplett, E . W . 1997. Molecular microbial diversity in soils from eastern Amazonia: evidence for unusual microorganisms and microbial population shifts associated with deforestation. App l . Environ. Microbiol . 63: 2647-2653. Bossio, D . A . and Scow, K . M . 1998. Impacts of carbon and flooding on soil microbial communities: phospholipid fatty acid profiles and substrate utilization patterns. Microb. Ecol . 35: 278. Bradley, R . L . and Fyles, J.W. 1995. Growth of paper birch (Betula papyrifera) seedlings increases soil available C and microbial acquisition of soil nutrients. Soil B i o l . Biochem. 27: 1565-1571. Bradley, R . L . , Titus, B . D . , and Preston, C P . 2000a. Changes to mineral N cycling and microbial communities in black spruce humus after additions of (NH4)2S04 and condensed tannins extracted from Kalmia angustifolia and balsam fir. Soil B i o l . Biochem. 2: 1227-1240. Bradley, R . L . , Titus, B . D . , and Preston, C P . , and Bennett, J. 20006 Improvement of nutritional site quality 13 years after single application of fertiliser N and P on regenerating cedar-hemlock cutovers on northern Vancouver Island, B . C . Plant Soil 223: 195-206. Brosius, J., Du l l , T.J., Sleeter, D .D . , and Noller, H.F . 1981. Gene organization and primary structure of a ribosomal operon from Escerichia coli. J. M o l . B io . 148: 107-127. Brussaard, L . 1998. Soil fauna, guilds, functional groups and ecosystem processes. App l . Soil Ecol . 9: 123-135. Bundt, M . , Widmer, F. , Pesaro, M . , Zeyer, J., and Blaser, P. 2001. Preferential flow-paths: biological 'hot spots' in soils. Soil B i o l . Biochem. 33: 729-738. Cairney, J .W.G. and Burke, R . M . 1998a. Do ecto- and ericoid mycorrhizal fungi produce peroxidase activity? Mycorrhiza 8: 61-65. Cairney, J .W.G. and Burke, R . M . 19986. Extracellular enzyme activities of the ericoid mycorrhizal endophyte Hymenoscyphus ericae (Read) K o r f & Kernan: their likely roles in decomposition of dead plant tissue in soil. Plant Soil 205: 181-192. 88 Carlile, M J . and Watkinson, S.C. 1994. The Fungi. Academic Press Inc., San Diego. Cavigelli , M . A . , Robertson, G.P. , and K l u g , M . J . 1995. Fatty acid methyl ester ( F A M E ) profiles as measures of soil microbial community structure. Plant Soil 170: 99-113. Chang, S.X. , Preston, C M . , and Weetman, G.F . 1995. Soil microbial biomass and microbial and mineralizable N in a clear-cut chronosequence on northern Vancouver Island, British Columbia. Can. J. For. Res. 25: 1595-1607. Chen, D . M . , Taylor, A . F . S . , Burke, R . M . , and Cairney, J .W.G. 2001. Identification of genes for lignin peroxidases and manganese peroxidases in ectomycorrhizal fungi. New Phytol. 152: 151-158. Clarholm, M . and Rosswall, T. 1980. Biomass and turnover of bacteria in a forest soil and a peat. Soil B i o l . Biochem. 12: 49-57. Collins, D . B . , Feller, M . C , Kl inka , K . , and de Montigny, L . E . 2001. Forest floor nutrient properties in single- and mixed-species, second growth stands of western hemlock and western redcedar. Northwest Sci. 75: 407-416. Cooke, R . C . and Raynor, A . D . M . 1984. Ecology of saprotrophic fungi. Longman Group Limited, New York. Degens, B .P . 1998. Decreases in microbial functional diversity do not result in corresponding changes in decomposition under different moisture conditions. Soil B i o l . Biochem. 30: 1989-2000. Degens, B .P . and Harris, J .A. 1997. Development of a physiological approach to measuring the catabolic diversity of soil microbial communities. Soil B i o l . Biochem. 29: 1309-1320. de Montigny, L . E . 