INQUA - COMMISSION FOR THE STUDY OF THE HOLOCENE Working Group on Data-handling Methods Newsletter 3, January, 1990 ECOLOGY OF PEATLAND BRYOPHYTES AND PALAEOENVIRONMENTAL RECONSTRUCTION OF PEATLANDS USING FOSSIL BRYOPHYTES SUMMARY Jan A. Janssens 1. Introduction: The composition of bryophyte communities correlates excellently with edaphic and microclimatic situations (Miller 1980). Dickson (1986) and Janssens (1990) reviewed the literature on palaeoenvironmental reconstruction with fossil bryophytes. This paper is a summary of the autoecological and stratigraphical theory and methods using fossil mosses for the reconstruction of environmental parameters important in the development of peatlands. Quantitative paleoenvironmental reconstruction (Birks and Birks 1980, Birks 1985, 1987, Birks et al. 1990) consists of the reconstruction (inference) of environmental variables such as pH or water-table height, from fossil assemblages. Initially, a modern reference data-set of living populations is needed, associated with environmental variables (Janssens 1989). These results can then be used to predict values for the environmental variables from the fossil assemblages. Techniques to analyze peats for fossil bryophytes and quantitative reconstruction of fossil assemblages are explained in detail in Janssens (1989). In the following the reference data-set on North American bryophyte populations as it presently exists is described and the inference technique is outlined. 2. The reference data-set: pH, and the height of the peat surface above the water table are major environmental variables explaining most biological relevant variability in the North American bryophyte reference data-set. The relationship among three major peatland gradients is illustrated in Figure 1: Ph, height above the mean water table (HMWT), and shade are represented as arrows on both a species-environment and sample-environment biplot (Jongman et al. 1987). Species and plot scores are the result of a canonical correspondence analysis (CCA) ordination obtained with the CANOCO (ter Braak 1987). Data input consists of cover values for all bryophytes of 226 plots collected in 5 major regions in North America (Gorham and Janssens, in press) and of measurements of pH, shade, and HMWT (see Janssens 1989 for details on the standardization to mean water table). No significant correlations exist among the three environmental variables, plotted as arrows on the ordinations. This supports the individual gradient concept for all three environmental variables (Janssens 1989). Relative length of the arrows on the biplots indicates approximately the importance of the variable in separating species and plots. pH and HMWT explain most of the biological relevant information in the data. The most important species outliers (common peatland species or common fossils in peat) have been labeled on the species-environment biplot (Figure 1, left-hand side). The environmental arrows can be extended backwards through the origin of the ordination and individual species scores can be projected perpendicularly onto each one of the environmental arrows (Jongman et al. 1987). For example, Drepanocladus fluitans and Polytrichum strictum (Figure 1, left-hand side) are similar in response to pH, but have very different optima along the HMWT gradient. 3. Paleoenvironmental Reconstruction: Birks (1985, 1987) reviews the numerical procedures to obtain the inferences and Birks et al. (1990) concluded that the statistical calibration technique of weighted averaging (WA) gives superior results in the prediction of pH. Line and Birks (1989) and Birks et al. (1990) discuss the underlying theory of WA regression and calibration, and compare its performance with computationally more difficult techniques such as Maximum [*p.1 / p.2*] Likelihood regression and calibration. ---- Figure 1 on p. 2 is not reproduced in this ASCII-text version. Figure 1. CCA species-environment and sample-environment biplots. The environmental variables are represented by arrows. (Eigenvalues of canonical (constrained) axes: lambda(1)=0.5474, lambda(2)=0.3946, and lambda(3)=0.2333.) ---- Below follows a summary of the major ecological assumptions underlying WA, used for quantitative paleoenvironmental reconstruction, from Birks et al. (1990), discussed with fossil bryophytes in mind. (1) The environmental variables to be reconstructed, pH and HMWT, have to be ecologically the most important variables or the best parameters to reconstruct gradients of interest in peatlands. This is supported with the literature review in Janssens (1989) and Gorham and Janssens (in press). Above it was also illustrated (Figure 1) by the results of the CCA ordination, constrained by pH, HMWT, and shade. They indicate that the two first environmental variables are the two major parameters explaining most biological variability in the data set. (2) Bryophyte species in the reference data-set have to be the same as in the fossil assemblages and their ecological responses have not changed significantly over the late Holocene time span represented by the peatland profiles. The entire topic of palaeoenvironmental reconstruction is discussed in Janssens (1990), in which the rare occurrence of ecotypes and non-analogous fossil assemblages is explained. It suffices to indicate here that no species are recorded as fossils in the Holocene peat deposits of North American that are not found as living populations. In addition, few fossil assemblages have no representatives among plots of the reference data-set. (3) The mathematical method used in WA should adequately model the biological response to the environmental variables of interest. Gorham and Janssens (in press) clearly indicate that most bryophyte species have a well defined optimum and a species-specific tolerance along the pH gradient. The WA estimate of a species' optimum (=abundance weighted mean) is ^ n n u = Sigma y x / Sigma y i=1 ik i i=1 ik [*p.2 / p.3*] Table 1. Optima, tolerances, and number of occurrences among all plots for pH and HMWT calculated with WACALIB 2.1 for the most common bryophyte species of the combined North American reference data-sets (total n=431, total m=217) ---------------------------------------------------------------------------- species pH height ^ ^ ^ ^ u t n u t m k k k k ---------------------------------------------------------------------------- Sphagnaceae S. fallax 4.27 0.68 32 12.5 10.2 22 S. fuscum 4.35 0.73 46 27.5 9.5 25 S. capillifolium 4.33 0.71 83 19.3 7.4 41 S. majus 4.40 0.44 20 6.6 4.7 18 S. angustifolium 4.69 0.89 99 19.5 8.9 58 S. papillosum 4.33 0.38 59 12.0 6.8 41 S. magellanicum 4.39 0.65 103 19.4 8.3 65 S. centrale 5.68 0.99 27 19.9 6.9 10 S. teres 5.78 0.88 24 12.7 8.5 6 S. subsecundum 5.29 0.43 31 10.0 4.9 22 S. warnstorfii 6.77 0.47 33 19.2 8.7 9 Amblystegiaceae Drepanocladus fluitans 4.66 0.70 12 3.7 7.0 8 Calliergon stramineum 5.03 0.72 46 11.9 8.2 25 C. cordifolium 6.17 0.62 15 7.4 7.5 4 Calliergonella cuspidata 6.19 1.05 14 - - 0 Calliergon giganteum* 6.91 0.35 11 6.0 3.1 7 Campylium stellatum 6.56 0.78 29 10.8 5.9 2 Scorpidium scorpioides* 6.65 0.71 16 3.3 2.4 8 Drepanocladus revolvens* 6.88 0.54 10 8.7 4.2 6 Amblystegium riparium 6.79 0.44 17 3.6 - 1 Others Mylia anomala 4.11 0.26 25 20.7 8.4 12 Pleurozium schreberi 4.95 1.34 73 26.9 9.7 37 Polytrichum strictum 4.71 1.01 36 29.7 9.2 18 Dicranum undulatum 4.91 1.31 22 24.3 8.2 8 Cladopodiella fluitans 4.39 0.30 32 6.5 3.6 22 Aulacomnium palustre 5.47 1.15 86 19.5 8.5 40 Helodium blandowii 6.96 0.41 26 11.2 11.7 3 Plagiomnium ellipticum 6.88 0.61 22 13.0 11.4 6 Hylocomium splendens 5.90 1.38 21 19.4 13.4 11 ------------- * The height optima and tolerances for these three species are above the local water table, because no plots with these species were available in the data- set for mean water table standardized to Sphagnum angustifolium. ------------------------------------------------------------------------------ [*p.3 / p.4*] and its tolerance (weighted standard deviation) is ^ n ^ 2 n 1/2 t = ( Sima y (x - u ) / Sigma y ) k i=1 ik i k i=1 ik in which x is the environmental variable (pH or height), x(1) is the value of X in sample i (a plot), and y(ik) is the abundance of species k in plot i, (i-1),..., n plots; k=1, ..., m moss species). Table 1 lists the WA optima and tolerances for pH and HMWT of the same set of species illustrated in Figure 2. Presently the entire North American reference data set has been used to infer pH and HMWT from fossil assemblages identified in 20 Holocene peat profiles from southeastern Alaska, Minnesota (Janssens et al. in press), New York, and Maine. A detailed description of stratigraphical techniques and inference results is available in Janssens (1989) and Janssens (1989) and Janssens et al. (in press). 4. References: Birks, H.J.B., 1985, Recent and possible future mathematical developments in quantitative paleoecology. Paleogeography, Paleoclimatology, Paleoecology 50:107-147. Birks, H.J.B., 1987, Methods for pH-calibration and reconstruction from paleolimnological data: procedures, problems, potential technique. Pp 370-380, in Proceedings of Surface Water Acidification Programme (SWAP). Mid-Term Review Conference Bergen 22-26 Jume, 1987. Birks, H.J.B. and H.H. Birks, 1980, Quaternary Palaeoecology. Edward Arnold Publishers, London, 289 p. Birks, H.J.B., J.M. Line, S. Juggins, A.C. Stevenson, and C.J.F. ter Braak, 1990, Diatoms and pH reconstruction. Philosophical Transactions of the Royal Society of London B, in press. Dickson, J.H., 1986, Chapter 30: Bryophyte analysis. Pp 627-643, in B.E. Berglund (ed.), Handbook of Holocene Paleoecology and Paleohydrology. Wiley & Sons, New York. Gorham, E. and J.A. Janssens, Concepts of fen and bog re-examined in relation to bryophyte cover and the acidity of surface waters. Acta Societatis Botanicorum Poloniae. Invited Paper, in press. Janssens, J.A., 1989, Ecology of Peatland Bryophytes and Paleoenvironmental Reconstruction of Peatlands using Fossil Bryophytes. Methods Manual, Update December, 1989. Janssens, J.A., 1990, Methods in Quaternary Ecology - Bryophytes. Geoscience Canada, in Press. Janssens, J.A., B.C.S. Hansen, P.H. Glaser, and C.W. Barnowsky. Development of a raised bog complex in northern Minnesota. In H.E. Write Jr., B. Coffin, and N. Aasing (eds). Patterned Peatlands of northern Minnesota, in press. Jongman, T.H.G., C.J.F. ter Braak, and O.F.R. van Tongeren, 1987, Data Analysis in Community and Landscape Ecology. Pudoc, Wageningen, 200 p. Line, J.M., and H.J.B. Birks, 1989, WACALIB 2.1 - a computer program to reconstruct environmental variables from fossil assemblages by weighted ave raging. Available for US $50.00 from H.J.B. Birks, Botanical Institute, University of Bergen, Allegaten 41, N-5007 Bergen, Norway. Miller, N.G., 1980, Mosses as paleoecological indicators of late-glacial terrestrial environments: some North American studies. Torrey Botanical Club, Bulletin 107:373-391. ter Braak, C.J.F., 1987, CANOCO -- a FORTRAN program for canonical community ordination by (partial) (detrended) (canonical) correspondence [*p.4 / p.5*] analysis, principal components analysis and redundancy analysis (version 2.1). ITU- TNO, Wageningen, 95 p. Jan A. Janssens, Department of Ecology, Evolution and Behaviour, University of Minnesota, 208 Zoology Building, 318 Church Street, SE, Minneapolis, Minnesota 55455, USA. (612) 625-9680. FAX: (612) 625-4490. Pollen Counting on a Pocket Computer K.D. Bennett Pollen counting is time-consuming, and the process is frequently lengthened by the necessity of transcribing and checking data for handling by computer programs. It has long been realized that computers can be programmed to act as counters, but the expense of providing each analyst in a laboratory with either their own PC or a terminal and line to a mainframe computer has limited the use of this possibility. Pocket computers are now available