1992. A n investigation into the factors contributing to the growth-check of conifer regeneration on northern Vancouver Island. Ph.D. thesis, Univeristy of British Columbia, Vancouver, B . C . de Montigny, L . E . , Preston, C M . , Hatcher, P .G. , and Kogel-Knabner, I. 1993. Comparison of humus horizons from two ecosystem types on northern Vancouver Island using C C P M A S N M R spectroscopy and CuO oxidation. Can. J. Soil Sci. 73: 9-25. Duineveld, B . M . , Kowalchuk, G . A . , Keijzer, A . , van Elsas, J.D., and van Veen, J .H. 2001. Analysis of bacterial communities in the rhizosphere of Chrysanthemum via denaturing gradient gel electrophoresis of PCR-amplified 16S r R N A as well as D N A fragments coding for 16S r R N A . App l . Environ. Microbiol . 67: 172-178. Dykhuizen, D . E . 1998. Santa Rosalina revisited: why are there so many species of bacteria? Antonie van Leeuwenhoek 73: 25-33. 89 Ekblad, A . and Nordgren, A . 2002. Is growth of soil microorganisms in boreal forests limited by carbon or nitrogen availability? Plant Soil 242: 115-122. Ekelund, F. , Rjann, R., and Christensen, S. 2001. Distribution with depth of protozoa, bacteria, and fungi in soil profiles from three Danish forest sites. Soil B i o l . Biochem. 33: 475-481. Expert Committee on Soil Survey (Canada) 1987. The Canadian system of soil classification. Canada Dept. of Agriculture, Research Branch, Ottawa. Farrelly, V . , Rainey, F . A . , and Stackebrandt, E . 1995. Effect of genome size and rrn gene copy number on P C R amplification of mixtures of 16S r R N A genes from a mixture of bacterial species. App l . Environ. Microbiol . 61: 2798-2801. Federle, T. W . 1986 Microbial distribution in soil - new techniques. In Perspectives in microbial ecology, Ljubljana. Edited by Megusar, F. and Gantar, M . Slovene Society for Microbiology, Ljubljana, pp. 493-498. Felske, A . and Akkermans, A . D . L . 1998. Spatial homogeneity of abundant bacterial 16S r R N A molecules in grassland soils. Microb. Ecol . 36: 31-36. Field, J .A. and Lettinga, G . 1992. Toxicity of tannin compounds to microorganisms. In Plant polyphenols: synthesis, properties, significance. Edited by Hemingway, R . W . and Laks, P .E. Plenum Press, New York. Fierer, N . , Schimel, J.P., Cates, R . G . , Zou, J. 2001. Influence of balsam poplar tannin fractions on carbon and nitrogen dynamics in Alaskan taiga floodplain soils. Soil B i o l . Biochem. 33: 1827-1839. Finlay, B . J . 2002. Global dispersal of free-living microbial eukaryote species. Science 296: 1061-1063. Fisher, M . M . and Triplett, E . W . 1999. Automated approach for ribosomal intergenic spacer analysis of microbial diversity and its application to freshwater bacterial communities. App l . Environ. Microbiol . 65: 4630-4636. Forge, T . A . and Simard, S.W. 2000. Trophic structure of nematode communities, microbial communities, and nitrogen mineralisation in soils of forests and clearcuts in the southern interior of British Columbia. Can. J. Soil Sci. 80: 401-410. Forge, T . A . and Simard, S.W. 2001. Short-term effects of nitrogen and phosphorous fertilizers on nitrogen mineralisation and trophic structure of the soil ecosystem in forest clearcuts in the southern interior of British Columbia. Can. J. Soil Sci. 81: 11-20. Fritze, H . , Pietikainen, J., and Pennanen, T. 2000. Distribution of microbial biomass and phospholipid fatty acids in Podzol profiles under coniferous forest. Eur. J. Soil Sci . 51: 565-573. 90 Frostegard, A. and Baath, E . 1996. The use of phospholipid fatty acid analysis to estimate bacterial and fungal biomass in soil. B i o l . Fertil. Soils 22: 59-65. Frostegard, A., Tunlid, A . , and Baath, E . 1991. Microbial biomass measured as total l ipid phosphate in soils of different organic content. J. Microbiol . Methods 14 : 151-163. Frostegard, A., Baath, E . , and Tunlid, A . 1993. Shifts in the structure of soil microbial communities in limed forests as revealed by phospholipid fatty acid analysis. Soil B i o l . Biochem. 25: 723-730. Frostegard, A., Peterson, S.O., Baath, E . , and Nielsen, T . H . 1997. Dynamics of a microbial community associated with manure hot spots as revealed by phospholipid fatty acid analyses. App l . Environ. Microbiol . 63: 2224-2231. Frostegard, A., Courtois, S., Ramisse, V . , Clerc, S., Bernillon, D . , Le Gal l , F. , Jeannin, P., Nesme, X . , and Simonet, P. 1999. Quantification of bias related to the extraction of D N A directly from soils. App l . Environ. Microbiol . 65: 5409-5420. Garcia-Martinez, J., Acinas, S.G., Anton, A . I . , and Rodriguez-Valera, F. 2002. Use of the 16S-23S ribosomal genes spacer region in studies of prokaryotic diversity. J. Microbiol . Methods 36: 55-64. Gardes, M . and Bruns, T .D. 1993. ITS primers with enhanced specificity for basidiomycetes -application to the identification of mycorrhizae and rusts. M o l . Ecol . 2: 113-118. Garland, J .L. and M i l l s , A . L . 1991. Classification and characterization of heterotrophic microbial communities on the basis of patterns of community-level sole-carbon-source utilization. App l . Environ. Microbiol . 57: 2351-2359. Gelsomino, A . , Keijzer-Wolters, A . , Cacco, G . , and van Elsas, J.D. 1999. Assessment of bacterial community structure in soil by polymerase chain reaction and denaturing gradient gel electrophoresis. J. Microbiol . Methods 38: 1-15. Godfray, H .C.J , and Lawton, J .H. 2001. Scale and species number. Trends Ecol . Evo l . 16: 400-404. Grayston, S.J. and Campbell, C D . 1996. Functional biodiversity of microbial communities in the rhizospheres of hybrid larch (Larix eurolepis) and Sitka spruce (Picea sitchensis). Tree Physiol. 16: 1031-1038. Grayston, S.J., Wang, S., Campbell, C D . , and Edwards, A . C 1998. Selective influence of plant species on microbial diversity in the rhizosphere. Soil B i o l . Biochem. 30: 369-378. 91 Grayston, S.J., Griffith, G.S. , Mawdsley, J .L. , Campbell, C D . , and Bardgett, R . D . 2001. Accounting for variability in soil microbial communities of temperate upland grassland ecosystems. Soil Biol .Biochem. 33: 533-551. Green, R . N . and Kl inka , K . 1994. A field guide to site identification and interpretation for the Vancouver forest region. Crown Publications Inc., Victoria, British Columbia. Green, R . N . , Trowbridge, R . L . , and Kl inka , K . 1993. Towards a taxonomic classification of humus forms. Forest Science Monograph 29. Supplement to Forest Science 39. Griffiths, B.S. , Bonkowski, M . , Dobson, G . , and Caul, S. 1999a. Changes in soil microbial community structure in the presence of microbial-feeding nematodes and protozoa. Pedobiologia 43: 297-304. Griffiths, B.S. , Ritz, K . , Ebblewhite, N . , and Dobson, G . 19996. Soil microbial community structure: effects of substrate loading rates. Soil B i o l . Biochem. 31: 145-153. Griffiths, B .S . , Ritz, K . , Bardgett, R .D. , Cook, R., Christensen, S., Ekelund, F. , S0renson, S.J., Baath, E . , Bloem, J., de Ruiter, P . C , Dolfing, J., and Nicolardot, B . 2000. Ecosystem response of pasture soil communities to fumigation-induced microbial diversity reductions: an examination of the biodiversity-ecosystem function relationship. Oikos 90: 294 . Griffiths, B .S . , Ritz, K . , Wheatley, R., Kuan, H . L . , Boag, B . , Christensen, S., Ekelund, F. , Sorenson, S.J., Muller , S. , and Bloem, J. 2001. A n examination of the biodiversity-ecosystem function relationship in arable soil microbial communities. Soil B i o l . Biochem. 33: 1713-1722. Guckert, J .B. , Hood, M . A . , and White, D . C . 1985. Phospholipid ester-linked fatty acid profile changes during nutrient deprivation of Vibrio cholerae: increases in the trans/cis ratio and proportions of cyclopropyl fatty acids. App l . Environ. Microbiol . 52: 794-801. Haack, S.H., Garchow, H . , Odelson, D . A . , Forney, L . J . , and K l u g , M . J . 