which can handle the task of acting as pollen counters and then transferring the results to a large computer for further handling. One such system is marketed in Europe and North America by Psion plc. The smallest model they produce (Psion Organizer CM) has 8 kbytes RAM and the possibility of adding two EPROM Datapaks, of up to 64 kbytes each. With mains adaptor and two 16 kbyte EPROMs, a model Cm costs about 135 Sterling (Jan. 1990). This is less than the cost of a 15-place mechanical counter. Psion also produce the hardware and software necessary to transfer data as ASCII files direct to an IBM-PC. I have written a suite of procedures using the Psion programming language (OPL) which enable this computer to act as a pollen counter, and also to tally charcoal particles using the point-count estimation method. The amount of memory available is more than adequate for it to handle the data collected during the course of one pollen count. These procedures have variables to store the names and counts of 100 taxa, and can produce results as raw data, percentages and concentrations. I normally transfer the raw data to a PC in a form that can be used immediately as input for a spreadsheet, and then work with the data from all levels of a site together. Increasing miniaturization of components has lead to the appearance of several other pocket computers, any of which could probably be programmed in the same way as I have programmed the Psion. As far as I am aware, however, the Psion is still the cheapest with a programming language as standard, and thus much the most suitable for use in a laboratory where several (or many?) may be desired. The ease of use and programming makes it possible for each analyst to have programs tailored to the individual needs of their type of data, or their preferred form of output. I will be happy to supply, free of charge, copies of the OPL procedures that I use to anyone interested, in a Datapak (at least 16 kbytes, provided by recipient), or an IBM-compatible 3.5" floppy disc, or on paper. A more detailed account of the use of the Psion Organizer as a pollen counter will be published in New Phytologist 114, Feb. 1990. K.D. Bennett, Subdepartment of Quaternary Research, University of Cambridge, Botany School, Downing Street, Cambridge CB2 3EA, England. THE AASP DATA COMMITTEE Owen K. Davis The AASP (American Association of Stratigraphic Palynologists) Data Committee was created in 1988 to promote and facilitate the exchange of palynological information and data via electronic media. Committee members now have contributed 9 "PALYDISKS" that reflect the breadth of the organization -- from archeological palynology to paleozoic acritarachs. The Palydisk files [*p.5 / p.6*] are on 5", 360K, IBM-format, floppy disks, but translation to and from other media is possible. The Palydisks accumulate as Data Committee members contribute diskettes of original material and receive comparable material in return. Of course, the material must be in the "public domain"; the author must formally state the material is available for non-profit distribution. We have provided palydisks to those who offered nothing in return, but most persons who use computers have had something of value to trade. If you are interested in exchanging data, write or call Owen Davis (602) 621- 7953, Telex: 187 167 ACUTUC, Fax: (602) 621-2672, Bitnet: PALYNOLO@ARIZRVAX. David, O.K., S.T. Hall, J. Schoenwetter, and G.J. West, 1988, An ASCII file bibliography of Southwestern Quaternary Palynology. AASP Data Committee, PALYDISK 1 (1 diskette). Elsik, W.C., 1989, An ASCII file bibliography of living and fossil fungi: annotated fungal references: AASP Data Committee, PALYDISK 2 (2 diskettes). Elsik, W.C., 1989, An ASCII file bibliography of Gulf Coast palynology, AASP Data Committee, PALYDISK 3 (1 diskette). Fredlund, G.G., 1989, PD5.EXE: pollen diagrams on an EGA monitor. AASP Data Committeem PALYDISK 4 (1 diskette). Fensome, R.A., 1989, AASPPALY.DBF: A dBase file of AASP members, 876 records, 378 characters in 17 fields. AASP Data Committee, PALYDISK 6 (1 diskette). Holloway, R.G., 1989, COPROLIT.DBF, MORPH.DBF, and STATS.DBF. Three dBase bibliographic files of Quaternary pollen literature concerning coprolite analysis, pollen morphology, and numerical methods. AASP Data Committee, PALYDISK 7 (1 diskette). Holloway, R.G., 1989, ABOT.DBF, NE.DBF, and NW.DBF: Three dBase bibliographic files of Quaternary pollen literature concerning archeological botany, northeastern U.S.A., and northwestern U.S.A. references. AASP Data Committee, PALYDISK 8 (1 diskette). Holloway, R.G., 1989, SE.DBF: A dBase bibliographic file of Quaternary pollen literature concerning southeastern U.S.A. AASP Data Committee, PALYDISK 9 (1 diskette). Owen K. Davis Department of Geosciences University of Arizona Tucson, AZ 85721 An Annotated Bibliography of Numerical Methods in Quaternary Pollen Analysis 1985 - 1989. H.J.B. Birks and Magne Saetersdal We are preparing an annotated bibliography of publications on the application of numerical methods to Quaternary pollen analysis. The publications included have either appeared in the 5-year period 1985-1989 or are not listed in bibliography of Birks and Gordon's (1985) Numerical Methods in Quaternary Pollen Analysis (Academic Press). Our bibliography attempts to cover all Quaternary palynological applications of statistical methods and numerical theory, and new quantitative procedures of direct or potential applicability to pollen analysis. It contains about 800 entries, annotated by subject. Besides pollen-stratigraphical and surface-sample studies, we have tried to include references on image analysis, automated counting, error estimation, compositional data analysis, constrained ordinations, pollen mapping, aerobiology, and new, relevant statistical developments. Copies will be available in April 1990 at cost price (about 3 pounds, $5) from H.J.B. Birks, Botanical Institute, University of Bergen, Allegaten 41, N-5007 Bergen, Norway (email birks@rose.uib.uninett). [*p.6 / p.7*] H.J.B. Birks and Magne Saetersdal, Botanical Institute, University of Bergen, Bergen, Norway. New Bookshelf H.J.B. Birks The aim of this item is to share information about new books that may be of interest to readers of this newsletter. I hope that this will be a two-way process and you will tell me about any recently published books that you think would also be of interest to readers. Here are some recent titles of books that I have found interesting and useful. Atkinson, A.C., 1987, Plots, transformations and regression. An introduction to graphical methods of diagnostic regression analysis. Clarendon Press, Oxford 282 pp. (paperback). Burrough, P.A., 1986, Principles of geographical information systems for land resources assessment. Clarendon Press, Oxford, 193 pp. (paperback). Chatfield, C., 1988, Problem solving. A statistician's guide. Chapman and Hall, London, 261 pp. (paperback). Dunn, G., 1989, Design and analysis of reliability studies - the statistical evaluation of measurement errors. Edward Arnold, London, 198 pp. Everitt, B.S., 1987, Introduction to optimization methods and their application in statistics. Chapman and Hall, London, 99 pp. (paperback). Ferguson, J., 1988, Mathematics in geology. Allen & Unwin, London, 299 pp. Gilbert, R.O., 1987, Statistical methods for environmental pollution monitoring. Van Nostrand Reinhold, New York, 320 pp. Hanley, J.T. and Merriam, D.F., 1986, Microcomputer applications in geology, Pergamon Press, Oxford, 258 pp. Hengeveld, R., 1989, Dynamics of biological invasions. Chapman and Hall, London, 160 pp. Jeffers, J.N.R., 1988, Practitioner's handbook on the modelling of dynamic change in ecosystems. Scope 34, Wiley, Chichester, 181 pp. Kirkby, M.J. et al., 1987, Computer simulation in physical geography. Wiley, Chichester, 227 pp. (computer diskette for IBM PC's also available). Krebs, C.J. 1989, Ecological Methodology. Harper and Row, New York, 654 pp. Krzanowski, W.J., 1988, Principles of multivariate analysis: a user's perspective. Oxford University Press, 563 pp. Ludwig, J.A. & Reynolds, J.F., 1988, Statistical ecology. A primer on methods and computing. Wiley, Chichester, 337 pp. (paperback with computer diskette for IBM PC's). Maguire, D.J., 1989, Computers in geography. Longman, Harlow, 348 pp. (paperback). Magurran, A.E., 1988, Ecological Diversity and its measurements. Chapman and Hall, 179 pp. (paperback). Mather, P.M., 1987, Computer processing of remotely-sensed images. Wiley, Chichester, 352 pp. Morrill, R., Gaile, G.L. and Thrall, G.I., 1988, Spatial diffusion. Sage Publications, Newbury Park, 86 pp. (paperback). Myers, A.A. and Giller, P.S. (eds.), 1988, Analytical biogeography. An integrated approach to the study of animal and plant distributions. Edward Arnold, London, 578 pp. Odland, J., 1988, Spatial autocorrelation. Sage Publications, Newbury Park, 87 pp. (paperback). [*p.7 / p.8*] Rock, N.M.S., 1988, Numerical geology. Springer-Verlag, Berlin, 427 pp. (paperback). Rosswall, T. et al., 1988, Scales and global change. Scope 35, Wiley, Chichester, 355 pp. Rousseeuw, P.J. and Leroy, A.M., 1987, Robust regression and outlier detection. Wiley and Sons, New York, 329 pp. Size, W.B., 1987, Use and abuse of statistical methods in the Earth Sciences, Oxford University Press, New York, 169 pp. Swartzman, G.L. and Kaluzny, S.P., 1987, Ecological simulation primer. Macmillan, New York, 370 pp. Unwin, D.J. and Dawson, J.A., 1985, Computer programming for geographers, Longman, London, 252 pp. (paperback). Upton, G.J.G. and Fingleton, B., 1985, Spatial data analysis by example 1. Point pattern and quantitative data, Wiley, Chichester, 410 pp. Upton, G.J.G. and Fingleton, B., 1989, Spatial data analysis by example 2. Categorical and directional data. Wiley, Chichester, 416 pp. H.J.B. Birks, Botanical Institute, University of Bergen, Bergen, Norway. Useful PC Software 1 H.J.B. Birks This article gives details of several moderately well-known and some less well-known IBM PC computer programs that in the last few years I have found to be particularly useful for multivariate data analysis, for printing or plotting, or for data (re)formatting. No attempt is made to provide full reviews. I simply summarize what the programs do, their availability, and my comments on their performance based on my experiences. FPLOT - This is an easy-to-use pen-plotter emulating program that allows the use of a dot-matrix or laser printer as a HPGL pen plotter. It produces high- resolution printer output of HPGL plot-files, as well as allowing the viewing of HPGL plot-files on Hercules, EGA, CGA, or VGA displays. Available from FPLOT Corporation, 24-16 Steinway Street, Suite 605, Astoria, New York 11103, USA. ($64). Extremely useful, and invaluable for debugging graphics programs generating HPGL output. PIZAZZ - This is an easy-to-use memory resident program that captures almost any image you can display on the computer's screen and prints it on dot-matrix or laser printers. It allows change of print colours, grey tones, sizes, proportions, rotations, image smoothing, print density, and position. Works with Hercules, EGA, CGA, and many other adapter cards. Available from Application Techniques, Inc., 10 Lomar Park Drive, Pepperell, Massachusetts 01463, USA (ca. $100). Excellent, very elegant and staggering versatility. PRINTALL - This is similar but much less versatile than PIZAZZ in that it allows printing of whatever image can be displayed on the computer's screen (EGA, CGA, Hercules, etc.). Available from MTR Software, P.O. Box 13, Islington 'A', Toronto, Ontario, CANADA M9A 4X1 ($40). SCI-GRAF - This is graphics program for producing a wide variety of high- quality graphs and scatter plots on dot-matrix printers. It is easy to use, is menu-driven, and can import data from several standard packages. Available from Microcomputer Systems Consultants, 32 West Anapamu, Suite 