1994. Accuracy, reproducibility, and interpretation of fatty acid methyl ester profiles of model bacterial communities. App l . Environ. Microbiol . 60: 2483-2493. Harris, D . 1994. Analyses of D N A extracted from microbial communities. In Beyond the biomass: compositional and functional analysis of soil microbial communities. Edited by K . Ritz, J. Dighton, and K . E . Giller, John Wiley and Sons, Chichester, pp. 111-118. Hart, S . C and Stark, J . M . 1997. Nitrogen limitation of the microbial biomass in an old-growth forest soil. Ecoscience 4: 91-98. Hattenschwiler, S. and Vitousek, P . M . 2000. The role of polyphenols in terrestrial ecosystem nutrient cycling. Trends Ecol . Evo l . 15: 238-242. 92 Heipieper, H.J . , Meulenbeld, G . , van Oirschot, Q., and de Bont, J .A. 1996. Effect of environmental factors on the trans/cis ratio of unsaturated fatty acids in Pseudomonas putida S12. App l . Environ. Microbiol . 62: 2773-2777. Hillebrand, H . , Watermann, F. , Karez, R., and Berninger, U . - G . 2001. Differences in species richness patterns between unicellular and multicellular organisms. Oecologia 126: 114-124. Hugenholtz, P., Goebel, B . M . , and Pace, N . R . 1998. Impact of culture-independent studies on the emerging phylogenetic view of bacterial diversity. J. Bacteriol. 180: 4765-4774. Hughes, J.B., Hellmann, J.J., Ricketts, T . H . , and Bohannan, B . J . M . 2001. Counting the uncountable: statistical approaches to estimating microbial diversity. App l . Environ. Microbiol . 67: 4399-4406. Kaye, J.P. and Hart, S.C. 1997. Competition for nitrogen between plants and soil microorganisms. Trends Ecol . Evol . 12: 139-143. Keenan, R.J . , Prescott, C .E . , and Kimmins, J.P. 1993. Mass and nutrient content of woody debris and forest floor in western red cedar and western hemlock forests on northern Vancouver Island. Can. J. For. Res. 23: 1052-1059. Ki l lham, K . 1994. Soil Ecology. Cambridge University Press, Cambridge. Kjoller, A . and Struwe, S. 1982. Microfungi in ecosystems: fungal occurrence and activity in litter and soil. Oikos 39: 391-422. Krave, A . S . , L i n , B . , Braster, M . , Laverman, A . M . , van Straalen, N . M . , Roling, W . F . M . , and van Verseveld, H . W . 2002. Stratification and seasonal stability of diverse bacterial communities in a Pinus merkusii (pine) forest soil in central Java, Indonesia. Environ.Microbiol . 4:361-373. Landeweert, R., Leeflang, P., Kuyper, T.W., Hoffland, E . , Rosling, A . , Wernars, K . , and Smit, E . 2003. Molecular identification of ectomycorrhizal mycelium in soil horizons. App l . Environ. Microbiol . 69: 327-333. Lechevalier, M . P . 1977. Lipids in bacterial taxonomy - a taxonomists view. Crit. Rev. Microbiol . 5: 109-210. Lechevalier, H . and Lechevalier, M . P . 1988. Chemotaxonomic use of lipids - an overview. In Microbial Lipids. Edited by C . Ratledge and S.G. Wilkinson, Academic Press, London, England, pp. 869-902. Lewis, T. 1985. Ecosystems of Quatsino Tree-Farm License (TFL 6). Western Forest Products Ltd. (internal report), British Columbia. Lindahl, B .O . , Taylor, A .F .S . , and Finlay, R . D . 2002. Defining nutritional constraints on carbon cycling in boreal forests - towards a less "phytocentric" perspective. Plant Soil 242: 123-135. 93 Marchesi, J.R., Sato, T., Weightman, A . J . , Martin, T .A . , Fry, J . C , Hiom, S.J., and Wade, W . G . 1998. Design and evaluation of useful bacterium-specific P C R primers that amplify genes coding for bacterial 16S r R N A . App l . Environ. Microbiol . 64: 795-799. Marstorp, H . , Guan, X . , and Gong, P. 2000. Relationship between d s D N A , chloroform labile C and ergosterol in soils of different organic matter contents and p H . Soil B i o l . Biochem. 32: 879-882. Martens, R. 1995. Current methods for measuring microbial biomas C in soil: potentials and limitations. B i o l . Fertil. Soils 19: 87-99. Martin-Laurent, F. , Philippot, L . , Hallet, S., Chaussod, R., Germon, J . C , Soulas, G . , and Catroux, G . 2001. D N A extraction from soils: old bias for new microbial diversity analysis methods. App l . Environ. Microbiol . 67: 2354-2359. M c C a i g , A . E . , Glover, L . A . , and Prosser, J.I. 1999. Molecular analysis of bacterial community structure and diversity of unimproved and improved upland grass pastures. App l . Environ. Microbiol . 65: 1721-1730. M c E v o y , C . R . E . , Seshadri, R., and Firgaira, F . A . 1998. Large D N A fragment sizing using native acrylamide gels on an automated D N A sequencer and G E N E S C A N ™ software. BioTechniques 25: 464-470. McLean , M . A . and Huhta, V . 2000. Temporal and spatial fluctuations in moisture affect humus microfungal community structure in microcosms. B i o l . Fertil. Soils 32: 114-119. Mil ler , D . N . , Bryant, J.E., Madsen, E . L . , and Ghiorse, W . C 1999. Evaluation and optimization of D N A extraction and purification procedures for soil and sediment samples. App l . Environ. Microbiol . 65:4715-4724. Mottonen, M . , Ohtonen, R., Hokkanen, T.J., and Kuuluvainen, T. 1999. Spatial distribution of soil ergosterol in the organic layer of a Scots pine (Pinus sylvestris L. ) forest. Soil B i o l . Biochem. 31: 503-516. Muyzer, G . and Smalla, K . 1998. Application of denaturing gradient gel electrophoresis ( D G G E ) and temperature gradient gel electrophoresis (TGGE) in microbial ecology. Antonie van Leeuwenhoek 73: 127-141. Muyzer, G . , de Waal, E.S. , and Uitterlinden, A . G . 1993. Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S r R N A . App l . Environ. Microbiol . 59: 695-700. Muyzer, G . , Teske, A . , Wirsen, C O . , and Jannasch, H . W . 1995. Phylogenetic relationship of Thiomicrospira species and their identification in deep-sea hydrothermal vent samples by denaturing gradient gel electrophoresis of 16S r D N A fragments. Arch . Microbiol . 164: 165-172., 94 Myers, R.T. , Zak, D.R. , White, D .C . , and Peacock, A . 2001. Landscape-level patterns of microbial community composition and substrate use in upland forest ecosystems. Soil Sci. Soc. A m . J. 65: 359-367. Nakatsu, C . H . , Torsvik, V . , and Ovreas, L . 2000. Soil community analysis using D G G E of 16S r D N A polymerase chain reaction products. Soil Sci . Soc. A m . J. 64: 1382-1388. Neely, C . L . , Beare, M . H . , Hargrove, W . L . , and Coleman, D . C . 1991. Relationships between fungal and bacterial substrate-induced respiration, biomass, and plant residue decomposition. Soil B i o l . Biochem. 23: 947-954. Newell , S .Y. 1992. Estimating fungal biomass and productivity in decomposing litter. In The fungal community: its organization and role in the ecosystem (2 n d edition). Edited by G . C . Carroll and D.T. Wicklow, Marcel Dekker Inc., New York. pp. 521-561. Nunan, N . , Wu , K . , Young, I . M . , Crawford, J.W., and Ritz, K . 2002. In situ spatial patterns of soil bacterial populations, mapped at multiple spatial scales, in an arable soil. Microb. Ecol . 44: 296-305. Niisslein, K . and Tiedje, J . M . 1998. Characterization of the dominant and rare members of a young Hawaiian soil bacterial community with small-subunit ribosomal D N A amplified from D N A fractionated on the basis of its guanine and cytosine composition. App l . Environ. Microbiol . 64: 1283-1289. Ogram, A . 2000. Soil molecular microbial ecology at age 20: methodological challenges for the future. Soil B i o l . Biochem. 32: 1499-1504. O'Leary, W . M . and Wilkinson, S.G. 1988. Gram-positive bacteria. In Microbial Lipids. Edited by C . Ratledge and S.G. Wilkinson, Academic Press, London, England, pp. 117-202. Olsen, G.J. , Lane, D.J . , Giovannoni, S.J., and Pace, N . R . 