190, Santa Barbara, California 93101, USA ($99.95). [*p.8 / p.9*] Useful, but now largely replaced by GRAPHER (see PC Magazine 14 March 1989 for a detailed review of scientific graphics packages). PC MAPICS - This is an extremely powerful package for capturing, managing, analyzing, manipulating, and plotting spatial information. It also includes versatile graphics facilities and can produce a wide range of point and area maps, scattergrams, barcharts, line graphs, and histograms on HPGL plotters. A program for incorporating digitized data (PC-ISIS) is also available. PC MAPICS is run through a series of easy commands, allowing very great flexibility in mapping and data handling. Available from MAPICS Ltd., 26 Bedford Way, London WC1H 0AP England (ca 450 pounds). Excellent, powerful, and remarkably good value-for-money. It takes some time to learn how to use it, but it is time well spent. Strongly recommended. SAAP - This is Danile Wartenberg's program for spatial autocorrelation analysis, using Moran's I and Geary's c statistics. It provides statistical summaries, graphical plots, and geographical maps for up to 400 localities and 25 variables using up to 25 correlogram classes. Available from Exeter Publishing Ltd., 100 North Country Rd., Bldg. B., Setauket, New York 11733, USA ($125). Easy to use and moderately versatile. BIOM-pc - This is a set of 22 computer programs to accompany R.R. Sokal's and F.J. Rohlf's Biometry (Second edition). The programs cover basic statistics, analysis of variance, differences among means, correlation and regression, non-parametric statistics, randomization tests, and jack-knife resampling. They are a useful adjunct to Sokal and Rohlf's classic biometric statistics text Biometry. The related program BIOMLAB written by F.J. Rohlf and Dennis Slice is a fascinating statistical simulation program in which you can set up different sampling models and investigate effects of changing sample size, distribution type, transformations, and alternative hypotheses. It covers topics in sampling, error types, correlation, regression, and confidence intervals. Superb for teaching and demonstrating basic statistical principles. Both are available from Exeter Publishing Ltd., 100 North Country Rd., Bldg. B., Setauket, New York 11733, USA (BIOM-pc $60; BIOMLAB $90). Very useful both for teaching and research. CODA - This is a series of BASIC programs written by John Aitchison, that implements nearly all the new and important statistical methods developed and described by Aitchison in his 1986 monograph The Statistical Analysis of Compositional Data (Chapman and Hall - see the review of it in Newsletter No. 1, 1988). Many of these procedures are very relevant to pollen analysis because our data are often closed percentage data. The CODA programs are easy-to-use and powerful, but output is rather limited and graphics are poor. Available from Chapman and Hall, 11 New Fetter Lane, London EC4P 4EE, England (95 pounds). A revised version of CODA is available - one should contact Professor John Aitchison directly (Department of Statistics, University of Hong Kong) for details about availability. Important package for people interested in statistical analysis of percentage data. NTSYS-pc - This is a PC version of F.J. Rohlf's well-known NTSYS main-frame package for numerical taxonomy. NTSYS-pc is a great improvement on the mainframe version, and covers similarity and dissimilarity measures, cluster techniques, minimum spanning trees, principal components analysis, non-metric scaling, principal coordinates analysis, correspondence analysis, cophenetic correlations, Mantel's test, Fourier [*p.9 / p.10*] transformations, dendrogram graphics, and two- and three-dimensional scatter plots. Easy-to- use, extremely fast, and versatile. Available from Exeter Publishing Ltd., 100 North Country Rd., Bldg. B, Setauket, New York 11733, USA ($130). Excellent both for teaching multivariate data analysis and for research purposes. (Other