1986. Microbial ecology and evolution: a ribosomal R N A approach. Annu. Rev. Microbiol . 40: 337-365. Olsson, P. A . 1999. Signature fatty acids provide tools for determination of the distribution and interactions of mycorrhizal fungi in soil. F E M S Microbiol . Ecol . 29: 303-310. Ovreas, L . , Jensen, S., Daae, F . L . , and Torsvik, V . 1998. Microbial community changes in a perturbed agricultural soil investigated by molecular and physiological approaches. App l . Environ. Microbiol . 64: 2739-2742. Pace, N .R . , Stahl, D . A . , Lane, D.J . , and Olsen, G.J . 1986. The analysis of natural microbial populations by ribosomal R N A sequences. Adv. Microb. Ecol . 9: 1-55. Pennanen, T., L i s k i , J., Baath, E . , Kitunin, V . , Uotila, J., Westman, C.J . , and Fritze, H . 1999. Structure of the microbial communities in coniferous forest soils in relation to site fertility and stand development. Microb. Ecol . 38: 168-179. 95 Pennanen, T., Strommer, R., Markkola, A . , and Fritze, H . 2001. Microbial and plant community structure across a primary succession gradient. Scan. J. For. Res.16: 37-43. Prescott, C E . 1995. Does nitrogen availability control rates of litter decomposition in forests? Plant Soil 168-169: 83-88. Prescott, C E . and Weetman, G.F. 1994. Salal Cedar Hemlock Integrated Research Program: a synthesis. Faculty of Forestry, University of British Columbia, Vancouver, B . C . Prescott, C .E . , McDonald, M . A . , and Weetman, G.F. 1993. Availabil i ty of N and P in the forest floors of adjacent stands of western red cedar-western hemlock and western hemlock-amabilis fir on northern Vancouver Island. Can. J. For. Res. 23: 605-610. Prescott, C .E . , Vesterdal, L . , Pratt, J., Venner, K . H . , de Montigny, L . E . , and Trofymow, J .A. 2000. Nutrient concentration and nitrogen mineralisation in forest floors of single species conifer plantations in coastal British Columbia. Can. J. For. Res. 30: 1352. Preston-Mafham, J., Boddy, L . , and Randerson, P.F. 2002. Analysis of microbial community functional diversity using sole-carbon-source utilisation profiles- a critique. F E M S Microbiol . Ecol . 42: 1-14. Priha, O. and Smolander, A . 1997. Microbial biomass and activity in soil and litter under Pinus sylvestris, Picea abies, and Betula pendula at originally similar field afforestation sites. B i o l . Fertil. Soils 24:45-51. Priha, O. and Smolander, A . 1999. Nitrogen transformations in soil under Pinus sylvestris, Picea abies, and Betula pendula at two forest sites. Soil B i o l . Biochem. 31: 965-977. Priha, O., Grayston, S.J., Pennanen, T., and Smolander, A . 1999. Microbial activities related to C and N cycling and microbial community structure in the rhizospheres of Pinus sylvestris, Picea abies, and Betula pendula seedlings in an organic and mineral soil. F E M S Microbiol . Ecol . 30: 187-199. Priha, O., Grayston, S.J., Hiukka, R., Pennanen, T., and Smolander, A . 2001. Microbial community structure and characteristics of the organic matter in soils under Pinus sylvestris, Picea abies, and Betula pendula at two forest sites. B i o l . Fertil. Soils 33: 17-24. Ranjard, L . , Poly, F. , Lata, J . - C , Mougel , C , Thioulouse, J., and Nazaret, S. 2001. Characterization of bacterial and fungal soil communities by automated ribosomal intergenic spacer analysis fingerprints: biological and methodological variability. App l . Environ. Microbiol . 67: 4479-4487. Ratledge, C. and Wilkinson, S.G. 1988. Microbial Lipids. Academic Press, London, England. 96 Ritz, K . , Dighton, J., and Giller, K . E . 1994. Beyond the biomass: compositional and functional analysis of soil microbial communities. John Wiley and Sons, Chichester Rogers, B . F . and Tate, R .L . I . 