programs available from Exeter Publishing Ltd., include software for Fractals, Image Analysis, and Chaos Modelling, and J. Podani's SYN-TAX III data-analysis package. I have no experience of any of these). CANOCO - Although described by ter Braak (1988 Vegetatio 75, 159-160) as "an extension of DECORANA to analyze species - environment relationships", it is much more than this. It is probably the most powerful multivariate ordination program currently available. It implements principal components analysis (including biplots), (detrended) correspondence analysis, and, if 'external' data are available, canonical correspondence analysis, redundancy analysis, and canonical variates analysis. It can also implement 'partial' analyses where the effects of a particular external variable are removed, and allows statistical testing of patterns and relationships by Monte Carlo permutation test. The PC version can handle up to 750 samples and 600 species. Available from Cajo ter Braak, Agricultural Mathematics Group, Box 100, 6700 AC Wageningen, The Netherlands ($150). Once the data are prepared, CANOCO is very easy to use. A truly excellent program. CORNELL CEP programs - PC versions, source code, and the original documentation for the Cornell Ecology Programs DECORANA ORDIFLEX, COMPCLUS, and TWINSPAN are available from Microcomputer Power, 111 Clover Lane, Ithaca, New York 14850, USA. Public-domain PC versions of TWINSPAN, DECORANA, COMPCLUS, DATAEDIT, ORDIFLEX, and CONDENSE are available from various sources. I can provide details for those interested. DBMS/COPY and DBMS/COPY Plus - These are two remarkable and extremely useful programs that allow one to move data from one supported software package to another supported package either in toto or as selected variables or objects only. Nearly all the standard packages, spreadsheets, and databases are supported, e.g. SAS, BMDP, dBase III, Excel, GLIM, Lotus 1-2-3, SOLO, SPSS/PC, Stratigraphics, Quattro, Paradox, Oracle, Reflex, PC-File, Minitab, SYSTAT, Symphony, etc. along with many more. It will also translate pure ASCII files (both fixed-format and free-format) into the required format for the supported packages. Available from Conceptual Software Inc., P.O. Box 56627, Houston, Texas 77256, USA ($195 for DBMS-copy, $295 for DBMS/Plus). Excellent and invaluable if you use more than one PC package. It really makes data transfer from one program to another easy, fast and trouble-free. PC MAGAZINE - the 14 March 1989 issue of PC Magazine (volume 8, number 5) may be of interest to many readers of this Newsletter. It contains detailed reviews of 15 programs for advanced statistical analysis, 14 programs for basic statistical analysis, 7 programs for two-dimensional scientific graphical plots, 4 programs for three-dimensional scientific graphics, 6 programs for econometric modelling, 8 for time-series analysis and forecasting, and 6 programs for quality control and the analysis of survey data. The recommended basic statistical packages are MINITAB and STATISTIX, the advanced statistical packages are SYSTAT (along with its graphical package SYGRAPH), SPSS/PC+, and STATGRAPHICS, the two-dimensional plotting packages are SIGMA-PLOT and GRAPHER, and the three-dimensional graphics packages are SURFER and SDS. [*p.10 / p.11*] Tailpiece I am an observation. I was captured in the field. My conscience said 'co-operate' My instinct said 'don't yield'. But I yielded up my data Now behold my sorry plight I'm part of a statistic Which is not a pretty sight. The Bootstrap and the Jackknife Oh, the tortures I've endured They analyze my variance Until my meaning is obscured. But I've a plan to beat them I'll climb up in the trees Pretend I am a chi-square And get freedom by degrees. From: Problem Solving. A Statistician's guide. Christopher Chatfield, 1988, Chapman & Hall. H.J.B. Birks, Botanical Institute, University of Bergen, Bergen, Norway. [ Email addresses (not reproduced) extend onto p. 12. ]