2001. Temporal analysis of the soil microbial community along a toposequence in Pineland soils. Soil B i o l . Biochem. 33: 1389-1401. Rann, R., M c C a i g , A . E . , Griffiths, B .S . , and Prosser, J.I. 2002. Impact of protozoan grazing on bacterial community structure in soil microcosms. App l . Environ. Microbiol . 68: 6094-6105. Ross, D.J . and Tate, K . R . 1993. Microbial C and N in litter and soil o f a southern beech (Northofagus) forest: comparison of measurement procedures. Soil B i o l . Biochem. 25: 467-475. Saetre, P. 1998. Decomposition, microbial community structure, and earthworm effects along a birch-spruce soil gradient. Ecology 79: 834-846. Saetre, P. and Baath, E . 2000. Spatial variation and patterns of soil microbial community structure in a mixed spruce-birch stand. Soil B i o l . Biochem. 32: 909-917. Schmalenberger, A . , Schwieger, F. , and Tebbe, C C . 2003. Effect of primers hybridizing to different evolutionarily conserved regions of the small-subunit r R N A gene in PCR-based microbial community analyses and genetic profiling. App l . Environ. Microbiol.67:3557-3563. Setala, H . and Aarnio, T. 2002. Vertical stratification and trophic interactions among organisms of a soil decomposer food web - a field experiment using 1 5 N as a tool. Eur. J. Soil B i o l . 38: 29-34. Siira-Pietikainen, A . , Pietikainen, J., Fritze, H . , and Haimi, J. 2001. Short-term response of soil decomposer communities to forest management: clear felling versus alternative forest harvesting methods. Can. J. For. Res. 31: 88-99. Smit, E . , Leeflang, P., Gommans, S., van den Broek, J., van M i l , S., and Wernars, K . 2001. Diversity and seasonal fluctuations of the dominant members of the bacterial community in a wheat field as determined by cultivation and molecular methods. App l . Environ. Microbiol . 67:2284-2291. Smolander, A . , Kurka, A . , Kitunen, V . , and Malkonen, E . 1994. Microbial biomass C and N , and respiratory activity in soil of repeatedly limed and N-fertilized Norway spruce stands. Soil B i o l . Biochem. 26: 957-962. Sokal, R .R. and Michener, C D . 1958. A statistical method for evaluating systematic relationships. University of Kansas bulletin 38: 1409-1438. Soderstrdm, B . , Baath, E . , and Lundgren, B . 1983. Decrease in microbial activity and biomasses owing to nitrogen amendments. Can. J. Microbiol . 29: 1500-1506. 97 Suzuki, M . T . and Giovannoni, S.J. 1996. Bias caused by template annealing in the amplification of mixtures of 16S r R N A genes by P C R . App l . Environ. Microbiol . 62: 625-630. Tate, K . R . , Ross, D.J . , and Feltham, C . W . 1988. A direct extraction method to estimate soil microbial C: effects of experimental variables and some different calibration procedures. Soil B i o l . Biochem. 20: 329-335. Thirukkumaran, C M . and Parkinson, D . 2000. Microbial respiration, biomass, metabolic quotient and litter decomposition in a lodgepole pine forest floor amended with nitrogen and phosphorous fertilizers. Soil B i o l . Biochem. 32: 59-66. Tiedje, J . M . , Asuming-Brempong, S., Niisslein, K . , Marsh, T .L . , and Flynn, S.J. 1999. Opening the black box of soil microbial diversity. App l . Soil Ecol . 13: 109-122. Torsvik, V . , Goksoyr.J., and Daae, F . L . 1990a. High diversity in D N A of soil bacteria. App l . Environ. Microbiol . 56: 782-787. Torsvik, V . , Salte, K . , Sorheim, R., and Goks0yr.J. 19906. Comparison of phenotypic diversity and D N A heterogeneity in a population of soil bacteria. J. Microbiol . Methods 56: 776-781. Torsvik, V . , Ovreas, L . , and Thingstad, T.F. 2002. Prokaryotic diversity - magnitude, dynamics, and controlling factors. Science 296: 1064-1067. Turner, D.P. and Franz, E . H . 1985. The influence of western hemlock and western redcedar on microbial numbers, nitrogen mineralisation, and nitrification. Plant Soil 8 8 : 259-267. Vance, E .D . , Brookes, P .C. , and Jenkinson, D.S. 1987. A n extraction method for measuring soil microbial biomass C. Soil B i o l . Biochem. 19: 703-707. van der Lee, G . E . M . , de Winder, B . , Bouten, W. , and Tietema, A . 1999. Anoxic microsites in Douglas-fir litter. Soil B i o l . Biochem. 31: 1295-1301. van Elsas, J.D., Duarte, G.F. , Keijzer-Wolters, A . , and Smit, E . 2000. Analysis of the dynamics of fungal communities in soil via fungal-specific P C R of soil D N A followed by denaturing gradient gel electrophoresis. J. Microbiol . Methods 43: 133-151. Vestal, J.R. and White, D . C 1989. L ip id analysis in microbial ecology: quantitative approaches to the study of microbial communities. Bioscience 39: 535-541. Wardle, D . A . 2002. Communities and ecosystems: linking the aboveground and belowground components. Princeton University Press, Princeton. Weetman, G.F. , Fournier, R., Barker, J., Schnorbus-Panozzo, E . , and Germain, A . 1989a. Foliar analysis and response of fertilized chlorotic Sitka spruce plantations on salal-dominated cedar-hemlock cutovers on Vancouver Island. Can. J. For. Res. 19: 1501-1511. 98 Weetman, G.F. , Fournier, R., Barker, J., and Schnorbus-Panozzo, E . 19896. Foliar analysis and response of fertilized chlorotic western hemlock and western red cedar reproduction on salal-dominated cedar-hemlock cutovers on Vancouver Island. Can. J. For. Res. 19: 1520. White, D .C . , Davis, W . M . , Nickels, J.S., K ing , J.D., and Bobbie, R.J . 1979. Determination of the sedimentary microbial biomass by extractable l ipid phosphate. Oecologia 40: 51-62. Wilkinson, S.C. and Anderson, J . M . 2001. Spatial patterns of soil microbial communities in a Norway spruce (Picea abies) plantation. Microb. Ecol . 42: 248-255. Wilkinson, S.G. 1988. Gram-negative bacteria. In Microbial lipids. Edited by C. Ratledge and S.C. Wilkindon, Academic Press, London, England, pp. 299-488. Wintzingerode, F. , Gobel, U . B . , and Stackebrandt, E . 1997. Determination of microbial diversity in environmental samples: pitfalls of PCR-based r R N A analysis. F E M S Microbiol . Rev.21: 213-229. Woese, C.R. , Kandler, O., and Wheelis, M . L . 1990. Towards a natural system of organisms: proposal for the domains Archaea, Bacteria, and Eucarya. Proc. Natn. Acad. Sci. U S A 87: 4576-4579. Zak, D.R. , Ringelberg, D . B . , Pregitzer, K . S . , Randlett, D . L . , White, D .C . , and Curtis, P.S. 1996. Soil microbial communities beneath Populus tremuloides grown under elevated atmospheric C 0 2 . Ecol . App l . 6: 257-262. Zak, D.R. , Pregitzer, K . S . , Curtis, P.S., and Holmes, W . E . 2000. Atmospheric C 0 2 and the composition and function of soil microbial communities. Ecol . App l . 10: 47-59. Zelles, L . 1997. Phospholipid fatty acid profiles in selected members of soil microbial communities. Chemosphere 35: 275-294. Zelles, L.1999. Fatty acid patterns of phospholipids and lipopolysaccharides in the characterisation of microbial communities in soil: a review. B i o l . Fertil. Soils. 29: 111-129. Zelles, L . , Ba i , Q . Y . , Beck, T., and Beese, F. 1992. Signature fatty acids in phospholipids and lipopolysaccharides as indicators of microbial biomass and community structure in agricultural soils. Soil B i o l . Biochem. 24: 317-323. Zhou, J., Bruns, M . A . , and Tiedje, J . M . 1996. D N A recovery from soils of diverse composition. App l . Environ. Microbiol . 62: 316-322. Zogg, G.P. ; Zak, D.R. ; Ringelberg, D . B . ; MacDonald, N . W . ; Pregitzer, K . S . ; White, D . C . 1997. Compositional and functional shifts in microbial communities due to soil warming. Soil Sci. Soc. of A m . J. 61: 475-481. 99 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0075021/manifest

Comment

Related Items